A system for treating a fan blade

Abstract

The application relates to the technical field of wind power equipment operation and maintenance, in particular to a wind turbine blade processing system, which comprises the following: a robot body adsorbed on the surface of a wind turbine blade; a damage identification module fixedly installed at the front end of the robot body and used for identifying multidimensional damage data of the blade according to a blade surface image; a repair operation module rotatably installed in the robot body and used for executing corresponding repair procedures after the robot body moves to a damage area, wherein the repair procedures include cleaning, polishing, drying, filling, solidification, spraying and re-inspection; and a control station connected with the robot body through a wireless communication link and used for generating a repair scheme according to the multidimensional damage data and issuing corresponding time sequence control instructions according to the repair scheme. Therefore, the problems of high operation risk and low efficiency in the related art that depend on manual climbing or a hanging basket for wind turbine blade detection and repair are solved.

Description

Technical Field

[0001] This application relates to the field of wind power equipment operation and maintenance technology, and in particular to a wind turbine blade processing system. Background Technology

[0002] Large wind turbine blades operate in complex high-altitude environments for extended periods, and are highly susceptible to damage from multiple factors, including lightning strikes, rain erosion, hail impacts, and fatigue under alternating loads. This makes them prone to leading-edge corrosion, coating peeling, surface cracks, internal delamination, and perforation. If these damages are not repaired promptly, they will severely reduce aerodynamic efficiency, increase operating noise, and may even lead to major safety accidents such as blade breakage, directly impacting the power generation efficiency and operational safety of wind farms.

[0003] In related technologies, the inspection and repair of wind turbine blades mainly rely on manual high-altitude operations. This involves workers accessing the blade surface via suspended platforms or ropes for inspection and repair. This method carries an extremely high risk of falls from heights, is inefficient, and has a long cycle time. Furthermore, the quality of the work is highly dependent on the experience of the personnel, resulting in poor consistency, and is strictly limited by weather conditions such as wind speed, rain, and snow, with a short effective working window. Summary of the Invention

[0004] This application provides a wind turbine blade processing system to solve the problems of high operational risk and low efficiency in related technologies that rely on manual climbing or suspended baskets for wind turbine blade inspection and repair.

[0005] The first aspect of this application provides a wind turbine blade processing system, comprising: a robot body adsorbed onto the surface of the wind turbine blade; a damage recognition module fixedly installed at the front end of the robot body for recognizing multi-dimensional damage data of the blade based on images of the blade surface; a repair operation module rotatably installed in the robot body for performing corresponding repair procedures after the robot body moves to the damaged area, wherein the repair procedures include: cleaning, grinding, drying, filling, curing, spraying, and re-inspection; and a control station connected to the robot body via a wireless communication link for generating a repair plan based on the multi-dimensional damage data and generating corresponding timing control commands based on the repair plan and sending them to the repair operation module.

[0006] Optionally, the robot body includes: a moving mechanism, an adsorption mechanism, and a controller. The moving mechanism consists of a composite structure of tracks and wheels. The track surface is covered with rubber material with a target coefficient of friction, and the wheels have independent lifting and adjustment functions to drive the robot body to move on the surface of the wind turbine blades. The adsorption mechanism includes a permanent magnet adsorption unit and a negative pressure vacuum adsorption unit, which are used to switch the adsorption mode according to the adsorption mode switching command. The controller is connected to the drive circuit of the moving mechanism and the control valve of the composite adsorption system. It is used to output the corresponding moving control command to the moving mechanism according to the target damage location of the wind turbine blades, determine the corresponding target adsorption mode according to the blade surface material, and generate an adsorption mode switching command based on the target adsorption mode and the current adsorption mode and send it to the adsorption mechanism.

[0007] Optionally, it also includes: an adsorption force monitoring unit and a positioning and navigation unit. The adsorption force monitoring unit, installed on the contact surface between the adsorption chamber of the robot body and the blade surface, is used to monitor the adsorption force between the robot body and the blade surface and to feed back the adsorption force to the controller of the robot body. The controller is used to output a locking command to the moving mechanism of the robot body when the adsorption force is lower than a preset adsorption force threshold, stop the robot's movement, and trigger an alarm signal. The positioning and navigation unit, connected to the controller of the robot body, is used to collect multi-source positioning data and transmit the multi-source positioning data to the controller of the robot body. The controller is used to fuse the multi-source positioning data to calculate the current pose of the robot body on the blade surface and drive the robot to move to the target position based on the deviation between the current pose and the coordinates of the target damage point.

[0008] Optionally, the damage identification module includes: a first detection unit, a second detection unit, a third detection unit, and a processing unit. The first detection unit is used to identify the type and location of damage on the blade surface based on the acquired high-definition images of the blade surface. The second detection unit is used to determine the damage parameters of the damage location based on point cloud scanning data. The third detection unit is used to determine the location and extent of internal defects in the blade based on changes in the temperature field on the blade surface. The processing unit is used to generate multi-source data based on the type and location of damage on the blade surface, the damage parameters of the damage location, and the location and extent of internal defects in the blade. It also performs timestamp and spatial coordinate alignment on the multi-source data to generate a damage model and transmits the damage model to the robot body controller and control station.

[0009] Optionally, the repair module includes: a turret base, a cleaning device, a grinding device, a drying device, a filling device, a curing device, a spraying device, and a re-inspection device. The turret base is fixedly installed in the middle of the robot body. The cleaning device, grinding device, drying device, filling device, curing device, spraying device, and re-inspection device are sequentially installed on the workstation mounting base of the turret base. After receiving the timing control command issued by the control station, the controller of the robot body sequentially outputs corresponding commands to the cleaning device, grinding device, drying device, filling device, curing device, spraying device, and re-inspection device to execute the corresponding repair process according to the timing control command.

[0010] Optionally, the cleaning device includes a water jet unit and a dry ice cleaning unit. The water jet unit is used to clean contaminants by the impact force of the water jet, and the impact force value of the water jet is adjusted according to the type of contaminant. The dry ice cleaning unit is used to remove contaminants from the blade surface by utilizing the low-temperature embrittlement effect and impact of dry ice particles. The water jet unit and the dry ice cleaning unit are connected to the controller of the robot body through a switching valve. The controller selects either the water jet unit or the dry ice cleaning unit to clean the blade surface according to the type of contaminant on the damaged surface.

[0011] Optionally, the grinding device includes: a grinding head and a target sensor, wherein the drive end of the grinding head is connected to the controller of the robot body; the target sensor is installed between the grinding head and the robot body to acquire grinding contact force and grinding position data; the controller calculates the grinding path and depth based on the damage data, and controls the grinding head to contact the blade surface with the target pressure for grinding based on the grinding path and depth.

[0012] Optionally, the drying device includes: a telescopic hood, a vacuum pump, and a heating element. The telescopic hood is installed at the corresponding workstation on the turret base and is used to extend under the control of the controller to form a closed space by fitting the damaged area of ​​the blade after grinding. The vacuum pump is installed on the robot body, and the suction port of the vacuum pump is connected to the internal chamber of the telescopic hood through a pipeline. It is used to extract air from the closed space after the telescopic hood fits the blade, so that the vacuum degree in the chamber reaches a preset threshold to remove moisture from the repair area. The heating element is installed inside the telescopic hood and is used to heat the temperature of the damaged area of ​​the blade to a preset construction temperature after the vacuum degree in the chamber reaches the preset threshold.

[0013] Optionally, the filling device includes an injection mechanism and a laying mechanism. The control end of the injection mechanism is connected to the controller of the robot body and is used to inject filler into the target damaged gap according to the filling ratio. The controller is used to calculate the filling ratio of the corresponding filler according to the damage data of the target damaged gap. The control end of the laying mechanism is connected to the controller of the robot body and is used to transfer the cut prepreg to the damaged area. The laying mechanism is controlled to position and lay the prepreg in the damaged area according to the number of layups and the layup angle. The rolling unit is controlled to roll and defoam at a set pressure. The controller is used to determine the number of layups and the layup angle according to the loss data of the damaged area.

[0014] Optionally, the spraying device includes a spray gun, a hopper, and a flow valve. The spray gun is installed at the corresponding station on the turret base. The drive end of the spray gun is connected to the controller of the robot body for spraying the damaged area under the control of the controller. The hopper is connected to the spray gun through a supply pipeline for storing and supplying primer and topcoat. The flow valve is installed on the supply pipeline for adjusting the paint output flow of the spray gun according to the spraying thickness parameters set by the controller of the robot body.

[0015] Therefore, this application has at least the following beneficial effects: This application embodiment achieves fully automated closed-loop operation of wind turbine blades from intelligent perception of damaged areas to precise repair of damaged areas by integrating damage identification and multi-functional repair operation modules. It effectively solves the problems of fragmented detection and repair, high risk and low efficiency of manual high-altitude operations in traditional operation and maintenance. Its rotatable repair module design ensures flexible switching and precise execution of complex processes such as cleaning, grinding, filling and curing in a confined space. Combined with the dynamic repair scheme generated by the control station based on multi-dimensional damage data, it significantly improves the consistency and reliability of the repair process, while greatly reducing operation and maintenance costs and extending the wind turbine downtime window.

[0016] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0017] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a block diagram of a wind turbine blade processing system according to an embodiment of this application; Figure 2 This is a flowchart illustrating a method for processing wind turbine blades according to an embodiment of this application. Detailed Implementation

[0018] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0019] Large wind turbine blades operate in complex high-altitude environments for extended periods, and are susceptible to damage such as leading-edge corrosion, coating peeling, surface cracks, internal delamination and debonding, and perforation due to factors such as lightning strikes, rain erosion, hail impacts, and alternating load fatigue. These damages directly affect the wind turbine's power generation efficiency and operational safety.

[0020] Existing blade inspection and repair technologies are mainly divided into two categories: one is the manual high-altitude operation mode, which involves suspending personnel in baskets or ropes to carry out inspection and repair. This mode carries extremely high safety risks and is strictly limited by weather conditions such as wind speed, rain, and snow. It also has a short window period, low efficiency, and poor consistency in repair quality due to the variability of manual operation. The other is the blade disassembly and repair mode, which requires disassembling the blades from the nacelle and transporting them to the ground for operation. This results in long-term downtime of the wind turbine, huge loss of power generation, and high costs for disassembly, transportation, and reassembly.

[0021] In related technologies, some wall-climbing robots only have a single detection function. After damage is detected, manual or other equipment still needs to be dispatched to carry out repair work, which cannot form a closed loop of detection and repair. A few robots that integrate repair functions have problems such as poor adsorption stability (difficult to adapt to changes in the twist angle of the blade surface), low efficiency of switching operation modules, and uncontrollable repair process parameters, which make it difficult to meet the needs of large-scale operation and maintenance.

[0022] Therefore, this application can move autonomously along the blade surface while the wind turbine is running or under short-term low-speed shutdown conditions, integrating the entire process of "inspection-cleaning-grinding-drying-filling-curing-spraying-re-inspection" to achieve intelligent identification, standardized repair, and digital traceability of blade damage, thereby significantly improving the safety, efficiency, and quality stability of operation and maintenance.

[0023] The following description, with reference to the accompanying drawings, describes a wind turbine blade processing system according to an embodiment of this application.

[0024] Specifically, Figure 1 This is a block diagram of a wind turbine blade processing system according to an embodiment of this application.

[0025] like Figure 1 As shown, the wind turbine blade processing system 10 includes: a robot body 100, a damage identification module 200, a repair operation module 300, and a control station 400.

[0026] The robot body 100 is attached to the surface of the wind turbine blade, and the damage identification module 200 is fixedly installed at the front end of the robot body 100 to identify multi-dimensional damage data of the blade based on the image of the blade surface. The repair operation module 300 is rotatably installed on the robot body and is used to perform corresponding repair procedures after the robot body moves to the damaged area. The repair procedures include cleaning, grinding, drying, filling, curing, spraying, and re-inspection. The control station 400 is connected to the robot body 100 through a wireless communication link and is used to generate a repair plan based on the multi-dimensional damage data and generate corresponding timing control commands based on the repair plan and send them to the repair operation module.

[0027] It is understood that the embodiments of this application can utilize a damage identification module fixed to the front end of the robot body to collect and analyze multi-dimensional damage data of the blade surface in real time, and then transmit it to the control station to dynamically generate targeted repair plans and precise timing control commands; thereby driving the rotatably mounted repair operation module to automatically switch and accurately execute a full set of processes such as cleaning, grinding, drying, filling, curing, spraying and re-inspection according to the command sequence. This working mechanism, which tightly couples front-end intelligent identification with back-end multi-station rotational execution, not only achieves seamless connection and full-process automation from damage discovery to standardized repair, but also ensures high consistency of process parameters and flexibility of spatial operation through timing control. Thus, while completely avoiding the risks of manual high-altitude operations, it significantly improves the operation efficiency, repair quality and structural reliability of wind turbine blade maintenance.

[0028] In this embodiment, the robot body 100 includes a moving mechanism, an adsorption mechanism, and a controller.

[0029] The mobile mechanism consists of a composite structure of tracks and wheels. The track surface is covered with rubber material with a target friction coefficient, and the wheels have independent lifting and adjustment functions to drive the robot body to move on the surface of the wind turbine blades. The adsorption mechanism includes a permanent magnet adsorption unit and a negative pressure vacuum adsorption unit, which are used to switch the adsorption mode according to the adsorption mode switching command. The controller is connected to the drive circuit of the mobile mechanism and the control valve of the composite adsorption system. It is used to output the corresponding movement control command to the mobile mechanism according to the target damage location of the wind turbine blades, and to determine the corresponding target adsorption mode according to the blade surface material. It also generates an adsorption mode switching command based on the target adsorption mode and the current adsorption mode and sends it to the adsorption mechanism.

[0030] It is understood that in this embodiment, the controller can intelligently determine the target adsorption mode based on the material characteristics of the blade surface (such as metal lightning protection strip or composite material body), and generate switching commands in real time to drive the adsorption mechanism to flexibly switch between permanent magnet and negative pressure vacuum unit, thereby ensuring that the robot can maintain a constant and safe adsorption force in different material areas. On this basis, the controller plans the path in combination with the target damage location and drives the composite movement mechanism composed of tracks and independent lifting wheels. The tracks covered with high friction coefficient rubber provide basic traction, while the independent lifting adjustment of the wheels compensates for the large curvature change and torsion deviation of the blade surface in real time. This control logic that deeply integrates adaptive adsorption switching and terrain active following not only effectively solves the problems of easy detachment at the interface of changing materials and poor fit of the single movement structure on complex curved surfaces in traditional single adsorption methods, but also realizes the stable residence and high-precision autonomous navigation of the robot on the entire surface of the wind turbine blade.

[0031] For example, the robot needs to move to repair a deep dent caused by a lightning strike on the leading edge of a blade. Path segment A (from the blade root to the damage point): The blade surface is a smooth glass fiber composite material (non-ferromagnetic) with a large curvature. Path segment B (damage point area): Due to previous repairs, a metal lightning protection guide mesh is embedded inside the lightning strike point, and the surface is covered with a thick layer of putty repair, exhibiting weak ferromagnetism locally but with a rough and uneven surface. The robot is located at the blade root, currently in negative pressure vacuum adsorption mode (default safe mode), with its wheels retracted and tracks in contact with the ground.

[0032] The controller receives instructions from the ground station, locks onto the target damage coordinates, and calls the built-in "blade digital twin model" to retrieve the path attributes from the current position to the target point. The determination result: most of the path area is "smooth composite material (non-ferromagnetic)," but there is a "region containing metal inserts" near the target point. Due to the large blade curvature, the controller decides to keep "independent wheel lifting adjustment" active during travel to compensate for surface gaps and ensure track contact. For section A (non-ferromagnetic): maintain the current negative pressure vacuum adsorption (permanent magnets are ineffective). For section B (containing metal / roughness): switch to a hybrid mode with permanent magnet adsorption as the primary method and negative pressure as a secondary method, utilizing the metal guide net to provide stronger grip, while the negative pressure fills the surface roughness gaps.

[0033] The controller sends pulse signals to the motion mechanism's drive circuit: the high-friction rubber track, covered with a high coefficient of friction, begins to rotate, driving the robot upwards. Laser sensors monitor the gap between the robot body and the blades in real time. When a change in blade curvature is detected (such as a transition from a flat surface to a leading-edge arc), the controller instantly adjusts the extension and retraction lengths of the four corner wheels. For example, the outer wheel extends by 5mm, and the inner wheel shortens by 3mm, ensuring the robot body remains parallel to the blade's cross-section and preventing the track from slipping.

[0034] When the robot approaches the target damage point (approximately 0.5 meters away): If the current mode is negative pressure vacuum adsorption (vacuum pump on, electromagnet off), and the target mode is permanent magnet and negative pressure composite adsorption (based on the prediction that the area in front contains metal inserts), a switching command is triggered. Specifically: The controller does not turn off the vacuum pump first, but instead sends a power-on command to the electromagnetic coil of the permanent magnet adsorption unit. The electromagnet generates a magnetic field, pushing the internal permanent magnet assembly to the working position (or activating the electromagnet itself), making it adhere tightly to the metal guide net on the blade surface. At this time, the robot is in a dual adsorption state of "vacuum + permanent magnet," and the suction force instantly doubles, ensuring that there is no risk of slippage during the switching process. The pressure sensor of the adsorption mechanism reports that the suction force value exceeds the threshold (e.g., a jump from -60 kPa in pure vacuum to the equivalent -90 kPa total suction force), and the controller confirms that the switching is successful. If the target area is mainly ferromagnetic and tightly bonded, the controller can selectively reduce the vacuum pump power to save energy, relying only on the permanent magnet to maintain the main adsorption force, or use negative pressure to fine-tune the sealing. The robot precisely stops above the lightning strike crater in a strong adsorption state. The controller locks the moving mechanism and starts the repair module (such as grinding or spraying). Even if the grinding generates vibration, the permanent magnet adsorption can effectively resist the shearing force and ensure the accuracy of the operation.

[0035] In this embodiment, the application also includes an adsorption force monitoring unit and a positioning and navigation unit.

[0036] The adsorption force monitoring unit, installed at the contact surface between the adsorption chamber of the robot body and the blade surface, monitors the adsorption force between the robot body and the blade surface and feeds back the adsorption force to the robot body's controller. The controller outputs a locking command to the robot body's movement mechanism when the adsorption force is lower than a preset adsorption force threshold, stopping the robot's movement and triggering an alarm signal. The positioning and navigation unit, connected to the robot body's controller, collects multi-source positioning data and transmits the multi-source positioning data to the robot body's controller. The controller fuses the multi-source positioning data to calculate the robot body's current pose on the blade surface and drives the robot to move to the target position based on the deviation between the current pose and the coordinates of the target damage point.

[0037] It is understood that the embodiments of this application can utilize the positioning and navigation unit to collect multi-source data and fuse and calculate the robot's precise current pose on the blade surface. The controller then drives the moving mechanism to correct the path in real time based on the deviation between the pose and the coordinates of the target damage point, ensuring that the robot arrives at the work area with high precision. At the same time, the adsorption force monitoring unit continuously monitors the adsorption state of the contact surface in real time. Once the adsorption force is detected to be lower than the preset safety threshold, the controller immediately triggers the safety interlock mechanism, forcibly locking the moving mechanism to stop moving and issuing an alarm signal. This working mechanism, which deeply couples high-precision navigation and positioning with real-time adsorption safety monitoring, not only effectively overcomes the positioning drift problem caused by the complex curved surface of the wind turbine blade, but also fundamentally eliminates the risk of falling due to adsorption failure, realizing the robot's precise accessibility and inherent safety in high-altitude variable working conditions.

[0038] For example, a wind turbine is located in an offshore wind farm, and a robot is performing an inspection task on the leading edge of the blade (the part with the greatest curvature). Suddenly, a strong crosswind occurs (the wind speed instantly reaches 4.5 m / s, close to the upper limit of operation), causing the blade to vibrate slightly, and the airflow generates a huge lateral shear force on the robot. The robot needs to move from its current position to a tiny crack 5 meters ahead. The robot is in "negative pressure vacuum adsorption" mode and is crawling steadily at a speed of 0.2 m / s.

[0039] (1) Multi-source fusion positioning and dynamic correction (positioning and navigation unit workflow): Visual sensor: The camera captures the texture features of the blade surface and the preset marker points in real time. Inertial measurement unit: The robot's acceleration and angular velocity are collected at high frequency to sense the attitude changes caused by blade vibration. Wheel odometer: Records the number of rotations of the track and calculates the relative displacement. Laser rangefinder: Measures the distance from the robot to the edge of the blade in real time to assist in lateral positioning.

[0040] The extended Kalman filter algorithm within the controller fuses the aforementioned data in real time. When strong winds cause slight slippage of the tracks, relying solely on the odometer will result in cumulative errors (displaying a movement of 4 meters when the actual movement is only 3.8 meters). At this point, the visual sensor identifies the unique spiral pattern features of the blades, the IMU detects an abnormal tilt in the robot's posture, the controller corrects the odometer drift, calculates the current true pose, and finds that the robot is lagging behind the expected position by 0.2 meters and laterally deviating from the centerline by 0.05 meters.

[0041] The controller calculates the deviation vector and immediately adjusts the movement mechanism commands, instructing the left track to accelerate and the right track to decelerate (differential steering), while simultaneously adjusting the wheel height to adapt to the current blade tilt angle, so that the robot is re-aligned with the target crack path and the lateral deviation is eliminated.

[0042] (2) Real-time monitoring and emergency locking of adsorption force (workflow of adsorption force monitoring unit) A high-sensitivity pressure sensor array (adsorption force monitoring unit) installed on the contact surface of the adsorption chamber samples at a frequency of 100Hz, providing real-time feedback on the negative pressure value and total adsorption force of each area. The preset safe adsorption force threshold is 1500N (sufficient to withstand the shear force under a level 5 wind). When a strong crosswind suddenly intensifies and blade vibration causes a tiny gap to appear at the edge of the adsorption chamber, the sensor reading instantly drops from 1800 N to 1350 N, indicating a risk of falling from a height.

[0043] Upon detecting a drop below the threshold (<10ms), the controller prioritizes all movement commands and sends an "emergency lock command" to the drive circuit of the mobile mechanism. The electromagnetic brakes within the mobile mechanism immediately engage the track wheel axles, while the hydraulic / electric push rods of the wheel legs lock at their current extension positions, "pinning" the robot to the blade surface to prevent uncontrollable slippage. The controller triggers the onboard audible and visual alarm (flashing red light, continuous buzzer) and sends an "adsorption failure warning" and the current precise coordinates to the ground maintenance center via wireless module. While locking, the controller can briefly increase the vacuum pump power or switch to a "permanent magnet + negative pressure" composite mode. If the adsorption force recovers above the safety threshold within 5 seconds, the lock is released and operation continues; if it remains below the threshold, the lock is maintained, awaiting manual rescue or further instructions.

[0044] In this embodiment, the damage identification module includes: a first detection unit, a second detection unit, a third detection unit, and a processing unit.

[0045] The system comprises the following components: a first detection unit for identifying the type and location of surface damage based on high-definition images of the blade; a second detection unit for determining damage parameters based on point cloud scanning data; a third detection unit for determining the location and extent of internal defects based on changes in the temperature field of the blade surface; and a processing unit for generating multi-source data based on the type and location of surface damage, damage parameters, and the location and extent of internal defects. The processing unit performs timestamp and spatial coordinate alignment on the multi-source data to generate a damage model and transmits the damage model to the robot controller and control station.

[0046] It is understood that the embodiments of this application can utilize the high-definition image recognition of the surface damage type and location of the first detection unit, the point cloud scanning of the second detection unit to quantify the damage geometric parameters, and the thermal imaging of the third detection unit to detect the range of internal defects in parallel, to achieve comprehensive data acquisition from macroscopic appearance to microscopic structure. Subsequently, the processing unit performs strict timestamp synchronization and spatial coordinate registration on these three heterogeneous data sources to eliminate the perspective and timing deviations between multiple sensors, and then fuses them to generate a high-precision three-dimensional damage model and transmits it to the control system in real time. This processing logic, which deeply couples visible light, three-dimensional geometry, and thermal field information, not only effectively solves the limitation that a single detection method cannot take into account both surface morphology and internal hidden dangers, but also provides a comprehensive, accurate, and spatially consistent decision basis for the formulation of subsequent repair plans through standardized damage models, significantly improving the completeness of damage assessment and the pertinence of repair operations.

[0047] For example, maintenance personnel receive a report that an offshore wind turbine blade appears to have been struck by lightning after a thunderstorm. A robot is dispatched to the suspected area (15 meters from the blade tip) to conduct a detailed diagnosis. Damage characteristics: obvious black ablation pits and paint peeling; the ablation pits vary in depth and have slightly raised edges; the high temperature from the lightning strike may have caused carbonization or delamination of the resin between composite layers, which is not visible to the naked eye, but there is abnormal heat conduction.

[0048] (1) Parallel acquisition of multi-source heterogeneous data The robot docks above the target area, and the three detection units operate according to a preset sequence or are triggered synchronously. 1) The first detection unit activates a high-resolution industrial camera (20 megapixels) and, in conjunction with a ring light, captures visible light images of the damaged area. The built-in deep learning model analyzes the images in real time, identifies the damage types as "lightning ablation" and "coating peeling," and initially defines the coordinates of the damage location, outputting damage category labels and a two-dimensional pixel coordinate mask. The deep learning model includes an input layer, a feature extractor, a feature fusionist, and an output layer. The input layer receives the high-resolution images captured by the camera of the first detection unit. The feature extractor transforms the original image into a multi-level feature map, extracting shallow (detail-rich) and deep (semantic-rich) information. The feature fusionist transmits the semantic information from the deep layer to the shallow layer to help locate small targets and transmits the location information from the shallow layer to the deep layer to improve the accuracy of the bounding box. The output layer outputs the corresponding damage type (crack, peeling, lightning strike, oil stain, bubble, etc.) based on the semantic information.

[0049] 2) The second detection unit activates the line laser scanner to perform high-density scanning of the damaged area locked by the first unit, generating three-dimensional point cloud data containing hundreds of thousands of points. Based on the point cloud data, the depth distribution of the ablation pit (maximum depth 12mm), volume loss (approximately 50 cubic centimeters), and the height gradient of the edge bulge are calculated to determine the three-dimensional geometric parameters (depth, area, volume, roughness).

[0050] 3) Third Detection Unit: The robot activates the active thermal excitation module (e.g., briefly turning on the heating lamp to irradiate the surface for 3 seconds), and then a high-sensitivity infrared thermal imager acquires a sequence of images showing the surface temperature field changing over time. Analyzing the temperature decay curve using the principle of heat wave propagation reveals a significant "thermal stagnation" phenomenon (temperature decreases more slowly than in normal areas) below and around the ablation pit. This indicates the presence of internal interlayer delamination or water accumulation, determining the planar projection range of the internal defect, estimating its depth (based on thermal diffusion time), and classifying the defect severity level.

[0051] (2) Spatiotemporal alignment and multi-source fusion Because the three sensors have different sampling frequencies (camera 30fps, laser 10Hz, thermal imaging 60Hz), the processing unit first extracts the high-precision hardware timestamps (accurate to microseconds) of each data packet. Using the laser scanning time as a reference, the visual image and thermal imaging sequence are resampled to the same time point through an interpolation algorithm to eliminate data asynchrony caused by slight robot jitter or differences in acquisition timing.

[0052] The external parameter matrices of the camera, laser, and thermal imager relative to the robot base coordinate system, which were calibrated at the factory, are called. The pixel coordinates of the visual image are combined with the depth information (from the point cloud) and back-projected into three-dimensional spatial points. The temperature pixels of the thermal image are mapped onto the corresponding three-dimensional point cloud surface to form a point cloud with temperature attributes. All data are uniformly transformed into the global coordinate system of the blade to ensure that the ablation pits on the surface (visual), the depth of the pits (point cloud), and the layering under the pits (thermal image) are strictly coincident in space, with the error controlled within the millimeter level.

[0053] The processing unit encapsulates the fused data into a multi-dimensional damage object and generates a visualized three-dimensional damage heat map model, where color represents temperature anomalies (internal defects) and geometric bumps represent surface damage.

[0054] (3) Data distribution The controller receives the damage model and immediately assesses the risk level. Upon detecting "severe internal delamination," the controller automatically adjusts subsequent actions: prohibiting grinding operations in the area (to prevent penetration), planning an alternative route, or performing only surface cleaning, while marking the point as "requiring manual re-inspection."

[0055] The control station's large screen instantly renders a 3D holographic model of the damage. Maintenance experts can rotate the view to see both the surface pit and the internal red layered areas through the semi-transparent layer. The system automatically generates a repair recommendation report: "Recommended solution: Remove the damaged layer and perform structural repair, not simple filling." In this embodiment, the repair operation module includes: a turret base, a cleaning device, a grinding device, a drying device, a filling device, a curing device, a spraying device, and a re-inspection device.

[0056] The turret base is fixedly installed in the middle of the robot body; the cleaning device, grinding device, drying device, filling device, curing device, spraying device, and re-inspection device are installed sequentially on the workstation mounting base of the turret base; after receiving the timing control command issued by the control station, the controller of the robot body outputs corresponding commands to the cleaning device, grinding device, drying device, filling device, curing device, spraying device, and re-inspection device in sequence to execute the corresponding repair process according to the timing control command.

[0057] It is understood that the embodiments of this application can utilize a turret base fixed in the middle of the robot body to highly integrate seven functional devices—cleaning, grinding, drying, filling, curing, spraying, and re-inspection—onto the same rotating workstation sequence, forming a compact multi-station work unit. Subsequently, after receiving the timing control instructions issued by the control station, the controller strictly follows the process sequence to drive each device to move precisely, ensuring a seamless transition from the previous process (such as cleaning and grinding) to the next process (such as drying and filling), until the final spraying and re-inspection are completed. This working mechanism, which deeply integrates the physical integration of multiple processes with logical timing control, not only completely eliminates the positioning errors and efficiency losses caused by frequent repositioning of traditional split equipment, but also ensures the consistency and continuity of process parameters in each repair stage through standardized process control, thereby achieving full-process automation, high precision, and quality traceability in wind turbine blade damage repair.

[0058] For example, the robot has located a lightning strike ablation pit with a depth of 12mm and a volume of 50cm³ (based on data from the preceding detection module). The ground control station issues a timing control command for a "standard lightning strike repair process package," which includes parameters for seven steps (such as cleaning pressure, grinding particle size, curing time, etc.). The turret base, as the core actuator, is like a precise "revolver" or "rotary tool magazine," responsible for sequentially switching different functional devices to the working position (facing the damage point).

[0059] 1) The robot is attached directly above the damage point, the turret base is in a "standby position" (usually unloaded or in a safety locked state), and all operating devices are distributed in a circle around the central axis. Command reception: The controller receives timing commands issued by the control station.

[0060] 2) Surface cleaning (cleaning equipment) The controller drives the servo motor of the turret base to rotate, precisely rotating the cleaning device to the working position (0°) and triggering the mechanical locking pin to prevent shaking during operation. The cleaning device activates the high-pressure gas-liquid mixing nozzle, spraying a special cleaning agent (to remove oil, dust, and loose carbon deposits). Simultaneously, the vacuum recovery port opens to suck up waste liquid and debris, preventing contamination of other areas of the blades. The flow sensor confirms that the cleaning fluid volume is up to standard, and the vision sensor confirms that there are no visible stains on the surface, reporting "cleaning complete".

[0061] 3) Damage repair (grinding device) The turret base rotates to the next station (e.g., 45°), and the grinding device is positioned and locked. Based on the previously detected "depth distribution" and "edge gradient" data, the controller dynamically adjusts the grinding head's pressure (Z-axis feed) and rotation speed. Chamfering is performed on the raised edges (high gradient areas) to eliminate stress concentration; the bottom of the pit is roughened to increase the adhesion of the filler. A built-in force control sensor monitors the grinding pressure in real time to prevent over-grinding and damage to the healthy fiber layer. A laser displacement sensor confirms that the post-grinding contour conforms to the preset "V-groove" standard and reports "grinding complete."

[0062] 4) Surface drying (drying equipment) The turret rotates to 90°, and the drying device is in position. The infrared heating lamp array and hot air nozzles are activated to quickly dry the polished area, removing residual moisture and solvents. A humidity sensor monitors the surface humidity in real time until it drops below 0.1%. Once the humidity target is met, a "drying complete" message is displayed.

[0063] 5) Material filling (filling device) The turret rotates to 135°, and the filling device is in place. The two-component dispensing pump starts, mixing epoxy resin and hardener in the correct proportions. Based on the calculated "volume loss (50cm³)," the controller precisely controls the dispensing volume, employing either a "layered filling" or "one-time injection" strategy to fill the pit with repair paste, slightly raising it above the surface (allowing for shrinkage allowance). The leveling mechanism then activates, initially smoothing the surface. The flow meter confirms that the dispensing volume has reached 105% of the theoretical value, and reports "filling complete."

[0064] 6) Rapid curing (curing device) Rotate the turret to 180° to position the curing unit. Activate the high-power ultraviolet (UV) lamp array or heating blanket to cover the filling area. Use a temperature sensor to control the heating temperature in a closed loop (e.g., maintain a constant temperature of 60°C) to accelerate the resin curing reaction. Start a countdown timer (e.g., 15 minutes) and continuously monitor the temperature profile. Once the curing time is up and the temperature profile is normal, display "Curing complete".

[0065] 7) Surface spraying (spraying equipment) The turret rotates to 225°, and the spraying device is in position. The automatic color matching system mixes the topcoat according to the original color code of the blades. The high-precision spray gun performs multi-layer spraying (primer + topcoat + clear coat), using atomization technology to ensure uniform coating thickness and a smooth transition with the surrounding old paint surface (without obvious seams). The film thickness gauge confirms that the dry film thickness meets the standard and reports "Spraying complete".

[0066] 8) Quality re-inspection (re-inspection device) The turret rotates to 270°, and the re-inspection device is in place. A high-definition camera captures images of the repaired area, and an artificial intelligence algorithm compares the image with a standard sample to check for bubbles, runs, and color differences. A laser scanner scans again, comparing the deviation between the repaired surface and the original blade design surface (flatness error required <0.5mm). Hardness / adhesion sampling is performed: if necessary, a micro probe is used for non-destructive testing. If successful, a "Repair Completion Report" is generated and uploaded to the control station, and the robot prepares to move to the next damage point. If unsuccessful, the defect type is automatically marked (e.g., "uneven surface"), and a "local rework" instruction is requested from the control station (the turret rotates to the grinding or spraying station for secondary processing).

[0067] In this embodiment, the cleaning device includes a water jet unit and a dry ice cleaning unit.

[0068] The water jet unit is used to clean contaminants by the impact force of the water jet, and the impact force value of the water jet is adjusted according to the type of contaminant. The dry ice cleaning unit is used to remove contaminants from the blade surface by utilizing the low-temperature embrittlement effect and impact of dry ice particles. The water jet unit and the dry ice cleaning unit are connected to the robot body's controller through a switching valve. The controller selects either the water jet unit or the dry ice cleaning unit to clean the blade surface according to the type of contaminant on the damaged surface.

[0069] It is understood that, according to the specific type of contaminants on the damaged surface (such as thick mud or stubborn oil), the controller can intelligently decide and drive the switching valve to flexibly switch between the water jet unit and the dry ice cleaning unit: for contaminants that require strong rinsing, the water jet unit with dynamically adjustable impact force is activated to efficiently peel them off using fluid kinetic energy; for scenarios requiring non-destructive treatment or water-restricted scenarios, the system switches to the dry ice cleaning unit, using the low-temperature embrittlement effect of dry ice particles to shrink and peel off the contaminants, supplemented by impact to remove residues; this control logic, which deeply couples the adjustability of hydraulic impact with the non-abrasive nature of dry ice's cold embrittlement, not only effectively solves the problem that a single cleaning method is difficult to take into account the characteristics of different contaminants and is prone to damaging the blade substrate, but also achieves the refined operation goal of maximizing the protection of the blade surface integrity while ensuring thorough cleaning.

[0070] In this embodiment of the application, the grinding device includes a grinding head and a target sensor.

[0071] The grinding head's drive end is connected to the robot's controller; the target sensor is installed between the grinding head and the robot to acquire grinding contact force and grinding position data; the controller calculates the grinding path and depth based on the damage data, and controls the grinding head to contact the blade surface with the target pressure for grinding based on the grinding path and depth.

[0072] It is understood that the embodiments of this application can utilize a target sensor installed between the grinding head and the body to collect grinding contact force and position data in real time, providing accurate feedback signals to the controller. Subsequently, the controller combines preset damage data to calculate the optimal grinding path and removal depth, and adjusts the grinding head drive end in real time based on the dynamic force and position information fed back by the sensor, ensuring that it adheres to the blade surface with a constant target pressure. This working mechanism, which deeply integrates real-time force and position monitoring with feedforward path planning, not only effectively overcomes the problem of contact force fluctuation caused by the complex curved surface of the wind turbine blade, avoiding damage to the substrate due to excessive pressure or incomplete grinding due to insufficient pressure, but also achieves adaptive and precise removal of damaged areas, significantly improving the uniformity and consistency of the grinding operation and the ability to protect the original profile of the blade.

[0073] In this embodiment, the drying device includes: a telescopic hood, a vacuum pump, and a heating element.

[0074] The telescopic cover is installed at the corresponding workstation on the turret base and extends under the control of the controller to form a closed space by fitting the damaged area of ​​the blade after grinding. The vacuum pump is installed on the robot body, and the suction port of the vacuum pump is connected to the internal chamber of the telescopic cover through a pipeline. It is used to extract air from the closed space after the telescopic cover fits the blade, so that the vacuum degree in the chamber reaches a preset threshold to remove moisture from the repair area. The heating element is installed inside the telescopic cover and is used to heat the damaged area of ​​the blade to the preset construction temperature after the vacuum degree in the chamber reaches the preset threshold.

[0075] The preset threshold and preset construction temperature can be set according to actual needs without specific limitations.

[0076] It is understood that in this embodiment, the telescopic cover can be extended by the controller and closely fit the damaged area after the blade is polished, thus constructing an independent closed working chamber. Then, the vacuum pump is started to extract the air in the chamber until the preset vacuum level is reached. The negative pressure environment is used to efficiently remove the residual moisture and micropore water on the repair interface. After the vacuum conditions are stable, the built-in heating element is activated to precisely heat the damaged area to the preset construction temperature. This working mechanism, which couples the space sealing, low-pressure dehumidification and heat curing pretreatment in a strict sequence, not only completely solves the problems of low efficiency and susceptibility to environmental humidity interference of traditional open drying, but also significantly improves the wettability and bonding strength of the subsequent filling material by creating an ideal interface environment with no water and constant temperature, thus ensuring the overall reliability of the repair process.

[0077] In this embodiment of the application, the filling device includes an injection mechanism and a laying mechanism.

[0078] The injection mechanism's control unit is connected to the robot's controller, used to inject filler into the target damaged crevices according to the filler ratio. The controller calculates the corresponding filler ratio based on the damage data of the target damaged crevices. The laying mechanism's control unit is also connected to the robot's controller, used to transfer the cut prepreg to the damaged area. The laying mechanism is controlled to position and lay the prepreg in the damaged area according to the number and angle of layup. The rolling unit is controlled to roll and defoam at a set pressure. The controller determines the number and angle of layup based on the loss data of the damaged area.

[0079] It is understood that, in this embodiment of the application, the controller can intelligently calculate the optimal filling ratio and layup parameters (quantity and angle) based on the damage data, and then drive the injection mechanism to inject the filler into the target gap according to the precise ratio to complete the matrix restoration; then, the layup mechanism is linked to precisely position the cut prepreg to the damaged area, strictly follow the preset layup angle and number of layers for orderly stacking, and simultaneously control the roller pressing unit to apply constant pressure to eliminate interlayer air bubbles; this collaborative working mechanism that couples quantitative injection with directional layup depth not only achieves a seamless connection from gap filling to structural reinforcement, but also ensures the density of the repair material, the mechanical matching of fiber orientation, and the quality of interlayer bonding through parameterized precise control, thereby significantly improving the structural integrity and service life of the blade after damage repair.

[0080] In this embodiment, the spraying device includes a spray gun, a hopper, and a flow valve.

[0081] The spray gun is installed at the corresponding station on the turret base. The drive end of the spray gun is connected to the controller of the robot body and is used to spray the damaged area under the control of the controller. The hopper is connected to the spray gun through the supply pipeline and is used to store and supply primer and topcoat. The flow valve is installed on the supply pipeline and is used to adjust the paint output flow of the spray gun according to the spray thickness parameters set by the controller of the robot body.

[0082] It is understood that in this embodiment, the controller can set the target coating thickness parameters according to the repair process requirements, and then drive the flow valve to adjust the output flow of the primer or topcoat in the supply pipeline in real time to ensure that the coating supply is strictly matched with the preset thickness. Subsequently, the controller links the spray gun drive end, and with the support of continuous material supply from the hopper, performs uniform spraying on the damaged area according to the planned path. This feedforward working mechanism, which directly converts the thickness target into flow control command, not only effectively solves the problem of uneven coating thickness caused by flow fluctuation in traditional spraying, but also realizes precise quantitative control of primer sealing and topcoat protection, ensuring a high degree of consistency in the density, smoothness and anti-corrosion and weather resistance of the coating in the repaired area.

[0083] In this embodiment, the curing device includes an ultraviolet curing unit and a hot air curing unit.

[0084] The UV curing unit is equipped with a UV lamp assembly, and its control terminal is connected to the robot's controller. It is activated when the repair material is UV-curable resin to cure the filled damaged area with UV light. The hot air curing unit is equipped with a hot air circulation system, and its control terminal is connected to the robot's controller. It is activated when the repair material is thermosetting epoxy resin to cure the filled damaged area with hot air. The controller selects to activate either the UV curing unit or the hot air curing unit according to the type of repair material to achieve adaptable curing for different repair materials.

[0085] It is understood that the embodiments of this application can be analyzed in real time by the controller to determine the chemical properties of the repair material (UV-curable resin or thermosetting epoxy resin), and make intelligent decisions and activate the corresponding execution units accordingly: for photosensitive resin, the UV lamp group is activated to initiate polymerization by high-energy light irradiation, and for thermosetting epoxy, the hot air circulation system is activated to implement constant temperature heating and crosslinking. This working mechanism of dynamically switching curing energy based on material characteristics not only completely solves the process limitation that a single curing method cannot be compatible with multiple repair materials, but also ensures that repair materials of different systems can complete deep curing under their optimal reaction conditions, thereby significantly improving the mechanical properties of the repair layer, the interfacial bonding strength, and the overall process's wide adaptability to complex damage scenarios.

[0086] In this embodiment, the re-inspection device includes a laser 3D scanner and a camera.

[0087] The corresponding detection unit in the laser 3D scanner and camera multiplexing integrated detection module has its signal output end connected to the robot's controller via a data bus. This is used to quickly re-inspect the repaired area after spraying and curing, collect data on the surface flatness and coating thickness after repair, and transmit the re-inspection data to the controller. The controller compares the re-inspection data with the damage parameters before repair to form a repair quality data comparison, and uploads the re-inspection results to the ground control station via a wireless communication module.

[0088] It is understood that the embodiments of this application can reuse existing laser 3D scanners and camera units to perform non-contact rapid scanning of the repair area immediately after spraying and curing, accurately collecting key geometric quantities such as surface flatness and coating thickness. Subsequently, the controller automatically compares and analyzes the real-time collected re-inspection data with the original damage parameters before repair, quantitatively generates a repair quality assessment report, and uploads the results to the ground control station in real time via a wireless communication module. This collaborative working mechanism, which efficiently reuses detection resources, accurately matches the before and after states, and remotely synchronizes detection results, not only realizes the closed-loop verification of the entire process from "damage input" to "repair output," effectively avoiding the subjective errors and lags of manual re-testing, but also ensures the traceability and transparency of repair quality, providing a reliable digital basis for the full life cycle management of blades.

[0089] According to the wind turbine blade processing system proposed in this application, a damage identification module fixed to the front end of the robot body collects and analyzes multi-dimensional damage data of the blade surface in real time, and then transmits it to the control station to dynamically generate targeted repair plans and precise timing control commands. This drives the rotatably mounted repair operation module to automatically switch and precisely execute a full set of processes such as cleaning, grinding, drying, filling, curing, spraying, and re-inspection according to the command sequence. This working mechanism, which tightly couples front-end intelligent identification with back-end multi-station rotational execution, not only achieves seamless connection and full-process automation from damage discovery to standardized repair, but also ensures high consistency of process parameters and flexibility of spatial operation through timing control. Thus, while completely avoiding the risks of manual high-altitude operations, it significantly improves the operation efficiency, repair quality, and structural reliability of wind turbine blade maintenance.

[0090] The wind turbine blade processing system of this application will be described in detail below: (1) Composite Adsorption Wall Climbing Robot Body Mobile mechanism: It adopts a track-wheel composite structure. The track surface is covered with high friction coefficient rubber material, and the wheel has independent lifting and adjustment function. It can adapt to the torsion angle change of the blade surface and the protruding structure such as blade rib. The obstacle crossing height is ≥20mm and the moving speed is ≥50mm / s.

[0091] Composite Adsorption System: Integrates a permanent magnet adsorption unit and a negative pressure vacuum adsorption unit, with independent control and automatic switching between the two systems. For the metal lightning protection mesh area at the leading edge of the blade, the permanent magnet adsorption unit is activated, and the adsorption force is adjusted by an electromagnetic control valve, with an adsorption force ≥80N / cm². For the composite material area of ​​the blade body, the vacuum adsorption unit is activated, and a negative pressure is generated by an onboard vacuum pump, forming a sealed chamber between the adsorption chamber and the blade surface, with an adsorption force ≥60N / cm². Equipped with a real-time adsorption force monitoring sensor, the system automatically alarms and locks the moving mechanism when the adsorption force falls below the threshold.

[0092] Power and communication unit: Built-in high-energy-density lithium battery pack, rated voltage 48V, battery life ≥4h, supports online fast charging; the communication module adopts 5G+microwave dual-mode redundant communication, communication distance with ground control station ≥500m, transmission rate ≥100Mbps, supports real-time interaction of high-definition images, sensor data and control commands; equipped with emergency wired communication interface to ensure communication reliability in extreme environments.

[0093] Positioning and navigation unit: It integrates LiDAR, inertial measurement unit, visual positioning camera and odometry encoder, and fuses the 3D model of the blade. It adopts the iterative nearest point registration algorithm to achieve centimeter-level accurate positioning of the robot on the blade surface (positioning error ≤ ±2mm). It supports two modes: autonomous navigation and remote control. In autonomous mode, it can complete full-coverage scanning according to the preset path. In remote control mode, the operator can accurately control the robot to the designated damage point.

[0094] (2) Intelligent damage recognition unit As a perception system for robots, it integrates a multi-sensor fusion detection device to achieve comprehensive and accurate identification of damage to the surface and interior of blades. Macroscopic visual inspection unit: Equipped with two high-definition industrial visible light cameras (resolution ≥ 12 million pixels), with automatic zoom and supplemental lighting functions, to acquire images of the blade surface and identify surface damage such as coating peeling, cracks, and holes through artificial intelligence algorithms, with an identification accuracy ≥ 0.1mm. The artificial intelligence algorithms include, but are not limited to, lightweight target detection networks.

[0095] 3D morphology inspection unit: Equipped with a line laser 3D scanner, with a scanning speed of ≥1000 points / second and a scanning accuracy of ≤±0.05mm, it can accurately obtain the 3D morphology parameters such as area, depth, and volume of the damaged area, providing data support for repair process planning.

[0096] Internal Defect Detection Unit: Integrates an infrared thermal imager and an ultrasonic probe array. The infrared thermal imager uses an active thermal excitation method (built-in heating source) to identify internal defects such as delamination and debonding by monitoring changes in the temperature field on the blade surface, with a detection depth ≥20mm; the ultrasonic probe array uses contact detection and is equipped with an automatic coupling agent application device to perform A / B scanning on key suspected areas, quantitatively analyze the size and location of defects, and has a detection resolution ≥0.5mm.

[0097] Airborne AI data processing unit: It has a built-in embedded processor and is equipped with a deep learning-based damage recognition model. It integrates multi-sensor data in real time, automatically classifies damage types (such as cracks, corrosion, delamination, and perforation), assesses damage levels (minor / moderate / severe), generates preliminary repair process plans, and transmits them to the ground control station for confirmation.

[0098] (3) Integrated repair operation module The system adopts a multi-station turret-type end effector design, with all operating devices integrated into the same turret. The turret is driven to rotate by a servo motor, enabling rapid switching of operating modules (switching time ≤ 5s). The following functional devices are integrated according to the process sequence: High-pressure cleaning device: Employs a dual-mode system of high-pressure micro-water jet cleaning and dry ice cleaning, which can be switched according to the type of contaminants on the damaged surface. The high-pressure micro-water jet pressure is adjustable (0-30MPa) and equipped with a pure water filtration system to avoid impurity residue; the dry ice cleaning device uses food-grade dry ice particles (1-3mm in diameter) to remove old coatings and stubborn stains through high-pressure airflow; an integrated waste liquid / dry ice residue collection device is included to prevent secondary contamination of the blade surface.

[0099] Adaptive constant force grinding device: Equipped with a high-speed electric spindle grinding head (speed adjustable from 0-10000r / min), and equipped with force and displacement sensors to achieve constant pressure grinding control (grinding pressure settable range 0.5-5N); based on the damage 3D data from the laser 3D scanner, it automatically plans the grinding path and depth to form a bevel that meets process requirements (bevel angle adjustable from 30°-45°); integrated negative pressure dust collection hood, dust collection efficiency ≥95%.

[0100] Vacuum drying and preheating device: It adopts a sealed telescopic cover structure to form a closed space by fitting the damaged area after grinding; the built-in vacuum pump extracts air to make the vacuum degree in the chamber reach above -0.08MPa, removing surface moisture; it is equipped with an infrared heating element to precisely preheat the repair area, with a temperature control range of 20-60℃ and a temperature control accuracy of ≤±2℃, ensuring the best adhesion performance of the repair material.

[0101] Composite material infill and layup device: Employing differentiated repair solutions for different damage sizes to achieve automated and precise construction operations. Repairing minor damage (damage diameter <10mm, depth <5mm): Equipped with a two-component epoxy resin automatic mixing injection head, the ratio of components A / B is precisely controlled by a metering pump (adjustable range 1:1-5:1), and the injection pressure is adjustable from 0.1 to 0.5MPa, ensuring that the resin is evenly filled into the damaged gap.

[0102] Repair of large damage (damage diameter ≥10mm, depth ≥5mm): The integrated prepreg fiber cloth automatic laying mechanism can automatically complete the cutting (cutting accuracy ≤±1mm), gripping, laying and roller defoaming of prepreg; it supports multi-layer laying, the laying sequence can be automatically adjusted according to the expert system scheme, and the roller pressure is adjustable from 0.5-2N.

[0103] Precision curing device: Adaptable to different types of repair materials, integrating UV curing and hot air curing dual systems: For UV-curable resins, it is equipped with a 365nm UV lamp assembly with adjustable light intensity and precise control of curing time (10-600s).

[0104] For thermosetting epoxy resin, a hot air circulation system is provided, with a temperature control range of 40-80℃ and a temperature control accuracy of ≤±2℃, to avoid local overheating damage to the blade substrate.

[0105] Automatic spraying device: integrates a miniaturized multi-axis linkage spray gun, equipped with dual material bins for primer and topcoat, with a material bin capacity of ≥500ml; adopts closed-loop flow control technology, and the spraying thickness is uniform and controllable (adjustable from 50-200μm); the spray gun can achieve ±15° oscillation spraying to ensure a smooth transition of the aerodynamic shape between the repair area and the surrounding blade surface.

[0106] In-situ re-inspection device: Reuses a laser 3D scanner and a high-definition camera to quickly re-inspect the repaired area after spraying and curing, and check whether the surface flatness and coating thickness meet the process requirements, forming a comparison of repair quality data.

[0107] (4) Ground control station and expert system Remote control interface: Adopting an industrial-grade touch screen, it provides an immersive visual operation interface that displays the robot status, adsorption force parameters, sensor detection data, blade 3D model and operation progress in real time; it supports one-click start / stop, emergency braking, operation module switching and other operations, and has automatic fault alarm and log recording functions.

[0108] Repair process expert database: Built-in knowledge base of wind turbine blade repair process, covering standard processes such as grinding parameters, material selection, layup scheme, and curing curves corresponding to different damage types and sizes; can automatically generate the optimal repair scheme based on the damage parameters uploaded by the detection module, and supports engineers to manually adjust and confirm before sending it to the robot for execution.

[0109] Digital document management system: Automatically records all data throughout the entire operation process, including detection images, damage parameters, process parameters, and before-and-after comparison data, generating a full life cycle health management file for the blade, supporting export and cloud storage, and providing data support for predictive maintenance of wind farms.

[0110] The following will combine Figure 2 The method for processing the wind turbine blades in this application is described in detail below: S101. Preparation and Task Planning Before the Operation (1) Adjust the fan to a short-term low-speed stop state (speed ≤ 2r / min) to ensure that the blades are in a safe working posture; the staff installs the robot in the designated area at the root of the blade and checks whether the adsorption system, power system and communication system are normal.

[0111] (2) The ground control station imports the three-dimensional model of the blade, sets the full-coverage scanning path parameters (such as scanning spacing and moving speed), and sends the task instructions to the robot.

[0112] S102. Global Scanning and Intelligent Damage Calibration (1) The robot starts autonomous navigation mode and moves along the preset path along the blade length direction. The composite adsorption system automatically switches the adsorption mode according to the blade area to ensure stable adsorption during the movement.

[0113] (2) During the movement, the integrated detection module is started simultaneously. The high-definition camera collects surface images, the laser 3D scanner acquires 3D morphology, and the infrared thermal imager detects internal defects. The multi-sensor data is transmitted to the airborne artificial intelligence processing unit in real time.

[0114] (3) Artificial intelligence algorithms perform data fusion analysis, automatically identify damage type, location, size and grade, generate blade health map, mark all damage points to be repaired, and transmit to ground control station.

[0115] S103. Point-by-point automated repair operation (1) The ground engineer reviews the blade health map and preliminary repair plan, and issues a repair instruction after confirming that there are no errors; the robot autonomously navigates to the first damage point to be repaired, locks the moving mechanism, and starts adsorption force monitoring.

[0116] (2) Cleaning process: The turret is switched to the high-pressure cleaning device. The micro water jet or dry ice cleaning mode is selected according to the type of contaminants on the damaged surface. The damaged area and the surrounding 50mm range are cleaned, and the waste liquid / residue is collected. After completion, the cleaning effect is confirmed by visual inspection.

[0117] (3) Grinding process: The turret is switched to the adaptive constant force grinding device, which automatically plans the grinding path and depth according to the laser three-dimensional scanning data, starts constant pressure grinding, monitors the grinding depth in real time, and forms a standard bevel; the dust collection device is turned on during the grinding process to avoid dust diffusion.

[0118] (4) Drying and preheating process: The turret is switched to the vacuum drying and preheating device. The telescopic cover fits into the grinding area to form a sealed chamber. Vacuum is drawn to remove moisture. At the same time, infrared heating is started for precise preheating. It stops automatically after reaching the preset temperature.

[0119] (5) Filling / Laying process: The repair mode is switched according to the size of the damage. For small damage, the two-component resin injection head is started, and the resin is accurately injected after being mixed in proportion. For large damage, the prepreg laying mechanism is started, and the cutting, laying and rolling defoaming are automatically completed. When laying multiple layers, the laying angle is automatically adjusted.

[0120] (6) Curing process: switch the curing system according to the type of repair material, either UV curing or hot air curing, strictly follow the time and temperature parameters set by the expert system, and monitor the temperature change during the curing process in real time.

[0121] (7) Spraying process: The turret is switched to an automatic spraying device, and the primer and topcoat are sprayed in sequence. The spraying thickness and uniformity are controlled to ensure a smooth transition between the repaired area and the surrounding blade surface, and the aerodynamic shape is restored to the design standard.

[0122] (8) In-situ re-inspection process: The integrated inspection module is reused to perform three-dimensional morphological scanning and visual inspection of the repair area, and the parameters before and after the repair are compared to confirm that the repair quality meets the requirements; if it does not meet the requirements, the secondary repair process is automatically started.

[0123] S104. Cyclic Operations and System Recovery (1) After the first damage point is repaired, the robot autonomously navigates to the next damage point to be repaired and repeats the cleaning-grinding-drying-filling-curing-spraying-re-inspection process in step 3 until all damage points are repaired.

[0124] (2) After all the work is completed, the robot returns to the starting position of the blade root along the preset path, the staff retrieves the robot, and the fan resumes normal operation.

[0125] S105. Digital Archiving The ground control station's file management system automatically organizes data from the entire operation process, generates digital maintenance files that include inspection reports, repair process parameters, and quality re-inspection results, and uploads them to the wind farm operation and maintenance management platform to complete the entire closed-loop process.

[0126] In summary, this application replaces traditional manual high-altitude and high-risk operations with fully automated robotic operations, fundamentally eliminating safety risks such as falls and object strikes. Simultaneously, leveraging the integrated "detection-repair" closed-loop technology, it reduces wind turbine downtime by over 70%, increases operational efficiency by 5 times, and lowers maintenance costs by 40%, achieving a significant leap in economic benefits. In terms of quality control, automated processes and parameter monitoring eliminate human error, ensuring highly standardized repair indicators to extend blade life. Furthermore, the composite adsorption system's excellent adaptability to curved surfaces and low-wind environments increases the effective operating window by 80%.

[0127] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0128] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0129] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0130] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

Claims

1. A wind turbine blade processing system, characterized in that, include: The robot body is attached to the surface of the wind turbine blades; A damage recognition module is fixedly installed at the front end of the robot body, which is used to identify multi-dimensional damage data of the blade based on the blade surface image. A repair module rotatably mounted in the robot body is used to perform corresponding repair procedures after the robot body moves to the damaged area. The repair procedures include: cleaning, grinding, drying, filling, curing, spraying, and re-inspection. The control station is connected to the robot body via a wireless communication link. It is used to generate a repair plan based on the multi-dimensional damage data and generate corresponding timing control commands based on the repair plan and send them to the repair operation module.

2. The wind turbine blade processing system according to claim 1, characterized in that, The robot body includes: a moving mechanism, an adsorption mechanism, and a controller, wherein, The moving mechanism is composed of a composite structure of tracks and wheels. The surface of the tracks is covered with rubber material with a target coefficient of friction, and the wheels have independent lifting and adjustment functions to drive the robot body to move on the surface of the wind turbine blades. The adsorption mechanism includes a permanent magnet adsorption unit and a negative pressure vacuum adsorption unit, which are used to switch the adsorption mode according to the adsorption mode switching command. The controller is connected to the drive circuit of the moving mechanism and the control valve of the composite adsorption system, respectively. It is used to output the corresponding moving control command to the moving mechanism according to the target damage location of the wind turbine blade, and to determine the corresponding target adsorption mode according to the blade surface material. It also generates an adsorption mode switching command based on the target adsorption mode and the current adsorption mode and sends it to the adsorption mechanism.

3. The wind turbine blade processing system according to claim 2, characterized in that, Also includes: Adsorption force monitoring unit and positioning and navigation unit, wherein, An adsorption force monitoring unit installed on the contact surface between the adsorption chamber and the blade surface of the robot body is used to monitor the adsorption force between the robot body and the blade surface and to feed back the adsorption force to the controller of the robot body. The controller is used to output a locking command to the moving mechanism of the robot body when the adsorption force is lower than a preset adsorption force threshold, so as to stop the robot movement and trigger an alarm signal. The positioning and navigation unit connected to the controller of the robot body is used to collect multi-source positioning data and transmit the multi-source positioning data to the controller of the robot body. The controller is used to fuse the multi-source positioning data to calculate the current pose of the robot body on the blade surface, and drive the robot to move to the target position based on the deviation between the current pose and the coordinates of the target damage point.

4. The wind turbine blade processing system according to claim 3, characterized in that, The damage recognition module includes: a first detection unit, a second detection unit, a third detection unit, and a processing unit, wherein, The first detection unit is used to identify the type and location of damage on the blade surface based on the high-definition images of the blade surface acquired; The second detection unit is used to determine the damage parameters of the damage location based on the point cloud scan data; The third detection unit is used to determine the location and extent of internal defects in the blade based on changes in the temperature field on the blade surface. The processing unit is used to generate multi-source data based on the damage type and location of the blade surface, the damage parameters of the damage location, and the location and range of the internal defects of the blade; to generate a damage model by timestamping and aligning the multi-source data with spatial coordinates; and to transmit the damage model to the robot body controller and control station.

5. The wind turbine blade processing system according to claim 4, characterized in that, The repair operation module includes: a turret base, a cleaning device, a grinding device, a drying device, a filling device, a curing device, a spraying device, and a re-inspection device, wherein... The turret base is fixedly installed in the middle of the robot body; the cleaning device, the grinding device, the drying device, the filling device, the curing device, the spraying device and the re-inspection device are sequentially installed on the workstation mounting base of the turret base; The robot's controller, upon receiving the timing control command from the control station, sequentially outputs corresponding commands to the cleaning device, the grinding device, the drying device, the filling device, the curing device, the spraying device, and the re-inspection device to execute the corresponding repair process according to the timing control command.

6. The wind turbine blade processing system according to claim 5, characterized in that, The cleaning device includes a water jet unit and a dry ice cleaning unit, wherein... The water jet unit is used to clean contaminants by the impact force of the water jet, and the impact force value of the water jet is adjusted according to the type of contaminant. The dry ice cleaning unit is used to remove contaminants from the surface of the blades by utilizing the low-temperature embrittlement effect and impact of dry ice particles. The water jet unit and the dry ice cleaning unit are connected to the robot's controller via a switching valve. The controller selects either the water jet unit or the dry ice cleaning unit to clean the blade surface based on the type of contaminants on the damaged surface.

7. The wind turbine blade processing system according to claim 5, characterized in that, The polishing device includes: a polishing head and a target sensor, wherein... The drive end of the grinding head is connected to the controller of the robot body; The target sensor is installed between the grinding head and the robot body to acquire grinding contact force and grinding position data; The controller calculates the grinding path and depth based on the damage data, and controls the grinding head to contact the blade surface with the target pressure to perform grinding based on the grinding path and depth.

8. The wind turbine blade processing system according to claim 5, characterized in that, The drying device includes: a telescopic hood, a vacuum pump, and a heating element, wherein, The telescopic cover is installed at the corresponding work position on the turret base and is used to extend under the control of the controller to form a closed space by fitting the damaged area after the blade is polished. The vacuum pump is installed on the robot body. The suction port of the vacuum pump is connected to the internal chamber of the telescopic cover through a pipeline. It is used to extract air from the closed space after the telescopic cover is attached to the blade, so that the vacuum degree in the chamber reaches a preset threshold to remove moisture from the repair area. The heating element is installed inside the telescopic cover and is used to heat the damaged area of ​​the blade to a preset construction temperature after the vacuum level in the chamber reaches a preset threshold.

9. The wind turbine blade processing system according to claim 5, characterized in that, The filling device includes an injection mechanism and a placement mechanism, wherein... The control end of the injection mechanism is connected to the controller of the robot body, and is used to inject the filler into the target damaged crevices according to the filling ratio. The controller is used to calculate the filling ratio of the corresponding filler according to the damage data of the target damaged crevices. The control terminal of the laying mechanism is connected to the controller of the robot body. It is used to transfer the cut prepreg to the damaged area, control the laying mechanism to position and lay the prepreg in the damaged area according to the number of layers and the layup angle, and control the rolling unit to roll and defoam at a set pressure. The controller is used to determine the number of layers and the layup angle according to the loss data of the damaged area.

10. The wind turbine blade processing system according to claim 5, characterized in that, The spraying device includes: a spray gun, a material hopper, and a flow valve, wherein, The spray gun is installed at the corresponding work station on the turret base. The drive end of the spray gun is connected to the controller of the robot body and is used to spray the damaged area under the control of the controller. The hopper is connected to the spray gun via a supply pipeline and is used to store and supply primer and topcoat; The flow valve is installed on the feed pipeline and is used to adjust the paint output flow of the spray gun according to the coating thickness parameters set by the controller of the robot body.

Images