A method and device for detecting surface defects of a wind turbine blade

Abstract

The present application relates to the technical field of defect detection, and discloses a wind turbine blade surface defect detection method and device, wherein the wind turbine blade surface defect detection method provided by the present application fuses multi-modal data, and improves defect detection precision and wind power generation efficiency.

Description

Technical Field

[0001] This invention relates to the field of defect detection technology, specifically to a method and apparatus for detecting surface defects on wind turbine blades. Background Technology

[0002] A wind turbine is a power generation device that converts wind energy into mechanical work. This mechanical work drives a rotor to rotate, ultimately outputting alternating current (AC) electricity. A wind turbine typically consists of components such as a wind rotor, a generator (including the generator itself), a directional control unit (tail fin), a tower, a speed limiting safety mechanism, and an energy storage device. To drive the wind rotor, the generator blades capture wind energy to propel it. Therefore, the quality of the generator blades affects the generator's efficiency. To ensure proper operation, the generator blades are inspected before installation to prevent defects on the blade surface.

[0003] There are currently two methods for inspecting generator blades:

[0004] Method 1: Visual Inspection. Visual inspection is widely used for inspecting large structural materials on spacecraft and bridges, and can also be applied to generator blades. Due to the large size of the structural materials, visual inspection is time-consuming, and its accuracy depends heavily on the experience of the inspectors. Since some materials are handled at heights, the work is dangerous. Inspectors are typically equipped with a long-lens digital camera, but prolonged inspections can cause eye fatigue. While visual inspection can directly detect surface defects, it cannot detect internal structural defects; therefore, other effective methods are needed to evaluate the internal structure of the material.

[0005] Method 2: Camera-based visual inspection. Visual inspection uses machine vision products to transmit images to a dedicated image processing system. It replaces the human eye in measurement and judgment. Specifically, machine vision products (image acquisition devices: CMOS and CCD) convert images of wind turbine blades into image signals, which are then transmitted to a dedicated image processing system. Based on pixel distribution and information such as brightness and color, these signals are converted into digital signals. The image system performs various calculations on these converted signals to extract target features, and then controls the on-site equipment based on the judgment results.

[0006] Compared to the two methods mentioned above, Method 1 is simple and practical, but it takes a long time to detect and is prone to omissions. Method 2 can efficiently detect generator blades and is less prone to omissions. However, both methods have a fatal flaw: they can only detect generator blades before installation. When surface defects such as cracks appear on the generator blades after installation and in operation, these methods are not applicable. Based on the above description, existing defect detection methods can only detect defects when the wind turbine has serious problems, resulting in low wind power generation efficiency. Summary of the Invention

[0007] In view of this, the present invention provides a method and apparatus for detecting surface defects of wind turbine blades to solve the problem of low efficiency in wind power generation.

[0008] In a first aspect, the present invention provides a method for detecting surface defects on wind turbine blades, the method comprising:

[0009] The detection device acquires images of the surface of the wind turbine blade and ultrasonic reflection signals inside the blade. The images of the blade surface include visible light images and infrared thermal images.

[0010] The visible light image and the infrared thermal image are preprocessed respectively, and the preprocessed visible light image and infrared thermal image are defect detected by a preset image analysis model to generate surface defect data and thermal anomaly data of wind turbine blades.

[0011] The ultrasonic reflection signal is preprocessed, and the preprocessed ultrasonic reflection signal is then subjected to defect detection to generate data on defects in the internal structure of the wind turbine blade.

[0012] Using a pre-defined algorithm, data on surface defects, thermal anomalies, and internal structural defects of wind turbine blades are fused together to complete the defect detection of wind turbine blade surfaces.

[0013] The wind turbine blade surface defect detection method provided by this invention integrates multi-modal data, thereby improving defect detection accuracy and wind power generation efficiency.

[0014] In one optional implementation, the step of detecting defects on the surface of a wind turbine blade further includes:

[0015] Based on preset warning levels and preset thresholds, the system issues warnings and alarms for the detection results of defects on the surface of wind turbine blades and sends warning signals to the operator.

[0016] This invention sends early warning signals to operators through early warning alarms, thereby enabling timely maintenance measures to prevent defects from escalating or malfunctions from occurring.

[0017] In one optional embodiment, the detection device includes: an unmanned helicopter, a high-definition camera, an infrared thermal imager, an ultrasonic detector, and cables, wherein,

[0018] High-definition cameras, infrared thermal imagers, and ultrasonic detectors are directly connected to the unmanned helicopter via cables.

[0019] High-definition cameras are used to collect visible light images;

[0020] Thermal imagers are used to collect infrared thermal images;

[0021] Ultrasonic detectors are used to collect ultrasonic reflected signals from the internal structure of blades.

[0022] The method provided by this invention enables more comprehensive and accurate detection of wind turbine blades through a detection device.

[0023] In one alternative implementation, the high-definition camera, infrared thermal imager, and ultrasonic detector are all fixed to the bottom of the unmanned helicopter.

[0024] The method provided by this invention, by fixing the high-definition camera, infrared thermal imager and ultrasonic detector to the bottom of the unmanned helicopter, ensures that the sensor can always face the blade during flight, ensuring the comprehensiveness and continuity of data acquisition. At the same time, this design also allows the sensor to achieve comprehensive detection of different parts of the blade as the unmanned helicopter turns and moves during flight.

[0025] In one alternative implementation, the unmanned helicopter is a six-axis stabilized multi-rotor unmanned helicopter.

[0026] The method provided by this invention enables stable flight in complex wind farm environments using a six-axis stabilized multi-rotor unmanned helicopter, ensuring the accuracy of the detection data.

[0027] In one alternative implementation, the unmanned helicopter flies around the wind turbine blades according to a preset flight path to complete the photography and scanning of the wind turbine blades.

[0028] In one optional implementation, the preprocessing of the visible light image and the infrared thermal image respectively includes:

[0029] The visible light image is denoised using a Gaussian filter to enhance its quality.

[0030] Histogram equalization is used to enhance the contrast of the infrared thermal image, and median filtering is applied to the enhanced infrared thermal image to determine the temperature difference region and thermal anomaly region of the blade.

[0031] The method provided by this invention enhances visible light images and infrared thermal images, employing different enhancement strategies for the two images to more accurately identify and locate defects in wind turbine blades.

[0032] In one alternative implementation, the visible light image is denoised using a Gaussian filter according to the following formula:

[0033] Ienh=G(I vis ,σ)

[0034] Where Ienh represents the denoised visible light image, G represents the Gaussian filter, σ represents the standard deviation, and I... vis Represents a visible light image.

[0035] The method provided by this invention reduces random noise in an image by using a Gaussian filter.

[0036] In one alternative implementation, the infrared thermal image is contrast-enhanced using histogram equalization according to the following formula:

[0037] I ir,contrast =HE(I ir )

[0038] Among them, I ir,contrast This represents the infrared thermal image after contrast enhancement, I ir This represents an infrared thermal image, and HE represents the histogram equalization algorithm.

[0039] The method provided by this invention enhances contrast through histogram equalization, clearly revealing temperature differences and thermal anomalies on the blades. Histogram equalization optimizes the grayscale histogram of the image, making its distribution more uniform, thereby improving the overall contrast of the image and making the temperature differences and thermal anomaly areas on the blades more noticeable.

[0040] In one alternative implementation, the contrast-enhanced infrared thermal image is subjected to median filtering using the following formula:

[0041] I ir,denoise =MF(I ir,contrast ,k)

[0042] Among them, I ir,denoise This represents the infrared thermal image after median filtering, where MF represents the median filtering algorithm and k represents the filter window size.

[0043] The method provided by this invention is for I ir,contrast The median filter is used for noise reduction. The median filter can effectively eliminate salt and pepper noise and preserve the details of the image.

[0044] In one optional implementation, a preset algorithm is used to fuse surface defect data, thermal anomaly data, and internal structural defect data of the wind turbine blade to complete the defect detection of the wind turbine blade surface, including:

[0045] The K-means clustering algorithm is used to fuse surface defect data, thermal anomaly data, and internal structural defect data of wind turbine blades to determine the type and depth of surface defects and complete the defect detection of wind turbine blades.

[0046] In one optional implementation, based on a preset warning level and a preset threshold, an early warning alarm is triggered based on the defect detection results on the surface of the wind turbine blade, and a warning signal is sent to the operator, including:

[0047] There are m cluster centers, each representing a defect of a preset defect type and depth;

[0048] Assign a weight value wi to each cluster center i, representing the severity of the defect of that type and depth;

[0049] The preset warning level L is calculated using the following formula:

[0050] L=∑wi*Ni

[0051] Where Ni represents the number of pixels belonging to the i-th cluster center;

[0052] Set a preset threshold T. When the preset warning level L exceeds the preset threshold T, a warning signal is sent to the operator.

[0053] Secondly, the present invention provides a device for detecting surface defects in wind turbine blades, the device comprising:

[0054] The acquisition module is used to acquire images of the surface of the wind turbine blade and ultrasonic reflection signals inside the blade collected by the detection device. The images of the blade surface include visible light images and infrared thermal images.

[0055] The first detection module is used to preprocess the visible light image and the infrared thermal image respectively, and use a preset image analysis model to perform defect detection on the preprocessed visible light image and infrared thermal image respectively, and generate surface defect data and thermal anomaly data of wind turbine blades.

[0056] The second detection module is used to preprocess the ultrasonic reflection signal and perform defect detection on the preprocessed ultrasonic reflection signal to generate data on defects in the internal structure of the wind turbine blade.

[0057] The fusion module is used to fuse data on surface defects, thermal anomalies, and internal structural defects of wind turbine blades using a preset algorithm, thereby completing the defect detection of the wind turbine blade surface.

[0058] Thirdly, the present invention provides a computer device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the wind turbine blade surface defect detection method described in the first aspect or any corresponding embodiment thereof.

[0059] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the wind turbine blade surface defect detection method of the first aspect or any corresponding embodiment described above. Attached Figure Description

[0060] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0061] Figure 1 This is a flowchart of a method for detecting surface defects on wind turbine blades according to an embodiment of the present invention;

[0062] Figure 2 This is another flowchart of a method for detecting surface defects on wind turbine blades according to an embodiment of the present invention;

[0063] Figure 3 This is a structural block diagram of a wind turbine blade surface defect detection device according to an embodiment of the present invention;

[0064] Figure 4 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] According to an embodiment of the present invention, a method for detecting surface defects of wind turbine blades is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0067] This embodiment provides a method for detecting surface defects on wind turbine blades, which can be used on the aforementioned mobile terminals, such as mobile phones and tablets. Figure 1 This is a flowchart of a method for detecting surface defects on wind turbine blades according to an embodiment of the present invention, as follows: Figure 1 As shown, the process includes the following steps:

[0068] Step S101: Acquire images of the wind turbine blade surface and ultrasonic reflection signals inside the blade collected by the detection device. The images of the blade surface include visible light images and infrared thermal images.

[0069] In this embodiment of the invention, the detection device includes: an unmanned helicopter, a high-definition camera, an infrared thermal imager, an ultrasonic detector, and cables. The high-definition camera, the infrared thermal imager, and the ultrasonic detector are directly connected to the unmanned helicopter via cables to achieve real-time data acquisition and transmission.

[0070] In one specific embodiment, a high-definition camera is used to collect visible light images, which reflect the surface physical structure of the blade and any obvious surface defects. An infrared thermal imager is used to collect infrared thermal images, which show the surface temperature distribution of the blade. By analyzing the temperature distribution, areas of potential surface defects can be identified. An ultrasonic detector is used to collect ultrasonic wave reflection signals from inside the blade. By analyzing the reflection signals, potential structural defects inside the blade can be detected.

[0071] In one specific embodiment, all three sensors are connected to the unmanned helicopter via cables, which serve both data transmission and power supply functions. One end of the cable is connected to the main body of the unmanned helicopter, and the other end is connected to each of the three sensors. This connection method enables real-time, multi-directional, and multi-dimensional monitoring of the wind turbine blades. In this embodiment, comprehensive monitoring is achieved by using an unmanned helicopter equipped with multiple sensors, including a high-definition camera, an infrared thermal imager, and an ultrasonic detector. This allows for all-around monitoring of the wind turbine blades, covering both the surface and interior of the blades, enabling real-time status detection. This overcomes the limitations of existing technologies that can only monitor blades before installation or when serious problems occur, thus improving the safety of the wind turbine. The unmanned aerial vehicle and sensors used in this embodiment are low-cost and cost-effective. The image analysis and ultrasonic detection technologies are simple, practical, and easy to implement, allowing for long-term monitoring of wind turbine blades, extending their service life, reducing maintenance costs, and demonstrating high cost-effectiveness.

[0072] In one specific embodiment, the unmanned helicopter is selected as a six-axis stabilized multi-rotor type. Based on the advantages of this type of unmanned helicopter, such as good stability and high hovering accuracy, it is particularly suitable for inspecting wind turbine blades. It can fly stably in the complex wind farm environment, ensuring the accuracy of the inspection data.

[0073] In one specific embodiment, the detection device is designed as an integrated unit, meaning that the three sensors share a single power supply and data transmission system. This design simplifies the system's complexity, reduces potential points of failure, and improves the reliability and stability of the detection system. Simultaneously, the sharing of the three sensors allows the system's weight and size to be kept within a reasonable range, ensuring the flight performance of the unmanned helicopter.

[0074] In one specific embodiment, the high-definition camera, infrared thermal imager, and ultrasonic detector are all fixed to the bottom of the unmanned helicopter. This design ensures that the sensors can always face the blade during flight, guaranteeing the comprehensiveness and continuity of data acquisition. Simultaneously, this design also allows the sensors to perform comprehensive detection of different parts of the blade as the unmanned helicopter turns and moves during flight.

[0075] In one specific embodiment, an unmanned helicopter conducts inspections near a wind turbine. A high-definition camera is used to collect visible light images, a thermal imager is used to collect infrared thermal images, and a sound frequency detection device is used to collect the sound frequency range of the wind turbine during operation. This data is transmitted in real time to a ground control station, and the high-definition camera is used to capture visible light images of the wind turbine blades. vis Infrared thermal images of the blades were acquired using an infrared thermal imager. ir .

[0076] In one specific embodiment, the unmanned helicopter flies according to a preset flight path, with a high-definition camera and an infrared thermal imager constantly aimed at the wind turbine blades to capture and scan them. The preset flight path is not limited here and can be set according to actual conditions. The flight speed and altitude of the unmanned helicopter, as well as the operating parameters of the high-definition camera and infrared thermal imager, can all be adjusted according to actual needs.

[0077] The primary function of the high-definition camera is to collect visible light images of the blades, which reveal the blade's surface physical structure and any obvious surface defects. Each image will be labeled with an I0. vis Value. Infrared thermal imagers are mainly used to collect infrared thermal images of leaves, which reflect the surface temperature distribution of the leaves. Each infrared thermal image is calibrated as an I value. ir value.

[0078] Through the embodiments of the present invention, visible light images and infrared thermal images of the blades can be acquired simultaneously, providing data support for subsequent image analysis and defect detection. At the same time, through real-time image acquisition, abnormal conditions of the blades can be quickly detected, improving the safe operation of wind turbines.

[0079] Step S102: The visible light image and the infrared thermal image are preprocessed respectively, and the preprocessed visible light image and infrared thermal image are defect detected by a preset image analysis model to generate surface defect data and thermal anomaly data of wind turbine blades.

[0080] In this embodiment of the invention, the visible light image and the infrared thermal image are preprocessed respectively, including: using a Gaussian filter to denoise the visible light image to enhance its quality; using histogram equalization to enhance the contrast of the infrared thermal image; and performing median filtering on the contrast-enhanced infrared thermal image to determine the temperature difference regions and thermal anomaly regions of the blades.

[0081] In this embodiment of the invention, during the image enhancement stage, for the visible light image I... vis and infrared thermal image I ir Different enhancement strategies were employed to more accurately identify and locate defects in wind turbine blades.

[0082] For visible light image I vis First, denoising is performed to enhance image quality and improve the accuracy of defect detection. In this step, a Gaussian filter is used to reduce random noise in the image. The Gaussian filter performs a convolution operation on the image, determining the new grayscale value of each pixel through a weight distribution (i.e., a Gaussian distribution).

[0083] A Gaussian filter is expressed by the following formula:

[0084] Ienh=G(I vis ,σ)

[0085] Where Ienh represents the denoised visible light image, G represents the Gaussian filter, and σ represents the standard deviation, which needs to be adjusted according to the noise level of the actual image.

[0086] For infrared thermal image I ir First, histogram equalization is used to enhance contrast, aiming to more clearly reveal temperature differences and thermal anomalies on the blades. Histogram equalization optimizes the grayscale histogram of the image, making its distribution more uniform, thereby improving the overall contrast of the image. After this step, temperature differences and thermal anomaly areas on the blades will be more noticeable. The following formula is used to enhance the contrast of infrared thermal images using histogram equalization:

[0087] I ir,contrast =HE(I ir )

[0088] Among them, I ir,contrast This represents the infrared thermal image after contrast enhancement, I ir This represents an infrared thermal image, and HE represents the histogram equalization algorithm.

[0089] While enhanced contrast infrared thermograms can better reveal temperature differences, they also introduce some noise. To address this issue, further research was conducted on I... ir,contrast A median filter was applied for noise reduction. The median filter effectively eliminates salt-and-pepper noise and other noises while preserving the image's detailed features. The processing formula is as follows:

[0090] I ir,denoise =MF(I ir,contrast ,k)

[0091] Where MF represents the median filtering algorithm, and k is the filter window size. Usually, a small filter window of 5×5 is chosen to preserve the detailed features of the image. The filter window is not limited here and can be set according to the actual situation.

[0092] Image enhancement processing reveals surface and internal information of the blades through visible light images and infrared thermal images. In subsequent defect detection and analysis, both types of images will be considered simultaneously to more comprehensively and accurately assess the overall condition and performance of the blades. The combined use of visible light and infrared technologies ensures a comprehensive and in-depth identification and analysis of wind turbine blade defects.

[0093] In the defect detection stage, a phased strategy is adopted, which processes visible light images and infrared thermal images separately. In this way, not only can surface defects be accurately identified, but also potential internal defects and temperature anomalies can be revealed, enabling comprehensive and multi-dimensional monitoring of wind turbine blades.

[0094] For visible light images, a Mask R-CNN model is employed. Feature extraction is performed using a pre-trained convolutional neural network, and a region proposal network is combined to generate candidate regions. ROI Align technology is then used to resize these candidate regions to a uniform size. Based on this, fully connected layers are used to perform classification tasks and bounding box regression, thereby accurately identifying and locating various surface defects on wind turbine blades.

[0095] For processing infrared thermal images, the Mask R-CNN model was continued, but with parameter and structural optimizations to make it more sensitive to temperature anomalies and internal defects. Infrared thermal images revealed complex temperature distributions on the blades, and through fine-tuning and training, abnormal hot zones could be accurately identified and located, revealing potential internal defects.

[0096] In this embodiment of the invention, the defect classification process uniformly adopts support vector machine to classify defects identified from two image sources. Multiple defect types are preset, and each defect region is processed to calculate a one-dimensional vector. Each dimension represents the probability that the region belongs to a certain type of defect. The final classification of the defect is determined by the index corresponding to the maximum value in the vector.

[0097] Through image analysis, potential surface defects and thermal anomalies can be effectively detected from visible light images and infrared thermal images of wind turbine blades. This provides important information and practically significant detection data for subsequent ultrasonic testing and data fusion, thereby enabling effective monitoring and maintenance of the blades.

[0098] Step S103: Preprocess the ultrasonic reflection signal and perform defect detection on the preprocessed ultrasonic reflection signal to generate data on defects in the internal structure of the wind turbine blade.

[0099] In this embodiment of the invention, an ultrasonic detector is used to scan the blade and record the ultrasonic wave reflection signal. Ultrasonic detection is a non-destructive testing method used to detect structural defects inside materials. The ultrasonic detector transmits high-frequency sound waves to detect the internal structure of the wind turbine blade. When the sound waves encounter defects inside the material (e.g., cracks or cavities), they are reflected and received by the ultrasonic detector. The propagation time and intensity of the ultrasonic waves are affected by these internal defects. Therefore, by analyzing this information, possible defects inside the blade can be detected.

[0100] In one specific embodiment, let t be the time from when the ultrasonic wave is emitted from the detector to when the reflected signal is received, and v be the propagation speed of the ultrasonic wave in the blade material. Then, the depth d of the defect can be calculated using the following formula:

[0101] d = v * t / 2

[0102] The propagation speed v of the ultrasonic wave is known and mainly depends on the material properties of the wind turbine blades. It is selected according to the actual situation. The depth d of the defect is calculated by measuring the propagation time t of the ultrasonic wave.

[0103] During ultrasonic testing, the unmanned helicopter needs to use a specific scanning pattern to ensure that the ultrasonic detector fully covers all areas of the blade. The design of the scanning pattern must take into account the shape and size of the blade, as well as the working distance and coverage area of ​​the ultrasonic detector. Furthermore, to improve the efficiency of ultrasonic testing, appropriate scanning speeds and frequencies can be designed.

[0104] After ultrasonic testing, the data collected by the ultrasonic detector needs further processing and analysis. First, noise filtering and signal enhancement are performed to improve data quality. Then, by comparing reflection signals from different regions, areas with potential defects can be identified. Finally, depth calculations and location determination are performed on these areas to obtain information about defects inside the blade.

[0105] In this embodiment of the invention, the ultrasonic testing step uses high-frequency sound waves to detect the internal structure of the wind turbine blades. By analyzing the reflected ultrasonic signals, possible structural defects inside the blades can be effectively detected, thereby further improving the safety performance of the wind turbine blades.

[0106] Step S104: Using a preset algorithm, the surface defect data, thermal anomaly data, and data on defects in the internal structure of the wind turbine blades are fused together to complete the defect detection of the wind turbine blade surface.

[0107] In this embodiment of the invention, the data fusion step is designed to fuse the image analysis results and the ultrasonic detection results to eliminate false detections and to more accurately locate and describe any possible blade defects.

[0108] First, assign an (n+1)-dimensional vector xp to each detected pixel, where the n-dimensional vector dp1 represents the probability that the pixel may belong to each defect type, and the (n+1)-dimensional value dp2 represents the depth of the defect. xp = [dp1, dp2] contains all the defect information for the pixel, including both the defect type and the defect depth.

[0109] Secondly, an unsupervised clustering algorithm, K-means, is used to cluster all xp data. The goal is to divide n data points into k clusters, minimizing the distance between data points within the same cluster and maximizing the distance between different clusters. Each cluster center can represent a specific type and depth of defect, and each pixel is assigned to the nearest cluster center, thus determining its defect type and depth.

[0110] In one specific embodiment, the working process of the K-means clustering algorithm can be divided into the following steps:

[0111] Step a: Randomly select k points as initial cluster centers;

[0112] Step b: For each data point, calculate its distance to each cluster center and assign it to the nearest cluster center;

[0113] Step c: For each cluster, calculate the average of all its data points and update the cluster centers;

[0114] Repeat steps b and c until the cluster centers no longer change, or the predetermined number of iterations is reached.

[0115] For each pixel, its distance can be defined as its Euclidean distance to the cluster center in the n+1 dimensional vector space. Furthermore, by choosing an appropriate k value, the sensitivity and discriminative ability of defects can be adjusted.

[0116] This invention fuses data obtained through image analysis and ultrasonic testing, which can effectively reduce false detections and accurately determine the type and depth of defects.

[0117] The wind turbine blade surface defect detection method provided in this embodiment effectively improves the system's detection accuracy and precision by integrating image analysis results and ultrasonic testing results. It features low human intervention, employing automated image analysis, ultrasonic testing, and data fusion technologies. Human involvement is primarily limited to system setup and parameter setting; the detection and early warning processes are automated, reducing operator workload and improving work efficiency.

[0118] This embodiment provides a method for detecting surface defects on wind turbine blades, which can be used on the aforementioned mobile terminals, such as mobile phones and tablets. Figure 2 This is a flowchart of a method for detecting surface defects on wind turbine blades according to an embodiment of the present invention, as follows: Figure 2 As shown, the process includes the following steps:

[0119] Step S201: Acquire images of the wind turbine blade surface and ultrasonic reflection signals from inside the blade, collected by the detection device. The blade surface images include visible light images and infrared thermal images. For details, please refer to [link to relevant documentation]. Figure 1 Step S101 of the illustrated embodiment will not be described again here.

[0120] Step S202 involves preprocessing the visible light image and the infrared thermal image, respectively, and then using a preset image analysis model to perform defect detection on the preprocessed visible light image and infrared thermal image, generating surface defect data and thermal anomaly data for the wind turbine blade. For details, please refer to... Figure 1 Step S102 of the illustrated embodiment will not be described again here.

[0121] Step S203 involves preprocessing the ultrasonic reflection signal and performing defect detection on the preprocessed ultrasonic reflection signal to generate data on defects in the internal structure of the wind turbine blade. For details, please refer to [link to relevant documentation]. Figure 1 Step S103 of the illustrated embodiment will not be described again here.

[0122] Step S204: Using a preset algorithm, data on surface defects, thermal anomalies, and internal structural defects of the wind turbine blades are fused to complete the defect detection of the wind turbine blade surface. For details, please refer to [link to relevant documentation]. Figure 1 Step S104 of the illustrated embodiment will not be described again here.

[0123] Step S205: Based on the preset warning level and preset threshold, issue a warning alarm for the defect detection results on the surface of the wind turbine blade and send a warning signal to the operator.

[0124] In this embodiment of the invention, the purpose of the early warning alarm is to send an early warning signal to the operator when a potential defect appears on the wind turbine blade, so as to take timely maintenance measures to prevent the defect from expanding further or causing a failure. The design of this step reflects the preventive and real-time characteristics of the invention.

[0125] The specific implementation method of the early warning and alarm steps is as follows:

[0126] First, we define m cluster centers, each representing a specific type and depth of defect. Then, we assign a weight value wi to each cluster center i, representing the severity of the defect at that type and depth. The weight values ​​are set according to the actual situation; they can be set by experts or learned and optimized through extensive experimental data.

[0127] Secondly, the preset warning level L is calculated. The preset warning level L is an index used to measure the overall degree of defect in the blade, and its calculation formula is as follows:

[0128] L=∑wi*Ni

[0129] Where Ni represents the number of pixels belonging to the i-th cluster center, that is, the number of pixels belonging to the i-th type and depth of defects. According to this formula, the warning level L is the weighted sum of the weight values ​​of all defects and the number of pixels, reflecting the overall situation of defects of various types and depths on the blade.

[0130] Then, a preset threshold T is set. When the preset warning level L exceeds the preset threshold T, a warning signal is sent to the operator. The setting of the preset threshold T needs to take into account various factors such as the material properties of the blade, the working environment, and safety requirements. It is set according to the actual situation and is generally determined by experts based on the actual situation.

[0131] The method provided in this invention, through setting warning levels and thresholds, enables real-time monitoring and early warning of defects in wind turbine blades, thereby improving the safety and stability of wind turbines and extending their service life.

[0132] The wind turbine blade surface defect detection method provided in this embodiment calculates the warning level based on image analysis and ultrasonic detection results. When the warning level exceeds the threshold, a warning signal is issued to the operator. In this way, potential defects can be warned in advance, and maintenance measures can be taken to achieve preventive maintenance and avoid the defects from expanding and causing serious consequences.

[0133] Three sets of experimental data were obtained through embodiments of the present invention, and the analysis of the three sets of experimental data is as follows:

[0134] Experimental data 1:

[0135] Ten inspections of a wind turbine blade were conducted using an inspection system mounted on a drone. Each inspection recorded the number of areas with potential defects obtained from image analysis and the number of areas with potential internal defects obtained from ultrasonic detection.

[0136] The results are recorded as follows:

[0137] First test: Image analysis revealed 15 defect areas, and ultrasonic detection revealed 12 internal defect areas.

[0138] Second time: Image analysis revealed 18 defect areas, and ultrasonic detection revealed 15 internal defect areas. ......

[0140] 10th time: Image analysis revealed 17 defect areas, and ultrasonic detection revealed 14 internal defect areas.

[0141] Results Analysis: The recorded results show that, in most cases, image analysis identified a greater number of potentially defective areas than ultrasonic detection. This indicates that image analysis has a certain number of false positives, while ultrasonic detection has fewer. Therefore, data fusion, combining the two detection results, can effectively reduce the overall number of false positives.

[0142] Experimental Result 2:

[0143] Ten blade samples with different types and depths of defects were inspected using an inspection system mounted on a drone. After each inspection, the actual types and depths of defects on the blade were compared with the results detected by the inspection system, and the average error rate was calculated.

[0144] The average error rate is calculated as follows:

[0145] Using image analysis alone: ​​the average error rate for defect type is 15%, and the average error rate for defect depth is 25%.

[0146] Using only ultrasonic detection: the average error rate for defect type is 8%, and the average error rate for defect depth is 7%.

[0147] Data fusion using image analysis and ultrasonic detection: average error rate of 5% for defect type and 4% for defect depth.

[0148] Results Analysis: Using only a single detection method introduces certain errors in defect identification. However, by fusing data and integrating the results of image analysis and ultrasonic detection, the judgment of defect type and depth can be significantly improved, thereby enhancing detection accuracy.

[0149] Experimental data 3:

[0150] Ten blade samples with different types of defects but the same depth were inspected using an inspection system mounted on a drone. The warning levels were 3.2, 2.8, 4.5, 6.1, 5.3, 7.8, 4.9, 6.7, 8.2 and 5.6, respectively.

[0151] If the threshold is set to 6, the warning level of the 4th and 6th samples will exceed the threshold, and a warning signal will be sent to the operator.

[0152] Results Analysis: At the same depth, different types of defects exhibit varying degrees of severity. Based on warning levels and thresholds, the detection system can distinguish which defects are more severe and require timely repair.

[0153] This embodiment also provides a wind turbine blade surface defect detection device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0154] This embodiment provides a device for detecting surface defects on wind turbine blades, such as... Figure 3 As shown, it includes:

[0155] The acquisition module 301 is used to acquire images of the surface of the wind turbine blade and ultrasonic reflection signals inside the blade collected by the detection device. The images of the blade surface include visible light images and infrared thermal images.

[0156] The first detection module 302 is used to preprocess the visible light image and the infrared thermal image respectively, and use a preset image analysis model to perform defect detection on the preprocessed visible light image and infrared thermal image respectively, and generate surface defect data and thermal anomaly data of wind turbine blades.

[0157] The second detection module 303 is used to preprocess the ultrasonic reflection signal and perform defect detection on the preprocessed ultrasonic reflection signal to generate data on defects in the internal structure of the wind turbine blade.

[0158] The fusion module 304 is used to fuse data on surface defects, thermal anomalies, and internal structural defects of wind turbine blades using a preset algorithm, thereby completing the defect detection of the wind turbine blade surface.

[0159] In some optional embodiments, the wind turbine blade surface defect detection device further includes:

[0160] The early warning module 305 is used to issue early warning alarms based on the preset early warning level and preset threshold for the defect detection results on the surface of the wind turbine blade, and to send early warning signals to the operator.

[0161] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0162] The wind turbine blade surface defect detection device in this embodiment is presented in the form of a functional unit. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0163] This invention also provides a computer device having the above-described features. Figure 3 The device shown is for detecting surface defects on wind turbine blades.

[0164] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 4 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 4 Take a processor 10 as an example.

[0165] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0166] The memory 20 stores instructions executable by at least one processor 10 to cause the at least one processor 10 to perform the method shown in the above embodiments.

[0167] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0168] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0169] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0170] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0171] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for detecting surface defects in wind turbine blades, characterized in that, The method includes: The detection device acquires images of the surface of the wind turbine blade and ultrasonic reflection signals inside the blade. The images of the blade surface include visible light images and infrared thermal images. The visible light image and the infrared thermal image are preprocessed respectively, and the preprocessed visible light image and infrared thermal image are defect detected by a preset image analysis model to generate surface defect data and thermal anomaly data of wind turbine blades. The ultrasonic reflection signal is preprocessed, and the preprocessed ultrasonic reflection signal is then subjected to defect detection to generate data on defects in the internal structure of the wind turbine blade. Using a preset algorithm, data on surface defects, thermal anomalies, and internal structural defects of wind turbine blades are fused together to complete the defect detection of wind turbine blade surfaces. Data on surface defects, thermal anomalies, and internal structural defects in wind turbine blades are fused to complete the defect detection of wind turbine blade surfaces, including: The K-means clustering algorithm is used to fuse surface defect data, thermal anomaly data, and internal structural defect data of wind turbine blades to determine the type and depth of surface defects and complete the defect detection of wind turbine blades. Based on preset warning levels and thresholds, the system issues warnings and alarms for defect detection results on the surface of wind turbine blades, and sends warning signals to the operator, including: There are m cluster centers, each representing a defect of a preset defect type and depth; Assign a weight value wi to each cluster center i, representing the severity of the defect of that type and depth; The preset warning level L is calculated using the following formula: L = ∑wi In Where Ni represents the number of pixels belonging to the i-th cluster center; Set a preset threshold T. When the preset warning level L exceeds the preset threshold T, a warning signal is sent to the operator.

2. The method according to claim 1, characterized in that, The steps for detecting defects on the surface of wind turbine blades also include: Based on preset warning levels and preset thresholds, the system issues warnings and alarms for the detection results of defects on the surface of wind turbine blades and sends warning signals to the operator.

3. The method according to claim 1, characterized in that, The detection device includes: an unmanned helicopter, a high-definition camera, an infrared thermal imager, an ultrasonic detector, and cables, wherein... High-definition cameras, infrared thermal imagers, and ultrasonic detectors are directly connected to the unmanned helicopter via cables. High-definition cameras are used to collect visible light images; Thermal imagers are used to collect infrared thermal images; Ultrasonic detectors are used to collect ultrasonic reflected signals from the internal structure of blades.

4. The method according to claim 3, characterized in that, High-definition cameras, infrared thermal imagers, and ultrasonic detectors are all fixed to the bottom of the unmanned helicopter.

5. The method according to claim 4, characterized in that, The unmanned helicopter is a six-axis stabilized multi-rotor unmanned helicopter.

6. The method according to claim 5, characterized in that, The unmanned helicopter flies around the wind turbine blades according to a preset flight path to complete the photography and scanning of the wind turbine blades.

7. The method according to claim 1, characterized in that, The preprocessing of the visible light image and the infrared thermal image includes: The visible light image is denoised using a Gaussian filter to enhance its quality. Histogram equalization is used to enhance the contrast of the infrared thermal image, and median filtering is applied to the enhanced infrared thermal image to determine the temperature difference region and thermal anomaly region of the blade.

8. The method according to claim 7, characterized in that, The visible light image is denoised using a Gaussian filter according to the following formula: Ienh = G( ,σ) Where Ienh represents the denoised visible light image, G represents the Gaussian filter, and σ represents the standard deviation. Represents a visible light image.

9. The method according to claim 7, characterized in that, The infrared thermal image is contrast-enhanced using histogram equalization according to the following formula: in, This represents an infrared thermal image with enhanced contrast. Represents an infrared thermal image. This represents the histogram equalization algorithm.

10. The method according to claim 9, characterized in that, The infrared thermal image after contrast enhancement is processed using the following formula for median filtering: in, This represents the infrared thermogram after median filtering. This represents the median filtering algorithm. This indicates the window size of the filter.

11. A device for detecting surface defects in wind turbine blades, characterized in that, The device includes: The acquisition module is used to acquire images of the surface of the wind turbine blade and ultrasonic reflection signals inside the blade collected by the detection device. The images of the blade surface include visible light images and infrared thermal images. The first detection module is used to preprocess the visible light image and the infrared thermal image respectively, and use a preset image analysis model to perform defect detection on the preprocessed visible light image and infrared thermal image respectively, and generate surface defect data and thermal anomaly data of wind turbine blades. The second detection module is used to preprocess the ultrasonic reflection signal and perform defect detection on the preprocessed ultrasonic reflection signal to generate data on defects in the internal structure of the wind turbine blade. The fusion module is used to fuse data on surface defects, thermal anomalies, and internal structural defects of wind turbine blades using a preset algorithm, thereby completing the defect detection of the wind turbine blade surface.

12. A computer device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the wind turbine blade surface defect detection method according to any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the wind turbine blade surface defect detection method according to any one of claims 1 to 10.

Images