A method and system for detecting defects in a wind turbine blade

Abstract

The application discloses a kind of wind power blade defect detection method and system, it is related to defect detection technical field, the system includes: external data acquisition module, state evaluation module of lightning arrester, inner cavity check mapping module, inner cavity image acquisition module, inner cavity defect evaluation module, blade health evaluation module.The application generates internal key check area instruction based on external defect coordinates, guides inner cavity detection equipment to carry out fixed-point review, improves the pertinence and efficiency of inner cavity detection;At the same time, by fusing lightning arrester electrical performance data and visual defect information, and the defects in space are associated with correction, the specificity of lightning damage is realized, the accuracy and reliability of wind power blade health state comprehensive evaluation are improved, the problems that the internal detection blindness is large, the efficiency is low in the prior art, and the defect cause judgment is single, and lightning damage and ordinary defects cannot be accurately distinguished.

Description

Technical Field

[0001] This invention relates to the field of defect detection technology, and in particular to a method and system for detecting defects in wind turbine blades. Background Technology

[0002] With the increasing size of wind turbine blades, their structural health directly affects the operational safety and power generation efficiency of wind turbine units. Therefore, efficient and accurate blade health monitoring has become an urgent need in the wind power operation and maintenance field. In wind farm operation, failure to promptly detect or accurately characterize blade damage can lead to catastrophic failures, resulting in significant safety risks and economic losses.

[0003] In related technologies, wind turbine blade inspection mainly relies on manual high-altitude inspections or single-function automated equipment. Manual inspection is inefficient and highly dangerous. Existing automated inspection technologies, such as independently operating endoscopic robots or drones used only for external inspection, suffer from fragmented inspection data and lack functional coordination. Endoscopic robots lack external guidance, resulting in numerous blind spots; drone-based external inspections cannot determine whether defects stem from internal damage or are related to electrical performance failure. Therefore, existing technologies suffer from the problems of fragmented internal and external inspection methods leading to significant blind spots and low efficiency in internal inspections, as well as limited defect cause identification and an inability to accurately distinguish between lightning damage and ordinary defects. Summary of the Invention

[0004] This application provides a method and system for detecting defects in wind turbine blades, which solves the problems of the prior art where the separation of internal and external detection methods leads to blindness and low efficiency in internal detection, as well as the single judgment of defect causes and the inability to accurately distinguish between lightning damage and ordinary defects. It achieves specific identification of lightning damage, thereby improving the accuracy and reliability of the comprehensive assessment of the health status of wind turbine blades.

[0005] This application provides a method for detecting defects in wind turbine blades. This method is applied to a wind turbine blade defect detection system and includes the following steps:

[0006] S1, control the image acquisition device to acquire the surface image of the target area of ​​the wind turbine blade, perform defect feature extraction processing on the surface image of the target area, and generate an external surface defect dataset containing defect type identifier and first spatial location information, wherein the first spatial location information is the spatial location information corresponding to the external surface defect.

[0007] The step of extracting defect features from the surface image of the target area to generate an external surface defect dataset containing defect type identifiers and first spatial location information includes:

[0008] The surface image of the target region is preprocessed by grayscale conversion and contrast enhancement to obtain the preprocessed surface image.

[0009] Edge contour extraction is performed on the preprocessed surface image to obtain discontinuous pixel regions in the image;

[0010] Based on the geometric morphological features of discontinuous pixel regions, the defect type identifier corresponding to the discontinuous pixel regions is determined.

[0011] By combining the device pose data when acquiring surface images of the target area, the three-dimensional coordinates of discontinuous pixel areas in the global coordinate system of the wind turbine blade are calculated, and the three-dimensional coordinates are marked as the first spatial position information.

[0012] S2, control the contact resistance measuring device to perform conduction performance test on the lightning rod component of the wind turbine blade, obtain the contact resistance value of the lightning rod, and generate lightning rod status assessment data based on the contact resistance value of the lightning rod.

[0013] S3, based on the first spatial location information in the external surface defect dataset, generates an instruction for the internal key inspection area mapped to the blade internal cavity coordinate system through a preset blade inner and outer wall coordinate mapping transformation logic;

[0014] S4, according to the internal key inspection area instruction, control the internal cavity detection equipment to move to the physical position indicated by the internal key inspection area instruction, acquire the internal cavity structure image at that position, and obtain local high-precision image data of the internal cavity;

[0015] S5, perform damage identification and analysis on the high-precision image data of the internal cavity, and generate an internal cavity defect assessment result containing information on the type of internal damage and the second spatial location;

[0016] S6 performs multi-source data association and matching processing on the external surface defect dataset, lightning arrester status assessment data, and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades.

[0017] The process of performing multi-source data association and matching on the external surface defect dataset, lightning arrester condition assessment data, and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades includes:

[0018] Read the lightning arrester status assessment data to determine if there is a lightning arrester failure.

[0019] If there is a lightning arrest failure, then the physical installation location of the lightning arrester is taken as the center, and the external surface defect dataset is searched to see if there is a first spatial location information within a preset range from the physical installation location.

[0020] If the first spatial location information that meets the conditions is retrieved, the corresponding defect type identifier will be corrected to lightning damage, and it will be merged with the lightning arrester status assessment data and written into the wind turbine blade comprehensive health status assessment report.

[0021] The multi-source data association and matching process also includes:

[0022] Calculate the Euclidean distance between the first spatial location information in the external surface defect dataset and the second spatial location information in the internal cavity defect evaluation result;

[0023] If the Euclidean distance is less than the preset wall thickness correlation threshold, then the external surface defect and the internal cavity defect are determined to be the same penetrating damage.

[0024] In the comprehensive health status assessment report of wind turbine blades, penetrating damage is marked as high-risk.

[0025] Furthermore, based on the first spatial location information in the external surface defect dataset, and through a preset blade inner and outer wall coordinate mapping transformation logic, an instruction for the internal key inspection area mapped to the blade inner cavity coordinate system is generated, including:

[0026] Extract the first spatial location information of the target defects labeled as crack types from the external surface defect dataset;

[0027] Call the pre-stored three-dimensional digital model of the wind turbine blade, and locate the first projection point corresponding to the first spatial location information of the target defect of the crack type on the outer surface of the three-dimensional digital model;

[0028] Extend inward along the normal direction of the first projection point to calculate the second projection point that intersects with the inner wall of the three-dimensional digital model;

[0029] The parameters of the second projection point in the internal cavity coordinate system are encapsulated into an internal key inspection area command that drives the movement of the internal cavity detection device.

[0030] Furthermore, according to the internal key inspection area command, the internal cavity detection device is controlled to move to the physical location indicated by the internal key inspection area command, and an image of the internal cavity structure at that location is acquired to obtain high-precision local image data of the internal cavity, including:

[0031] Analyze the instructions for the key internal inspection area to obtain the target coordinate parameters of the second projection point;

[0032] The internal cavity detection device is controlled to move axially along the inner cavity of the blade, and the current mileage code data of the internal cavity detection device is acquired in real time;

[0033] When the current mileage encoding data matches the target coordinate parameters, the internal cavity detection device is controlled to stop moving;

[0034] The supplementary lighting unit and imaging unit on the internal cavity detection device are activated to perform fixed-point focusing and imaging on the inner wall area where the second projection point is located, thereby acquiring high-precision image data of the local area of ​​the internal cavity.

[0035] Furthermore, the control contact resistance measuring device performs continuity performance testing on the lightning arrester components of the wind turbine blade, obtains the lightning arrester contact resistance value, and generates lightning arrester status assessment data based on the lightning arrester contact resistance value, including:

[0036] The probe module of the control contact resistance measuring device establishes physical contact with the surface of the lightning arrester;

[0037] Apply a test current and collect the feedback voltage to calculate the contact resistance value of the lightning arrester.

[0038] Compare the contact resistance value of the lightning arrester with the preset safety threshold;

[0039] If the contact resistance of the lightning arrester exceeds the preset safety threshold, lightning arrester status assessment data representing lightning arrester failure will be generated, and a command to acquire surface images of the surrounding area where the lightning arrester is located will be triggered.

[0040] Furthermore, the damage identification and analysis of the high-precision local image data of the cavity to generate a cavity defect assessment result containing internal damage type and second spatial location information includes:

[0041] Texture feature segmentation is performed on high-precision local image data of the cavity to extract suspected damage areas;

[0042] Calculate the aspect ratio and grayscale mean of the suspected damaged area;

[0043] Based on aspect ratio feature value and grayscale mean feature value, suspected damaged areas are classified into layered, missing glue or wrinkled types to obtain internal damage type;

[0044] The laser ranging value is read when acquiring high-precision image data of the internal cavity, and combined with the coordinate parameters in the command of the key internal inspection area, the second spatial position information is calculated.

[0045] Furthermore, after generating the comprehensive health status assessment report for wind turbine blades, the following steps are also included:

[0046] Analysis of high-risk damage data in the comprehensive health status assessment report of wind turbine blades;

[0047] Extract the spatial coordinates corresponding to high-risk damage data;

[0048] Based on spatial location coordinates, a re-inspection waypoint planning path file is generated for the wind turbine blades, and the re-inspection waypoint planning path file is transmitted to the remote monitoring terminal.

[0049] This application provides a wind turbine blade defect detection system to implement a wind turbine blade defect detection method, including:

[0050] External data acquisition module, lightning strike status assessment module, internal cavity verification and mapping module, internal cavity image acquisition module, internal cavity defect assessment module, blade health assessment module;

[0051] The external data acquisition module is used to control the image acquisition device to acquire the surface image of the target area of ​​the wind turbine blade, perform defect feature extraction processing on the surface image of the target area, and generate an external surface defect dataset containing defect type identifier and first spatial location information, wherein the first spatial location information is the spatial location information corresponding to the external surface defect.

[0052] The step of extracting defect features from the surface image of the target area to generate an external surface defect dataset containing defect type identifiers and first spatial location information includes:

[0053] The surface image of the target region is preprocessed by grayscale conversion and contrast enhancement to obtain the preprocessed surface image.

[0054] Edge contour extraction is performed on the preprocessed surface image to obtain discontinuous pixel regions in the image;

[0055] Based on the geometric morphological features of discontinuous pixel regions, the defect type identifier corresponding to the discontinuous pixel regions is determined.

[0056] By combining the device pose data when acquiring surface images of the target area, the three-dimensional coordinates of discontinuous pixel areas in the global coordinate system of the wind turbine blade are calculated, and the three-dimensional coordinates are marked as the first spatial position information.

[0057] The lightning status assessment module is used to control the contact resistance measuring device to perform a continuity performance test on the lightning rod component of the wind turbine blade, obtain the contact resistance value of the lightning rod, and generate lightning status assessment data based on the contact resistance value of the lightning rod.

[0058] The internal cavity inspection mapping module is used to generate an instruction for the internal key inspection area mapped to the blade internal cavity coordinate system based on the first spatial location information in the external surface defect dataset and through a preset blade inner and outer wall coordinate mapping transformation logic.

[0059] The internal cavity image acquisition module is used to control the internal cavity detection device to move to the physical position indicated by the internal key inspection area instruction according to the internal key inspection area instruction, and to acquire the internal cavity structure image at that position to obtain local high-precision image data of the internal cavity;

[0060] The cavity defect assessment module is used to perform damage identification and analysis on high-precision image data of the cavity, and generate cavity defect assessment results containing internal damage type and second spatial location information.

[0061] The blade health assessment module is used to perform multi-source data association and matching processing on the external surface defect dataset, lightning arrester status assessment data and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades.

[0062] The process of performing multi-source data association and matching on the external surface defect dataset, lightning arrester condition assessment data, and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades includes:

[0063] Read the lightning arrester status assessment data to determine if there is a lightning arrester failure.

[0064] If there is a lightning arrest failure, then the physical installation location of the lightning arrester is taken as the center, and the external surface defect dataset is searched to see if there is a first spatial location information within a preset range from the physical installation location.

[0065] If the first spatial location information that meets the conditions is retrieved, the corresponding defect type identifier will be corrected to lightning damage, and it will be merged with the lightning arrester status assessment data and written into the wind turbine blade comprehensive health status assessment report.

[0066] The multi-source data association and matching process also includes:

[0067] Calculate the Euclidean distance between the first spatial location information in the external surface defect dataset and the second spatial location information in the internal cavity defect evaluation result;

[0068] If the Euclidean distance is less than the preset wall thickness correlation threshold, then the external surface defect and the internal cavity defect are determined to be the same penetrating damage.

[0069] In the comprehensive health status assessment report of wind turbine blades, penetrating damage is marked as high-risk.

[0070] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0071] This application provides a method and system for detecting defects in wind turbine blades. First, an image acquisition device is controlled to acquire surface images of the target area of ​​the wind turbine blade and perform defect feature extraction processing to generate an external surface defect dataset containing defect type identifiers and first spatial location information. Next, a contact resistance measuring device is controlled to perform conductivity testing and generate lightning arrester status assessment data. Based on the first spatial location information, an internal key inspection area instruction is generated through a coordinate mapping transformation logic of the blade's inner and outer walls. An internal cavity detection device is controlled to move to a designated position to acquire internal cavity structure images, obtaining high-precision local image data of the internal cavity. Damage identification analysis is performed on the high-precision local image data of the internal cavity to generate an internal cavity defect assessment result. Finally, the external surface defect dataset, lightning arrester status assessment data, and internal cavity defect assessment result are processed through multi-source data association and matching to generate a comprehensive health status assessment report for the wind turbine blade.

[0072] In this process, by establishing a direct mapping relationship between the spatial location of external defects and internal detection instructions, the operation path of internal equipment is intelligently guided by the external detection results, which effectively overcomes the shortcomings of blind detection and insufficient coverage in internal cavity detection and improves the detection efficiency of internal defects.

[0073] Furthermore, when generating a comprehensive health status assessment report, the system provides multi-dimensional causal criteria for visually identified physical defects based on electrical performance test data. When electrical detection indicates abnormalities, the system can automatically correlate and correct the qualitative assessment of surrounding physical defects, thereby accurately distinguishing between ordinary structural damage and damage caused by specific events such as lightning strikes under complex operating conditions, improving the accuracy and reliability of the comprehensive health status assessment. Attached Figure Description

[0074] Figure 1 A flowchart of a wind turbine blade defect detection method provided in this application embodiment;

[0075] Figure 2 This is a schematic diagram of a wind turbine blade defect detection system provided in an embodiment of this application. Detailed Implementation

[0076] This application provides a method and system for detecting defects in wind turbine blades. It addresses the problems of existing technologies where the separation of internal and external detection methods leads to blind spots and low efficiency in internal detection, as well as the limitations of single-source defect cause analysis and the inability to accurately distinguish between lightning damage and ordinary defects. By generating internal key inspection area instructions based on external defect coordinates, the method guides the internal cavity detection equipment for targeted verification, improving the specificity and efficiency of internal cavity detection. Simultaneously, by integrating lightning arrester electrical performance data and visual defect information, and by correlating and correcting spatially adjacent defects, the method achieves specific identification of lightning damage, improving the accuracy and reliability of comprehensive health status assessment.

[0077] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0078] like Figure 1 As shown in the figure, this application provides a method for detecting defects in wind turbine blades. This method is applied to a wind turbine blade defect detection system and includes:

[0079] S1, control the image acquisition device to acquire the surface image of the target area of ​​the wind turbine blade, perform defect feature extraction processing on the surface image of the target area, and generate an external surface defect dataset containing defect type identifier and first spatial location information, wherein the first spatial location information is the spatial location information corresponding to the external surface defect.

[0080] The step of extracting defect features from the surface image of the target area to generate an external surface defect dataset containing defect type identifiers and first spatial location information includes:

[0081] The surface image of the target region is preprocessed by grayscale conversion and contrast enhancement to obtain the preprocessed surface image.

[0082] Edge contour extraction is performed on the preprocessed surface image to obtain discontinuous pixel regions in the image;

[0083] Based on the geometric morphological features of discontinuous pixel regions, the defect type identifier corresponding to the discontinuous pixel regions is determined.

[0084] By combining the device pose data when acquiring surface images of the target area, the three-dimensional coordinates of discontinuous pixel areas in the global coordinate system of the wind turbine blade are calculated, and the three-dimensional coordinates are marked as the first spatial position information.

[0085] S2, control the contact resistance measuring device to perform conduction performance test on the lightning rod component of the wind turbine blade, obtain the contact resistance value of the lightning rod, and generate lightning rod status assessment data based on the contact resistance value of the lightning rod.

[0086] S3, based on the first spatial location information in the external surface defect dataset, generates an instruction for the internal key inspection area mapped to the blade internal cavity coordinate system through a preset blade inner and outer wall coordinate mapping transformation logic;

[0087] S4, according to the internal key inspection area instruction, control the internal cavity detection equipment to move to the physical position indicated by the internal key inspection area instruction, acquire the internal cavity structure image at that position, and obtain local high-precision image data of the internal cavity;

[0088] S5, perform damage identification and analysis on the high-precision image data of the internal cavity, and generate an internal cavity defect assessment result containing information on the type of internal damage and the second spatial location;

[0089] S6 performs multi-source data association and matching processing on the external surface defect dataset, lightning arrester status assessment data, and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades.

[0090] The process of performing multi-source data association and matching on the external surface defect dataset, lightning arrester condition assessment data, and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades includes:

[0091] Read the lightning arrester status assessment data to determine if there is a lightning arrester failure.

[0092] If there is a lightning arrest failure, then the physical installation location of the lightning arrester is taken as the center, and the external surface defect dataset is searched to see if there is a first spatial location information within a preset range from the physical installation location.

[0093] If the first spatial location information that meets the conditions is retrieved, the corresponding defect type identifier will be corrected to lightning damage, and it will be merged with the lightning arrester status assessment data and written into the wind turbine blade comprehensive health status assessment report.

[0094] The multi-source data association and matching process also includes:

[0095] Calculate the Euclidean distance between the first spatial location information in the external surface defect dataset and the second spatial location information in the internal cavity defect evaluation result;

[0096] If the Euclidean distance is less than the preset wall thickness correlation threshold, then the external surface defect and the internal cavity defect are determined to be the same penetrating damage.

[0097] In the comprehensive health status assessment report of wind turbine blades, penetrating damage is marked as high-risk.

[0098] Furthermore, based on the first spatial location information in the external surface defect dataset, and through a preset blade inner and outer wall coordinate mapping transformation logic, an instruction for the internal key inspection area mapped to the blade inner cavity coordinate system is generated, including:

[0099] Extract the first spatial location information of the target defects labeled as crack types from the external surface defect dataset;

[0100] Call the pre-stored three-dimensional digital model of the wind turbine blade, and locate the first projection point corresponding to the first spatial location information of the target defect of the crack type on the outer surface of the three-dimensional digital model;

[0101] Extend inward along the normal direction of the first projection point to calculate the second projection point that intersects with the inner wall of the three-dimensional digital model;

[0102] The parameters of the second projection point in the internal cavity coordinate system are encapsulated into an internal key inspection area command that drives the movement of the internal cavity detection device.

[0103] Furthermore, according to the internal key inspection area command, the internal cavity detection device is controlled to move to the physical location indicated by the internal key inspection area command, and an image of the internal cavity structure at that location is acquired to obtain high-precision local image data of the internal cavity, including:

[0104] Analyze the instructions for the key internal inspection area to obtain the target coordinate parameters of the second projection point;

[0105] The internal cavity detection device is controlled to move axially along the inner cavity of the blade, and the current mileage code data of the internal cavity detection device is acquired in real time;

[0106] When the current mileage encoding data matches the target coordinate parameters, the internal cavity detection device is controlled to stop moving;

[0107] The supplementary lighting unit and imaging unit on the internal cavity detection device are activated to perform fixed-point focusing and imaging on the inner wall area where the second projection point is located, thereby acquiring high-precision image data of the local area of ​​the internal cavity.

[0108] Furthermore, the control contact resistance measuring device performs continuity performance testing on the lightning arrester components of the wind turbine blade, obtains the lightning arrester contact resistance value, and generates lightning arrester status assessment data based on the lightning arrester contact resistance value, including:

[0109] The probe module of the control contact resistance measuring device establishes physical contact with the surface of the lightning arrester;

[0110] Apply a test current and collect the feedback voltage to calculate the contact resistance value of the lightning arrester.

[0111] Compare the contact resistance value of the lightning arrester with the preset safety threshold;

[0112] If the contact resistance of the lightning arrester exceeds the preset safety threshold, lightning arrester status assessment data representing lightning arrester failure will be generated, and a command to acquire surface images of the surrounding area where the lightning arrester is located will be triggered.

[0113] Furthermore, the damage identification and analysis of the high-precision local image data of the cavity to generate a cavity defect assessment result containing internal damage type and second spatial location information includes:

[0114] Texture feature segmentation is performed on high-precision local image data of the cavity to extract suspected damage areas;

[0115] Calculate the aspect ratio and grayscale mean of the suspected damaged area;

[0116] Based on aspect ratio feature value and grayscale mean feature value, suspected damaged areas are classified into layered, missing glue or wrinkled types to obtain internal damage type;

[0117] The laser ranging value is read when acquiring high-precision image data of the internal cavity, and combined with the coordinate parameters in the command of the key internal inspection area, the second spatial position information is calculated.

[0118] Furthermore, after generating the comprehensive health status assessment report for wind turbine blades, the following steps are also included:

[0119] Analysis of high-risk damage data in the comprehensive health status assessment report of wind turbine blades;

[0120] Extract the spatial coordinates corresponding to high-risk damage data;

[0121] Based on spatial location coordinates, a re-inspection waypoint planning path file is generated for the wind turbine blades, and the re-inspection waypoint planning path file is transmitted to the remote monitoring terminal.

[0122] like Figure 2 As shown, this application provides a wind turbine blade defect detection system to implement the wind turbine blade defect detection method, including: an external data acquisition module, a lightning strike status assessment module, an internal cavity verification and mapping module, an internal cavity image acquisition module, an internal cavity defect assessment module, and a blade health assessment module.

[0123] The external data acquisition module is used to control the image acquisition device to acquire the surface image of the target area of ​​the wind turbine blade, perform defect feature extraction processing on the surface image of the target area, and generate an external surface defect dataset containing defect type identifier and first spatial location information, wherein the first spatial location information is the spatial location information corresponding to the external surface defect.

[0124] The step of extracting defect features from the surface image of the target area to generate an external surface defect dataset containing defect type identifiers and first spatial location information includes:

[0125] The surface image of the target region is preprocessed by grayscale conversion and contrast enhancement to obtain the preprocessed surface image.

[0126] Edge contour extraction is performed on the preprocessed surface image to obtain discontinuous pixel regions in the image;

[0127] Based on the geometric morphological features of discontinuous pixel regions, the defect type identifier corresponding to the discontinuous pixel regions is determined.

[0128] By combining the device pose data when acquiring surface images of the target area, the three-dimensional coordinates of discontinuous pixel areas in the global coordinate system of the wind turbine blade are calculated, and the three-dimensional coordinates are marked as the first spatial position information.

[0129] The lightning status assessment module is used to control the contact resistance measuring device to perform a continuity performance test on the lightning rod component of the wind turbine blade, obtain the contact resistance value of the lightning rod, and generate lightning status assessment data based on the contact resistance value of the lightning rod.

[0130] The internal cavity inspection mapping module is used to generate an instruction for the internal key inspection area mapped to the blade internal cavity coordinate system based on the first spatial location information in the external surface defect dataset and through a preset blade inner and outer wall coordinate mapping transformation logic.

[0131] The internal cavity image acquisition module is used to control the internal cavity detection device to move to the physical position indicated by the internal key inspection area instruction according to the internal key inspection area instruction, and to acquire the internal cavity structure image at that position to obtain local high-precision image data of the internal cavity;

[0132] The cavity defect assessment module is used to perform damage identification and analysis on high-precision image data of the cavity, and generate cavity defect assessment results containing internal damage type and second spatial location information.

[0133] The blade health assessment module is used to perform multi-source data association and matching processing on the external surface defect dataset, lightning arrester status assessment data and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades.

[0134] The process of performing multi-source data association and matching on the external surface defect dataset, lightning arrester condition assessment data, and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades includes:

[0135] Read the lightning arrester status assessment data to determine if there is a lightning arrester failure.

[0136] If there is a lightning arrest failure, then the physical installation location of the lightning arrester is taken as the center, and the external surface defect dataset is searched to see if there is a first spatial location information within a preset range from the physical installation location.

[0137] If the first spatial location information that meets the conditions is retrieved, the corresponding defect type identifier will be corrected to lightning damage, and it will be merged with the lightning arrester status assessment data and written into the wind turbine blade comprehensive health status assessment report.

[0138] The multi-source data association and matching process also includes:

[0139] Calculate the Euclidean distance between the first spatial location information in the external surface defect dataset and the second spatial location information in the internal cavity defect evaluation result;

[0140] If the Euclidean distance is less than the preset wall thickness correlation threshold, then the external surface defect and the internal cavity defect are determined to be the same penetrating damage.

[0141] In the comprehensive health status assessment report of wind turbine blades, penetrating damage is marked as high-risk.

[0142] Example: S1: Control an industrial drone equipped with a three-axis active stabilization gimbal to fly along a preset trajectory to a distance of 10-15 meters from the outer surface of the wind turbine blade. Acquire surface images of the target area using a 40-megapixel high-definition industrial camera mounted on the drone. Obtain the drone's pose data corresponding to each frame of the image using an RTK-GNSS and IMU fusion positioning system. ,in Three-dimensional spatial coordinates (unit: meters). Pitch angle, roll angle, and yaw angle (unit: degrees); the surface image of the target area is preprocessed by grayscale conversion and contrast enhancement. Canny edge detection is then used to extract discontinuous pixel regions from the preprocessed surface image. Based on the geometric features of these discontinuous pixel regions, a trained ResNet-50 and MobileNetV3 dual-network collaborative deep learning model is used to determine the defect type corresponding to each discontinuous pixel region. Defect types include at least one of cracks, spalling, bubbles, corrosion, and lightning strike damage. This is combined with UAV pose data. With camera intrinsic parameter matrix The following coordinate transformation is used to calculate the three-dimensional coordinates of discontinuous pixel regions in the global coordinate system of the wind turbine blade. The three-dimensional coordinates are marked as the first spatial location information: ;in For the first in the image Homogeneous coordinates of the defective pixels For rotation matrix, As a translation vector; associate the defect type identifier with the three-dimensional coordinates. Encapsulated as an external surface defect dataset ,in, This represents the number of external defects. It is an external defect data index. Use the type identifier for the j-th external defect in the external surface defect dataset; This represents the three-dimensional spatial coordinates of the j-th external defect in the global coordinate system of the wind turbine blade within the external surface defect dataset.

[0143] S2: Control the industrial drone equipped with a lightning protection detection device to fly to the location of the lightning arrester on the wind turbine blade. The lightning protection detection device includes a flexible docking mechanism, a metal polishing brush, and a four-probe resistance measurement module. Using a model predictive control (MPC) and multi-level PID control algorithm, establish physical contact between the probe module of the lightning protection detection device and the surface of the lightning arrester, with the contact pressure controlled within the range of 5-10N. Apply a test current with a frequency of 1kHz and an amplitude of 10mA. Collect feedback voltage Calculate the contact resistance of the lightning arrester according to Ohm's law. :

[0144] ;

[0145] The contact resistance value of the lightning arrester With preset safety threshold Perform a comparison; if This generates lightning arrester condition assessment data characterizing lightning failure. ,in Here are the coordinates of the lightning arrester's position. This indicates that the lightning protection test result is invalid; simultaneously triggering a lightning protection test targeting... A command to acquire high-definition images of an area within a 3-meter radius centered on the target.

[0146] S3: From external surface defect dataset Extract the first spatial location information of the target defect labeled as crack type. ; Call the pre-stored 3D digital model of wind turbine blades ,exist External surface positioning and The first projection point with the smallest Euclidean distance ;along normal direction Extending inwards, among which, On the three-dimensional digital model of wind turbine blades The unit normal vector at a certain point is the direction perpendicular to the outer surface of the blade and pointing towards the inner cavity; the calculation of... The second projection point where the inner wall surfaces intersect : ;

[0147] in, The blade wall thickness parameter has a range of 30-100 mm; The coordinate parameters in the blade internal cavity coordinate system are encapsulated into internal key inspection area instructions. , This indicates that the defect type is a crack.

[0148] S4: The internal cavity inspection equipment is a highly mobile endoscopic robot with a multi-wheeled wheel structure, equipped with a high-definition vision sensor, a 650nm laser ranging module, and an odometer; it can analyze instructions for key internal inspection areas. Get target coordinates The endoscopic robot is controlled to move axially along the inner cavity of the blade, acquiring real-time readings from the odometer encoder. and laser ranging value When satisfied ( When the position tolerance is reached, the endoscopic robot is controlled to stop moving. It provides the real-time position coordinates of the endoscopic robot; it also activates the supplementary lighting unit and imaging unit to... Images of the internal cavity structure were captured at fixed points within the area to obtain high-precision local images of the cavity with a resolution of no less than 40 megapixels. .

[0149] S5: High-precision image data of the internal cavity Semantic segmentation based on the U-Net architecture is performed to extract suspected damaged regions; the aspect ratio features of the suspected damaged regions are calculated. With grayscale mean characteristic value Classified according to the following rules:

[0150] like and If so, it is determined to be a crack;

[0151] like and If so, it is determined to be a layer;

[0152] like and If so, it is determined to be an air bubble;

[0153] Obtain the internal damage type identifier; combine with laser ranging values Endoscopic robot odocoding data The second spatial location information is obtained by solving the problem using the Extended Kalman Filter (EKF) algorithm. ; Generate internal cavity defect assessment results ,in, This represents the number of internal defects. It is an internal defect data index. Represents the three-dimensional spatial coordinates of the k-th internal defect in the internal cavity defect assessment results; This indicates the type identifier of the k-th internal defect in the internal cavity defect assessment results.

[0154] S6: Transfer the external surface defect dataset Lightning arrester status assessment data and the results of the assessment of internal defects Perform multi-source data association and matching processing: If If the status is failed, then in Searching in the middle satisfies The defect was corrected to be lightning damage; calculation Each and Each Euclidean distance between ,like , If the blade wall thickness parameter defined in S3 is used, it is determined to be penetrating damage and marked as high-risk. The first spatial location information (external defect location) and the second spatial location information (internal defect location) corresponding to the penetrating damage are associated to obtain the spatial coordinates corresponding to the high-risk damage, so as to generate a re-inspection waypoint planning path. A comprehensive health status assessment report of the wind turbine blade is generated, which includes a defect distribution heatmap, a damage correlation matrix, and maintenance recommendations. .

[0155] S7: Analysis of the Comprehensive Health Status Assessment Report for Wind Turbine Blades Set of spatial coordinates corresponding to intermediate and high-risk injuries ;based on Generate re-inspection waypoint planning route file The file format is JSON and contains waypoint sequences. and the detection parameters for each waypoint, This represents the three-dimensional coordinates of a waypoint in the re-examination waypoint planning path; r is the waypoint number; Transmitted to the remote monitoring terminal via 4G / 5G network.

[0156] Specifically, the deep learning model training process in step S1 includes: collecting 100,000 images of wind turbine blade surfaces to construct a training set, labeling each image with defect type and location bounding boxes; initializing the ResNet-50 and MobileNetV3 dual networks using ImageNet pre-trained weights; and setting the initial learning rate. Batch size Training cycle The cross-entropy loss function is used. and FocalLoss (Focal Loss Function) Joint optimization loss value: ;

[0157] in, During training, the learning rate is evaluated on the validation set every 10 epochs, and the learning rate is reduced when the accuracy no longer improves.

[0158] The semantic segmentation model in step S5 adopts the U-Net architecture. The encoder part uses VGG16 pre-trained weights, and the decoder part is randomly initialized. The training data includes 5000 intracavitary images, which are augmented through rotation, scaling, and brightness adjustment. The loss value is obtained using the Dice coefficient loss function. The Adam optimizer is used during training, and the learning rate is set to [value missing]. .

[0159] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0160] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0161] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0162] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0163] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0164] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method of wind turbine blade defect detection, characterized in that, Includes the following steps: S1, control the image acquisition device to acquire the surface image of the target area of ​​the wind turbine blade, perform defect feature extraction processing on the surface image of the target area, and generate an external surface defect dataset containing defect type identifier and first spatial location information, wherein the first spatial location information is the spatial location information corresponding to the external surface defect. The step of extracting defect features from the surface image of the target area to generate an external surface defect dataset containing defect type identifiers and first spatial location information includes: The surface image of the target region is preprocessed by grayscale conversion and contrast enhancement to obtain the preprocessed surface image. Edge contour extraction is performed on the preprocessed surface image to obtain discontinuous pixel regions in the image; Based on the geometric morphological features of discontinuous pixel regions, the defect type identifier corresponding to the discontinuous pixel regions is determined. By combining the device pose data when acquiring surface images of the target area, the three-dimensional coordinates of discontinuous pixel areas in the global coordinate system of the wind turbine blade are calculated, and the three-dimensional coordinates are marked as the first spatial position information. S2, control the contact resistance measuring device to perform conduction performance test on the lightning rod component of the wind turbine blade, obtain the contact resistance value of the lightning rod, and generate lightning rod status assessment data based on the contact resistance value of the lightning rod. S3, based on the first spatial location information in the external surface defect dataset, generates an instruction for the internal key inspection area mapped to the blade internal cavity coordinate system through a preset blade inner and outer wall coordinate mapping transformation logic; S4, according to the internal key inspection area instruction, control the internal cavity detection equipment to move to the physical position indicated by the internal key inspection area instruction, acquire the internal cavity structure image at that position, and obtain local high-precision image data of the internal cavity; S5, perform damage identification and analysis on the high-precision image data of the internal cavity, and generate an internal cavity defect assessment result containing information on the type of internal damage and the second spatial location; S6 performs multi-source data association and matching processing on the external surface defect dataset, lightning arrester status assessment data, and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades. The process of performing multi-source data association and matching on the external surface defect dataset, lightning arrester condition assessment data, and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades includes: Read the lightning arrester status assessment data to determine if there is a lightning arrester failure. If there is a lightning arrest failure, then the physical installation location of the lightning arrester is taken as the center, and the external surface defect dataset is searched to see if there is a first spatial location information within a preset range from the physical installation location. If the first spatial location information that meets the conditions is retrieved, the corresponding defect type identifier will be corrected to lightning damage, and it will be merged with the lightning arrester status assessment data and written into the wind turbine blade comprehensive health status assessment report. The multi-source data association and matching process also includes: Calculate the Euclidean distance between the first spatial location information in the external surface defect dataset and the second spatial location information in the internal cavity defect evaluation result; If the Euclidean distance is less than the preset wall thickness correlation threshold, then the external surface defect and the internal cavity defect are determined to be the same penetrating damage. In the comprehensive health status assessment report of wind turbine blades, penetrating damage is marked as high-risk.

2. The method of claim 1, wherein the method further comprises: The first spatial location information in the external surface defect dataset, through a preset blade inner and outer wall coordinate mapping transformation logic, generates an internal key inspection area instruction mapped to the blade inner cavity coordinate system, including: Extract the first spatial location information of the target defects labeled as crack types from the external surface defect dataset; Call the pre-stored three-dimensional digital model of the wind turbine blade, and locate the first projection point corresponding to the first spatial location information of the target defect of the crack type on the outer surface of the three-dimensional digital model; Extend inward along the normal direction of the first projection point to calculate the second projection point that intersects with the inner wall of the three-dimensional digital model; The parameters of the second projection point in the internal cavity coordinate system are encapsulated into an internal key inspection area command that drives the movement of the internal cavity detection device.

3. The method of claim 2, wherein the method further comprises: The process involves controlling the internal cavity detection device to move to the physical location indicated by the internal key inspection area command, acquiring an image of the internal cavity structure at that location, and obtaining high-precision local image data of the internal cavity, including: Analyze the instructions for the key internal inspection area to obtain the target coordinate parameters of the second projection point; The internal cavity detection device is controlled to move axially along the inner cavity of the blade, and the current mileage code data of the internal cavity detection device is acquired in real time; When the current mileage encoding data matches the target coordinate parameters, control the internal cavity detection device to stop moving; The supplementary lighting unit and imaging unit on the internal cavity detection device are activated to perform fixed-point focusing and imaging on the inner wall area where the second projection point is located, thereby acquiring high-precision image data of the local area of ​​the internal cavity.

4. The method of claim 1, wherein the method further comprises: The controlled contact resistance measuring device performs continuity performance tests on the lightning arrester components of the wind turbine blades, obtains the contact resistance value of the lightning arrester, and generates lightning arrester status assessment data based on the contact resistance value, including: The probe module of the control contact resistance measuring device establishes physical contact with the surface of the lightning arrester; Apply a test current and collect the feedback voltage to calculate the contact resistance value of the lightning arrester. Compare the contact resistance value of the lightning arrester with the preset safety threshold; If the contact resistance of the lightning arrester exceeds the preset safety threshold, lightning arrester status assessment data representing lightning arrester failure will be generated, and a command to acquire surface images of the surrounding area where the lightning arrester is located will be triggered.

5. The method of claim 1, wherein the method further comprises: The process of performing damage identification and analysis on high-precision local image data of the cavity to generate a cavity defect assessment result containing internal damage type and second spatial location information includes: Texture feature segmentation is performed on high-precision local image data of the cavity to extract suspected damage areas; Calculate the aspect ratio and grayscale mean of the suspected damaged area; Based on aspect ratio feature value and grayscale mean feature value, suspected damaged areas are classified into layered, missing glue or wrinkled types to obtain internal damage type; The laser ranging value is read when acquiring high-precision image data of the internal cavity, and combined with the coordinate parameters in the command of the key internal inspection area, the second spatial position information is calculated.

6. The method of claim 1, wherein the method further comprises: After generating the comprehensive health status assessment report for wind turbine blades, the following are also included: Analysis of high-risk damage data in the comprehensive health status assessment report of wind turbine blades; Extract the spatial coordinates corresponding to high-risk damage data; Based on spatial location coordinates, a re-inspection waypoint planning path file is generated for the wind turbine blades, and the re-inspection waypoint planning path file is transmitted to the remote monitoring terminal.

7. A wind turbine blade defect detection system for implementing the method of any one of claims 1-6, characterized by, include: External data acquisition module, lightning strike status assessment module, internal cavity verification and mapping module, internal cavity image acquisition module, internal cavity defect assessment module, blade health assessment module; The external data acquisition module is used to control the image acquisition device to acquire the surface image of the target area of ​​the wind turbine blade, perform defect feature extraction processing on the surface image of the target area, and generate an external surface defect dataset containing defect type identifier and first spatial location information, wherein the first spatial location information is the spatial location information corresponding to the external surface defect. The step of extracting defect features from the surface image of the target area to generate an external surface defect dataset containing defect type identifiers and first spatial location information includes: The surface image of the target region is preprocessed by grayscale conversion and contrast enhancement to obtain the preprocessed surface image. Edge contour extraction is performed on the preprocessed surface image to obtain discontinuous pixel regions in the image; Based on the geometric morphological features of discontinuous pixel regions, the defect type identifier corresponding to the discontinuous pixel regions is determined. By combining the device pose data when acquiring surface images of the target area, the three-dimensional coordinates of discontinuous pixel areas in the global coordinate system of the wind turbine blade are calculated, and the three-dimensional coordinates are marked as the first spatial position information. The lightning status assessment module is used to control the contact resistance measuring device to perform a continuity performance test on the lightning rod component of the wind turbine blade, obtain the contact resistance value of the lightning rod, and generate lightning status assessment data based on the contact resistance value of the lightning rod. The internal cavity inspection mapping module is used to generate an instruction for the internal key inspection area mapped to the blade internal cavity coordinate system based on the first spatial location information in the external surface defect dataset and through a preset blade inner and outer wall coordinate mapping transformation logic. The cavity image acquisition module is used to control the cavity detection device to move to the physical position indicated by the internal key inspection area instruction according to the internal key inspection area instruction, and to acquire the cavity structure image at that position to obtain local high-precision image data of the cavity. The cavity defect assessment module is used to perform damage identification and analysis on high-precision image data of the cavity, and generate cavity defect assessment results containing internal damage type and second spatial location information. The blade health assessment module is used to perform multi-source data association and matching processing on the external surface defect dataset, lightning arrester status assessment data and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades. The process of performing multi-source data association and matching on the external surface defect dataset, lightning arrester condition assessment data, and internal cavity defect assessment results to generate a comprehensive health status assessment report for wind turbine blades includes: Read the lightning arrester status assessment data to determine if there is a lightning arrester failure. If there is a lightning arrest failure, then the physical installation location of the lightning arrester is taken as the center, and the external surface defect dataset is searched to see if there is a first spatial location information within a preset range from the physical installation location. If the first spatial location information that meets the conditions is retrieved, the corresponding defect type identifier will be corrected to lightning damage, and it will be merged with the lightning arrester status assessment data and written into the wind turbine blade comprehensive health status assessment report. The multi-source data association and matching process also includes: Calculate the Euclidean distance between the first spatial location information in the external surface defect dataset and the second spatial location information in the internal cavity defect evaluation result; If the Euclidean distance is less than the preset wall thickness correlation threshold, then the external surface defect and the internal cavity defect are determined to be the same penetrating damage. In the comprehensive health status assessment report of wind turbine blades, penetrating damage is marked as high-risk.

Images