Inspection methods for wind turbine blades and handling robots

Abstract

This invention relates to the field of wind power equipment testing technology, and provides a method for inspecting wind turbine blades and a handling robot. The method includes: acquiring a wind turbine blade inspection request; responding to the wind turbine blade inspection request, controlling and adjusting the posture of the handling robot so that the ultrasonic probe in the handling robot is in contact with the target wind turbine blade to be inspected; controlling the ultrasonic probe to scan the target wind turbine blade to determine the corresponding ultrasonic image set; analyzing and processing the ultrasonic image set based on a defect detection model to determine the defect detection result corresponding to the target wind turbine blade; the defect detection model adopts a deep learning model, and the training sample set of the defect detection model includes multiple ultrasonic image samples with corresponding defect labels. This achieves fully automated processing from image scanning to defect monitoring result output, improving the inspection efficiency of wind turbine blades and reducing inspection labor costs.

Description

Technical Field

[0001] This invention relates to the field of wind power equipment testing technology, and in particular to a method for testing wind turbine blades and a handling robot. Background Technology

[0002] Wind power is one of the most promising renewable energy technologies for the future. With the continuous construction of large-scale wind farms, the demand for wind turbines (especially the core component of wind turbines, wind turbine blades) is also increasing day by day.

[0003] Due to the large size of wind turbine blades, defects in the connection structure are inevitable during the manufacturing process, reducing the service life of the blades. To ensure the quality of the wind turbine blades, non-destructive testing is necessary.

[0004] Currently, when performing non-destructive testing on wind turbine blades, it is usually necessary to manually fix the scanner to the wind turbine blade, then collect images, and then have experts analyze and identify them. This results in low blade inspection efficiency and requires huge manpower costs.

[0005] Currently, the industry has not proposed a better technical solution to the above problems. Summary of the Invention

[0006] This invention provides a method for inspecting wind turbine blades, a handling robot, and a non-transitory computer-readable storage medium, which at least solves the shortcomings of low inspection efficiency and high labor costs in the prior art for wind turbine blades.

[0007] This invention provides a method for detecting wind turbine blades, applied to a handling robot. The method includes: acquiring a wind turbine blade detection request; responding to the wind turbine blade detection request, controlling and adjusting the pose of the handling robot so that an ultrasonic probe in the handling robot is in contact with the target wind turbine blade to be detected; controlling the ultrasonic probe to scan the target wind turbine blade to determine a corresponding ultrasonic image set; analyzing and processing the ultrasonic image set based on a defect detection model to determine the defect detection result corresponding to the target wind turbine blade; the defect detection model adopts a deep learning model, and the training sample set of the defect detection model includes multiple ultrasonic image samples with corresponding defect labels.

[0008] According to the present invention, a method for detecting wind turbine blades is provided, wherein, in response to a wind turbine blade detection request, the posture of a handling robot is controlled and adjusted so that an ultrasonic probe in the handling robot is in contact with a target wind turbine blade to be detected. The method includes: acquiring environmental data of the surrounding environment based on the wind turbine blade detection request; determining the fixture position of the target fixture based on the environmental data if a target fixture for placing the target wind turbine blade to be detected is identified from the environmental data; and controlling and adjusting the posture of the handling robot according to the fixture position so that the ultrasonic probe in the handling robot is in contact with the target wind turbine blade.

[0009] According to the present invention, a method for detecting wind turbine blades is provided. Before controlling an ultrasonic probe to scan the target wind turbine blade to determine the corresponding ultrasonic image set, the method further includes: determining blade identification information corresponding to the target wind turbine blade; wherein, after analyzing and processing the ultrasonic image set based on a defect detection model to determine the defect detection result corresponding to the target wind turbine blade, the method further includes: associating the blade identification information and the defect detection result, and sending the associating blade identification information and the defect detection result to a management server.

[0010] According to the present invention, a method for detecting wind turbine blades is provided, wherein associating the blade identification information and the defect detection result, and sending the associated blade identification information and the defect detection result to a management server, comprises: determining, in the ultrasonic image set, an ultrasonic image corresponding to the defect detection result; associating the blade identification information, the defect detection result, and the ultrasonic image corresponding to the defect detection result, and sending the associated blade identification information, the defect detection result, and the ultrasonic image corresponding to the defect detection result to a management server.

[0011] According to the present invention, a method for detecting wind turbine blades is provided, wherein controlling the ultrasonic probe to scan the target wind turbine blade to determine a corresponding ultrasonic image set includes: acquiring a preset probe scanning speed for the ultrasonic probe; determining a robot moving speed according to the probe scanning speed; moving along the target wind turbine blade according to the robot moving speed, and controlling the ultrasonic probe to scan the target wind turbine blade while the transport robot is moving, thereby determining a corresponding ultrasonic image set.

[0012] According to the present invention, a method for detecting wind turbine blades is provided. The step of determining the robot's moving speed based on the probe scanning speed includes: determining the probe's lateral displacement speed based on the probe scanning speed; the lateral displacement speed being the speed at which the ultrasonic probe moves laterally along the target wind turbine blade; obtaining the blade width at the point where the target wind turbine blade contacts the ultrasonic probe, and determining the robot's moving speed based on the blade width and the probe scanning speed; the robot's moving speed being the speed at which the handling robot moves longitudinally along the target wind turbine blade.

[0013] According to the present invention, a method for detecting wind turbine blades is provided, wherein controlling the ultrasonic probe to adhere to the target wind turbine blade and scan it while the transport robot is moving to determine a corresponding ultrasonic image set includes: detecting the real-time adhesion force between the ultrasonic probe and the target wind turbine blade while the transport robot is moving; controlling and adjusting the distance of the ultrasonic probe relative to the target wind turbine blade according to the real-time adhesion force and a preset adhesion force threshold; and controlling the adjusted ultrasonic probe to adhere to the target wind turbine blade and scan it while the transport robot is moving to determine a corresponding ultrasonic image set.

[0014] According to the present invention, a method for detecting wind turbine blades is provided. The transport robot further includes a probe water boot disposed around the ultrasonic probe and a water tank for holding water. The method further includes, before controlling the ultrasonic probe to scan the target wind turbine blade to determine the corresponding ultrasonic image set, turning on the water supply switch of the transport robot to control the water tank to supply water to the probe water boot, so as to provide coupling water for the ultrasonic probe to scan the target wind turbine blade.

[0015] According to the present invention, a method for detecting wind turbine blades is provided. The handling robot further includes a water collection module and a water filtration module, wherein the water collection module is used to recycle the coupling water, and the water filtration module is used to filter the coupling water recycled by the water collection module and store the filtered coupling water in the water tank.

[0016] The present invention also provides a handling robot, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of any of the wind turbine blade detection methods described above.

[0017] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the wind turbine blade detection method as described above.

[0018] The wind turbine blade inspection method, handling robot, and storage medium provided by this invention enable the AGV (Automated Guided Vehicle) to adjust its posture to align the ultrasonic probe with the target wind turbine blade when it receives a wind turbine blade inspection request. The AGV controls the ultrasonic probe to scan the target wind turbine blade and uses a deep learning model to identify the blade defects corresponding to the ultrasonic scan image. This achieves fully automated processing from image scanning to defect monitoring result output, effectively improving the inspection efficiency of wind turbine blades and reducing inspection labor costs. Attached Figure Description

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

[0020] Figure 1 A flowchart illustrating an example of a method for detecting wind turbine blades according to an embodiment of the present invention is shown;

[0021] Figure 2 It shows that according to Figure 1 An example operation flowchart of step S120 in the process;

[0022] Figure 3 It shows that according to Figure 1 An example operation flowchart of step S130 in the process;

[0023] Figure 4 A flowchart illustrating an example of a method for detecting wind turbine blades according to an embodiment of the present invention is shown;

[0024] Figure 5 A structural block diagram of an example of a handling robot according to an embodiment of the present invention is shown;

[0025] Figure 6 A signal timing flowchart of an example of a wind turbine blade detection method according to an embodiment of the present invention is shown;

[0026] Figure 7 This is a structural schematic diagram of the handling robot provided by the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0028] It should be noted that phased array ultrasonic testing technology is a non-destructive testing technology applied to the injection and bonding quality testing of wind turbine blades. Compared with conventional ultrasonic testing technology, it has higher testing efficiency and accuracy.

[0029] In current related technologies, some industry experts and scholars have proposed using phased array scanners for non-destructive testing of wind turbine blades. However, phased array scanners require manual attachment to the wind turbine blades via pneumatic suction cups. Power outages or air supply failures can cause the scanner to detach from the blade, damaging the equipment. Furthermore, phased array scanners consist of two sets of toothed belt drives, one longitudinal and one transverse, connected at right angles. Since the relative angle between the transverse drive and the wind turbine blade under inspection is fixed, the scanner probe's compensation on the transverse drive cannot adapt to the significant curvature changes on the blade surface during longitudinal movement. This results in poor coupling between the scanner probe and the blade, affecting the accuracy of the scan results. Additionally, the low transmission precision of the toothed belts also impacts the accuracy of the scan results. Furthermore, after obtaining the scanning results, professionals with mature analytical skills are needed to analyze the results. Manual analysis involves a large workload, resulting in low efficiency in the detection of wind turbine blades, and manual analysis is also prone to errors and omissions.

[0030] Figure 1 A flowchart illustrating an example of a method for detecting wind turbine blades according to an embodiment of the present invention is shown.

[0031] Regarding the execution subject of the method in the embodiments of the present invention, it can be any controller or processor with computing or processing capabilities to achieve the goal of controlling the AGV to intelligently detect the wind turbine blades. In some examples, it can be integrated and configured in the AGV through software, hardware, or a combination of software and hardware, and there should be no limitation thereto.

[0032] The following description uses an AGV as an exemplary implementation to illustrate the details of the technical solution involved in this invention. However, it should be understood that one or more steps involved in the following processes can be implemented by one or more controllers or software installed and deployed in the AGV.

[0033] like Figure 1 As shown, in step S110, a wind turbine blade detection request is obtained.

[0034] In one example of this invention, the AGV is equipped with interactive controls, allowing operators to interact with the AGV on-site to trigger the generation of a wind turbine blade detection request. In another example of this invention, the AGV can also receive wind turbine blade detection requests remotely, enabling remote triggering of wind turbine blade detection.

[0035] In some implementations, wind turbine blade inspection requests may be triggered via a WMS (Warehouse Management System) or MOM (Manufacturing Operations Management) system. For example, a management user may trigger an AGV to inspect the wind turbine blades by operating a client integrated with a WMS or MOM system.

[0036] In step S120, in response to the wind turbine blade detection request, the pose of the handling robot is adjusted so that the ultrasonic probe in the handling robot is in contact with the target wind turbine blade to be detected. It should be understood that pose is a technical term in robot control theory, which can represent the position and orientation of the AGV in a specified coordinate system.

[0037] In one example of this invention, the AGV can be adjusted based on preset pose parameters to achieve the bonding of the ultrasonic probe with the target wind turbine blade at a fixed blade inspection station. In another example of this invention, the AGV can also intelligently locate one or more wind turbine blades with defects to be identified and bond them using various intelligent control methods.

[0038] In step S130, the ultrasonic probe is controlled to scan the target wind turbine blades to determine the corresponding ultrasonic image set.

[0039] It should be understood that ultrasound probes can use various optimal scanning resolutions. For example, the scanning resolution used should meet the design requirements of phased array scanning technology, and there should be no restrictions on it.

[0040] In one example of this invention, only the core area (e.g., the main beam) of the target wind turbine blade needs to be scanned to determine whether a specific local defect exists on the blade. In another example of this invention, most of the target wind turbine blade or the entire blade needs to be scanned to identify whether a defect exists on the blade.

[0041] It should be noted that wind turbine blades are relatively large and their surfaces have various curvatures in localized areas. Therefore, multiple shots from the ultrasonic probe are typically required to obtain a complete set of ultrasonic images. In some implementations, an AGV can acquire ultrasonic images through multiple specific pose adjustments and corresponding scanning operations. For example, the AGV can move along the wind turbine blades and simultaneously perform scanning operations.

[0042] In step S140, the ultrasonic image set is analyzed and processed based on the defect detection model to determine the defect detection result corresponding to the target wind turbine blade.

[0043] Here, the defect detection model employs a deep learning model, and the training sample set for the defect detection model includes multiple ultrasound image samples with corresponding defect labels. It should be understood that the deep learning model can use various model structures for image target classification, such as DNN or CNN neural network model structures, etc., and no restrictions are imposed here. Furthermore, the defect labels can be general classification labels, such as "defective" or "no defect," or more detailed classification labels based on defect indicators, such as "injection defective" and "injection defect-free," "bonding defective" and "bonding defect-free," "web offset defective" and "web offset-free," "structural adhesive width defective" and "structural adhesive width-free," and so on.

[0044] In some implementations, the defect detection model can inspect each image in the ultrasonic image set of the blades one by one. When a defect is identified in any ultrasonic image of the target wind turbine blade, the defect in the target wind turbine blade can be determined. Preferably, when the defect detection model determines that the target wind turbine blade has a defect, it can further output the category of the corresponding defect index.

[0045] In this embodiment of the invention, the AGV adjusts its posture to bring the ultrasonic probe into contact with the wind turbine blade for scanning. After scanning, a defect detection model based on deep learning is used to identify the corresponding defect detection results. Therefore, compared to scanning wind turbine blades with a phased array scanner, this embodiment eliminates the need for manual fixation of the phased array scanner to the blade, saving labor costs and completely avoiding the risk of the scanner detaching from the blade. Furthermore, the deep learning model identifies blade defects in the scanned image, eliminating the need for professional personnel, improving identification efficiency, and lowering the operational threshold for non-destructive testing of blades.

[0046] Figure 2 It shows that according to Figure 1 An example operation flowchart of step S120 in the process.

[0047] like Figure 2As shown, in step S210, environmental data for the surrounding environment is collected based on the wind turbine blade detection request.

[0048] For example, when a wind turbine blade inspection request is received, the AGV vehicle body distance sensor and safety obstacle avoidance system are used to scan the surrounding environment to obtain the corresponding environmental data.

[0049] In step S220, if a target tooling for placing the target wind turbine blade to be inspected is identified from the environmental data, the tooling position of the target tooling is determined based on the environmental data.

[0050] In some examples of embodiments of the present invention, a pre-trained tooling recognition model is used to identify environmental data to determine whether a target tooling exists in the environmental data. The tooling recognition model may also employ a deep learning model structure. Furthermore, the identification of the target tooling and its location can be achieved through various ultrasonic target recognition and ultrasonic positioning techniques, without limitation.

[0051] In step S230, the pose of the transport robot is adjusted according to the tooling position, so that the ultrasonic probe in the transport robot is in contact with the target fan blade. Specifically, the pose of the AGV can be adjusted by controlling the torque detection and closed-loop control system of the scanning mechanism.

[0052] In some implementations, the AGV can calculate coordinates based on the tooling position and its current position to plan its movement path. Then, when the AGV moves to the tooling position, its posture is adjusted using preset posture parameters for aligning with the wind turbine blades, ensuring the ultrasonic probe is in contact with the target wind turbine blade. This allows the AGV to automatically sense the position of the target wind turbine blades and adjust its direction and distance of travel as it moves forward.

[0053] In addition, a winding and unwinding structure is deployed on the AGV body to automatically wind up and unwind the cable, so that the AGV body will not be tripped by the cable when performing posture adjustment operations.

[0054] In some business scenarios, AGVs are also equipped with distance sensors and safety obstacle avoidance systems, which automatically adjust the distance between the AGV and the blades, automatically identify obstacles, avoid interference from objects on site, and achieve automatic safe path planning.

[0055] In this embodiment of the invention, the AGV collects and analyzes environmental images, automatically locates the target wind turbine blades that need to be ultrasonically inspected, and adjusts its posture to fit the target wind turbine blades. Thus, the AGV can automatically identify each wind turbine blade to be inspected in the environment and complete the defect inspection of each blade one by one, eliminating the need for manual placement of each blade to be inspected sequentially at a fixed blade inspection station. This effectively reduces labor costs and improves inspection efficiency.

[0056] Figure 3 It shows that according to Figure 1 An example operation flowchart for step S130.

[0057] In step S310, a preset probe scanning speed for the ultrasound probe is obtained. Here, the probe scanning speed can be preset to ensure that the probe acquires a sufficiently clear image.

[0058] In some examples of embodiments of the present invention, the ultrasonic probe can be mounted on the lifting axis of the handling robot. Specifically, the lifting axis is relative to the axes of the robot body. For example, for a six-axis AGV, the lifting axis can be an additional, separately set seventh axis, which works with the AGV to increase the scanning height of the equipment, meet the scanning requirements of the large and small webs of wind turbine blades of different sizes, reduce the equipment changeover cost and the development cost of different AGV models.

[0059] In step S320, the robot's moving speed is determined based on the probe scanning speed.

[0060] In one example of this invention, the ultrasonic probe is relatively stationary to the AGV during ultrasonic scanning. For instance, the probe's scanning width is sufficient to cover the width of the blade's main beam, and it can be moved by the AGV to scan different positions on the wind turbine blade. Accordingly, the AGV's movement speed can be set based on the probe's required scanning speed.

[0061] In another example of the present invention, the ultrasonic probe moves relative to the AGV during the ultrasonic scanning process performed by the AGV. For example, if the scanning width of the ultrasonic probe is insufficient to cover the width of the blade main beam or if the ultrasonic probe is required to perform a global scan of the blade, the probe needs to move laterally to achieve a wider range of scanning of the target wind turbine blade.

[0062] Specifically, the AGV determines the lateral displacement speed of the ultrasonic probe based on the probe scanning speed. Here, the lateral displacement speed is the speed at which the ultrasonic probe moves laterally along the target wind turbine blade, thereby ensuring the quality of the ultrasonic images acquired during the probe's lateral movement. Furthermore, the AGV obtains the blade width at the point where the ultrasonic probe contacts the target wind turbine blade, and determines the robot's longitudinal movement speed along the target wind turbine blade based on the blade width and the probe scanning speed.

[0063] It should be noted that the longitudinal width of the wind turbine blades varies; for example, the blade width near the blade tip is smaller than that near the blade root. Specifically, when using an AGV to scan the target wind turbine blade, the current blade width corresponding to the current position of the probe is collected, and the lateral movement time required for the ultrasonic probe to complete the current blade width is calculated. This, in turn, yields the longitudinal movement speed of the AGV. For example, the AGV needs to complete the lateral scan of the blade before performing a longitudinal step movement.

[0064] In step S330, the robot moves along the target wind turbine blade according to its moving speed, and the ultrasonic probe is controlled to scan the target wind turbine blade while the robot is moving to determine the corresponding ultrasonic image set. Therefore, maintaining the ultrasonic probe in contact with the target wind turbine blade during AGV movement ensures high-resolution ultrasonic images acquired through the ultrasonic probe scan.

[0065] It should be noted that the curvature of some local areas of wind turbine blades varies. In current related technologies, the lateral transmission device of the phased array scanner can only achieve scanning at a fixed relative distance to the wind turbine blade, which cannot adapt to the curvature changes of the wind turbine blade, resulting in unsatisfactory results.

[0066] Therefore, regarding step S140 above, in some embodiments, when the AGV moves, the AGV can detect the real-time contact force between the ultrasonic probe and the target fan blade using a pressure sensor, and control the adjustment of the distance between the ultrasonic probe and the target fan blade based on the real-time contact force and a preset contact force threshold. For example, when the real-time contact force is less than the contact force threshold, the distance between the ultrasonic probe and the target fan blade is controlled to decrease; conversely, when the real-time contact force is greater than the contact force threshold, the distance between the ultrasonic probe and the target fan blade is controlled to increase. Furthermore, the adjusted ultrasonic probe is controlled to contact the target fan blade and scan it while the transport robot moves to determine the corresponding ultrasonic image set.

[0067] Therefore, when the AGV moves to scan the target wind turbine blade, the ultrasonic probe can maintain constant force contact with the target wind turbine blade in real time in the blade area with diverse curvature, which can ensure the high definition of the acquired ultrasonic images and provide a basis for the high accuracy of the defect detection results of the wind turbine blade.

[0068] Figure 4 A flowchart illustrating an example of a method for detecting wind turbine blades according to an embodiment of the present invention is shown.

[0069] like Figure 4 As shown, in step S410, a wind turbine blade detection request is received from the management server.

[0070] Here, the management server can be a server corresponding to a WMS or MOM system, which can maintain the inspection process and results of the wind turbine blades before they leave the factory. In some implementations, the management server can also generate inspection reports for each wind turbine blade to ensure the quality of the wind turbine blades leaving the factory.

[0071] In step S420, in response to the wind turbine blade detection request, the position of the handling robot is controlled and adjusted so that the ultrasonic probe in the handling robot is in contact with the target wind turbine blade to be detected.

[0072] In step S430, the blade identification information corresponding to the target wind turbine blade is determined.

[0073] In one example of this invention, each wind turbine blade has a unique blade identifier, and the AGV can obtain the blade identifier information corresponding to the target wind turbine blade to be detected by scanning. In another example of this invention, the AGV is equipped with an automated blade identifier generation algorithm, which can automatically assign blade identifier information to the target wind turbine blade to be detected.

[0074] In step S440, the ultrasonic probe is controlled to scan the target wind turbine blades to determine the corresponding ultrasonic image set.

[0075] In step S450, the ultrasonic image set is analyzed and processed based on the defect detection model to determine the defect detection result corresponding to the target wind turbine blade.

[0076] In step S460, the blade identification information and defect detection results are associated, and the associated blade identification information and defect detection results are sent to the management server.

[0077] Therefore, the management server can accurately obtain the defect detection results of each blade identifier, which is beneficial for management users to implement targeted management strategies. For example, when multiple blades are found to have the same defect, rectification suggestions can be generated to urge the wind turbine blade production line to rectify the production of the blades.

[0078] In some examples of embodiments of the present invention, the AGV can also determine the ultrasonic image corresponding to the defect detection result in the ultrasonic image set, and then associate the blade identification information, the defect detection result and the ultrasonic image corresponding to the defect detection result, and send the associated blade identification information, the defect detection result and the ultrasonic image corresponding to the defect detection result to the management server.

[0079] Therefore, in addition to accurately knowing the defect detection results of each blade identification, the management server can also obtain ultrasonic images with corresponding defect detection results, which makes it convenient for management users to review the intelligently identified defect detection results, avoid the occasional calculation errors of the intelligent algorithm, and further ensure the reliability of the final wind turbine blade detection results.

[0080] Figure 5 A structural block diagram of an example of a handling robot according to an embodiment of the present invention is shown.

[0081] like Figure 5 As shown, the transport robot 500 includes a probe water boot 510, a water tank 520 for holding water, a water collection module 530, and a water filtration module 540. Specifically, the probe water boot 510 is positioned around the ultrasonic probe. Specifically, before controlling the ultrasonic probe to scan the target wind turbine blades, the water supply switch (not shown) of the transport robot is turned on to control the water tank to supply water to the probe water boot 510, providing coupling water for the ultrasonic probe to scan the target wind turbine blades.

[0082] Therefore, by using the probe water boot 510 to provide coupling water around the ultrasonic probe, providing a coupling agent for ultrasonic scanning, the coupling effect between the fan blade and the probe is enhanced, which can further improve the high definition of the ultrasonic image determined by the ultrasonic probe scanning.

[0083] Furthermore, the water collection module 530 and the water filtration module 540 are connected in series with the water tank 520. Specifically, the water collection module 530 is used to recycle coupling water. For example, the water collection module 530 can be located below the target fan blades to collect coupling water flowing from the fan blades. The water filtration module 540 is used to filter the coupling water recycled by the water collection module 530 and store the filtered coupling water in the water tank 520 to optimize the water quality of the recycled water flowing into the water tank 520. Thus, the AGV integrates a water supply system and a circulating water system. By recycling the ultrasonic coupling water, it realizes automatic water supply during the scanning process, as well as water filtration and recycling, avoiding water waste and preventing coupling water from spilling onto the ground. This effectively solves the problem of coupling water contaminating the ground in the blade inspection workshop, ensuring that the workshop can achieve 6S hygiene standards.

[0084] Figure 6 A signal timing flowchart of an example of a wind turbine blade detection method according to an embodiment of the present invention is shown.

[0085] like Figure 6 As shown, the entire process of automated inspection of wind turbine blades is achieved through data interaction between the MOM system 601, WMS system 603, AGV scheduling system 605 and target AGV 607.

[0086] Specifically, upon detecting a blade inspection requirement, the MOM system 601, WMS system 603, and AGV scheduling system 605 synchronize the wind turbine blade inspection task. Furthermore, the AGV scheduling system 605 can determine the target AGV 607 to perform the wind turbine blade task and control the target AGV 607 to move to the vicinity of the blade to be inspected.

[0087] As the target AGV 607 approaches the blade, it scans the blade number and feeds it back to the MOM system 601, thus binding the detection result with the blade number. At this point, the target AGV 607 controls and adjusts its six and seventh axes to the preset posture according to the preset posture parameters. Then, the target AGV 607 can activate the water supply system and the circulating water system with water filtration function to apply coupled water to the detection area of ​​the wind turbine blade. Next, the target AGV 607 can control the ultrasonic probe to continuously scan the work area, and the scanning results and images are automatically recorded in the database and bound to the blade number. Furthermore, the target AGV 607 automatically adjusts its movement speed according to the probe scanning speed and moves along the blade. For example, a distance sensor on the vehicle body measures the distance and uses an algorithm to adjust the displacement in real time. In addition, during movement, the scanning mechanism automatically adjusts the distance between the scanning probe and the blade based on the pressure sensor, maintaining constant force contact at all times. Optionally, the target AGV 607 can also identify the blade storage fixture and travel along a planned path based on an obstacle avoidance algorithm. Furthermore, after completing a scan of one side of the blade, the target AGV 607 can automatically move to the initial position on the other side of the blade and begin automatic scanning.

[0088] After the target AGV 607 completes the scanning of the entire blade, interactive items (thresholds, data rulers, buttons, etc.) on the interface can be automatically dragged and dropped by the drive analysis software. With the video stream capture function, video frame data is generated, and deep learning algorithms are used for defect identification, generating defect detection results and publishing them to the MOM system 601. Furthermore, for abnormal defect detection results, the target AGV 607 can automatically upload corresponding images for manual review and analysis by quality inspectors. Once confirmed, the MOM system 601 publishes an inspection report. Subsequently, the target AGV 607 can automatically move to a preset storage location and upload a standby status to the system, awaiting the next inspection trigger.

[0089] This enables real-time automatic storage of scanning imaging results, while the data platform facilitates real-time data transmission. Automatic analysis software performs automatic analysis on the images, and images with anomalies are saved separately. The analysis results are then manually reviewed.

[0090] Figure 7 An example is a schematic diagram of the physical structure of a transport robot, such as... Figure 7As shown, the handling robot may include a processor 710, a communication interface 720, a memory 730, and a communication bus 740. The processor 710, communication interface 720, and memory 730 communicate with each other via the communication bus 740. The processor 710 can call logical instructions in the memory 730 to execute a wind turbine blade detection method. This method includes: acquiring a wind turbine blade detection request; responding to the wind turbine blade detection request, controlling and adjusting the pose of the handling robot so that the ultrasonic probe in the handling robot is in contact with the target wind turbine blade to be detected; controlling the ultrasonic probe to scan the target wind turbine blade to determine a corresponding ultrasonic image set; and analyzing and processing the ultrasonic image set based on a defect detection model to determine the defect detection result corresponding to the target wind turbine blade. The defect detection model employs a deep learning model, and the training sample set of the defect detection model includes multiple ultrasonic image samples with corresponding defect labels.

[0091] Furthermore, the logical instructions in the aforementioned memory 730 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0092] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, wherein when the program instructions are executed by a computer, the computer is able to execute the wind turbine blade detection method provided by the above methods, the method comprising: acquiring a wind turbine blade detection request; responding to the wind turbine blade detection request, controlling and adjusting the pose of a handling robot such that an ultrasonic probe in the handling robot is in contact with the target wind turbine blade to be detected; controlling the ultrasonic probe to scan the target wind turbine blade to determine a corresponding ultrasonic image set; analyzing and processing the ultrasonic image set based on a defect detection model to determine the defect detection result corresponding to the target wind turbine blade; the defect detection model employs a deep learning model, and the training sample set of the defect detection model includes multiple ultrasonic image samples with corresponding defect labels.

[0093] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the aforementioned methods for detecting wind turbine blades. The method includes: acquiring a wind turbine blade detection request; responding to the wind turbine blade detection request, controlling and adjusting the pose of a handling robot such that an ultrasonic probe in the handling robot is in contact with a target wind turbine blade to be detected; controlling the ultrasonic probe to scan the target wind turbine blade to determine a corresponding ultrasonic image set; and analyzing and processing the ultrasonic image set based on a defect detection model to determine a defect detection result corresponding to the target wind turbine blade. The defect detection model employs a deep learning model, and the training sample set of the defect detection model includes multiple ultrasonic image samples with corresponding defect labels.

[0094] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0095] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting wind turbine blades, characterized in that, Applied to a handling robot, the method includes: Get wind turbine blade inspection request; In response to the wind turbine blade detection request, the position of the handling robot is controlled and adjusted so that the ultrasonic probe in the handling robot is in contact with the target wind turbine blade to be detected; The ultrasonic probe is controlled to scan the target wind turbine blades to determine the corresponding ultrasonic image set; The ultrasonic image set is analyzed and processed based on a defect detection model to determine the defect detection result corresponding to the target wind turbine blade; the defect detection model adopts a deep learning model, and the training sample set of the defect detection model includes multiple ultrasonic image samples with corresponding defect labels; The step of controlling and adjusting the pose of the handling robot in response to the wind turbine blade detection request, so that the ultrasonic probe in the handling robot is in contact with the target wind turbine blade to be detected, includes: Based on the wind turbine blade detection request, environmental data of the surrounding environment is collected. If a target tooling for placing a target wind turbine blade to be inspected is identified from the environmental data, the tooling position of the target tooling is determined based on the environmental data. Based on the tooling position, the pose of the handling robot is controlled and adjusted so that the ultrasonic probe in the handling robot fits into the target wind turbine blade; The process of controlling the ultrasonic probe to scan the target wind turbine blades to determine the corresponding ultrasonic image set includes: Obtain the preset probe scanning speed for the ultrasound probe; The robot's moving speed is determined based on the probe's scanning speed; The robot moves along the target wind turbine blade at the specified moving speed, and the ultrasonic probe is controlled to scan the target wind turbine blade while the transport robot is moving, so as to determine the corresponding ultrasonic image set. Determining the robot's moving speed based on the probe's scanning speed includes: Based on the probe scanning speed, the probe lateral displacement speed is determined; the probe lateral displacement speed is the speed at which the ultrasonic probe moves laterally along the target wind turbine blade. The width of the target wind turbine blade at the point where it contacts the ultrasonic probe is obtained. The lateral movement time required for the ultrasonic probe to complete the current blade width is calculated. Based on the blade width, the probe scanning speed, and the lateral movement time, the robot's moving speed is determined. The robot's moving speed is the speed at which the handling robot moves longitudinally along the target wind turbine blade. The control of the ultrasonic probe to scan the target wind turbine blade while the transport robot moves, in order to determine the corresponding ultrasonic image set, includes: As the transport robot moves, the real-time contact force between the ultrasonic probe and the target wind turbine blade is detected; Based on the real-time adhesion force and a preset adhesion force threshold, the distance between the ultrasonic probe and the target wind turbine blade is controlled and adjusted; wherein, when the real-time adhesion force is less than the adhesion force threshold, the distance between the ultrasonic probe and the target wind turbine blade is controlled to decrease; when the real-time adhesion force is greater than the adhesion force threshold, the distance between the ultrasonic probe and the target wind turbine blade is controlled to increase. The adjusted ultrasonic probe is controlled to scan the target wind turbine blade as the transport robot moves to determine the corresponding set of ultrasonic images.

2. The method for detecting wind turbine blades according to claim 1, characterized in that, Before controlling the ultrasonic probe to scan the target wind turbine blades to determine the corresponding ultrasonic image set, the method further includes: Determine the blade identification information corresponding to the target wind turbine blade; The method further includes, after analyzing and processing the ultrasonic image set based on the defect detection model to determine the defect detection result corresponding to the target wind turbine blade: The blade identification information and the defect detection result are associated, and the associated blade identification information and the defect detection result are sent to the management server.

3. The method for detecting wind turbine blades according to claim 2, characterized in that, The process of associating the blade identification information and the defect detection result, and sending the associated blade identification information and the defect detection result to the management server, includes: In the set of ultrasonic images, determine the ultrasonic image corresponding to the defect detection result; The blade identification information, the defect detection result, and the corresponding ultrasonic image are associated, and the associated blade identification information, the defect detection result, and the corresponding ultrasonic image are sent to the management server.

4. The method for detecting wind turbine blades according to claim 1, characterized in that, The transport robot also includes a water tank for holding water and probe water boots arranged around the ultrasonic probe. The method further includes, prior to controlling the ultrasonic probe to scan the target wind turbine blades to determine the corresponding ultrasonic image set: Turn on the water supply switch of the transport robot to control the water tank to supply water to the probe water boot, so as to provide coupling water for the ultrasonic probe to scan the target wind turbine blade.

5. The method for detecting wind turbine blades according to claim 4, characterized in that, The transport robot also includes a water collection module and a water filtration module, wherein the water collection module is used to recycle the coupling water, and the water filtration module is used to filter the coupling water recycled by the water collection module and store the filtered coupling water in the water tank.

6. A handling robot, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the wind turbine blade detection method as described in any one of claims 1 to 5.

Images