Methods, electronic equipment and media for detecting cracks in hybrid towers of wind turbines
By collecting and processing images and multi-frequency signal data of mixed towers using drones, and combining this data with analysis from a ground control terminal, the problems of low efficiency and insufficient accuracy in detecting mixed tower cracks in wind turbines have been solved, achieving efficient and accurate online detection and report generation.
Patent Information
- Application Number
- CN202611017226.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-09
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies for detecting cracks in mixed towers of wind turbines are inefficient, with scattered detection data, a lack of unified data analysis capabilities, difficulty in identifying minute cracks and blind spots in the tower's visual field, and insufficient detection accuracy.
By collecting images of the concrete tower, measuring distances, multi-frequency ultrasonic signals of the concrete, electromagnetic induction signals of the steel bars, and environmental parameters using drones, and combining this data with a ground control terminal for data processing and analysis, a crack detection report is generated, enabling online detection and data analysis.
It improves the efficiency and accuracy of crack detection in wind turbine towers, can identify minute cracks, eliminate blind spots, generate detailed inspection reports, and support operation and maintenance decisions.
Smart Images

Figure CN123115806A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of wind power equipment maintenance and inspection, specifically to a method for detecting cracks in the mixed tower of a wind turbine, electronic equipment, and a medium. Background Technology
[0002] In related technologies, the detection of cracks in the mixed-stove structure of wind turbines mainly employs techniques such as manual inspection, ultrasonic testing, and infrared thermal imaging. These methods are mostly offline, resulting in scattered data and a lack of unified data analysis capabilities, leading to low efficiency in crack detection. Summary of the Invention
[0003] This application provides a method, electronic device, and medium for detecting cracks in the hybrid tower of wind turbines. By receiving detection data transmitted from a drone, processing and analyzing the data, and obtaining a crack detection report, online detection and data analysis of cracks in the hybrid tower of wind turbines are achieved, thereby improving the detection efficiency of cracks in the hybrid tower of wind turbines.
[0004] In a first aspect, embodiments of this application provide a method for detecting cracks in the hybrid tower of a wind turbine, applied to a ground control terminal, comprising: receiving detection data transmitted by a drone during flight, the detection data including: a first hybrid tower image, the measured distance from the drone to the surface of the hybrid tower, multi-frequency ultrasonic signals of concrete inside the hybrid tower, electromagnetic induction signals of steel bars inside the hybrid tower, and environmental parameters; preprocessing the first hybrid tower image to extract a target crack region; determining the crack width corresponding to the target crack region on a second hybrid tower image based on the target crack region, the second hybrid tower image being obtained after distortion correction of the first hybrid tower image; performing fusion analysis on the multi-frequency ultrasonic signals of concrete and the electromagnetic induction signals of steel bars, and determining the crack depth corresponding to the target crack region in conjunction with the environmental parameters; and generating a detection report based on the comparison results of the crack width, the crack depth, and a preset threshold.
[0005] Secondly, this application provides a method for detecting cracks in the hybrid tower of a wind turbine, applied to an unmanned aerial vehicle (UAV), comprising: acquiring a preset detection path and flying along the detection path, wherein the detection path is generated by a ground control terminal to maintain a preset distance between the UAV and the surface of the hybrid tower; acquiring a first hybrid tower image of the surface of the hybrid tower through a visual detection unit, and simultaneously acquiring the measured distance from the UAV to the surface of the hybrid tower through a laser ranging unit; simultaneously acquiring multi-frequency ultrasonic signals of concrete and electromagnetic induction signals of steel bars in the hybrid tower at the same target point through a composite detection unit; acquiring environmental parameters, including temperature, humidity and wind speed, through an environmental sensing unit; and transmitting the first hybrid tower image, the measured distance, the multi-frequency ultrasonic signals of concrete, the electromagnetic induction signals of steel bars and the environmental parameters to the ground control terminal.
[0006] Thirdly, embodiments of this application provide an electronic device including a processor and a memory, wherein the memory stores a program or instructions that run on the processor, and the program or instructions, when executed by the processor, implement the steps of the methods described in the first and second aspects.
[0007] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the methods described in the first and second aspects.
[0008] The wind turbine tower crack detection method, electronic equipment, and medium provided in this application improve the detection efficiency by processing and analyzing data such as the first tower image transmitted by a drone, the measured distance from the drone to the tower surface, the multi-frequency ultrasonic signals of the concrete inside the tower, the electromagnetic induction signals of the reinforcing steel inside the tower, and environmental parameters. The detection report of wind turbine tower cracks is obtained through these data processing and analysis. Furthermore, the environmental parameters considered in the wind turbine tower crack detection report enhance the accuracy of crack detection. Attached Figure Description
[0009] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0010] Figure 1 This is one of the flowcharts illustrating the wind turbine tower crack detection method provided in the embodiments of this application; Figure 2This is the second flowchart illustrating the wind turbine tower crack detection method provided in this application embodiment. Detailed Implementation
[0011] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0012] The terms "setup" and "connection" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral structure; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium, or an internal connection between two devices, components, or parts. Those skilled in the art can understand the specific meaning of these terms in this application based on the specific circumstances.
[0013] Furthermore, the terms "first," "second," etc., are primarily used to distinguish different locations and are not intended to indicate or imply the relative importance or quantity of the indicated devices, elements, or components. Unless otherwise stated, "a plurality of" means two or more.
[0014] With the rapid development of the wind power industry, the installed capacity and service life of wind turbines are constantly increasing. The hybrid tower is a key supporting structure for wind turbines, and its operational safety directly affects the stable operation of the entire wind farm. Constructed from a composite of high-strength concrete and steel, the hybrid tower is prone to cracking under long-term exposure to complex conditions such as wind loads, gravity loads, and temperature changes. If these cracks are not detected and addressed in a timely manner, they will gradually expand, leading to a decrease in structural strength and, in severe cases, causing tower collapse and other safety accidents, resulting in huge economic losses and the risk of personal injury. Therefore, crack detection in the hybrid tower is one of the core aspects of wind farm equipment management.
[0015] In related technologies, the detection of cracks in wind turbine towers mainly employs manual inspection, ultrasonic testing, and infrared thermal imaging. Manual inspection relies on the experience of the inspectors, who observe through binoculars or climb to the tower surface for close-up inspection. This method struggles to identify minute cracks and cracks in blind spots such as the back and top of the tower, leading to missed or false detections. Ultrasonic testing requires close contact with the tower surface, which often has poor flatness, necessitating pre-cleaning and coupling treatment. Furthermore, its detection range is limited, hindering large-area rapid detection and resulting in low efficiency. Infrared thermal imaging's accuracy is affected by ambient temperature and lighting conditions, and its ability to identify crack depth is insufficient, leading to low accuracy. Moreover, most of these methods are offline, resulting in scattered data and a lack of unified data analysis.
[0016] This application provides a method for detecting cracks in the hybrid tower of a wind turbine. Figure 1 A flowchart of a method for detecting cracks in the hybrid tower of a wind turbine provided in an embodiment of this application is shown. This method can be applied to a ground control terminal.
[0017] The ground control terminal can be an industrial tablet PC or embedded industrial control device equipped with dedicated detection and analysis software. The ground control terminal is located in the ground-based operation and maintenance area of the wind farm. It possesses full data processing capabilities, including 5G data reception, image processing, signal analysis, report generation, and trend prediction. Operators can complete the processing and analysis of detection data on cracks in the wind turbine towers from the ground.
[0018] like Figure 1 As shown, the wind turbine tower crack detection method provided in this application embodiment may include the following steps 101, 102, 103, 104 and 105.
[0019] Step 101: Receive the detection data transmitted by the drone during flight.
[0020] The detection data includes: images of the first concrete mixing tower, the measured distance from the drone to the surface of the concrete mixing tower, multi-frequency ultrasonic signals of the concrete inside the concrete mixing tower, electromagnetic induction signals of the steel bars inside the concrete mixing tower, and environmental parameters.
[0021] The detection data can be a complete set of raw sensor data with accompanying spatial coordinates and timestamps. This data is collected by a detection module mounted on the UAV during flight. The spatial coordinates are obtained using GPS and BeiDou positioning systems on the UAV.
[0022] The first image of the mixing tower can be an image of the surface appearance of the mixing tower. The first image of the mixing tower was obtained directly by a 4K high-definition industrial camera mounted on a drone, without geometric correction.
[0023] The measured distance is the vertical straight-line distance from the drone to the surface of the mixing tower, measured in real time by a laser ranging unit installed on the drone.
[0024] Multi-frequency ultrasonic signals for concrete can be presented as time-series waveforms. Ultrasonic sensors ranging from 1MHz to 10MHz can generate time-series waveforms after penetrating the concrete in a concrete mixing tower. Multi-frequency ultrasonic signals can reflect internal defects and crack depths within the concrete. However, multi-frequency ultrasonic signals cannot detect the corrosion state of the reinforcing steel bars inside the concrete mixing tower.
[0025] Electromagnetic induction signals from reinforcing bars can be presented as magnetic field distortion time-series signals. The alternating electromagnetic field in an electromagnetic induction sensor with a detection range of 0.01mm-100mm penetrates the concrete cover of a concrete tower based on the electromagnetic eddy current effect and electromagnetic coupling principle, generating reverse eddy currents in the reinforcing bars inside the tower, thus forming a magnetic field distortion time-series signal. By analyzing the impedance, phase shift, and electromagnetic amplitude of the electromagnetic induction signal from the reinforcing bars and comparing it with standard values, structural problems such as the degree of corrosion of the reinforcing bars, the thickness of the concrete cover, internal voids and defects, and abnormal reinforcement can be determined, solving the problem that ultrasonic sensors cannot detect reinforcing bar damage.
[0026] Environmental parameters include three types of data: real-time temperature, ambient humidity, and on-site wind speed. These parameters are collected by an environmental sensing unit mounted on the drone. The environmental sensing unit includes a temperature sensor, a humidity sensor, and a wind speed sensor. The temperature sensor's measurement range is -40℃ to 85℃, the humidity sensor's range is 0%RH to 100%RH, and the wind speed sensor's range is 0m / s to 20m / s. Temperature alters the propagation speed of multi-frequency ultrasonic signals, humidity attenuates the amplitude of electromagnetic induction signals from reinforcing bars, and wind speed causes drone vibration, leading to target point displacement. All three types of parameters interfere with the detection accuracy of cracks in the tower.
[0027] In practical implementation, the ground terminal receives in real-time the first image of the concrete tower transmitted wirelessly by the UAV, the measured distance from the UAV to the surface of the concrete tower, the multi-frequency ultrasonic signals of the concrete inside the concrete tower, the electromagnetic induction signals of the steel bars inside the concrete tower, and environmental parameters, thus solving the problems of missed detection and insufficient detection accuracy caused by single crack detection methods. At the same time, the UAV can automatically complete a full-coverage scan of the entire surface of the concrete tower, eliminating blind spots in the detection line of sight and improving detection efficiency.
[0028] Step 102: Preprocess the first mixed tower image to extract the target crack region.
[0029] Image preprocessing can perform basic image processing operations such as denoising, grayscale enhancement, and contrast stretching on the first mixed tower image. After image preprocessing, the first mixed tower image is clearer, reducing the probability of misidentification of the target crack region in subsequent image segmentation algorithms.
[0030] The target crack region can be the entire pixel range of the suspected crack. However, the target crack region is only a pixel-level outline and has not undergone distortion correction; therefore, it cannot be directly used to calculate the actual width of the crack.
[0031] In practice, preprocessing the first mixed tower image can improve the accuracy of extracting the target crack region, thereby improving the accuracy of determining the crack width.
[0032] Step 103: Based on the target crack region, determine the crack width corresponding to the target crack region on the second mixed tower image. The second mixed tower image is obtained after distortion correction of the first mixed tower image.
[0033] The second mixed-tower image can be obtained by correcting the distortion of the preprocessed first mixed-tower image. The pixel size of the second mixed-tower image corresponds one-to-one with the actual physical size of the mixed-tower surface. The second mixed-tower image serves as the reference image for calculating the crack width.
[0034] When the measured distance from the drone to the surface of the mixing tower fluctuates, the lens of the 4K high-definition industrial camera tilts, requiring distortion correction of the first mixing tower image. The distortion correction rule is: True pixel coordinates = Measured pixel coordinates × (Measured distance / Standard distance).
[0035] In practical implementation, by performing distortion correction on the second mixed tower image and then determining the crack width corresponding to the target crack area, minute cracks ≥0.05mm can be identified, improving the accuracy of crack width calculation.
[0036] For example, the standard distance is 40cm, the actual measured distance is 45cm, the image is compressed, and stretched proportionally to restore the true size.
[0037] The above step 103 can be implemented through the following steps 103a, 103b and 103c.
[0038] Step 103a: Extract the pixel length of the target crack region from the second mixed tower image according to the image segmentation algorithm.
[0039] Image segmentation algorithms can automatically identify the outlines of linear defect pixels in the second mixed tower image after distortion correction, whose grayscale and texture differ from those of the concrete matrix, thereby extracting the crack outline of the target crack area.
[0040] Pixel length is the total number of pixels along the extension direction of the complete crack contour extracted by the image segmentation algorithm.
[0041] In practical implementation, an image segmentation algorithm is used to extract the pixel length of the target crack region from the second mixed tower image, which can reduce the error of manually marking cracks and improve the efficiency of crack detection. At the same time, the image segmentation algorithm only retains the real crack region, solving the problem that including irrelevant pixels in the pixel length calculation leads to inaccurate crack width calculation results.
[0042] Step 103b: Determine the physical size of a single pixel based on the measured distance from the UAV to the surface of the mixing tower, the focal length and resolution of the first mixing tower image.
[0043] Focal length and resolution together determine the spatial field of view corresponding to a single pixel.
[0044] The physical size of a single pixel can be defined as the actual millimeter length of a single pixel on the surface of the mixing tower in the second mixing tower image. The value of the physical size of a single pixel changes dynamically with the measured distance. The greater the measured distance, the larger the actual size corresponding to the physical size of a single pixel; the closer the measured distance, the smaller the actual size corresponding to the physical size of a single pixel.
[0045] In practical implementation, the physical size of a single pixel in the second mixing tower image is determined by dynamically converting the measured distance, focal length, and resolution. This solves the problem of image distortion caused by the measured distance and improves the accuracy of determining the crack width.
[0046] Step 103c: Determine the crack width based on the pixel length and the physical size of a single pixel.
[0047] The crack width is the millimeter-scale opening of a real crack, calculated by multiplying the pixel length by the physical size of a single pixel.
[0048] In practical implementation, the crack width determined by pixel length and single pixel physical size has higher accuracy and can stably identify minute cracks ≥0.05mm.
[0049] Step 104: Perform fusion analysis on the multi-frequency ultrasonic signals of concrete and the electromagnetic induction signals of steel bars, and determine the crack depth corresponding to the target crack area by combining environmental parameters.
[0050] Fusion analysis is the process of analyzing multi-frequency ultrasonic signals from concrete and electromagnetic induction signals from reinforcing steel. Fusion analysis incorporates environmental parameters to compensate for these signals in real time. Through the fusion analysis of these signals, crack depth and the degree of steel corrosion inside the concrete structure can be determined.
[0051] In practical implementation, multi-frequency ultrasonic signals from concrete can be used to calculate crack depth, while electromagnetic induction signals from reinforcing bars can be used to determine the damage state of the bars. This distinguishes between surface cracks in concrete and high-risk through cracks induced by steel corrosion, improving the accuracy of crack grade determination. When determining the crack depth corresponding to the target crack area, incorporating environmental parameters can eliminate the drift issues of multi-frequency ultrasonic signals from concrete and electromagnetic induction signals from reinforcing bars caused by environmental parameters, thus improving the detection accuracy of cracks under adverse weather conditions.
[0052] The above step 104 can be implemented through the following steps 104a, 104b, 104c, 104d, 104e and 104f.
[0053] Step 104a: Preprocess the multi-frequency ultrasonic signal of concrete and the electromagnetic induction signal of steel reinforcement to obtain standardized multi-frequency ultrasonic signal of concrete and standardized electromagnetic induction signal of steel reinforcement.
[0054] Preprocessing includes noise reduction, environmental drift correction, and normalization conversion.
[0055] The multi-frequency ultrasonic signal of standardized concrete and the electromagnetic induction signal of standardized steel bars are signals in a unified interval with a mean of zero and a variance of one.
[0056] In practical implementation, preprocessing can filter out on-site wind noise, circuit noise, and sensor jitter interference, eliminate drifting concrete multi-frequency ultrasonic signals and drifting steel bar electromagnetic induction signals caused by temperature, humidity, and wind speed, and at the same time eliminate the dimensional and numerical differences between concrete multi-frequency ultrasonic signals and standardized steel bar electromagnetic induction signals, so as to make subsequent feature fusion accurate and improve the accuracy of crack depth detection.
[0057] The above step 104a can be implemented through the following steps 104a1, 104a2 and 104a3.
[0058] Step 104a1: Perform noise reduction processing on the multi-frequency ultrasonic signal of concrete and the electromagnetic induction signal of steel reinforcement, remove invalid noise signals, and generate noise-reduced multi-frequency ultrasonic signal of concrete and electromagnetic induction signal of steel reinforcement.
[0059] Noise reduction can be achieved by using wavelet filtering algorithms to filter out high-frequency noise in the multi-frequency ultrasonic signals of concrete and the electromagnetic induction signals of steel bars.
[0060] Invalid noise can be caused by drone body vibration, wind electromagnetic interference, and inherent noise in sensor circuits, which can affect the signals collected by ultrasonic and electromagnetic sensors.
[0061] In practical implementation, noise reduction processing of multi-frequency ultrasonic signals from concrete and electromagnetic induction signals from reinforcing bars can solve the problem of inaccurate crack depth calculation caused by invalid noise.
[0062] Step 104a2: Based on the quantitative mapping relationship between environmental parameters and signal drift, the noise-reduced concrete multi-frequency ultrasonic signal and steel bar electromagnetic induction signal are corrected according to the preset correction coefficient to obtain the corrected concrete multi-frequency ultrasonic signal and steel bar electromagnetic induction signal.
[0063] The preset correction factors include temperature correction factor, humidity correction factor, and wind speed correction factor. Different correction factor values correspond to different environmental ranges. After the ground control terminal reads the environmental parameters in real time, it matches the corresponding correction factors to complete the correction of the noise-reduced multi-frequency ultrasonic signal of concrete and electromagnetic induction signal of steel reinforcement.
[0064] In practical implementation, a preset quantitative mapping model of temperature, humidity, wind speed and signal drift is used, combined with real-time environmental parameters, to correct the noise-reduced concrete multi-frequency ultrasonic signal and steel bar electromagnetic induction signal according to the preset correction coefficient set in the quantitative mapping model. This can eliminate the interference of environmental parameters on the noise-reduced concrete multi-frequency ultrasonic signal and steel bar electromagnetic induction signal, improve the detection accuracy of crack depth, and improve the detection accuracy of crack depth under harsh working conditions.
[0065] For example, for every 10°C increase in temperature, the sound velocity of multi-frequency ultrasound decreases by approximately 2%. A correction is applied using Vtrue = Vmeasured × (1 + 0.02 × ΔT / 10), where ΔT = real-time ambient temperature - calibration reference temperature. If ΔT > 0, it indicates that the real-time ambient temperature is higher than the calibration reference temperature, and the measured sound velocity is too low. A correction factor greater than 1 is needed to increase the sound velocity to the true speed. If ΔT < 0, it indicates that the real-time ambient temperature is lower than the calibration reference temperature, and the measured sound velocity is too high. A correction factor less than 1 is needed to decrease the sound velocity to the true speed.
[0066] Increased humidity causes the electromagnetic induction signal of the steel bar to decrease by about 3%, so multiplying it by a compensation coefficient of 1.03 restores the true electromagnetic induction signal of the steel bar.
[0067] When wind speed causes vibrations in the ultrasonic and electromagnetic induction sensors, the multi-frequency ultrasonic signals of concrete and the electromagnetic induction signals of the reinforcing bars are corrected according to the displacement deviation. After correction, the detection accuracy of crack depth remains ≥90% under severe weather conditions.
[0068] Step 104a3: Standardize the corrected concrete multi-frequency ultrasonic signal and the steel bar electromagnetic induction signal to generate standardized concrete multi-frequency ultrasonic signal and standardized steel bar electromagnetic induction signal.
[0069] Standardization conversion can perform uniform dimensionless processing on the corrected concrete multi-frequency ultrasonic signals and steel bar electromagnetic induction signals, mapping the corrected concrete multi-frequency ultrasonic signals and steel bar electromagnetic induction signals with different units, magnitudes, and numerical ranges to a unified interval with a mean of zero and a variance of one.
[0070] In practical implementation, by standardizing and converting the corrected concrete multi-frequency ultrasonic signal and the steel bar electromagnetic induction signal, the differences in dimensions and magnitudes between the concrete multi-frequency ultrasonic signal and the steel bar electromagnetic induction signal can be eliminated, thus solving the problem that the two heterogeneous signals cannot be directly fused.
[0071] Step 104b: Analyze the standardized concrete multi-frequency ultrasonic signal and the standardized steel reinforcement electromagnetic induction signal point by point to extract the multi-dimensional concrete multi-frequency ultrasonic features and the multi-dimensional steel reinforcement electromagnetic induction features.
[0072] Point-by-point analysis can be a calculation method that separately separates the standardized concrete multi-frequency ultrasonic signal and the standardized steel bar electromagnetic induction signal according to each target point.
[0073] The characteristics of multi-dimensional concrete multi-frequency ultrasonic waves include sound velocity, attenuation coefficient, first wave time, frequency, amplitude, and fractal dimension.
[0074] The speed of sound can characterize how fast multi-frequency ultrasound waves propagate in concrete. When cracks appear in concrete, the multi-frequency ultrasound waves must travel around the cracks, reducing the speed of sound. The deeper the crack, the lower the speed of sound.
[0075] The attenuation coefficient describes the degree of energy loss during the propagation of multi-frequency ultrasound. In crack regions, a large number of multi-frequency ultrasound waves are reflected and scattered, resulting in energy attenuation. The attenuation coefficient is highly sensitive to microcracks and network cracks, compensating for the limitation of sound velocity being insensitive to minute defects.
[0076] The initial propagation time can be defined as the initial travel time of the multi-frequency ultrasonic wave from the transmitting probe to the receiving probe. When cracks exist in the concrete, the propagation of the multi-frequency ultrasonic wave is blocked, the initial wave arrives later, and the propagation time increases.
[0077] The frequency range is 1MHz-10MHz.
[0078] The amplitude can be the attenuated amplitude collected from the crack area, which is the pre-calibrated value of the mixed tower concrete material.
[0079] Fractal dimension describes the irregularity of the waveform of multi-frequency ultrasound caused by defect disturbances. The more chaotic the crack distribution, the higher the fractal dimension. Fractal dimension can be used to distinguish crack types. Crack types include surface cracks, deep cracks, network cracks, and through cracks.
[0080] Multidimensional electromagnetic induction characteristics of reinforcing steel bars include impedance amplitude, phase, resonant frequency, corrosion rate, and steel bar embedment depth, used to characterize damage to the internal steel reinforcement structure of concrete towers. These multidimensional electromagnetic induction characteristics can distinguish defects in concrete that cannot be identified by multi-frequency ultrasonic signals, such as steel bar corrosion, internal voids, and protective layer detachment.
[0081] The impedance amplitude represents the total impedance of the electromagnetic circuit formed by the excitation coil and the internal reinforcing steel. If the reinforcing steel is corroded, its resistivity increases, and the impedance amplitude increases. The impedance amplitude can determine whether a crack extends to the reinforcing steel layer, and can distinguish between concrete cracks and cracks extending to the reinforcing steel layer.
[0082] Phase can be the phase shift between voltage and current in an electromagnetic signal. When the reinforcing steel undergoes local deformation or corrosion, the phase will shift.
[0083] The resonant frequency can be the characteristic frequency at which an electromagnetic system resonates. When a crack penetrates concrete, the reinforcing steel undergoes localized deformation or corrosion, causing the resonant frequency to drift. The resonant frequency can serve as a long-term monitoring characteristic and can be used to track the development trend of concrete cracks.
[0084] The corrosion rate can be used to quantify the severity of steel reinforcement corrosion. The wider and deeper the crack, the faster the steel reinforcement corrodes, and the higher the corrosion rate. The corrosion rate can be used to assess the severity level of cracks.
[0085] The depth of rebar embedment can be used to determine the actual thickness of the protective layer from the top of the rebar to the concrete surface. The depth of rebar embedment can also determine the upper limit of crack depth; if the measured rebar embedment depth is small, it indicates that the crack has penetrated deep into the protective layer.
[0086] In practical implementation, standardized concrete multi-frequency ultrasonic signals and standardized steel reinforcement electromagnetic induction signals are analyzed point by point to extract multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel reinforcement electromagnetic induction features. Each concrete multi-frequency ultrasonic feature and steel reinforcement electromagnetic induction feature corresponds to the same target point, which can solve the problem of signal cross-interference between different target points.
[0087] Step 104c: Perform multi-source feature fusion preprocessing on the multi-dimensional concrete multi-frequency ultrasonic features and the multi-dimensional steel reinforcement electromagnetic induction features to obtain a joint feature vector.
[0088] Multi-source feature fusion preprocessing includes spatiotemporal registration, serial feature splicing, and redundant feature removal. The joint feature vector can be a set of low-dimensional features obtained by serially splicing multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel reinforcement electromagnetic induction features in a preset order and then filtering for correlation.
[0089] In practical implementation, the computational burden of the multi-source feature fusion model can be reduced by performing multi-source feature fusion preprocessing on the multi-frequency ultrasonic features of multi-dimensional concrete and the electromagnetic induction features of multi-dimensional steel bars.
[0090] The above step 104c can be implemented through the following steps 104c1, 104c2 and 104c3.
[0091] Step 104c1: Associate the multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel reinforcement electromagnetic induction features at the same target location and at the same time using the same spatial coordinates to obtain the associated feature sequence.
[0092] The same target location can be the mixed tower detection area defined by the path planning algorithm of the ground control terminal.
[0093] Associating the same spatial coordinates can match the three-dimensional coordinates and collection timestamps output by the UAV's GPS and BeiDou with the multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel reinforcement electromagnetic induction features at the same time and target location.
[0094] The associated feature sequence can be a set of features arranged in the time sequence of acquisition, with one pair of multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel bar electromagnetic induction features, and each set of features is bound to the three-dimensional coordinates of the same target point.
[0095] In practical implementation, the three-dimensional coordinates and timestamps of the detection data are registered with the multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel reinforcement electromagnetic induction features. This solves the problem of misalignment of the concrete multi-frequency ultrasonic features and steel reinforcement electromagnetic induction features caused by the displacement of the UAV, so that the multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel reinforcement electromagnetic induction features correspond to the same target point, thereby improving the accuracy of crack depth detection.
[0096] Step 104c2: The multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel bar electromagnetic induction features in the associated feature sequence are sequentially spliced in a preset order to obtain the initial joint feature vector.
[0097] Serial splicing is used to integrate multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel reinforcement electromagnetic induction features in a fixed preset order. The preset order can be concrete multi-frequency ultrasonic features first, followed by steel reinforcement electromagnetic induction features.
[0098] The initial joint feature vector can be a set of multidimensional features obtained by serial concatenation.
[0099] In practical implementation, serial splicing preserves the complete multi-dimensional concrete multi-frequency ultrasonic characteristics and multi-dimensional steel reinforcement electromagnetic induction characteristics, thereby improving the accuracy of crack depth detection and the accuracy of steel reinforcement corrosion level determination.
[0100] Step 104c3: Based on the correlation between the features in the initial joint feature vector, remove redundant features to obtain the joint feature vector.
[0101] In practical implementation, the correlation coefficients between features within the initial joint feature vector are calculated. Through principal component analysis (PCA) or independent component analysis (ICA), redundant features with linear correlation and information repetition are eliminated, while the joint feature vector is retained, thereby improving the feature fusion accuracy of the multi-source feature fusion model.
[0102] Step 104d: Simultaneously input the standardized concrete multi-frequency ultrasonic signal, the standardized steel bar electromagnetic induction signal, and the joint feature vector into the pre-set multi-source feature fusion model.
[0103] Multi-source feature fusion models can be deep learning models built primarily with LSTM neural networks. These models can autonomously uncover nonlinear correlations between multiple signal sources, achieving feature fusion.
[0104] In practical implementation, standardized concrete multi-frequency ultrasonic signals, standardized steel bar electromagnetic induction signals, and joint feature vectors are simultaneously input into a pre-set multi-source feature fusion model to solve the problem of loss of defect information caused by using only a single type of data, and to improve the ground control terminal's ability to identify minute cracks and hidden steel bar corrosion.
[0105] Step 104e: Extract the nonlinear correlation between standardized concrete multi-frequency ultrasonic signals, standardized steel bar electromagnetic induction signals and joint feature vectors through a pre-set multi-source feature fusion model, and perform feature fusion based on the nonlinear correlation to obtain fused features.
[0106] Nonlinear correlation can reveal the inherent data patterns among standardized concrete multi-frequency ultrasonic signals, standardized steel bar electromagnetic induction signals, and joint eigenvectors, where there is no fixed linear correspondence.
[0107] Fusion features can be a unified feature set output by a multi-source feature fusion model.
[0108] In practical implementation, a pre-set multi-source feature fusion model is used to extract the nonlinear correlation between standardized concrete multi-frequency ultrasonic signals, standardized steel reinforcement electromagnetic induction signals, and joint feature vectors. This solves the problem of inaccurate crack depth calculation caused by features fused with linear correlations. Furthermore, the fused features include concrete crack information and steel reinforcement corrosion damage information, improving the stability and accuracy of crack depth calculation.
[0109] Step 104f: Determine the crack depth corresponding to the target crack region based on the fusion characteristics.
[0110] In practical implementation, based on the fusion features, combined with the crack depth calculation formula, environmental parameter compensation algorithm and concrete-reinforcement coupling coefficient, the accurate crack depth can be obtained, thus solving the problem of low crack depth detection accuracy.
[0111] The above step 104f can be implemented through the following steps 104f1, 104f2, 104f3 and 104f4.
[0112] Step 104f1: Extract the reference amplitude, measured amplitude, frequency, and attenuation coefficient of multi-frequency ultrasonic waves in concrete from the fusion features.
[0113] The reference amplitude for multi-frequency ultrasonic testing of concrete is the ultrasonic echo amplitude of an intact concrete matrix.
[0114] The measured amplitude of the multi-frequency ultrasonic wave in concrete is the amplitude after the multi-frequency ultrasonic wave has attenuated as it passes through the crack.
[0115] The frequency range for multi-frequency ultrasonic testing of concrete is 1MHz-10MHz.
[0116] The attenuation coefficient of the multi-frequency ultrasonic in concrete was pre-calibrated and confirmed based on the concrete material of the mixing tower.
[0117] Step 104f2: Through Determine the depth of the first crack, where h is the depth of the first crack, k is the attenuation coefficient of the concrete multi-frequency ultrasound, A0 is the reference amplitude of the concrete multi-frequency ultrasound, A is the measured amplitude of the concrete multi-frequency ultrasound, and f is the frequency of the concrete multi-frequency ultrasound.
[0118] The ratio of the reference amplitude of multi-frequency ultrasonic waves in concrete to the measured amplitude of multi-frequency ultrasonic waves in concrete can reflect the degree of energy loss of multi-frequency ultrasonic waves penetrating cracks.
[0119] In practice, frequency deviations can be eliminated by fitting multiple sets of multi-frequency ultrasonic amplitudes at different frequencies, and the calculation error of the first crack depth can be controlled within ±0.05mm.
[0120] Step 104f3: Correct the first crack depth using an environmental parameter compensation algorithm to obtain the second crack depth.
[0121] The first crack depth can be the numerical value corresponding to the depth of loss of multi-frequency ultrasonic waves in concrete.
[0122] The environmental parameter compensation algorithm can use a preset quantitative mapping model of temperature, humidity, wind speed and signal drift, combined with real-time environmental parameters, to reverse the depth of the first crack according to the preset correction coefficient set in the quantitative mapping model.
[0123] The second crack depth can be the intermediate depth value after the first crack depth has been compensated for environmental parameters.
[0124] In practical implementation, the first crack depth is corrected in reverse by using an environmental parameter compensation algorithm, which can eliminate the interference of environmental parameters on the first crack depth and improve the detection accuracy of crack depth.
[0125] Step 104f4: Extract the electromagnetic induction features of the reinforcing bars from the fusion features, correct the second crack depth according to the concrete-reinforcing bar coupling coefficient, and determine the crack depth.
[0126] The concrete-reinforcement coupling coefficient can be a correction parameter obtained from calibration tests. It quantifies the extent to which the degree of reinforcement corrosion corrects for the actual crack penetration depth. In practical implementation, the electromagnetic induction characteristics of the reinforcing steel bars are extracted from the fusion features. These characteristics include the corrosion rate and the embedment depth of the reinforcing steel bars, and the corresponding concrete-reinforcing steel coupling coefficient is matched. The depth of the second crack is then adjusted based on the degree of steel bar corrosion: if the corrosion is severe, the depth of the second crack is increased; otherwise, it remains unchanged. In this way, the crack depth of the target crack area is reliably determined.
[0127] Step 105: Generate an inspection report based on the comparison results of crack width, crack depth and preset threshold.
[0128] The preset threshold can be a pre-defined standard for judging crack risk. For example, cracks with a width ≥ 0.3 mm and a depth ≥ 10 mm are judged as high-risk structural cracks. The ground control terminal compares the calculated crack width and crack depth with the preset threshold to automatically distinguish between high-risk structural cracks, pure concrete cracks, and potential cracks.
[0129] The inspection report is an automatically generated document from the ground control terminal. It includes information on the hybrid tower foundation, inspection data for each crack, crack depth, crack width, rebar corrosion level, and risk markers. The report can be directly uploaded to the wind farm equipment management platform to update the hybrid tower's health record, providing quantitative data support for the development of operation and maintenance plans.
[0130] In practical implementation, by comparing the preset threshold with the crack width and crack depth, the crack type of each crack is identified, and each crack is marked with a risk. This not only allows us to understand the crack situation of the mixed tower and take appropriate measures, but also allows us to predict the crack expansion trend of the mixed tower.
[0131] The method for detecting cracks in the hybrid tower of a wind turbine can also be achieved through the following steps 106 and 107.
[0132] Step 106: Obtain the test data, test report, and operating conditions of the mixing tower.
[0133] Operating conditions include long-term wind load, diurnal temperature variation, and time-series data on the service life of the hybrid tower. In practice, combining the hybrid tower's test data, test reports, and operating conditions can improve the accuracy of predicting crack propagation trends.
[0134] Step 107: Based on the test data, test report and operating conditions of the mixing tower, analyze the development trend of cracks through the crack prediction model, and predict the crack propagation within a preset period.
[0135] The crack prediction model can learn from multi-dimensional time-series data such as detection data, detection reports, and the operating conditions of the mixing tower, and can capture the nonlinear time-series pattern of slow crack propagation.
[0136] The development trend of cracks can be represented by the rate at which crack width and crack depth expand over time.
[0137] The preset cycle can be 1-3 months.
[0138] In practical implementation, the crack propagation predicted by the crack prediction model within a preset period is compared with the early warning threshold, resulting in a three-level risk warning system: slow crack propagation, moderate crack propagation, and rapid crack propagation. The corresponding warning signals are blue, yellow, and red. For rapidly propagating red-warning cracks, priority is given to maintenance to address the issue of tower collapse caused by continuous crack propagation, thereby improving the safety of wind farm operation and maintenance. Furthermore, the three-level risk warning system allows for refined allocation of operation and maintenance resources, eliminating the need for unified shutdown and maintenance of all towers, reducing power generation losses caused by shutdowns of risk-free towers, and lowering the manpower and material costs of wind farm operation and maintenance.
[0139] Each of the above-described method embodiments, or various possible implementations of each method embodiment, can be executed individually or in combination of any two or more. The specific implementation can be determined according to actual usage requirements, and this application does not impose any restrictions on this.
[0140] The wind turbine tower crack detection method provided in this application embodiment can be applied to drones. Figure 2 A flowchart of a method for detecting cracks in the hybrid tower of a wind turbine provided in an embodiment of this application is shown.
[0141] like Figure 2 As shown, the wind turbine tower crack detection method provided in this application embodiment may include the following steps 201, 202, 203, 204 and 205.
[0142] Step 201: Obtain the preset detection path and fly according to the detection path. The detection path is generated by the ground control terminal to keep the UAV at a preset distance from the surface of the mixing tower.
[0143] The preset detection path is automatically generated by importing the hybrid tower 3D model from the ground control terminal and inputting parameters such as the tower's height, diameter, and geographical location. The path spacing of the preset detection path can be set to 5cm.
[0144] The preset distance can be 30cm-50cm.
[0145] In practical implementation, the drone flies along a preset detection path, eliminating blind spots in the field of vision for manual inspections, such as the back and top of the mixing tower, thus improving detection efficiency. Maintaining a preset distance during drone flight ensures stable signals from the ultrasonic and electromagnetic sensors, stable measured distances from the laser ranging unit, and uniform size of the first mixing tower image captured by the high-definition industrial camera. This solves the problem of reduced accuracy in detection data caused by fluctuations in the preset distance between the drone and the mixing tower surface.
[0146] Step 202: Acquire the first image of the mixing tower surface through the visual inspection unit, and simultaneously acquire the measured distance from the UAV to the mixing tower surface through the laser ranging unit.
[0147] The vision inspection unit can integrate a 20-megapixel 4K high-definition industrial camera and an infrared fill light. The infrared fill light can improve the image brightness in low-light environments such as cloudy days or dusk, enabling the 4K high-definition industrial camera to clearly inspect the first mixing tower image.
[0148] In practical implementation, the laser ranging unit outputs the vertical distance from the lens of the high-definition industrial camera to the surface of the mixing tower in real time, and the vision inspection unit acquires the first image of the mixing tower surface in real time. Each frame of the image is synchronously bound to a set of measured distances, which can make the calculation results of the crack width accurate.
[0149] Step 203: Using the composite detection unit, simultaneously collect multi-frequency ultrasonic signals of the concrete and electromagnetic induction signals of the steel bars inside the concrete tower at the same target location. The composite detection unit integrates an ultrasonic sensor and an electromagnetic induction sensor, both fixed on the same bracket, allowing the ultrasonic and electromagnetic induction sensors to synchronously acquire data at the same target point. The ultrasonic sensor operates at a frequency of 1MHz–10MHz. The magnetic induction sensor has a detection range of 0.01mm–100mm.
[0150] In practical implementation, a composite detection unit is used to acquire multi-frequency ultrasonic signals from the concrete and electromagnetic induction signals from the reinforcing bars inside the concrete tower, improving detection efficiency. Simultaneously, the spatiotemporal alignment of the multi-frequency ultrasonic signals from the concrete and the electromagnetic induction signals from the reinforcing bars resolves detection errors caused by misaligned data.
[0151] Step 204: Collect environmental parameters through the environmental sensing unit. The environmental parameters include temperature, humidity and wind speed.
[0152] The environmental sensing unit can integrate temperature sensors, humidity sensors, and wind speed sensors.
[0153] In practical implementation, the environmental sensing unit synchronously collects environmental parameters at the same target location. The ground control terminal can then use these synchronously collected environmental parameters from the same target location to correct the concrete multi-frequency ultrasonic signal, the steel reinforcement electromagnetic induction signal, and the depth of the first crack, thereby improving the crack detection accuracy.
[0154] Step 205: Transmit the first mixing tower image, measured distance, concrete multi-frequency ultrasonic signal, steel bar electromagnetic induction signal, and environmental parameters to the ground control terminal.
[0155] In practice, the UAV transmits images of the first concrete tower, measured distances, multi-frequency ultrasonic signals from the concrete, electromagnetic induction signals from the reinforcing bars, and environmental parameters to the ground control terminal in real time via a wireless communication module. This enables real-time acquisition, transmission, and analysis of detection data, improving the detection efficiency of cracks in wind turbine towers.
[0156] This application also provides an electronic device, which includes a processor and a memory. The memory stores a program or instructions that run on the processor. The program or instructions are executed by the processor to implement the various processes of the above-described wind turbine tower crack detection method embodiments and can achieve the same technical effect. To avoid repetition, they will not be described again here.
[0157] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described wind turbine tower crack detection method embodiment and achieve the same technical effect. To avoid repetition, they will not be described again here.
[0158] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.
Claims
1. A method for detecting cracks in the hybrid tower of a wind turbine, applied to a ground control terminal, characterized in that, include: The system receives detection data transmitted by the drone during flight. The detection data includes: a first image of the concrete mixing tower, the measured distance from the drone to the surface of the concrete mixing tower, multi-frequency ultrasonic signals of the concrete inside the concrete mixing tower, electromagnetic induction signals of the steel bars inside the concrete mixing tower, and environmental parameters. The first mixed tower image is preprocessed to extract the target crack region; Based on the target crack region, the crack width corresponding to the target crack region is determined on the second mixed tower image, wherein the second mixed tower image is obtained after distortion correction of the first mixed tower image. The multi-frequency ultrasonic signal of the concrete and the electromagnetic induction signal of the steel bar are fused and analyzed, and the crack depth corresponding to the target crack area is determined in combination with the environmental parameters. A detection report is generated based on the comparison results of the crack width, crack depth and preset threshold.
2. The method for detecting cracks in the hybrid tower of a wind turbine according to claim 1, characterized in that, The step of fusing and analyzing the multi-frequency ultrasonic signals of the concrete and the electromagnetic induction signals of the reinforcing steel, and combining this with environmental parameters to determine the crack depth corresponding to the target crack area, includes: The concrete multi-frequency ultrasonic signal and the steel bar electromagnetic induction signal are preprocessed to obtain standardized concrete multi-frequency ultrasonic signal and standardized steel bar electromagnetic induction signal. The standardized concrete multi-frequency ultrasonic signal and the standardized steel reinforcement electromagnetic induction signal are analyzed point by point to extract multi-dimensional concrete multi-frequency ultrasonic features and multi-dimensional steel reinforcement electromagnetic induction features. The multi-source feature fusion preprocessing is performed on the multi-frequency ultrasonic features of the multi-dimensional concrete and the electromagnetic induction features of the multi-dimensional steel reinforcement to obtain a joint feature vector. The standardized concrete multi-frequency ultrasonic signal, the standardized steel bar electromagnetic induction signal, and the joint feature vector are simultaneously input into a pre-set multi-source feature fusion model. The nonlinear correlation between the standardized concrete multi-frequency ultrasonic signal, the standardized steel bar electromagnetic induction signal, and the joint feature vector is extracted by the pre-set multi-source feature fusion model, and feature fusion is performed based on the nonlinear correlation to obtain fused features; Based on the fusion features, the crack depth corresponding to the target crack region is determined.
3. The method for detecting cracks in the hybrid tower of a wind turbine according to claim 2, characterized in that, The preprocessing of the multi-frequency ultrasonic signal of the concrete and the electromagnetic induction signal of the reinforcing steel includes: The concrete multi-frequency ultrasonic signal and the steel bar electromagnetic induction signal are subjected to noise reduction processing to remove invalid noise signals and generate the noise-reduced concrete multi-frequency ultrasonic signal and steel bar electromagnetic induction signal. Based on the quantitative mapping relationship between environmental parameters and signal drift, the noise-reduced concrete multi-frequency ultrasonic signal and the steel bar electromagnetic induction signal are corrected according to a preset correction coefficient to obtain the corrected concrete multi-frequency ultrasonic signal and the steel bar electromagnetic induction signal. The corrected concrete multi-frequency ultrasonic signal and the steel bar electromagnetic induction signal are standardized and converted to generate standardized concrete multi-frequency ultrasonic signal and standardized steel bar electromagnetic induction signal.
4. The method for detecting cracks in the hybrid tower of a wind turbine according to claim 2, characterized in that, The multi-source feature fusion preprocessing of the multi-dimensional concrete multi-frequency ultrasonic features and the multi-dimensional steel reinforcement electromagnetic induction features includes: The multi-dimensional concrete multi-frequency ultrasonic features and the multi-dimensional steel electromagnetic induction features at the same target point and at the same time are associated with the same spatial coordinates to obtain an associated feature sequence. The multidimensional concrete multi-frequency ultrasonic features and the multidimensional steel reinforcement electromagnetic induction features in the associated feature sequence are sequentially spliced together in a preset order to obtain an initial joint feature vector. Based on the correlation between the features in the initial joint feature vector, redundant features are removed to obtain the joint feature vector.
5. The method for detecting cracks in the hybrid tower of a wind turbine according to claim 2, characterized in that, Determining the crack depth corresponding to the target crack region based on the fusion features includes: The reference amplitude, measured amplitude, frequency, and attenuation coefficient of multi-frequency ultrasonic waves in concrete are extracted from the fusion features. pass Determine the depth of the first crack, where h is the depth of the first crack, k is the attenuation coefficient of the concrete multi-frequency ultrasound, A0 is the reference amplitude of the concrete multi-frequency ultrasound, A is the measured amplitude of the concrete multi-frequency ultrasound, and f is the frequency of the concrete multi-frequency ultrasound. The first crack depth is corrected by an environmental parameter compensation algorithm to obtain the second crack depth; The electromagnetic induction characteristics of the reinforcing steel are extracted from the fusion characteristics, and the second crack depth is corrected according to the concrete-reinforcing steel coupling coefficient to determine the crack depth.
6. The method for detecting cracks in the hybrid tower of a wind turbine according to claim 1, characterized in that, Determining the crack width corresponding to the target crack region on the second mixing tower image includes: According to the image segmentation algorithm, the pixel length of the target crack region is extracted from the second mixed tower image; Based on the measured distance from the UAV to the surface of the mixing tower, the focal length and resolution of the detected first mixing tower image, the physical size of a single pixel is determined; The crack width is determined based on the pixel length and the physical size of the single pixel.
7. The method for detecting cracks in the hybrid tower of a wind turbine according to claim 1, characterized in that, The method further includes: Acquire the detection data, detection report, and operating status of the mixing tower; Based on the detection data of the mixing tower, the detection report, and the operating conditions of the mixing tower, the development trend of the cracks is analyzed by the crack prediction model, and the crack propagation within a preset period is predicted.
8. A method for detecting cracks in the hybrid tower of a wind turbine, applied to unmanned aerial vehicles (UAVs), characterized in that... include: The UAV obtains a preset detection path and flies along the detection path, which is generated by the ground control terminal to keep the UAV at a preset distance from the surface of the mixing tower. The first image of the mixing tower surface is acquired by the visual inspection unit, and the measured distance from the UAV to the surface of the mixing tower is acquired simultaneously by the laser ranging unit. The composite detection unit simultaneously collects multi-frequency ultrasonic signals of concrete and electromagnetic induction signals of steel bars in the concrete tower at the same target location. Environmental parameters, including temperature, humidity, and wind speed, are collected through an environmental sensing unit. The first mixed tower image, the measured distance, the concrete multi-frequency ultrasonic signal, the steel bar electromagnetic induction signal, and the environmental parameters are transmitted to the ground control terminal.
9. An electronic device, characterized in that, It includes a processor and a memory, the memory storing a program or instructions that run on the processor, the program or instructions being executed by the processor to implement the steps of the wind turbine tower crack detection method as described in any one of claims 1 to 8.
10. A readable storage medium having a program or instructions stored thereon, characterized in that, When the program or instructions are executed by the processor, they implement the steps of the wind turbine tower crack detection method as described in any one of claims 1 to 8.
