Device and method for detecting welding seam of tower drum of wind generating set

Abstract

The invention discloses a wind generating set tower tube welding seam detection device and method, and relates to the technical field of welding seam detection, the wind generating set tower tube welding seam detection device comprises an annular frame main body, the annular frame main body is correspondingly sleeved on a tower tube, and the inner ring diameter of the annular frame main body is greater than the maximum diameter of the tower tube; a driving mechanism is mounted on the inner ring surface of the ring frame main body and is used for driving the detection equipment to move along the circumference of the ring frame main body; the detection equipment is composed of a camera, an infrared sensor and an ultrasonic sensor; at least three walking mechanisms are hinged to the top of the ring frame body, the walking mechanisms are used for moving along the outer wall of the tower drum, and the multiple walking mechanisms are evenly distributed; the problem that traditional welding seam detection is low in precision is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of weld detection, and in particular to a device and method for detecting welds on a tower of a wind turbine generator set. Background Art

[0002] With the increasing installed capacity of wind power, the operational safety of wind turbines is receiving increasing attention, particularly the potential safety hazards posed by defects in wind turbine tower welds. As the sole supporting component of a wind turbine, the wind turbine tower's safe operation is paramount. The welds on the tower's exterior typically include circumferential welds between segments and linear welds within each segment. Over years of operation, these welds are prone to cracks, erosion, and other defects, posing a serious threat to the tower's operational safety. Therefore, nondestructive testing of tower exterior welds has become a key focus.

[0003] The key measure to solve the above problems is to conduct regular non-destructive testing of tower welds to prevent accidents. However, since tower welds are located on vertical outer walls at high altitudes, their detection is relatively difficult. Currently, most people use manual high-altitude handheld detectors for this purpose. However, due to the strong crosswinds at high altitudes, this method is extremely dangerous and can easily cause safety accidents. In addition, manual inspections are also prone to missed inspections. In addition, there are currently plans to use wall-climbing robots instead of manual labor. However, when inspecting tower girth welds, single-unit wall-climbing robots still have the problem of deviating from their trajectory due to gravity, making it impossible to detect long girth welds.

[0004] The Chinese patent with publication number CN118533975A discloses a device and method for detecting welds on wind turbine towers. By installing a non-destructive testing unit on an annular slide and using a walking device to replace the existing technology of manual inspection of wind turbine tower welds, it effectively solves a series of problems that may be caused by manual high-altitude operations, such as personnel safety, low efficiency, high missed detection rate, and high cost. Installing the non-destructive testing unit on the annular slide avoids the problem of the probe being unable to align with the weld during inspection by the walking device, thereby improving the inspection efficiency and inspection area. However, the device uses a walking device to drive the annular slide, resulting in uneven force on the walking device, thereby affecting the circumferential inspection accuracy of the non-destructive testing unit.

[0005] Therefore, we propose a weld detection device and method with high detection accuracy. Summary of the Invention

[0006] The object of the present invention is to provide a wind turbine tower weld detection device and method, which are used to solve the problem of low accuracy of traditional weld detection.

[0007] The present invention is achieved through the following technical solutions:

[0008] A wind turbine tower weld inspection device includes a ring frame body, the ring frame body being sleeved on the tower, and the inner diameter of the ring frame body being larger than the maximum diameter of the tower; a driving mechanism being mounted on the inner surface of the ring frame body, the driving mechanism being used to drive a detection device to move along the circumference of the ring frame body;

[0009] The detection device is composed of a camera, an infrared sensor and an ultrasonic sensor;

[0010] At least three walking mechanisms are hinged on the top of the ring frame body. The walking mechanisms are used to move along the outer wall of the tower, and the multiple walking mechanisms are evenly distributed.

[0011] Furthermore, the inner ring surface of the ring frame body is provided with an annular track, and the driving mechanism includes a driving trolley installed in the annular track, and a detection device is installed on one end of the driving trolley corresponding to the ring frame body.

[0012] Furthermore, the walking mechanism includes an articulated block, an articulated shaft, a mounting plate and a permanent magnet wheel. The articulated block is fixedly mounted on the bottom of the ring frame body, the articulated shaft passes through the articulated block, and both ends of the articulated shaft are fixedly connected to a mounting plate respectively. Both mounting plates are rotatably connected to the permanent magnet wheel, and any mounting plate is provided with a walking motor for driving the permanent magnet wheel to rotate.

[0013] Furthermore, the number of the walking mechanisms is four.

[0014] Furthermore, an encoder is installed on the axle of the permanent magnet wheel.

[0015] A method for detecting weld seams of a wind turbine tower, comprising:

[0016] Obtain all weld data, as well as motion data of the traveling mechanism and driving mechanism;

[0017] Preprocess all acquired weld data;

[0018] Using the data processing model, the defective welds are judged based on all the pre-processed weld data;

[0019] Construct a 3D model of the tower structure;

[0020] Combine the motion data of the traveling mechanism and the driving mechanism, and mark the location of the defective weld in the 3D model.

[0021] Furthermore, the method for obtaining all weld seam data and motion data of the walking mechanism and the driving mechanism is specifically as follows:

[0022] The control drive mechanism drives the detection equipment to move, and the camera locates the weld position;

[0023] Start the walking mechanism and record its motion data synchronously. At the same time, infrared sensors and ultrasonic sensors acquire the detection data of the linear weld.

[0024] When the end of the straight weld is reached and the data of a straight weld is obtained, the traveling mechanism stops;

[0025] Then the drive mechanism is started, and the motion data of the walking mechanism is recorded synchronously, and the infrared sensor and ultrasonic sensor are controlled to obtain the detection data of the annular weld;

[0026] After the driving mechanism moves around the ring frame body for one circle, the data of a circular weld is obtained;

[0027] Finally, repeat the above steps until you reach the top of the tower and obtain all the weld data.

[0028] Furthermore, all weld data obtained by the preprocessing specifically include:

[0029] The infrared sensor raw data and ultrasonic sensor raw data in each section of weld data are filtered separately;

[0030] Normalize each weld seam data after filtering;

[0031] Time synchronization is performed on all normalized weld data.

[0032] Furthermore, the specific steps of using the data processing model to determine defective welds based on the pre-processed linear weld data and circular weld data are as follows:

[0033] Extract the temperature distribution characteristics of infrared temperature data in each weld segment to identify local thermal anomaly areas caused by cracks or corrosion;

[0034] Use dynamic time warping algorithm to compare the difference between current infrared temperature data and standard template;

[0035] Based on historical data statistics, a classification threshold is set for the echo amplitude of the ultrasonic data;

[0036] Defective welds are identified based on the difference results and threshold categories.

[0037] Furthermore, a three-dimensional model of the tower structure was constructed by combining SolidWorks and MATLAB.

[0038] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0039] The present invention discloses a wind turbine tower weld detection device and method. The device uses infrared and ultrasonic sensors to perform comprehensive weld detection, effectively improving the accuracy of weld detection results. Furthermore, with the cooperation of a drive mechanism and a travel mechanism, the detection device can detect circular welds and linear welds separately, thereby determining the specific locations of defective welds and improving the efficiency of subsequent manual review.

[0040] In addition, marking the location of defective welds in the 3D model allows staff to see it more intuitively, so that during subsequent manual review, staff can quickly locate defective welds, thereby improving detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A structural schematic diagram of the present invention;

[0042] Figure 2 This is a schematic structural diagram of a detection device of the present invention;

[0043] Figure 3 The present invention is a flow chart of a method.

[0044] Figure numerals: 1. Ring frame body; 2. Driving mechanism; 3. Detection equipment; 4. Traveling mechanism; 41. Articulated block; 42. Articulated shaft; 43. Mounting plate; 44. Permanent magnet wheel; 45. Traveling motor; 5. Tower. DETAILED DESCRIPTION

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0046] Example 1

[0047] like Figure 1-Figure 2 A wind turbine tower weld detection device shown includes a ring frame body 1, which is correspondingly mounted on the tower 5, and the inner diameter of the ring frame body 1 is larger than the maximum diameter of the tower 5. Since the tower 5 generally has a frustum structure with an upper diameter larger than the lower diameter, a frustum structure with an upper diameter smaller than the lower diameter, and a structure with the same upper and lower diameters, the use of a large-diameter ring frame body 1 can ensure that it is mounted on the tower 5 without interfering with the tower 5, and is compatible with towers 5 of different structures, thereby improving the applicability of the detection device;

[0048] The inner ring surface of the ring frame body 1 is equipped with a driving mechanism 2, which is used to drive the detection device 3 to move along the circumference of the ring frame body 1; that is, when the driving mechanism 2 drives the detection device 3 to move, it can detect the annular weld data between the two cylinder sections and can find the starting position of the straight weld; in particular, the inner ring surface of the ring frame body 1 is provided with an annular track, and the driving mechanism 2 includes a driving trolley, which is installed in the annular track, and the detection device 3 is installed on one end of the driving trolley corresponding to the ring frame body 1. The sliding installation between the driving trolley and the annular track is used to achieve a stable installation relationship between the detection device 3 and the driving trolley and the ring frame body 1, thereby preventing the detection device 3 or the driving trolley from falling from the ring frame body 1; in addition, the running distance of the driving trolley can be obtained according to the running speed and time of the driving trolley, and the length of the annular track is constant. Therefore, whether the detection of the annular weld is completed can be judged based on whether the running distance is consistent with the length of the annular track;

[0049] The detection device 3 is composed of a camera, an infrared sensor and an ultrasonic sensor; the camera mainly locates the starting position of the straight weld by image recognition technology. After locating the starting position of the straight weld, the walking mechanism 4 moves along the tower 5. When the camera locates the end position of the straight weld, that is, the intersection of the straight weld and the circular weld, it means that the detection of a section of straight weld data is completed, and then the walking mechanism 4 is stopped and the driving mechanism 2 is started to move one circle, that is, the detection of a section of circular weld data is completed. At the same time, during the detection process of the circular weld data, the camera will also synchronously locate the starting position of the next section of straight weld; the infrared sensor and the ultrasonic sensor respectively emit infrared light and ultrasonic waves to the weld during the detection process, and receive feedback infrared light data and ultrasonic data to determine whether the weld is a defective weld;

[0050] At least three walking mechanisms 4 are hinged on the top of the ring frame body 1. The walking mechanisms 4 are used to move along the outer wall of the tower 5, and the multiple walking mechanisms 4 are evenly distributed. By utilizing the walking mechanisms 4 to move along the outer wall of the tower 5, the ring frame body 1 can be driven to move synchronously. The evenly distributed multiple walking mechanisms 4 can keep the tower 5 always located at the center of the ring frame body 1, thereby ensuring the inspection quality of the subsequent inspection equipment 3; in particular, the number of the walking mechanisms 4 is four, which can achieve a balance between production cost and stability, and has a high cost performance.

[0051] Example 2

[0052] As an embodiment, the walking mechanism 4 includes a hinge block 41, a hinge shaft 42, a mounting plate 43 and a permanent magnet wheel 44, the hinge block 41 is fixedly mounted on the bottom of the ring frame body 1, the hinge shaft 42 passes through the hinge block 41, and both ends of the hinge shaft 42 are fixedly connected to a mounting plate 43, both mounting plates 43 are rotatably connected to the permanent magnet wheel 44, and any mounting plate 43 is provided with a walking motor 45 for driving the permanent magnet wheel 44 to rotate;

[0053] Since the tower 5 is made of metal material, the permanent magnet wheel 44 can be adsorbed on the tower 5, and then driven by the travel motor 45, the permanent magnet wheel 44 can be driven to rotate. In addition, since the tower 5 is a frustum structure, that is, the outer wall of the tower 5 has a certain inclination angle, the hinge structure between the hinge block 41 and the mounting plate 43 allows the permanent magnet wheel 44 to move along the tower 5, and the mounting plate 43 will rotate along the hinge axis 42, so that the ring frame body 1 maintains the axial position during the movement.

[0054] In addition, an encoder is installed at the axle of the permanent magnet wheel 44 , and the encoder records the rotation angle of the permanent magnet wheel 44 , so that the movement distance of the permanent magnet wheel 44 can be calculated.

[0055] Example 3

[0056] like Figure 3 A wind turbine tower weld detection method is shown, specifically comprising:

[0057] Obtain all weld data, as well as motion data of the traveling mechanism 4 and the driving mechanism 2;

[0058] The total weld data includes the linear weld data of each cylinder segment and the annular weld data between two adjacent cylinder segments. Each weld data segment includes the detection data of the infrared sensor and the detection data of the ultrasonic sensor. The motion data of the walking mechanism 4 is the rotation angle of the permanent magnet wheel 44 recorded by the encoder, and the motion data of the driving mechanism 2 is calculated based on the running speed and time of the driving trolley.

[0059] Preprocess all acquired weld data to provide high-quality input for subsequent defect classification models and ensure the reliability of detection results;

[0060] Defective welds are judged based on all pre-processed weld data using a data processing model. The data processing model is used to analyze and process the detection data from infrared sensors and ultrasonic sensors, and then make judgments separately, thereby improving the accuracy of defective weld judgment.

[0061] Construct a 3D model of the tower structure;

[0062] Combine the motion data of the walking mechanism 4 and the driving mechanism 2, and mark the location of the defective weld in the three-dimensional model;

[0063] In particular, a 3D model of the tower structure was constructed by combining SolidWorks and MATLAB. SolidWorks was used to draw the 3D simulation model of the tower structure, while MATLAB was used to mark the locations of defective welds on the simulation model. It should be noted that when marking defective welds, different colors were used in the model according to the type of defective welds in the data processing model, such as cracks and corrosion, e.g., red for cracks and yellow for corrosion.

[0064] According to the severity of the defective welds in the data processing model, such as minor, serious, and critical, the icon size or transparency is used to distinguish them. For example, a large icon indicates a critical defect.

[0065] In addition, in MATLAB, the defect coordinate data is superimposed on the 3D model exported by SolidWorks to generate a visual interface with defect marks;

[0066] In addition, when the inspection device is re-inspected, the defect status in the model is updated through real-time data stream, such as marking it green after repair.

[0067] In addition, based on the detection data of each weld, the corresponding drive mechanism 2 motion data or walking mechanism 4 motion data can be recorded synchronously, so that it is possible to directly determine which weld is the defective weld, and then the MATLAB simulation software can generate a defect mark based on the three-dimensional simulation model.

[0068] Example 4

[0069] As an embodiment, the method for obtaining all weld seam data and motion data of the walking mechanism 4 and the driving mechanism 2 is specifically as follows:

[0070] Control the driving mechanism 2 to drive the detection device 3 to move, and the camera locates the weld position;

[0071] Start the walking mechanism 4 and synchronously record the movement data of the walking mechanism 4. At the same time, the infrared sensor and the ultrasonic sensor obtain the detection data of the linear weld;

[0072] When the end of the straight weld is reached and the data of a straight weld is obtained, the traveling mechanism 4 is stopped;

[0073] Then the driving mechanism 2 is started, and the motion data of the walking mechanism 4 is recorded synchronously, and the infrared sensor and ultrasonic sensor are controlled to obtain the detection data of the annular weld;

[0074] After the driving mechanism 2 moves around the ring frame body 1 for one circle, the data of a circular weld is obtained;

[0075] Finally, the above steps are repeated until the top of the tower 5 is reached to obtain all the weld data.

[0076] In addition, all the weld data obtained by this method are multi-segment data, and the method can synchronously record the motion data of the walking mechanism 4 when obtaining a certain segment of linear weld data, and synchronously record the motion data of the driving mechanism 2 when obtaining a certain segment of annular weld data, so that each segment of weld data can be matched with the corresponding motion data, and the current cylinder section can be determined based on the number of times the walking mechanism 4 is started;

[0077] It should be noted that after the driving mechanism 2 moves around the ring frame body 1 for one circle, in the step of obtaining the data of a circular weld, the camera is also in a synchronously turned-on state, that is, while obtaining the circular weld, the camera also searches for the starting position of the next straight weld. After obtaining the complete circular weld data, the driving mechanism 2 is controlled to drive the detection equipment 3 to move to the position of the next straight weld for subsequent weld detection.

[0078] Example 5

[0079] As an embodiment, the preprocessing of all weld data obtained specifically includes:

[0080] The infrared sensor raw data and ultrasonic sensor raw data in each section of weld data are filtered separately;

[0081] For the filtering of the original data of the infrared sensor, first set the current pixel to I(x,y) and the filter window size to (2K+1) 2 ,

[0082] Using the filtering formula:

[0083]

[0084] Among them, S xy The infrared sensor raw data set within the filter window, if I (x, y) is a noise point, then median (S xy ), in other cases, I(x,y) is used; and the noise judgment is |I(x,y)-median(S xy )|>3σ, if it holds, it is marked as a noise point; the purpose is to scale the temperature data to the interval [0,1] to eliminate the dimensional difference and facilitate subsequent model processing.

[0085] The raw data of the ultrasonic sensor is filtered using a transfer function, and the calculation formula is:

[0086]

[0087] Where, f is the ultrasonic frequency, f low and f high is the passband cutoff frequency, and n is the filter order, usually 4-6. The purpose is to make the data conform to the distribution with a mean of 0 and a standard deviation of 1, thereby improving the convergence speed of the model.

[0088] Normalization is performed on each weld seam data segment after filtering to unify sensor data of different dimensions to the same scale, facilitating feature fusion and model input.

[0089] The normalized calculation formula for the infrared temperature data of the infrared sensor is:

[0090]

[0091] Among them, I is the original temperature value of the infrared sensor, I min and I max They are the minimum and maximum values of the temperature data of the current weld respectively;

[0092] Ultrasonic sensor data is normalized using Z-score standardization, and the calculation formula is:

[0093]

[0094] Among them, A is the echo amplitude of the ultrasonic sensor, μ A and σ A are the mean and standard deviation of the current weld echo amplitude respectively;

[0095] Time synchronization is performed on all normalized weld data.

[0096] The interpolation formula is used to align the ultrasonic data with the infrared timestamp. The specific calculation formula is:

[0097]

[0098] Among them, A sync (t IR ) is the synchronized ultrasonic sensor data at time t IR The interpolation result at t IR is the timestamp of the infrared data, A(t1) is the echo amplitude collected by the ultrasonic sensor at time t1, and A(t2) is the echo amplitude collected by the ultrasonic sensor at time t2. The purpose is to align the data time axes of different sensors to ensure that the infrared, ultrasonic, and camera data at the same time can be fused and analyzed.

[0099] Example 6

[0100] As an embodiment, the specific steps of using the data processing model to determine defective welds based on the pre-processed linear weld data and circular weld data are as follows:

[0101] Extract the temperature distribution characteristics of the infrared temperature data in each weld data segment to identify local thermal anomaly areas caused by cracks or corrosion, specifically;

[0102]

[0103] Among them, G(x,y) is the temperature gradient amplitude, which reflects the severity of temperature change;

[0104]

[0105] Using dynamic time warping algorithm, compare the current infrared temperature data T current ={T1,T2,···,T n} with the standard template T template ={T1 ′ ,T2 ′ ,···,T n ′} difference;

[0106] Calculate the cumulative distance matrix, specifically:

[0107]

[0108] Among them, D(i,j) is the cumulative distance of the path;

[0109] Take the minimum cumulative distance of the path as the difference value ΔDTW:

[0110] ΔDTW=D(n,m);

[0111] Based on historical data statistics, a classification threshold is set for the echo amplitude of the ultrasonic data;

[0112] The classification thresholds include slight defects, moderate defects and severe defects, and the judgment rules for the three are: μ A -σ A ≤A norm <μ A 、μ A -2σ A ≤A norm <μ A -σ A and A norm <μ A -2σ A

[0113] Determine defective welds based on the difference results and threshold categories;

[0114] Defective welds are determined based on the defect confidence level N, specifically:

[0115]

[0116] Among them, ω1 and ω2 are weights, δ th It is the critical value used to determine the difference between the infrared temperature sequence and the standard template. If the calculated defect confidence N is greater than the threshold, it is marked as a defective weld.

[0117] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A wind turbine tower weld detection device, characterized in that: The invention comprises a ring frame body (1), wherein the ring frame body (1) is sleeved on a tower (5), and the inner diameter of the ring frame body (1) is larger than the maximum diameter of the tower (5); a driving mechanism (2) is installed on the inner surface of the ring frame body (1), and the driving mechanism (2) is used to drive the detection device (3) to move along the circumference of the ring frame body (1); The detection device (3) is composed of a camera, an infrared sensor and an ultrasonic sensor; At least three walking mechanisms (4) are hinged on the top of the ring frame body (1), and the walking mechanisms (4) are used to move along the outer wall of the tower (5), and the multiple walking mechanisms (4) are evenly distributed.

2. The wind turbine tower weld detection device according to claim 1, characterized in that: The inner ring surface of the ring frame body (1) is provided with an annular track, and the driving mechanism (2) includes a driving trolley installed in the annular track, and a detection device (3) is installed at one end of the driving trolley corresponding to the ring frame body (1).

3. The wind turbine tower weld detection device according to claim 1, characterized in that: The number of the walking mechanisms (4) is four.

4. The wind turbine tower weld detection device according to claim 1, characterized in that: The walking mechanism (4) comprises an articulated block (41), an articulated shaft (42), a mounting plate (43) and a permanent magnet wheel (44); the articulated block (41) is fixedly mounted on the bottom of the ring frame body (1); the articulated shaft (42) passes through the articulated block (41) and the two ends of the articulated shaft (42) are respectively fixedly connected to a mounting plate (43); both mounting plates (43) are rotatably connected to the permanent magnet wheel (44); and any mounting plate (43) is provided with a walking motor (45) for driving the permanent magnet wheel (44) to rotate.

5. The wind turbine tower weld detection device according to claim 4, characterized in that: An encoder is installed on the wheel shaft of the permanent magnet wheel (44).

6. A method for detecting weld seams of a wind turbine tower, used for implementing the device for detecting weld seams of a wind turbine tower as claimed in any one of claims 1 to 5, characterized in that: Specifically include: Acquiring all weld seam data, as well as motion data of the walking mechanism (4) and the driving mechanism (2); Preprocess all acquired weld data; Using the data processing model, the defective welds are judged based on all the pre-processed weld data; Construct a 3D model of the tower structure; The motion data of the walking mechanism (4) and the driving mechanism (2) are combined, and the position of the defective weld is marked in the three-dimensional model.

7. The wind turbine tower weld detection method according to claim 6, characterized in that: The method for obtaining all weld seam data and motion data of the walking mechanism (4) and the driving mechanism (2) is specifically as follows: The driving mechanism (2) is controlled to drive the detection device (3) to move, and the position of the weld is located by the camera; The walking mechanism (4) is started, and the movement data of the walking mechanism (4) is synchronously recorded, while the detection data of the linear weld is obtained by the infrared sensor and the ultrasonic sensor; When the end of the straight weld is reached and data of a straight weld is obtained, the traveling mechanism is stopped (4); Then, the driving mechanism (2) is started, and the motion data of the walking mechanism (4) is recorded synchronously, and the infrared sensor and the ultrasonic sensor are controlled to obtain the detection data of the annular weld; After the driving mechanism (2) moves around the ring frame body (1) for one circle, data of a circular weld is obtained; Finally, the above steps are repeated until the top of the tower (5) is reached to obtain all the weld data.

8. The wind turbine tower weld detection method according to claim 6, characterized in that: All weld data obtained by the preprocessing specifically include: The infrared sensor raw data and ultrasonic sensor raw data in each section of weld data are filtered separately; Normalize each weld seam data after filtering; Time synchronization is performed on all normalized weld data.

9. The wind turbine tower weld detection method according to claim 6, characterized in that: The specific steps of using the data processing model to determine defective welds based on the pre-processed linear weld data and circular weld data are as follows: Extract the temperature distribution characteristics of infrared temperature data in each weld segment to identify local thermal anomaly areas caused by cracks or corrosion; Use dynamic time warping algorithm to compare the difference between current infrared temperature data and standard template; Based on historical data statistics, a classification threshold is set for the echo amplitude of the ultrasonic data; Defective welds are identified based on the difference results and threshold categories.

10. The wind turbine tower weld detection method according to claim 6, characterized in that: A three-dimensional model of the tower structure is constructed by combining SolidWorks and MATLAB.

Images