Fan blade inspection method, system and equipment based on unmanned aerial vehicle and medium

Abstract

The invention relates to the technical field of fan defect detection, in particular to a fan blade inspection method based on an unmanned aerial vehicle, which comprises the following steps: acquiring static basic data of a fan; constructing a blade reference model based on the static basic data of the fan, and collecting dynamic vibration response data according to a preset short-time dynamic vibration excitation rule; identifying a blade high-risk area according to the dynamic vibration response data, and generating a layered surrounding route in combination with a blade reference model; the unmanned aerial vehicle is controlled to collect blade multi-angle images according to the layered surrounding route, the blade multi-angle images are processed to recognize surface defects, and visual confidence is generated; and mapping the dynamic vibration response data to a blade reference model by using a space-time alignment model, extracting power spectral density characteristics of corresponding positions, calculating a damage index by fusing visual confidence, and outputting an inspection report. The objective of the invention is to solve the technical problem that blade internal damage is difficult to identify in a static inspection mode.

Description

Technical Field

[0001] The present invention relates to the technical field of wind turbine defect detection, and in particular to a wind turbine blade inspection method, system, equipment and medium based on an unmanned aerial vehicle (UAV). Background Art

[0002] Currently, drone-based wind turbine blade inspection technology primarily monitors blade condition through two modes: dynamic inspection and static inspection. In dynamic inspection mode, the drone, equipped with an onboard computing unit and attitude sensing equipment, uses artificial intelligence algorithms to analyze the turbine's operating posture in real time, dynamically planning a flight path around the turbine. This allows for rapid, non-stop blade image acquisition while the turbine maintains normal power generation. Static inspection requires the turbine to be shut down and locked, allowing the drone to approach the blade surface to capture high-resolution images and identify subtle defects using computer vision technology.

[0003] Among the above-mentioned existing technologies, although dynamic inspection can obtain blade operating status information without affecting power generation, it is difficult to accurately identify subtle defects due to the limitations of flight distance and motion blur interference; although static inspection has the ability to detect surface defects with high precision, it cannot associate and identify internal structural damage. Operation and maintenance personnel make maintenance decisions based on the inspection results under the static inspection mode, which easily leads to the neglect of internal risks of the blades, resulting in the lack of ability to predict the evolution trend of damage in operation and maintenance decisions. Summary of the Invention

[0004] In order to solve the technical problem of difficulty in identifying internal blade damage in static inspection mode, the present invention provides a wind turbine blade inspection method, system, equipment and medium based on drones. The technical solutions adopted are as follows: The technical solution of the first aspect of the present invention provides a wind turbine blade inspection method based on a drone, the method comprising: Obtain static basic data of the wind turbine, including spatial positioning data, structural dimension data, and blade material property data; Build a blade benchmark model based on the static basic data of the wind turbine, and collect dynamic vibration response data according to the preset short-term dynamic vibration excitation rules; Identify high-risk areas of the blade based on dynamic vibration response data, and generate a layered circumferential route including a basic route and encrypted waypoints in the high-risk area in combination with the blade benchmark model; In the static inspection mode, the drone is controlled to collect multi-angle images of the blade according to a layered circling route, and the multi-angle images of the blade are processed to identify surface defects and generate visual confidence. The dynamic vibration response data is mapped to the blade benchmark model using a spatiotemporal alignment model, the power spectrum density features at the corresponding positions are extracted, and the damage index is calculated by integrating the visual confidence score. The damage level is divided based on the damage index and an inspection report is output.

[0005] Furthermore, a blade benchmark model is constructed based on the static basic data of the wind turbine, and dynamic vibration response data is collected according to the preset short-term dynamic vibration excitation rules, including: Plan a layered circling modeling route based on structural dimension data, and control the drone to collect multi-angle images of the blade along the layered circling modeling route; The blade benchmark model is generated by fusing multi-angle images through motion recovery algorithm; The fan is controlled to execute the preset short-time dynamic vibration excitation rules, starting from the shutdown state to the preset speed and running stably for the preset time, and then gradually shutting down, forming a vibration excitation process including startup, steady state and shutdown stages; During the vibration excitation process, vibration signals of the blade start-stop transient process are collected, and the SCADA system time stamp is synchronized to form dynamic vibration response data.

[0006] Furthermore, the high-risk areas of the blade are identified based on the dynamic vibration response data, and a layered circling route including a basic route and encrypted waypoints in the high-risk areas is generated in combination with the blade benchmark model, including: Perform spectrum analysis on dynamic vibration response data to extract characteristic frequencies of different blade regions and the power spectrum density at corresponding frequencies; The power spectrum density of each area is compared with the power spectrum density threshold of the healthy area of ​​the leaf, and the area exceeding the power spectrum density threshold is marked as a high-risk area of ​​the leaf; Generate a basic flight path covering the windward side, leeward side and leading edge of the blade based on the surface curvature and dimensional characteristics of the blade benchmark model; For high-risk areas of blades, the waypoint density is increased on the basis of the basic route to form a layered circular route with encrypted waypoints.

[0007] Furthermore, in the static inspection mode, the UAV is controlled to collect multi-angle images of the blade according to the layered circling route, and the multi-angle images of the blade are processed to identify surface defects and generate visual confidence, including: In static inspection mode, the drone is controlled to collect multi-angle images of the windward side, leeward side and leading edge of the blade along a layered circular route, and the spatial position information corresponding to each image is simultaneously recorded; Based on the inherent texture period and reflectivity characteristics in the blade material characteristic data, the areas consistent with the inherent texture of the blade are filtered out, and the abnormal texture areas are retained; Using multi-scale filtering to process the abnormal texture area, extract the defect contour and quantify the defect geometric features; The visual confidence is calculated based on the contrast between the edge sharpness of the defect area and the background area, the edge sharpness of the defect outline, and the geometric characteristics of the defect.

[0008] Furthermore, the dynamic vibration response data is mapped to the blade benchmark model using a spatiotemporal alignment model, including: Constructing the blade space rotation matrix based on the blade attitude parameters during dynamic vibration response data acquisition; According to the spatial coordinates of the blade reference model and the installation position of the vibration sensor, the dynamic vibration response data is mapped to the spatial coordinates of the blade reference model through the blade space rotation matrix; The dynamic vibration response data and the position information of the blade benchmark model are synchronized based on the timestamp to achieve the correlation and alignment between the dynamic vibration response data and the blade spatial position.

[0009] Furthermore, the power spectrum density features of the corresponding positions are extracted and integrated with the visual confidence to calculate the damage index, including: In the blade characteristic frequency range, the power spectrum density integral values ​​of the damaged area corresponding to the defect position on the blade surface and the preset healthy area are calculated respectively to obtain the vibration energy ratio; Adjust the weight coefficient of visual confidence according to the type of surface defects; The visual confidence and vibration energy ratio are weighted and fused according to the adjusted weight coefficient to generate a damage index that reflects the degree of surface and internal damage.

[0010] Furthermore, the damage index can be expressed as:

[0011] Where, represents the damage index; represents the visual confidence weight factor; represents the visual confidence correction term based on the defect area; Indicates the defect area; Indicates the preset correction factor, which is used to adjust the sensitivity of the defect area to the visual confidence level; represents the blade characteristic frequency; represents the characteristic frequency bandwidth; represents the power spectral density function of the damaged area; Represents the power spectral density function of the healthy area.

[0012] The technical solution of the second aspect of the present invention provides a wind turbine blade inspection system based on a drone, which adopts the wind turbine blade inspection method based on a drone described in the technical solution of the first aspect of the present invention, and the system includes: A data acquisition module configured to acquire static basic data of the wind turbine, including spatial positioning data, structural dimension data, and blade material characteristic data; A short-time dynamic vibration excitation module is configured to construct a blade benchmark model based on static basic data of the wind turbine and collect dynamic vibration response data according to preset short-time dynamic vibration excitation rules; a route planning module configured to identify high-risk areas of the blade based on the dynamic vibration response data and generate a layered circumferential route including a basic route and encrypted waypoints in the high-risk areas in combination with the blade benchmark model; a visual inspection module configured to, in a static inspection mode, control the drone to collect multi-angle images of the blade according to a layered circling route, process the multi-angle images of the blade to identify surface defects and generate a visual confidence score; a fusion diagnosis module configured to map the dynamic vibration response data to a blade benchmark model using a spatiotemporal alignment model, extract the power spectral density features at the corresponding positions, and fuse the visual confidence to calculate the damage index; The decision output module is configured to classify damage levels based on the damage index and output an inspection report.

[0013] The technical solution of the third aspect of the present invention provides an electronic device, which includes: a processor and a memory communicatively connected to the processor; wherein the memory stores instructions that can be executed by the processor, and the instructions are executed by the processor so that the processor can execute the steps of the drone-based wind blade inspection method described in the technical solution of the first aspect of the present invention.

[0014] The technical solution of the fourth aspect of the present invention provides a computer-readable storage medium, on which is stored a program for implementing a drone-based wind blade inspection method. The program for implementing a drone-based wind blade inspection method is executed by a processor to implement the steps of the drone-based wind blade inspection method described in the technical solution of the first aspect of the present invention.

[0015] The present invention has the following beneficial effects: The wind turbine blade inspection method based on drones provided by the present invention integrates short-time dynamic vibration excitation and static visual inspection in a time-series manner, thereby breaking through the technical bottleneck of internal damage identification while retaining the high-precision advantage of static inspection. First, a blade benchmark model is constructed based on the static basic data of the wind turbine, and dynamic vibration response data is collected by pre-setting short-time dynamic vibration excitation rules to accurately locate the high-risk areas of the blades; then, a layered circular route is generated in combination with the benchmark model, and multi-angle images are collected for encrypted waypoints in high-risk areas in the static inspection mode to achieve surface defect recognition and visual confidence quantification; finally, the dynamic vibration data is mapped to the benchmark model using a spatiotemporal alignment model, and the power spectrum density features and visual confidence are integrated to calculate the damage index. This method establishes a correlation diagnosis mechanism between surface defects and internal damage, so that static inspections have the ability to perceive internal structural damage without modifying wind turbine equipment and without increasing significant power generation losses, which helps to improve the comprehensiveness of operation and maintenance decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only 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.

[0017] Figure 1 A flowchart of a wind turbine blade inspection method based on a drone according to an embodiment of the present invention; Figure 2 A schematic structural diagram of a wind turbine blade inspection system based on a drone according to one embodiment of the present invention; Figure 3 A panoramic schematic diagram of the leading edge of a blade provided by one embodiment of the present invention; Figure 4 A panoramic schematic diagram of the windward side of a blade provided by one embodiment of the present invention; Figure 5 A panoramic schematic diagram of the leeward side of a blade provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0018] To further illustrate the technical means and effects employed by the present invention to achieve the intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail a method, system, device, and medium for inspecting wind turbine blades using a drone, as well as its specific implementation, structure, features, and effects. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0019] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0020] The following describes in detail a specific solution of a wind turbine blade inspection method, system, equipment and medium based on a drone provided by the present invention in conjunction with the accompanying drawings.

[0021] See also Figure 1 , which shows a method flow chart of a wind turbine blade inspection method based on a drone provided by one embodiment of the present invention, the method comprising: Step S100: Obtain static basic data of the wind turbine, including spatial positioning data, structural dimension data and blade material characteristic data; wherein the spatial positioning data includes the longitude and latitude coordinates of the wind turbine center and the elevation of the wind turbine hub center; the structural dimension data includes blade length, curvature radius, nacelle size and tower diameter; the blade material characteristic data includes blade surface reflectivity and inherent texture period.

[0022] In some embodiments, the wind turbine hub center elevation It can be obtained by RTK elevation measurement combined with tower height conversion. ,in is the ground elevation measured by RTK, is the tower height, i.e. the vertical distance from the ground to the hub center; the structural dimension data is obtained through layered scanning by drones; the blade length is the straight-line distance from the blade root to the blade tip; the curvature radius Through the calculation of blade surface point cloud fitting surface, the cabin size and tower diameter are obtained by extracting the point cloud boundary, providing geometric parameters for the construction of the blade benchmark model; in the blade material characteristic data, the surface reflectivity can be obtained by band collection and radiation calibration processing; the inherent texture period By performing Fourier transform on the high-definition image of the defect-free area of ​​the blade , extract the spatial frequency corresponding to the maximum energy in the Fourier spectrum , which is used to distinguish intrinsic texture from defect features, and can be expressed as:

[0023] in, is the intrinsic texture period, i.e., the minimum physical size of the repeating texture unit on the blade surface; It is a two-dimensional Fourier transform operator used to convert spatial domain images into frequency domain; It is the grayscale image of the defect-free area; Step S200: constructing a blade benchmark model based on static basic data of the wind turbine, and collecting dynamic vibration response data according to preset short-time dynamic vibration excitation rules; Step S200 specifically includes: Step S210: planning a layered circling modeling route based on the structural dimension data, and controlling the UAV to collect multi-angle images of the blade along the layered circling modeling route; In some embodiments, the blade is divided into a root section, a middle section, and a tip section in a ratio of 3:4:3 along the length direction of the blade, and each section is provided with an independent wrapping layer; the wrapping radius varies with the curvature of the blade. Dynamic adjustment is performed to ensure a constant distance between the drone and the blade surface; the course overlap rate is ≥80% and the lateral overlap rate is ≥70% to ensure image stitching continuity; In some embodiments, see Figures 3 to 5 As shown, a drone equipped with a camera is used to fly along a planned route to collect multi-angle images of the windward side, leeward side, and leading edge of the blade, and the GPS coordinates and camera attitude of each image are simultaneously recorded; Step S220: Generate a blade reference model by fusing multi-angle images using a structure-from-motion algorithm. Specifically, the ORB algorithm is used to extract feature points such as blade edge inflection points, bolt holes, and leading edge ridges in each image, and the RANSAC algorithm is used to remove mismatched point pairs to improve the matching degree of homonymous points. Then, using the feature point matching results as input, the camera rotation matrix and translation vector are optimized using the bundle adjustment method. The objective function can be expressed as:

[0024] Where, is the camera rotation matrix; is the translation vector, describing the camera position; is the perspective projection function, used to transform the three-dimensional point Projection to 2D image points ; is the image Laplace operator, that is, the image at the feature point The second-order differential at can reflect the intensity of texture changes; is the regularization coefficient; In some embodiments, based on the optimized camera pose, a three-dimensional point cloud of the blade is generated through triangulation calculation; the texture information of the multi-angle image is fitted to the surface of the triangular mesh model using a UV parameterization method to generate a blade reference model with high-definition texture; Step S230: Controlling the fan to execute a preset short-time dynamic vibration excitation rule, starting from a shutdown state to a preset speed and running stably for a preset time, and then gradually shutting down, forming a vibration excitation process including startup, steady state, and shutdown stages; In some embodiments, the preset short-term dynamic vibration excitation rules include: During the startup phase, the fan is controlled to start at a preset acceleration from the shutdown state, gradually increasing the speed to the rated speed, covering the transient response from low speed to rated speed; In the steady-state stage, the rated speed is maintained for a preset period of time to capture the steady-state vibration characteristics of the blade under the design load; During the shutdown phase, the speed is gradually reduced to a stop according to the preset deceleration rate, covering the transition process from rated speed to standstill.

[0025] The entire process must ensure that it includes the complete stages of startup, steady state, and shutdown to avoid power generation loss caused by long-term operation.

[0026] In some embodiments, the rate of change of the rotational speed satisfies a preset acceleration constant. ; Step S240: During the vibration excitation process, vibration signals of the blades including the start-stop transient process are collected and synchronized with the SCADA system timestamp to form dynamic vibration response data. Specifically, a vibration sensor pre-installed in the wind turbine nacelle, such as an IEPE acceleration sensor installed on the main shaft bearing seat, is used to directly collect structural vibration signals transmitted from the blades to the nacelle. In some embodiments, the real-time rotation speed and timestamp are obtained through the SCADA system, the vibration signal is bound to the timestamp, and the vibration time domain signal and the start-up, steady-state, and shutdown status labels are stored to form dynamic vibration response data; In some embodiments, the vibration data is time-stamped with the SCADA system , real-time speed , blade rotation angle Binding to form dynamic vibration response data:

[0027] Where, is the vibration sensor data collection time; 、 、 It is three-cycle vibration acceleration; is the yaw angle, usually the direction angle of the wind turbine nacelle relative to due north; is the pitch angle, i.e. the rotation angle of the blade around its own axis; This step generates a blade benchmark model through layered surround modeling, providing an accurate three-dimensional space and texture benchmark for static inspection, ensuring the accuracy of surface defect positioning and feature extraction; through preset short-term dynamic vibration excitation and response data collection, the structural vibration characteristics of the blade under all working conditions are captured without affecting the normal power generation of the wind turbine, providing a dynamic response basis for internal damage identification.

[0028] Step S300: identifying high-risk areas of the blade based on the dynamic vibration response data, and generating a layered circling route including a basic route and encrypted waypoints in the high-risk areas in combination with the blade benchmark model; Step S300 specifically includes: Step S310: performing spectrum analysis on the dynamic vibration response data to extract characteristic frequencies of different regions of the blade and the power spectrum density at the corresponding frequencies; In some embodiments, based on the time domain vibration acceleration signal collected in step S240 , through cubic spline interpolation, the non-uniform sampling signal is unified to a fixed sampling rate , and the Hanning window function is used to suppress spectrum leakage; In some embodiments, the root of the blade , Ye Zhong , leaf tip Leaf area , calculate the PSD region by region:

[0029] Where, Leaf area In frequency The power spectral density of For the region Neidi The vibration acceleration sampling value of the sampling point, , is the sampling time; is the Hanning window function; To analyze the frequency; In some embodiments, the blade root area : Positioning within the 80~120HZ frequency band The maximum peak value, corresponding to the characteristic frequency ; Tip area : Positioning within the 30~50HZ frequency band The maximum peak value, corresponding to the characteristic frequency ; Yezhong District :Based on the modal data of blades of the same model, calibrate the characteristic frequency band and extract the peak frequency ; Step S320: Compare the power spectrum density of each area with the power spectrum density threshold of the blade's preset healthy area, and mark the area exceeding the power spectrum density threshold as a high-risk area of ​​the blade; call the historical health database of the same model blade to extract the benchmark PSD of the corresponding area: Blade root benchmark: , is the number of healthy samples, and the median is taken to avoid interference from outliers; for the leaf tip and leaf center benchmarks, the same calculation is performed 、 .

[0030] In some embodiments, a regional damage index is defined to quantify the degree of vibration energy abnormality, which can be expressed as:

[0031] Where, is the regional damage indicator; is the characteristic frequency band boundary; To extract the maximum PSD value in the characteristic frequency band; based on the threshold ,like , then mark the area High-risk areas and output their spatial coordinate sets; Step S330: Based on the surface curvature and dimensional characteristics of the blade reference model, a basic route covering the windward side, leeward side and leading edge of the blade is generated; specifically, based on the three-dimensional reference model of the blade constructed in step S220, combined with the blade length, curvature and other structural dimensions obtained in step S100, the blade is first divided into a root section, a middle section and a tip section along the length direction. Subsequently, the safe distance between the drone and the blade is dynamically set according to the maximum curvature of the blade surface. It should be noted that when planning waypoints along the length direction of the blade, the interval between adjacent waypoints needs to be controlled to ensure that the image stitching overlap rate is ≥80%. Finally, by adjusting the pitch angle and yaw angle of the drone, a circumferential path is generated around the blade, and the surface direction information of the reference model is used to ensure that the route covers the windward side, leeward side and leading edge of the blade.

[0032] Step S340: For the high-risk area of ​​the blade, the waypoint density is increased on the basis of the basic route to form a layered circular route with encrypted waypoints; In some embodiments, based on the high-risk areas marked in step S320, the nearest waypoint of each high-risk area is first located on the basic route, and a spatial search algorithm can be used to match the waypoints to ensure that the encrypted area is connected to the basic route; In some embodiments, the waypoint density is increased in a local area around the waypoint, and the waypoints are arranged more densely near high-risk areas on the blade surface, so that the image resolution is improved to a higher accuracy to adapt to the needs of fine defect detection such as microcracks; In some embodiments, a path optimization algorithm is used, such as a traveling salesman problem solution, to plan the sequence of waypoints, constraining the total number of waypoints and the turning angles between adjacent waypoints; In summary, step S300 quantifies the spectrum characteristics Accurately locate the abnormal vibration areas of the blade in the full span, generate a layered basic route based on the geometric parameters of the blade benchmark model, and implement waypoint encryption in high-risk areas. This improves the micro-crack recognition rate without significantly increasing the inspection time. At the same time, the TSP algorithm is used to optimize the route to ensure the flight safety and trajectory smoothness of the UAV, providing high-confidence images and spatial correlation data for subsequent defect quantitative analysis, and ultimately realizing the detection logic of dynamic vibration warning and static visual verification.

[0033] Step S400: In a static inspection mode, the UAV is controlled to collect multi-angle images of the blade according to a layered circling route, and the multi-angle images of the blade are processed to identify surface defects and generate visual confidence. Step S400 specifically includes: Step S410: In the static inspection mode, the drone is controlled to collect multi-angle images of the windward side, leeward side, and leading edge of the blade along a layered circular route, and the spatial position information corresponding to each image is simultaneously recorded; In some implementations, the drone is controlled to fly along a layered, circular trajectory, with the high-resolution camera dynamically adjusting exposure parameters to accommodate variations in the leaf surface reflectivity. Each image is simultaneously recorded with its acquisition position and pose data, including the drone's rotation matrix and the camera's optical center world coordinates, ensuring a strict binding between the image and its spatial position.

[0034] In some embodiments, according to the blade surface reflectivity data of step S100, the camera exposure time, gain and other parameters are dynamically adjusted to avoid overexposure or underexposure to ensure clear image details.

[0035] Step S420: Based on the inherent texture period and reflectivity characteristics in the blade material characteristic data, filter out the area consistent with the inherent texture of the blade and retain the abnormal texture area; In some embodiments, the inherent texture period and direction of the blade in step S100 are used to construct a Gabor filter. Furthermore, a convolution operation is performed to suppress background areas that match the texture period. Abnormal areas where the reflectivity difference exceeds a set threshold are extracted, and a binary mask is generated to separate potential defects. Step S430: using multi-scale filtering to process the abnormal texture area, extracting the defect contour and quantifying the defect geometric features; In some embodiments, a Gaussian pyramid is constructed to implement multi-scale analysis, combined with a Laplacian edge enhancement operator to enhance weak defect edges. A direction-sensitive Canny algorithm is used to extract continuous contours and quantify the physical size of the defect: length is calculated using the contour polygon perimeter, width is calculated using the projected boundary difference, and actual damage area is converted from pixel area.

[0036] Step S440: Calculating visual confidence based on the edge sharpness of the defect area and the contrast between the background area and the edge sharpness of the defect outline, combined with the defect geometric features; In some embodiments, the gradient mean of the defect edge and the gradient mean of the background area are calculated, and the difference between the edge of the defect and the background is quantified by the ratio of the two. The local illumination variation coefficient is calculated to compensate for the effect of uneven illumination on the gradient. Combined with the area influence coefficient, the formula is:

[0037] Where, is the visual confidence; is the mean gradient of the defect edge pixels; is the mean gradient of pixels in the background area; is the area influence coefficient; is the defect projection area; is the illumination uniformity compensation factor; the ratio of the numerator to the denominator of this formula reflects the clarity of the defect edge; the exponential term reduces the confidence of large-sized pseudo defects through area penalty; the illumination factor compensates for environmental interference and ultimately quantitatively evaluates the reliability of defect recognition.

[0038] In summary, step S400 filters out inherent texture interference based on the material properties of S100, and combined with the layered path of S300, achieves precise image capture of the entire surface. Multi-scale analysis and geometric quantification are used to extract defect details, and a confidence model is used to quantify defect authenticity. This method overcomes the interference of inherent blade texture on defect identification, providing highly reliable surface defect representation for subsequent dynamic and static data fusion, supporting hierarchical diagnosis and decision-making for wind turbine blade defects.

[0039] Step S500: Mapping the dynamic vibration response data to the blade reference model using the spatiotemporal alignment model, extracting the power spectrum density features at the corresponding positions, and fusing the visual confidence to calculate the damage index; Step S500 specifically includes: Step S510: constructing a blade space rotation matrix based on the blade attitude parameters during the dynamic vibration response data acquisition; specifically, obtaining the blade rotation angle acquired by the wind turbine SCADA system , which is used to reflect the blade spatial posture during vibration collection and determine the spatial mapping relationship of the vibration signal; construct a spatial rotation matrix for the plane rotation of the blade around the root point , the relative position of the vibration sensor is converted into the coordinates of the blade local coordinate system, which can be expressed as:

[0040] Step S520: mapping the dynamic vibration response data to the spatial coordinates of the blade reference model through the blade space rotation matrix according to the spatial coordinates of the blade reference model and the installation position of the vibration sensor; In some embodiments, the longitude and latitude of the center of the wind turbine are converted into the root coordinates of the world coordinate system in step S100. And the installation location of the vibration sensor , the spatial mapping can be expressed as:

[0041] Where, is the blade spatial coordinate corresponding to the dynamic vibration response, which maps the vibration signal from the sensor coordinate system to the blade local coordinate system; Step S530: Synchronizing the dynamic vibration response data with the position information of the blade reference model based on the timestamp to achieve correlation and alignment between the dynamic vibration response data and the blade spatial position; In some embodiments, by obtaining the timestamp of the vibration data in step S240 and the timestamp of the image in step S410, and aligning the asynchronous timestamps through linear interpolation, it is ensured that each vibration data point corresponds to a unique spatial position of the blade reference model, such as the S330 basic waypoint or the S340 encrypted waypoint, thereby solving the problem of spatiotemporal separation of dynamic and static data.

[0042] Step S540: Within the blade characteristic frequency range, the power spectrum density integral values ​​of the damaged area at the position corresponding to the defect on the blade surface and the preset healthy area are calculated to obtain the vibration energy ratio; the integral of the damaged area is expressed as: ,in is the dynamic vibration power spectrum density of the defect position after mapping in step S530; the integral of the healthy area is expressed as: ,in It is the reference power spectrum density of the healthy area in S320. It is the ratio of the integral of the damaged area to the integral of the healthy area of ​​vibration anomaly, which is used to reflect the change of vibration energy in the damaged area. The larger the ratio, the more serious the internal damage.

[0043] Step S550: adjusting the weight coefficient of the visual confidence according to the type of surface defect; In some embodiments, for visual confidence , introducing defect area , i.e. the actual area quantified in step S430; calculated based on the UAV binocular camera; the corrected visual confidence can be expressed as:

[0044] Where, Indicates the preset correction factor, which is used to adjust the sensitivity of the defect area to the visual confidence level; In some embodiments, the visual confidence weighting factor Can be pre-set according to the defect type of S430, such as crack ,corrosion , balancing the contributions of visual and vibration features.

[0045] Step S560: Perform weighted fusion of the visual confidence and the vibration energy ratio according to the adjusted weight coefficient to generate a damage index reflecting the degree of surface and internal damage. The damage index can be expressed as:

[0046] Where, represents the damage index; represents the visual confidence weight factor; represents the visual confidence correction term based on the defect area; Indicates the defect area; Indicates the preset correction factor, which is used to adjust the sensitivity of the defect area to the visual confidence level; represents the blade characteristic frequency; represents the characteristic frequency bandwidth; represents the power spectral density function of the damaged area; Represents the power spectral density function of the healthy region. This method achieves precise binding of vibration signals to the blade's spatial position through a spatiotemporal alignment model. By integrating the visual confidence level from step S400 with the vibration spectrum characteristics from step S310, a bimodal damage index is constructed. This not only quantifies the morphology and credibility of surface defects, but also characterizes the extent of internal structural damage. This provides a comprehensive, quantitative basis for damage diagnosis of wind turbine blades, improving the accuracy and reliability of defect diagnosis.

[0047] Step S600: Classify the damage level based on the damage index and output an inspection report; In some embodiments, a mapping relationship between damage index and damage level can be established through data calibration and threshold derivation. Specifically, a historical fault database of the same model of wind turbine blades is collected, including the damage index (DI) calculated in step S500, the actual damage level, and maintenance records. The correlation between DI and damage risk is fitted through statistical regression analysis to determine the classification threshold. In some embodiments, low risk corresponds to minor surface defects and internal vibration energy changes of ≤30%, which are controllable risks; medium risk corresponds to deeper surface defects or internal vibration energy changes of 30% to 100%, which require timely treatment; high risk corresponds to severe surface and internal coordinated damage, with the risk of fracture, requiring urgent intervention; the specific threshold can be dynamically adjusted based on the wind turbine operating environment, such as offshore or onshore, and the degree of blade aging; In some embodiments, the triggering conditions for periodic re-inspection can be configured as follows: if the DI reflects that the surface defect is a microcrack or slight corrosion, and the internal vibration energy ratio is less than or equal to 1.3, then a monthly re-inspection plan is initiated, and the layered circumferential route of step S300 is reused to focus on monitoring the DI changes in the area; the triggering conditions for time-limited maintenance can be configured as follows: if the DI reflects that the surface defect is a through crack or large-area corrosion, or the internal vibration energy ratio is in the interval [1.3, 2], then a maintenance work order is issued within 72 hours, and a maintenance plan is planned based on the defect spatial coordinates of step S410 and the blade structural dimensions of step S100; the triggering conditions for emergency shutdown maintenance can be configured as follows: if the DI is close to 1, the surface defect is a through crack, or the internal vibration energy ratio is greater than 2, then the wind turbine is immediately triggered to shut down, and a maintenance plan is generated simultaneously. At the same time, based on the defect geometric characteristics of step S430, the feasibility of blade replacement or overall repair is evaluated to prioritize equipment safety.

[0048] In some embodiments, data integration and visualization include a surface defect module for integrating the defect type, spatial position, size, and visual confidence of step S400; a vibration feature module for extracting the characteristic frequency and power spectrum density ratio of step S540; a damage assessment module for displaying the DI value, damage level, and risk description of step S560; a processing suggestion module for generating a maintenance strategy of the corresponding level; and finally generating a structured PDF report that embeds a three-dimensional view of the blade benchmark model and annotates the defect location and vibration spectrum diagram and highlights the characteristic frequency range, so that operation and maintenance personnel can intuitively understand the damage status.

[0049] In summary, the UAV-based wind turbine blade inspection method provided by the present invention collects dynamic vibration response data by presetting short-term dynamic vibration excitation rules. Under the premise of avoiding long-term impact on wind turbine power generation, it captures the vibration characteristics of the blades in the entire startup and steady-state shutdown stages, providing a dynamic structural response basis for identifying internal damage of the blades; based on the dynamic vibration data, high-risk areas are identified and a layered circular route containing basic routes and encrypted waypoints is generated, so that static inspections can not only achieve full coverage of the blade surface, but also focus on high-risk areas for inspection, thereby improving the pertinence and efficiency of the inspection; at the same time, in the static inspection, surface defects are identified and visual confidence is generated in combination with the blade material characteristics, thereby achieving accurate identification and reliability quantification of surface defects; the damage index is calculated by fusing dynamic vibration response data and visual confidence using a spatiotemporal alignment model, and the surface defect characteristics are associated with the internal vibration characteristics, breaking through the limitation of static inspection that can only detect surface defects, and realizing the coordinated evaluation of blade surface and internal damage; finally, the damage index is divided into levels and a report is output based on the damage index, providing a comprehensive and accurate status basis for blade operation and maintenance, effectively solving the technical problem that static inspections are difficult to identify internal damage, and improving the integrity and accuracy of wind turbine blade inspections.

[0050] See also Figure 2 , which shows a schematic structural diagram of a wind turbine blade inspection system based on a drone provided by one embodiment of the present invention, the system comprising: A data acquisition module configured to acquire static basic data of the wind turbine, including spatial positioning data, structural dimension data, and blade material characteristic data; A short-time dynamic vibration excitation module is configured to construct a blade benchmark model based on static basic data of the wind turbine and collect dynamic vibration response data according to preset short-time dynamic vibration excitation rules; a route planning module configured to identify high-risk areas of the blade based on the dynamic vibration response data and generate a layered circumferential route including a basic route and encrypted waypoints in the high-risk areas in combination with the blade benchmark model; a visual inspection module configured to, in a static inspection mode, control the drone to collect multi-angle images of the blade according to a layered circling route, process the multi-angle images of the blade to identify surface defects and generate a visual confidence score; a fusion diagnosis module configured to map the dynamic vibration response data to a blade benchmark model using a spatiotemporal alignment model, extract the power spectral density features at the corresponding positions, and fuse the visual confidence to calculate the damage index; The decision output module is configured to classify damage levels based on the damage index and output an inspection report.

[0051] The technical solution of the third aspect of the present invention provides an electronic device, which includes: a processor and a memory communicatively connected to the processor; wherein the memory stores instructions that can be executed by the processor, and the instructions are executed by the processor so that the processor can execute the steps of the drone-based wind blade inspection method described in the technical solution of the first aspect of the present invention.

[0052] The technical solution of the fourth aspect of the present invention provides a computer-readable storage medium, on which is stored a program for implementing a drone-based wind blade inspection method. The program for implementing a drone-based wind blade inspection method is executed by a processor to implement the steps of the drone-based wind blade inspection method described in the technical solution of the first aspect of the present invention.

[0053] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0054] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A wind turbine blade inspection method based on drones, characterized in that: The method comprises: Obtain static basic data of the wind turbine, including spatial positioning data, structural dimension data, and blade material property data; Build a blade benchmark model based on the static basic data of the wind turbine, and collect dynamic vibration response data according to the preset short-term dynamic vibration excitation rules; Identify high-risk areas of the blade based on dynamic vibration response data, and generate a layered circumferential route including a basic route and encrypted waypoints in the high-risk area in combination with the blade benchmark model; In the static inspection mode, the drone is controlled to collect multi-angle images of the blade according to a layered circling route, and the multi-angle images of the blade are processed to identify surface defects and generate visual confidence. The dynamic vibration response data is mapped to the blade benchmark model using a spatiotemporal alignment model, the power spectrum density features at the corresponding positions are extracted, and the damage index is calculated by integrating the visual confidence score. The damage level is divided based on the damage index and an inspection report is output.

2. The wind turbine blade inspection method based on a drone according to claim 1, characterized in that: A blade benchmark model is constructed based on the static basic data of the wind turbine, and dynamic vibration response data is collected according to the preset short-term dynamic vibration excitation rules, including: Plan a layered circling modeling route based on structural dimension data, and control the drone to collect multi-angle images of the blade along the layered circling modeling route; The blade benchmark model is generated by fusing multi-angle images through motion recovery algorithm; The fan is controlled to execute the preset short-time dynamic vibration excitation rules, starting from the shutdown state to the preset speed and running stably for the preset time, and then gradually shutting down, forming a vibration excitation process including startup, steady state and shutdown stages; During the vibration excitation process, vibration signals of the blade start-stop transient process are collected, and the SCADA system time stamp is synchronized to form dynamic vibration response data.

3. The wind turbine blade inspection method based on a drone according to claim 1, characterized in that: Identify high-risk areas of the blade based on the dynamic vibration response data, and generate a layered circumferential route containing a basic route and encrypted waypoints in the high-risk areas in combination with the blade benchmark model, including: Perform spectrum analysis on dynamic vibration response data to extract characteristic frequencies of different blade regions and the power spectrum density at corresponding frequencies; The power spectrum density of each area is compared with the power spectrum density threshold of the healthy area of ​​the leaf, and the area exceeding the power spectrum density threshold is marked as a high-risk area of ​​the leaf; Generate a basic flight path covering the windward side, leeward side and leading edge of the blade based on the surface curvature and dimensional characteristics of the blade benchmark model; For high-risk areas of blades, the waypoint density is increased on the basis of the basic route to form a layered circular route with encrypted waypoints.

4. The wind turbine blade inspection method based on a drone according to claim 1, characterized in that: In static inspection mode, the drone is controlled to collect multi-angle images of the blade according to a layered circling route. The multi-angle images of the blade are processed to identify surface defects and generate visual confidence, including: In static inspection mode, the drone is controlled to collect multi-angle images of the windward side, leeward side and leading edge of the blade along a layered circular route, and the spatial position information corresponding to each image is simultaneously recorded; Based on the inherent texture period and reflectivity characteristics in the blade material characteristic data, the areas consistent with the inherent texture of the blade are filtered out, and the abnormal texture areas are retained; Using multi-scale filtering to process the abnormal texture area, extract the defect contour and quantify the defect geometric features; The visual confidence is calculated based on the contrast between the edge sharpness of the defect area and the background area, the edge sharpness of the defect outline, and the geometric characteristics of the defect.

5. The wind turbine blade inspection method based on a drone according to any one of claims 1 to 4, characterized in that: Mapping dynamic vibration response data to a blade benchmark model using a spatiotemporal alignment model, including: Constructing the blade space rotation matrix based on the blade attitude parameters during dynamic vibration response data acquisition; According to the spatial coordinates of the blade reference model and the installation position of the vibration sensor, the dynamic vibration response data is mapped to the spatial coordinates of the blade reference model through the blade space rotation matrix; The dynamic vibration response data and the position information of the blade benchmark model are synchronized based on the timestamp to achieve the correlation and alignment between the dynamic vibration response data and the blade spatial position.

6. The wind turbine blade inspection method based on a drone according to claim 5, characterized in that: Extract the power spectrum density features of the corresponding position and fuse the visual confidence to calculate the damage index, including: In the blade characteristic frequency range, the power spectrum density integral values ​​of the damaged area corresponding to the defect position on the blade surface and the preset healthy area are calculated respectively to obtain the vibration energy ratio; Adjust the weight coefficient of visual confidence according to the type of surface defects; The visual confidence and vibration energy ratio are weighted and fused according to the adjusted weight coefficient to generate a damage index that reflects the degree of surface and internal damage.

7. The wind turbine blade inspection method based on a drone according to claim 6, characterized in that: The damage index can be expressed as: Where, represents the damage index; represents the visual confidence weight factor; represents the visual confidence correction term based on the defect area; Indicates the defect area; Indicates the preset correction factor, which is used to adjust the sensitivity of the defect area to the visual confidence level; represents the blade characteristic frequency; represents the characteristic frequency bandwidth; represents the power spectral density function of the damaged area; Represents the power spectral density function of the healthy area.

8. The wind turbine blade inspection system based on drones is characterized by: The wind turbine blade inspection method based on a drone according to any one of claims 1 to 7 is adopted, wherein the system comprises: A data acquisition module configured to acquire static basic data of the wind turbine, including spatial positioning data, structural dimension data, and blade material characteristic data; A short-time dynamic vibration excitation module is configured to construct a blade benchmark model based on static basic data of the wind turbine and collect dynamic vibration response data according to preset short-time dynamic vibration excitation rules; a route planning module configured to identify high-risk areas of the blade based on the dynamic vibration response data and generate a layered circumferential route including a basic route and encrypted waypoints in the high-risk areas in combination with the blade benchmark model; a visual inspection module configured to, in a static inspection mode, control the drone to collect multi-angle images of the blade according to a layered circling route, process the multi-angle images of the blade to identify surface defects and generate a visual confidence score; a fusion diagnosis module configured to map the dynamic vibration response data to a blade benchmark model using a spatiotemporal alignment model, extract the power spectral density features at the corresponding positions, and fuse the visual confidence to calculate the damage index; The decision output module is configured to classify damage levels based on the damage index and output an inspection report.

9. An electronic device, characterized in that: The electronic device includes: a processor and a memory communicatively connected to the processor; wherein the memory stores instructions executable by the processor, and the instructions are executed by the processor so that the processor can execute the steps of the drone-based wind turbine blade inspection method described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program for implementing a drone-based wind turbine blade inspection method, and the program for implementing a drone-based wind turbine blade inspection method is executed by a processor to implement the steps of the drone-based wind turbine blade inspection method described in any one of claims 1 to 7.

Images