Fan blade intelligent inspection method and system based on unmanned aerial vehicle

Abstract

The invention provides an intelligent inspection method and system for fan blades based on an unmanned aerial vehicle. The method comprises the steps that a visible light camera, an infrared thermal imager and a laser radar are integrated on the unmanned aerial vehicle; generating a flight route covering the whole surface based on the fan blade three-dimensional model; when the unmanned aerial vehicle is controlled to fly along the flight route, visible light images, infrared temperature data and laser point clouds are synchronously collected; the visible light image, the infrared temperature data and the laser point cloud are input into a defect analysis model, the defect type and grade are identified through conjoint analysis, a temperature abnormal area is positioned, and geometric deformation is quantified; a health assessment report is generated, the health assessment report comprises a defect positioning map and a graded maintenance decision matched with defect grades, multi-source data are collected through integration of a visible light camera, an infrared thermal imager and a laser radar, and then fusion analysis and recognition are carried out, so that compared with a manual inspection mode, the intelligent level of fan blade inspection is improved, and the inspection efficiency is improved. And the inspection efficiency of the fan blade is also improved.

Description

Technical Field

[0001] The present invention relates to the technical field of blade inspection, and in particular to a method and system for intelligent inspection of wind turbine blades based on a drone. Background Art

[0002] As the core load-bearing component of wind turbines, wind turbine blades are subjected to complex, alternating loads over long periods of time, making them susceptible to damage such as surface cracks, structural delamination, and geometric deformation. Traditional inspection methods have the following limitations: Manual inspections rely on close observation by technicians using high-powered telescopes or hanging baskets. This leads to large blind spots, high subjectivity, and high-altitude inspection risks, particularly in hard-to-reach areas like blade tips.

[0003] Therefore, how to improve the intelligent level of wind turbine blade inspection has become a technical problem that technical personnel in this field urgently need to solve. Summary of the Invention

[0004] The present invention provides a method and system for intelligent inspection of wind turbine blades based on a drone, so as to solve the defect of poor intelligence level of wind turbine blade inspection in the prior art.

[0005] In a first aspect, the present invention provides a method for intelligent inspection of wind turbine blades based on a drone, comprising: Integrate visible light cameras, infrared thermal imagers and lidar on drones; Generate a flight path covering the entire surface based on the three-dimensional model of the wind turbine blade, responding to the sensor configuration requirements of the visible light camera, infrared thermal imager and lidar; When controlling the UAV to fly along the flight route, the visible light image, infrared temperature data and laser point cloud are synchronously collected and bound to the time and space coordinate information; Inputting the visible light image, the infrared temperature data and the laser point cloud into a defect analysis model, identifying the defect type and grade, locating the temperature anomaly area, and quantifying the geometric deformation through joint analysis; Based on the defect type and level, the temperature anomaly area and the geometric deformation, a health assessment report is generated, which includes a defect location map and a graded maintenance decision matching the defect level.

[0006] According to the present invention, a method for intelligent inspection of wind turbine blades based on a drone is provided, wherein a flight route covering the entire surface is generated based on a three-dimensional model of the wind turbine blade, comprising: Deconstruct the 3D model of the wind turbine blade to divide the aerodynamic topology areas; Generating a scanning trajectory with no blind spot coverage based on the features of each of the aerodynamic topological regions; Dynamically calculating the optimal flight altitude based on the blade length and the field of view angles of the visible light camera, infrared thermal imager, and lidar; The scanning trajectory and the optimal flight altitude are used as the flight route.

[0007] According to the present invention, a method for intelligent inspection of wind turbine blades based on a drone further includes: Calculating waypoint density in response to field of view angle parameters of the visible light camera, infrared thermal imager, and lidar; Dynamically adjust the route starting point based on the real-time yaw data of the wind turbine; Based on the waypoint density and the route starting point, the route flight speed is optimized according to the weather station wind speed forecast.

[0008] According to the present invention, a method for intelligent inspection of wind turbine blades based on a drone, before synchronously collecting visible light images, infrared temperature data, and laser point clouds, further includes: Real-time collection of atmospheric transmittance and solar irradiance; Compensating for infrared temperature measurement deviation based on the atmospheric transmittance; The visible light dynamic range is optimized according to the solar irradiance.

[0009] According to a drone-based intelligent inspection method for wind turbine blades provided by the present invention, the binding of spatiotemporal coordinate information includes: Synchronize visible light images and UAV six-degree-of-freedom pose data to establish a spatial mapping benchmark; Correlate infrared data with the coordinates of blade stress concentration areas to locate abnormal temperatures; Motion distortion correction is performed on the laser point cloud to ensure the accuracy of deformation analysis.

[0010] According to the present invention, a drone-based intelligent inspection method for wind turbine blades is provided, wherein the joint analysis is used to identify defect types and levels, locate temperature anomaly areas, and quantify geometric deformations, including: Fusion of visible light images and infrared data to detect temperature anomalies associated with surface defects; Combine laser point cloud and visible light image to locate apparent damage in geometric deformation areas; Integrate multimodal outputs to generate defect-deformation-temperature correlation assessments.

[0011] According to a drone-based intelligent inspection method for wind turbine blades provided by the present invention, generating a health assessment report includes: Analyzing the correlation assessment results and defining a three-dimensional damage index; matching a gradient repair scheme according to a damage index threshold of the three-dimensional damage index; When the three-dimensional damage index exceeds a limit, an emergency shutdown alarm is generated.

[0012] According to the present invention, a method for intelligent inspection of wind turbine blades based on a drone further includes: Mapping the three-dimensional damage index to a three-dimensional blade model to generate a thermal map; Correlating the thermal map with a historical damage index to draw a degradation trajectory curve; The remaining service life is predicted based on the degradation trajectory curve.

[0013] According to the present invention, a method for intelligent inspection of wind turbine blades based on a drone is provided, wherein the method fuses visible light images with infrared data to detect temperature anomalies associated with surface defects, and further comprises: When the geometric deformation exceeds the threshold, the infrared data of the corresponding space-time coordinates are retrieved; Establishing a damage evolution map by correlating the deformation area with the visible light image defect distribution using the infrared data; Based on the evolution map, the root cause category of material fatigue or impact damage is identified.

[0014] In a second aspect, the present invention further provides a wind turbine blade intelligent inspection system based on a drone, comprising: Integration module for integrating visible light camera, infrared thermal imager and lidar on drone; A response module, configured to generate a flight path covering the entire surface based on the three-dimensional model of the wind turbine blade, and respond to sensor configuration requirements of the visible light camera, infrared thermal imager, and lidar; An acquisition module, configured to control the UAV to fly along the flight route, synchronously acquire visible light images, infrared temperature data, and laser point clouds, and bind the time and space coordinate information; an identification module, configured to input the visible light image, the infrared temperature data, and the laser point cloud into a defect analysis model, and identify the defect type and level, locate the temperature anomaly area, and quantify the geometric deformation through joint analysis; A generation module is used to generate a health assessment report based on the defect type and level, the temperature anomaly area and the geometric deformation, which includes a defect location map and a graded maintenance decision matching the defect level.

[0015] In a third aspect, the present invention also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the intelligent inspection method for wind turbine blades based on a drone as described above is implemented.

[0016] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described drone-based intelligent inspection methods for wind turbine blades.

[0017] In a fifth aspect, the present invention further provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-described drone-based intelligent inspection methods for wind turbine blades.

[0018] The present invention provides a drone-based intelligent inspection method and system for wind turbine blades, which includes integrating a visible light camera, an infrared thermal imager and a laser radar on the drone; generating a flight route covering the entire surface based on a three-dimensional model of the wind turbine blade, responding to the sensor configuration requirements of the visible light camera, infrared thermal imager and laser radar; controlling the drone to fly along the flight route, synchronously collecting visible light images, infrared temperature data and laser point clouds, and binding time and space coordinate information; inputting the visible light images, infrared temperature data and laser point clouds into a defect analysis model, identifying the defect type and level, locating temperature anomaly areas, and quantifying geometric deformations through joint analysis; generating a health assessment report based on the defect type and level, temperature anomaly areas and geometric deformations, which includes a defect location map and graded maintenance decisions matching the defect level. By integrating the visible light camera, infrared thermal imager and laser radar to collect multi-source data, and then performing fusion analysis and identification, compared with the manual inspection method, the intelligent level of wind turbine blade inspection is improved, as well as the inspection efficiency of wind turbine blades is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 This is a flow chart of the intelligent inspection method for wind turbine blades based on drones provided in this embodiment; Figure 2 Schematic diagram of the structure of the wind turbine blade intelligent inspection system based on drones provided in this embodiment; Figure 3 Schematic diagram of the structure of the electronic device provided in this embodiment. DETAILED DESCRIPTION

[0021] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0022] Figure 1 This is a flow chart of the intelligent inspection method for wind turbine blades based on drones provided in this embodiment.

[0023] like Figure 1 As shown, the intelligent inspection method for wind turbine blades based on a drone provided in an embodiment of the present invention mainly includes the following steps: 101. Integrate visible light cameras, infrared thermal imagers and lidar on drones.

[0024] Specifically, a high-load, highly stable drone platform was selected, with the lidar fixed at the center of the base to ensure a 360-degree scanning angle. The visible light camera and infrared thermal imager were mounted on an adjustable gimbal, allowing the camera's position to be adjusted based on inspection needs, ensuring that each device was securely mounted and did not interfere with each other.

[0025] Data transmission and power supply lines are rationally laid out to avoid entanglement. A power management system is established to prioritize power supply to key equipment, such as lidar and data transmission modules, based on differences in device power consumption, ensuring complete data collection. A unified software system coordinates drone flight control and device acquisition commands, setting operating parameters for each device as needed. The system also integrates lidar 3D point cloud data and visible light image data in real time to generate visual inspection results.

[0026] Through multi-device integration, visible light cameras capture surface defects such as cracks and wear; infrared thermal imagers detect internal structural anomalies or localized overheating; and lidar precisely measures blade deformation and constructs a three-dimensional model. This combination enables comprehensive, high-precision defect detection, significantly improving detection efficiency and accuracy compared to manual inspections. It is particularly suitable for rapid inspections of large-scale wind farms.

[0027] Generate a flight path covering the entire surface based on the three-dimensional model of the wind turbine blade, responding to the sensor configuration requirements of visible light cameras, infrared thermal imagers and lidar.

[0028] Specifically, the three-dimensional model of the wind turbine blade is deconstructed to divide the aerodynamic topology areas; a scanning trajectory with no blind spot coverage is generated based on the characteristics of each aerodynamic topology area; the optimal flight altitude is dynamically calculated based on the blade length and the field of view of the visible light camera, infrared thermal imager and lidar; the scanning trajectory and optimal flight altitude are used as the flight route to respond to the sensor configuration requirements of the visible light camera, infrared thermal imager and lidar.

[0029] Specifically, deconstructing the 3D model of a wind turbine blade involves scanning the blade with LiDAR to obtain high-precision point cloud data, which is then imported into 3D modeling software to construct the blade model. The model is then meshed using an algorithm, and based on aerodynamic principles, the blade surface is deconstructed into different aerodynamic topological regions, such as the leading edge, trailing edge, pressure surface, and suction surface.

[0030] Generating a scanning trajectory with complete coverage involves analyzing the shape, curvature, and key inspection locations of each aerodynamic topology area, and employing a path planning algorithm to generate a scanning trajectory tailored to the characteristics of each area. For example, for leading edge areas with large curvature variations, the scanning path is intensified; for flat pressure surface areas, parallel scanning trajectories are used to ensure complete coverage of all areas.

[0031] Dynamically calculating the optimal flight altitude includes: obtaining the actual length data of the wind turbine blades, combining the field of view parameters of the visible light camera, infrared thermal imager, and lidar, and dynamically calculating the optimal flight altitude through trigonometric function relationships. For example, according to formula (1): (1) Where H is the flight height, L is the blade width, and θ is the sensor field of view angle.

[0032] Ensure that at this height the sensor can completely cover the blade detection area and obtain clear images and accurate measurement data.

[0033] Determine the flight path, which involves combining the generated scanning trajectory with the calculated optimal flight altitude to form the drone's flight path. Simultaneously, based on the route characteristics and sensor performance, appropriately configure parameters such as the visible light camera's shooting frequency, the infrared thermal imager's temperature acquisition interval, and the lidar's scanning resolution to ensure the sensors capture high-quality inspection data.

[0034] By deconstructing the blade's three-dimensional model to divide the aerodynamic topology area and generate targeted scanning trajectories, the key inspection parts of the blade can be accurately covered, blind spots can be avoided, and the defect detection rate can be improved; the optimal flight altitude can be dynamically calculated to put the sensor in the best working state, ensuring the accuracy and completeness of the acquired data; the sensor parameters and flight routes can be reasonably configured to optimize the drone inspection process, improve inspection efficiency, and reduce the risk of repeated operations and invalid data caused by improper flight altitude or unreasonable scanning trajectory. It is especially suitable for intelligent inspection of wind turbine blades with complex structures.

[0035] 103. When controlling the drone to fly along the flight route, it simultaneously collects visible light images, infrared temperature data and laser point clouds, and binds the time and space coordinate information.

[0036] Specifically, the visible light camera, infrared thermal imager, and lidar are connected to the drone's data acquisition module via a data transmission line to ensure stable signal transmission. Simultaneously, a high-precision GNSS positioning module and inertial measurement unit (IMU) are installed on the drone to obtain the drone's real-time position, attitude, and time information. As the drone flies along its flight path, a unified clock signal synchronizes the visible light camera's capture, the infrared thermal imager's temperature data, and the lidar's point cloud data acquisition, based on a preset acquisition frequency. This ensures the temporal consistency of data collected by each sensor. During data acquisition, the drone's acquisition module acquires the position, attitude, and time information output by the GNSS positioning module and IMU in real time and binds this spatiotemporal coordinate information to the simultaneously collected visible light imagery, infrared temperature data, and lidar point cloud data. This spatiotemporal coordinate information is embedded in each set of collected data using data tags, facilitating subsequent data processing and analysis.

[0037] Synchronously collecting multi-source data and binding it to time-space coordinates can avoid data misalignment caused by asynchronous collection time or position deviation, ensuring that visible light images, infrared temperature data, and laser point cloud data all correspond to the actual status of the same blade position, improving the correlation between data and facilitating multi-dimensional analysis. Data with time-space coordinate information can be efficiently managed and quickly retrieved based on time and space dimensions. When a defect is found in a blade, operation and maintenance personnel can use the time-space coordinate information to accurately locate the defect and trace the status changes of the location at different times, providing strong support for fault diagnosis and maintenance decisions, and significantly improving the intelligence level and operation and maintenance efficiency of wind turbine blade inspections.

[0038] 104. Input visible light images, infrared temperature data and laser point clouds into the defect analysis model, and through joint analysis, identify the defect type and level, locate temperature anomaly areas, and quantify geometric deformation.

[0039] Specifically, the collected visible light images are subjected to noise reduction and contrast enhancement to improve image clarity; the infrared temperature data is converted into a standard temperature matrix to remove ambient temperature interference; the laser point cloud data is denoised and smoothed, and the coordinate system is unified through point cloud registration technology to align the three types of data in the spatial dimension.

[0040] Based on a deep learning framework, a multimodal defect analysis model was built, integrating a convolutional neural network (CNN) with a Transformer architecture to extract texture features from visible light images, thermal distribution features from infrared temperature data, and 3D geometric features from laser point clouds. The model was trained in a supervised manner using a large amount of annotated historical inspection data, including defect cases of varying types and levels, optimizing its parameters to accurately identify various defect patterns.

[0041] The three types of preprocessed data are simultaneously fed into a trained defect analysis model. The model then uses feature fusion and cross-validation to jointly analyze the correlation features between the data. Based on the analysis results, the model identifies defect types (such as cracks, wear, and corrosion) and their severity (minor, moderate, and severe). Abnormal temperature values ​​in the infrared temperature data are used to locate areas of temperature anomaly. The laser point cloud data is used to quantify the degree of blade geometric deformation (such as twist angle and thickness change), ultimately producing a detailed defect diagnosis report.

[0042] By jointly analyzing multi-source data and integrating the advantages of visible light images, infrared temperature, and laser point clouds, the system overcomes the limitations of single-source diagnosis, more accurately identifies various defects, comprehensively assesses blade condition, and reduces the probability of missed and false detections. It quickly outputs defect type, level, location, and quantitative data, providing operators with intuitive and accurate decision-making information, enabling them to quickly develop targeted maintenance plans, rationally allocate maintenance resources and time, minimize turbine downtime, reduce operation and maintenance costs, and ensure stable and efficient operation of wind farms.

[0043] 105. Generate a health assessment report based on defect type and level, temperature anomaly area and geometric deformation, which includes defect location map and graded maintenance decision matching defect level.

[0044] Specifically, the defect analysis model outputs information on defect type and level, location data for temperature anomaly zones, and quantified geometric deformations. Using Geographic Information System (GIS) technology, the defect locations and temperature anomaly zone coordinates are matched and annotated with a three-dimensional model or two-dimensional plan view of the wind turbine blade. This generates a visual defect location map that intuitively displays the distribution of anomaly points on the blade.

[0045] A rule base has been developed to map defect levels to repair strategies. This prioritizes repairs based on the severity of the defect (e.g., minor, moderate, or severe), the impact of temperature anomalies on blade performance, and the impact of geometric deformation on aerodynamic performance. For example, severe defects accompanied by significant temperature anomalies and geometric deformation are designated as emergency repairs, while minor defects are scheduled for routine maintenance.

[0046] Based on the integrated data and assessment rules, a templated report generator automatically generates a health assessment report. In addition to a defect location map, the report also details repair decision recommendations for different defect types and levels, including repair time, repair method, and required parts. A comprehensive score is also provided for the overall health of the blade, providing a direct reflection of its current condition.

[0047] Through systematic health assessment reports, complex inspection data is transformed into intuitive and actionable maintenance decisions, eliminating the subjectivity and experience limitations of human judgment. This makes operations and maintenance more scientific and standardized, ensuring that wind turbine blade repair plans are precisely tailored to the actual defect conditions. Tiered maintenance decisions help operators rationally allocate maintenance resources, prioritize high-risk defects, avoid excessive or delayed repairs, reduce unnecessary downtime and maintenance costs, improve the overall operational efficiency and economic benefits of wind farms, and enhance their sustainable operation capabilities.

[0048] Furthermore, based on the above embodiment, after generating the flight route, this embodiment also includes calculating the waypoint density in response to the field of view angle parameters of the visible light camera, infrared thermal imager and lidar; dynamically adjusting the route starting point based on the real-time yaw data of the wind turbine; and optimizing the route flight speed based on the waypoint density and the route starting point and the wind speed forecast of the weather station.

[0049] Specifically, the field of view parameters of the visible light camera, infrared thermal imager, and lidar are obtained. A waypoint density calculation model is established based on the preset image overlap ratio and data acquisition accuracy requirements. Using the length and width of the wind turbine blades as boundary conditions, the distance between adjacent waypoints is calculated using trigonometric relationships to determine the waypoint density. For example, sensors with a narrow field of view require a higher waypoint density to ensure complete data coverage.

[0050] Yaw sensors installed on the wind turbines capture real-time turbine yaw data and transmit it to the drone control system. The system calculates the optimal flight path starting point based on the turbine's current orientation and the drone's takeoff position. If the turbine yaws, the drone control system promptly updates the route starting point, ensuring the drone enters the inspection area parallel to the blades and avoiding ineffective flight paths.

[0051] The system receives wind speed forecasts from weather stations and, based on waypoint density and route starting point planning, develops a flight speed optimization strategy. When wind speeds are low, the system increases flight speed to improve inspection efficiency. If wind speeds exceed a set threshold, the system reduces flight speed to ensure drone stability and data collection accuracy. Furthermore, the system automatically reduces flight speed in critical areas, such as curves and near blade edges, to ensure high-quality sensor data.

[0052] The waypoint density is calculated based on the sensor's field of view to ensure that inspection data fully covers the blade surface and avoid data omissions. The starting point of the route is dynamically adjusted to allow the drone to approach the blade at the optimal angle, improving the targeted data collection and enhancing the accuracy of defect identification. The flight speed is optimized based on meteorological information, and weather conditions are rationally utilized to improve inspection efficiency while ensuring safety. At the same time, the speed is intelligently adjusted according to the environment and route characteristics to reduce the risk of interference from strong winds, ensure equipment safety, reduce inspection interruptions caused by bad weather or improper speed, and improve the reliability of the overall inspection operation.

[0053] Furthermore, based on the above embodiments, this embodiment also includes: real-time collection of atmospheric transmittance and solar irradiance before synchronously collecting visible light images, infrared temperature data and laser point clouds; compensating for infrared temperature measurement deviation based on atmospheric transmittance; and optimizing the visible light dynamic range based on solar irradiance.

[0054] Specifically, the drone is equipped with an atmospheric transmittance sensor and a solar irradiance sensor. The former uses the principle of spectral radiation measurement to calculate the atmospheric transmittance by analyzing the attenuation degree of light of a specific wavelength; the latter uses a photoelectric detector to measure the solar radiation power received per unit area in real time, ensuring that the atmospheric transmittance and solar irradiance data are continuously obtained during the drone's takeoff and inspection flight.

[0055] A mathematical model was established to correlate atmospheric transmittance with infrared temperature measurement deviations. This model, based on Planck's law and Beer-Lambert's law, uses real-time atmospheric transmittance data to calculate the infrared radiation energy attenuation caused by atmospheric absorption and scattering. Based on these results, the raw temperature data collected by the infrared thermal imager is corrected to compensate for the measurement deviation and obtain a value closer to the true leaf temperature.

[0056] A mapping table is created to map solar irradiance to the visible light camera's dynamic range. Real-time solar irradiance data is compared with the mapping table to determine the optimal dynamic range parameters for the current lighting conditions. The drone's control system sends commands to the visible light camera, automatically adjusting parameters such as aperture, shutter speed, and ISO sensitivity. This ensures the camera captures clear, detailed visible light images under varying light intensities, avoiding overexposure or underexposure.

[0057] By compensating for infrared temperature measurement deviations through atmospheric transmittance, atmospheric interference with temperature measurement is effectively eliminated, allowing infrared temperature data to more accurately reflect the heating status of the blades. By optimizing the visible light dynamic range based on solar irradiance, the visible light image clearly presents blade surface details, providing a reliable data foundation for subsequent defect analysis and reducing data errors and misjudgment risks caused by environmental factors. Accurate data collection helps defect analysis models more accurately identify blade defects, reducing missed detections and false detections. Furthermore, optimized multi-source data can better verify each other, enhancing the credibility of inspection results, providing strong support for wind turbine blade health assessments and maintenance decisions, and ensuring the efficient operation and maintenance of wind farms.

[0058] Furthermore, based on the above embodiment, the binding of spatiotemporal coordinate information in this embodiment includes: synchronizing visible light images with the six-degree-of-freedom posture data of the drone to establish a spatial mapping benchmark; associating infrared data with the coordinates of the blade stress concentration area to locate abnormal temperature; and performing motion distortion correction on the laser point cloud to ensure the accuracy of deformation analysis.

[0059] Specifically, the drone's six-degree-of-freedom inertial measurement unit (IMU) and global navigation satellite system (GNSS) are used to acquire the drone's pose data in real time, including position (X, Y, Z coordinates) and attitude (roll, pitch, and yaw). Simultaneously, as the visible light camera captures the image, the pose data is recorded at the moment of capture. Based on the principle of collinearity equations and a feature point matching algorithm, a correspondence is established between the feature points in the visible light image and the drone's pose data. A spatial mapping model is constructed to accurately correlate the visible light image with the drone's pose data, establishing a spatial mapping benchmark.

[0060] The coordinates of the stress concentration areas on the wind turbine blades are pre-determined through finite element analysis or historical monitoring data. When collecting infrared data, the temperature data acquired by the infrared thermal imager is correlated and matched with the coordinates of the stress concentration areas on the blades. When a temperature anomaly is detected, the coordinate correspondence is used to quickly locate the specific stress concentration area, determine whether the temperature increase is caused by the stress anomaly, and accurately pinpoint the abnormal temperature.

[0061] LiDAR generates motion distortion as it moves with the drone. By collecting the LiDAR's scan timestamps and the drone's motion trajectory data, point cloud registration algorithms, such as the Iterative Closest Point (ICP) algorithm and its improved algorithms, are used to register and fuse point cloud data from adjacent scan cycles. By establishing a motion compensation model, the laser point cloud data is synchronized in time and spatially calibrated to eliminate distortion caused by drone motion. This ensures that the laser point cloud data accurately reflects the true geometry of the wind turbine blades, providing high-precision data for deformation analysis.

[0062] Establishing a spatial mapping benchmark between visible light images and drone pose data ensures the spatial accuracy of the image data, facilitating the precise location of subsequent defects. Associating infrared data with the coordinates of the blade stress concentration area quickly identifies the source of abnormal temperatures and avoids misjudgment. Motion distortion correction is performed on the laser point cloud to ensure the reliability of blade geometric deformation analysis and provide more accurate data support for blade health assessment. Accurate data processing enables better collaborative analysis of multi-source data. For example, by combining the surface features of visible light images, temperature anomalies in infrared data, and geometric deformations of laser point clouds, the type, extent, and location of blade defects can be more comprehensively and accurately identified, reducing missed detections and misdiagnoses. This provides a scientific basis for maintenance decisions on wind turbine blades and effectively improves the efficiency and safety of wind farm operations.

[0063] Furthermore, this embodiment uses joint analysis to identify defect types and levels, locate temperature anomalies, and quantify geometric deformation. This includes: fusing visible light images with infrared data to detect temperature anomalies associated with surface defects. When geometric deformation exceeds a threshold, the infrared data corresponding to the time and space coordinates is retrieved. The infrared data is used to correlate the deformed areas with the defect distribution in the visible light image to create a damage evolution map. Based on this evolution map, the root cause of material fatigue or impact damage is identified. Laser point clouds and visible light images are combined to locate apparent damage in geometrically deformed areas. Multimodal outputs are integrated to generate a defect-deformation-temperature correlation assessment.

[0064] Specifically, the visible light image and infrared data are spatially registered to correspond to the same blade region. Using an image fusion algorithm, the texture details of the visible light image are combined with the temperature information from the infrared data. This allows the system to detect surface defects (such as cracks and wear) while also simultaneously identifying temperature anomalies associated with these defects, marking the location and temperature characteristics of the abnormal area.

[0065] Real-time monitoring of blade geometric deformation parameters (such as twist angle and thickness change) obtained from laser point cloud data. When the deformation value exceeds the preset threshold, the infrared data at the corresponding moment is traced back based on the time-space coordinate binding information, and the temperature changes in the area before and after the deformation are analyzed to determine whether the temperature anomaly is related to the deformation.

[0066] Based on the correspondence between temperature anomalies and deformation areas in infrared data, combined with the defect distribution in visible light images, the blade damage development process was analyzed in a time series. Using machine learning algorithms, the data was clustered and analyzed to construct a damage evolution map, showing the dynamic process of defect initiation, development, and deterioration, as well as the characteristics of each stage.

[0067] Using damage evolution maps, combined with material mechanical properties and wind turbine operating conditions, we analyze the inherent connections between defects, deformation, and temperature changes. If an abnormal temperature area gradually expands over time and is accompanied by crack growth, combined with stress concentration analysis, we determine whether it is material fatigue damage. If there is localized high temperature and sudden deformation, we combine historical meteorological data and operating records to identify whether it is caused by external impact and thus determine the root cause.

[0068] The 3D model constructed from the laser point cloud is integrated with the visible light image to precisely locate the apparent damage location and morphology of geometrically deformed areas in three-dimensional space. Finally, the defect information from the visible light image, the deformation data from the laser point cloud, and the temperature characteristics of the infrared data are integrated to generate a defect-deformation-temperature correlation assessment report, visually demonstrating the interrelationship between the three and the degree of impact on leaf health.

[0069] Multimodal data fusion analysis overcomes the diagnostic limitations of single data points and reveals the nature of blade damage from multiple dimensions. It not only accurately identifies surface defects but also deeply explores the inherent connections between defects, deformation, and temperature anomalies, reducing missed detections and misjudgments and improving diagnostic accuracy. Damage evolution maps and root cause identification provide maintenance personnel with a clear understanding of blade damage development and the root causes of failures. Correlation assessment reports provide a more comprehensive understanding of blade health, enabling them to develop more targeted maintenance strategies, rationally allocate maintenance resources and time, reduce maintenance costs, and ensure the safe and efficient operation of wind farms.

[0070] Furthermore, generating a health assessment report in this embodiment includes: analyzing the correlation assessment results to define a three-dimensional damage index; matching a gradient maintenance plan based on the damage index threshold of the three-dimensional damage index; generating an emergency shutdown alarm when the three-dimensional damage index exceeds the limit; mapping the three-dimensional damage index to a three-dimensional blade model to generate a heat map; correlating the heat map with historical damage indices to create a degradation trajectory curve; and predicting the remaining service life based on the degradation trajectory curve.

[0071] Specifically, a quantitative analysis of the defect-deformation-temperature correlation assessment report was conducted, comprehensively considering factors such as defect type, area, and depth; deformation degree and range; and the magnitude and location of temperature anomalies. A mathematical model was then established. Each factor was assigned a corresponding weight, and a three-dimensional damage index was calculated through weighted summation to comprehensively characterize the blade's damage status.

[0072] Multiple damage index thresholds of varying levels are pre-defined, such as mild, moderate, and severe. The calculated three-dimensional damage index is compared with each threshold, and a gradient maintenance plan is assigned to each threshold. For example, if the index falls below the mild damage threshold, conventional maintenance is implemented; if it falls between mild and moderate, a partial repair plan is implemented; and if it exceeds the severe threshold, a comprehensive overhaul is initiated.

[0073] An upper limit value of the three-dimensional damage index for emergency shutdown is set. When the calculated three-dimensional damage index exceeds this limit, the system immediately generates an emergency shutdown alarm message and sends it to the operation and maintenance personnel through various means such as text messages and emails. At the same time, the alarm signal is transmitted to the wind turbine control system to trigger the shutdown protection mechanism to prevent further deterioration of blade damage.

[0074] The 3D damage index is mapped onto the 3D model of the wind turbine blade, and different colors are assigned according to the damage index, generating an intuitive heat map that clearly shows the damage distribution of various parts of the blade. Simultaneously, historical 3D damage index data is collected and linked to the heat map in chronological order to plot the blade damage degradation trajectory curve, visually demonstrating the damage development trend over time.

[0075] Machine learning algorithms, such as neural networks and support vector regression, are used to analyze and model the degradation trajectory curve. Using historical damage data, operating condition data (such as speed and load), and environmental data (such as wind speed and temperature) as input, the model is trained to predict future damage trends in the blades and, in turn, estimate the remaining useful life of the blades, providing a forward-looking reference for operation and maintenance decisions.

[0076] By quantifying blade damage status through a three-dimensional damage index, maintenance plans can be automatically matched and emergency alarms can be intelligently triggered, reducing manual judgment errors. This makes operation and maintenance management more scientific and efficient, and enhances the intelligence and automation of operation and maintenance work. Timely emergency shutdown alarms effectively prevent serious failures caused by excessive blade damage, ensuring the safety of wind turbine equipment. Based on the remaining service life prediction, operation and maintenance personnel can plan maintenance and replacement plans in advance, rationally allocate resources, reduce unplanned downtime, lower operation and maintenance costs, and equipment replacement costs, thereby improving the economic benefits and operational reliability of wind farms.

[0077] Based on the same general inventive concept, the present invention also protects a drone-based intelligent inspection system for wind turbine blades. The drone-based intelligent inspection system for wind turbine blades described below and the drone-based intelligent inspection method for wind turbine blades described above can be referenced to each other.

[0078] Figure 2 Schematic diagram of the structure of the wind turbine blade intelligent inspection system based on drones provided in this embodiment.

[0079] like Figure 2 As shown, this embodiment provides a wind turbine blade intelligent inspection system based on a drone, including: Integration module 201, used to integrate a visible light camera, an infrared thermal imager, and a laser radar on a UAV; A response module 202 is used to generate a flight path covering the entire surface based on the three-dimensional model of the wind turbine blade, and respond to the sensor configuration requirements of the visible light camera, infrared thermal imager and lidar; The acquisition module 203 is used to control the UAV to fly along the flight route, synchronously collect visible light images, infrared temperature data and laser point clouds, and bind the time and space coordinate information; Identification module 204 is used to input visible light images, infrared temperature data and laser point cloud into the defect analysis model, identify defect types and levels, locate temperature anomaly areas, and quantify geometric deformation through joint analysis; The generation module 205 is used to generate a health assessment report based on the defect type and level, temperature anomaly area and geometric deformation, which includes a defect location map and a graded maintenance decision matching the defect level.

[0080] Figure 3 Schematic diagram of the structure of the electronic device provided in this embodiment.

[0081] like Figure 3 As shown, the electronic device may include: a processor (processor) 310, a communication interface (Communications Interface) 320, a memory (memory) 330 and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call the logic instructions in the memory 330 to execute a drone-based intelligent inspection method for wind turbine blades, which includes: integrating a visible light camera, an infrared thermal imager and a lidar on the drone; generating a flight route covering the entire surface based on a three-dimensional model of the wind turbine blade, responding to the sensor configuration requirements of the visible light camera, infrared thermal imager and lidar; controlling the drone to fly along the flight route, synchronously collecting visible light images, infrared temperature data and laser point clouds, and binding time and space coordinate information; inputting the visible light images, the infrared temperature data and the laser point clouds into a defect analysis model, identifying the defect type and level, locating the temperature anomaly area, and quantifying the geometric deformation through joint analysis; generating a health assessment report based on the defect type and level, the temperature anomaly area and the geometric deformation, which includes a defect location map and a graded maintenance decision matching the defect level.

[0082] Furthermore, the logic instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0083] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the drone-based intelligent inspection method for wind turbine blades provided by the above methods, the method including: integrating a visible light camera, an infrared thermal imager and a lidar on a drone; generating a flight route covering the entire surface based on a three-dimensional model of the wind turbine blade, responding to the sensor configuration requirements of the visible light camera, infrared thermal imager and lidar; controlling the drone to fly along the flight route, synchronously collecting visible light images, infrared temperature data and laser point clouds, and binding time and space coordinate information; inputting the visible light images, the infrared temperature data and the laser point clouds into a defect analysis model, identifying the defect type and level, locating the temperature anomaly area, and quantifying the geometric deformation through joint analysis; generating a health assessment report based on the defect type and level, the temperature anomaly area and the geometric deformation, which includes a defect location map and a graded maintenance decision matching the defect level.

[0084] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the drone-based intelligent inspection method for wind turbine blades provided by the above-mentioned methods, the method comprising: integrating a visible light camera, an infrared thermal imager and a lidar on a drone; generating a flight route covering the entire surface based on a three-dimensional model of the wind turbine blade, responding to the sensor configuration requirements of the visible light camera, infrared thermal imager and lidar; controlling the drone to fly along the flight route, synchronously collecting visible light images, infrared temperature data and laser point clouds, and binding time and space coordinate information; inputting the visible light images, the infrared temperature data and the laser point clouds into a defect analysis model, identifying the defect type and level, locating the temperature anomaly area, and quantifying the geometric deformation through joint analysis; generating a health assessment report based on the defect type and level, the temperature anomaly area and the geometric deformation, which includes a defect location map and a graded maintenance decision matching the defect level.

[0085] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0086] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A wind turbine blade intelligent inspection method based on drone, characterized in that: include: Integrate visible light cameras, infrared thermal imagers and lidar on drones; Generate a flight path covering the entire surface based on the three-dimensional model of the wind turbine blade, responding to the sensor configuration requirements of the visible light camera, infrared thermal imager and lidar; When controlling the UAV to fly along the flight route, the visible light image, infrared temperature data and laser point cloud are synchronously collected and bound to the time and space coordinate information; Inputting the visible light image, the infrared temperature data and the laser point cloud into a defect analysis model, identifying the defect type and grade, locating the temperature anomaly area, and quantifying the geometric deformation through joint analysis; Based on the defect type and level, the temperature anomaly area and the geometric deformation, a health assessment report is generated, which includes a defect location map and a graded maintenance decision matching the defect level.

2. The intelligent inspection method for wind turbine blades based on drones according to claim 1 is characterized in that: The generating of a flight route covering the entire surface based on the three-dimensional model of the wind turbine blade includes: Deconstruct the 3D model of the wind turbine blade to divide the aerodynamic topology areas; Generating a scanning trajectory with no blind spot coverage based on the features of each of the aerodynamic topological regions; Dynamically calculating the optimal flight altitude based on the blade length and the field of view angles of the visible light camera, infrared thermal imager, and lidar; The scanning trajectory and the optimal flight altitude are used as the flight route.

3. The intelligent inspection method for wind turbine blades based on drones according to claim 2 is characterized in that: Also includes: Calculating waypoint density in response to field of view angle parameters of the visible light camera, infrared thermal imager, and lidar; Dynamically adjust the route starting point based on the real-time yaw data of the wind turbine; Based on the waypoint density and the route starting point, the route flight speed is optimized according to the weather station wind speed forecast.

4. The intelligent inspection method for wind turbine blades based on drones according to claim 1 is characterized in that: Before the synchronous collection of visible light images, infrared temperature data and laser point clouds, the method further includes: Real-time collection of atmospheric transmittance and solar irradiance; Compensating for infrared temperature measurement deviation based on the atmospheric transmittance; The visible light dynamic range is optimized according to the solar irradiance.

5. The intelligent inspection method for wind turbine blades based on drone according to claim 1 is characterized in that: The binding space-time coordinate information includes: Synchronize visible light images and UAV six-degree-of-freedom pose data to establish a spatial mapping benchmark; Correlate infrared data with the coordinates of blade stress concentration areas to locate abnormal temperatures; Motion distortion correction is performed on the laser point cloud to ensure the accuracy of deformation analysis.

6. The intelligent inspection method for wind turbine blades based on drones according to claim 1 is characterized in that: The joint analysis to identify defect types and levels, locate temperature anomaly areas, and quantify geometric deformation includes: Fusion of visible light images and infrared data to detect temperature anomalies associated with surface defects; Combine laser point cloud and visible light image to locate apparent damage in geometric deformation areas; Integrate multimodal outputs to generate defect-deformation-temperature correlation assessments.

7. The intelligent inspection method for wind turbine blades based on drones according to claim 6 is characterized in that: The generating of the health assessment report includes: Analyzing the correlation assessment results and defining a three-dimensional damage index; matching a gradient repair scheme according to a damage index threshold of the three-dimensional damage index; When the three-dimensional damage index exceeds a limit, an emergency shutdown alarm is generated.

8. The intelligent inspection method for wind turbine blades based on drones according to claim 7 is characterized in that: Also includes: Mapping the three-dimensional damage index to a three-dimensional blade model to generate a thermal map; Correlating the thermal map with a historical damage index to draw a degradation trajectory curve; The remaining service life is predicted based on the degradation trajectory curve.

9. The intelligent inspection method for wind turbine blades based on drones according to claim 6 is characterized in that: The method of fusing visible light images and infrared data to detect temperature anomalies associated with surface defects further includes: When the geometric deformation exceeds the threshold, the infrared data of the corresponding space-time coordinates are retrieved; Establishing a damage evolution map by correlating the deformation area with the visible light image defect distribution using the infrared data; Based on the evolution map, the root cause category of material fatigue or impact damage is identified.

10. An intelligent inspection system for wind turbine blades based on drones, characterized in that: include: Integration module for integrating visible light camera, infrared thermal imager and lidar on drone; a response module for generating a flight path covering the entire surface based on the three-dimensional model of the wind turbine blade, responding to the sensor configuration requirements of the visible light camera, infrared thermal imager, and lidar; An acquisition module, configured to control the UAV to fly along the flight route, synchronously acquire visible light images, infrared temperature data, and laser point clouds, and bind the time and space coordinate information; an identification module, configured to input the visible light image, the infrared temperature data, and the laser point cloud into a defect analysis model, and identify the defect type and level, locate the temperature anomaly area, and quantify the geometric deformation through joint analysis; A generation module is used to generate a health assessment report based on the defect type and level, the temperature anomaly area and the geometric deformation, which includes a defect location map and a graded maintenance decision matching the defect level.

Images