A UAV intelligent identification system for wind turbine blade defects

Abstract

This invention provides an intelligent UAV identification system for wind turbine blade defects. The system includes an inspection and tracking module, a data acquisition module, a defect identification module, and a data integration module. The inspection and tracking module is used to plan and track the inspection route of the UAV. The data acquisition module is used to collect image information of the wind turbine blades. The defect identification module is used to analyze and identify defect information of the wind turbine blades. The data integration module is used to integrate the defect information of the wind turbine blades to generate visual results and maintenance suggestions. This invention achieves efficient detection, accurate identification, and visual analysis of wind turbine blade defects, significantly improving inspection efficiency and the scientific nature of maintenance decisions.

Description

Technical Field

[0001] This invention relates to the field of wind power equipment testing system technology, and in particular to an intelligent UAV identification system for wind turbine blade defects. Background Technology

[0002] With the rapid development of the wind power industry, wind turbine blades, as key components of wind power generation equipment, directly affect the safety and efficiency of wind turbine units. However, due to their long-term exposure to complex natural environments, wind turbine blades are susceptible to defects such as cracks, corrosion, and blade detachment caused by strong winds, dust storms, rain, snow, and ultraviolet radiation. If these defects are not detected and addressed in a timely manner, they may lead to a decline in blade performance or even structural damage, thereby affecting the operational reliability of the entire wind turbine unit. Traditional wind turbine blade inspection methods mainly rely on manual inspections or image acquisition using ground-based equipment. These methods suffer from low efficiency, limited coverage, and reliance on human experience, making it difficult to meet the large-scale, high-frequency inspection needs of wind farms. Furthermore, the analysis of inspection data typically relies on manual judgment, lacking efficient and standardized processing methods, further increasing maintenance difficulty and costs.

[0003] A review of publicly available technical solutions reveals that CN118188349A proposes a method, device, and system for detecting wind turbine blade defects based on unmanned aerial vehicles (UAVs). The UAV is equipped with a camera and a spectrometer. The detection method includes: acquiring images of the wind turbine blades taken by the camera and obtaining the wind turbine blade spectra by scanning them with the spectrometer; identifying the wind turbine blade images and spectra using a deep learning-based defect detection model to obtain surface defect information; and using a deep learning-based defect detection model to identify the wind turbine blade images and spectra allows for rapid and accurate acquisition of surface defect information. This addresses the problem that current UAV-based wind turbine blade detection technologies primarily rely on manual image analysis, lacking effective automatic identification and analysis. While this solution achieves automatic identification of wind turbine blade surface defects, it relies solely on image and spectral data for detection, without addressing the severity assessment and causal analysis of defects, resulting in a vague analysis of the criticality and impact range of defects. Furthermore, this solution does not provide functions for defect priority ranking and maintenance suggestion generation, making it difficult to support subsequent maintenance optimization and resource allocation. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of current systems by proposing an intelligent UAV identification system for wind turbine blade defects.

[0005] The present invention adopts the following technical solution:

[0006] A drone-based intelligent identification system for wind turbine blade defects includes an inspection and tracking module, a data acquisition module, a defect identification module, and a data integration module. The inspection and tracking module is used to plan and track the inspection route of the drone. The data acquisition module is used to acquire image information of the wind turbine blades. The defect identification module is used to analyze and identify defect information of the wind turbine blades. The data integration module is used to integrate the defect information of the wind turbine blades to generate visualization results and maintenance suggestions.

[0007] The inspection and tracking module includes a route planning unit and a trajectory tracking unit; the route planning unit is used to generate inspection routes based on wind turbine distribution and inspection requirements; the trajectory tracking unit is used to monitor the flight status of the UAV in real time and control the UAV to travel along the inspection route.

[0008] The data acquisition module includes an image acquisition unit and a data transmission unit; the image acquisition unit is used to acquire image information of the wind turbine blades, and the data transmission unit is used to transmit the image information to a subsequent processing module on the ground via wireless communication technology.

[0009] The defect identification module and data integration module are located on the ground. The defect identification module includes an image preprocessing unit and a defect identification unit. The image preprocessing unit is used to preprocess the raw wind turbine blade image information collected by the UAV. The defect identification unit is used to analyze and identify the specific defect type and defect location of the wind turbine blade from the preprocessed wind turbine blade image information.

[0010] Furthermore, the data integration module includes a data storage unit, a maintenance suggestion generation unit, a defect cause analysis unit, and a display unit; the data storage unit is used to store image information and defect information of the wind turbine blades; the maintenance suggestion generation unit is used to generate maintenance suggestions for the wind turbine blades; the defect cause analysis unit is used to analyze the cause patterns of wind turbine blade defects; and the display unit is used to visualize the content collected and analyzed by the system.

[0011] Furthermore, the defect identification unit completes the classification and priority ranking of wind turbine blade defects through a multi-task learning model; the framework of the multi-task learning model includes:

[0012] Shared feature layer: used to acquire input image information and extract global features related to wind turbine blade defect classification and priority ranking from the input image information;

[0013] Defect classification branch layer: used to analyze global features, obtain the specific type of wind turbine blade defect and output the corresponding category probability;

[0014] Priority ranking branch layer: used to assess the severity of defects and output a priority score based on the severity of the defect type. The priority score means the urgency of defect repair, and a higher score indicates that it needs to be dealt with first.

[0015] Furthermore, the multi-task learning model employs a joint loss function to optimize the defect classification and prioritization tasks; the joint loss function is expressed as follows:

[0016] L=α·L cls +β·L bank ;

[0017] Where L is the joint loss function; α is the weight of the defect classification task; β is the weight of the priority ranking task; and L cls L is the loss function value for the defect classification task. bank The loss function value for the priority ranking task; for α and β, the following holds:

[0018]

[0019] Furthermore, the defect identification unit inputs the wind turbine blade image information corresponding to each wind turbine blade position into the multi-task learning model, thereby outputting the specific defect type of the wind turbine blade at that position and the priority score corresponding to the defect type.

[0020] The beneficial effects achieved by this invention are as follows:

[0021] This solution, by combining a multi-task learning model, achieves accurate classification and priority assessment of wind turbine blade defects, providing a quantitative analysis basis for the urgency of repairs, significantly improving inspection efficiency and reducing maintenance costs. By introducing a dynamic weight allocation mechanism in the joint loss function design of the multi-task learning model, the weights of classification and ranking tasks are automatically adjusted according to the task difficulty, achieving balanced optimization of tasks and further improving the overall performance of the model and the reliability of the output results. In addition, the solution also includes generating maintenance suggestions for wind turbine blades and analyzing the causal patterns of defects, thereby optimizing maintenance strategies and improving the reliability and operational efficiency of wind turbines. Attached Figure Description

[0022] The invention will be further understood from the following description taken in conjunction with the accompanying drawings. The components in the drawings are not necessarily drawn to scale, but rather the emphasis is on illustrating the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0023] Figure 1 This is a schematic diagram of the overall modules of the present invention.

[0024] Figure 2 This is a schematic diagram of the multi-task learning model framework of the present invention.

[0025] Figure 3 This is a schematic diagram of the workflow of the defect cause analysis unit of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to its embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention. Other systems, methods, and / or features of this embodiment will become apparent to those skilled in the art after reviewing the following detailed description. It is intended that all such additional systems, methods, features, and advantages are included within this specification, are included within the scope of the present invention, and are protected by the appended claims. Further features of the disclosed embodiments are described in the following detailed description, and these features will be apparent from the following detailed description.

[0027] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present patent. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0028] Example 1:

[0029] like Figure 1 As shown in the figure, this embodiment provides an intelligent UAV identification system for wind turbine blade defects. The system includes an inspection and tracking module, a data acquisition module, a defect identification module, and a data integration module. The inspection and tracking module is used to plan and track the inspection route of the UAV. The data acquisition module is used to acquire image information of the wind turbine blades. The defect identification module is used to analyze and identify defect information of the wind turbine blades. The data integration module is used to integrate the defect information of the wind turbine blades to generate visualization results and maintenance suggestions.

[0030] The inspection and tracking module includes a route planning unit and a trajectory tracking unit; the route planning unit is used to generate inspection routes based on wind turbine distribution and inspection requirements; the trajectory tracking unit is used to monitor the UAV's flight status in real time and control the UAV to travel along the inspection routes.

[0031] The data acquisition module includes an image acquisition unit and a data transmission unit; the image acquisition unit is used to acquire image information of the wind turbine blades, and the data transmission unit is used to transmit the image information to a subsequent processing module on the ground via wireless communication technology.

[0032] The defect identification module and data integration module are located on the ground. The defect identification module includes an image preprocessing unit and a defect identification unit. The image preprocessing unit is used to preprocess the raw wind turbine blade image information collected by the UAV. The defect identification unit is used to analyze and identify the specific defect type and defect location of the wind turbine blade from the preprocessed wind turbine blade image information.

[0033] Furthermore, the data integration module includes a data storage unit, a maintenance suggestion generation unit, a defect cause analysis unit, and a display unit; the data storage unit is used to store image information and defect information of the wind turbine blades; the maintenance suggestion generation unit is used to generate maintenance suggestions for the wind turbine blades; the defect cause analysis unit is used to analyze the cause patterns of wind turbine blade defects; and the display unit is used to visually display the content collected and analyzed by the system.

[0034] Furthermore, the image preprocessing unit specifically includes the following preprocessing operations on the original wind turbine blade image information:

[0035] Image denoising: removing noise introduced during image capture due to sensor noise or environmental interference;

[0036] Image enhancement: Improves the brightness and contrast of the image, making defects on the surface of the wind turbine blades clearer;

[0037] Image geometric correction: Corrects image distortion caused by drone shooting angle or lens distortion, so that the wind turbine blades present the correct proportions in the image;

[0038] Image segmentation and region extraction: Extract the main regions of the wind turbine blades in the image, remove background information, and reduce the complexity of subsequent processing;

[0039] Image normalization: Normalizing the size of an image;

[0040] Furthermore, the defect identification unit uses a multi-task learning model to classify wind turbine blade defects and prioritize them; for example... Figure 2 As shown, the framework of the multi-task learning model includes:

[0041] Shared feature layer: Used to acquire input image information and extract global features related to wind turbine blade defect classification and priority ranking from the input image information; its form is expressed as follows:

[0042] F = CNN(X; θ) s );

[0043] Where F represents the extracted global features, and X represents the input image information, i.e., the wind turbine blade image information after preprocessing by the image preprocessing unit; θ s These are the model parameters for the shared feature layers; CNN() is the operation function for a deep convolutional neural network.

[0044] Defect classification branch layer: used to analyze global features, obtain the specific types of wind turbine blade defects, and output the corresponding category probabilities; its form is expressed as follows:

[0045]

[0046] in, The predicted probability distribution for defect types; W cls and b cls These are the model parameters for the defect branch classification layer; Softmax() is the model activation function.

[0047] Priority ranking branch layer: Used to assess the severity of defects and output a priority score based on the severity corresponding to the defect type. The priority score represents the urgency of defect repair; a higher score indicates that it needs to be addressed first. Its form is expressed as follows:

[0048]

[0049] in, Defect priority is scored, with values ​​being continuous; W rank and b rank Prioritize the model parameters of the branch layer;

[0050] Furthermore, the multi-task learning model employs a joint loss function to optimize the defect classification and prioritization tasks; the joint loss function is expressed as follows:

[0051] L=α·L cls +β·L bank ;

[0052] Where L is the joint loss function; α is the weight of the defect classification task; β is the weight of the priority ranking task; and L cls L is the loss function value for the defect classification task. bank The loss function value for the priority ranking task; for α and β, the following holds:

[0053]

[0054] Furthermore, the loss function for the defect classification task can be obtained through cross-entropy calculation;

[0055] Furthermore, the loss function of the priority ranking task can be obtained through mean squared error calculation;

[0056] Furthermore, the defect identification unit inputs the wind turbine blade image information corresponding to each wind turbine blade position into the multi-task learning model, thereby outputting the specific defect type of the wind turbine blade at that position and the priority score corresponding to the defect type.

[0057] This solution combines a multi-task learning model to classify wind turbine blade defects and assess their priority, achieving accurate identification of defect types and quantitative analysis of repair urgency. This significantly improves inspection efficiency and reduces inspection and maintenance costs. By setting a shared feature layer in the multi-task learning model, the model shares global features in classification and ranking tasks, greatly reducing computational redundancy and ensuring efficient collaboration between tasks. By dynamically allocating the weights of defect classification and priority ranking tasks according to task difficulty in the multi-task joint loss function, balanced optimization of classification and ranking tasks is achieved, further improving the overall performance and reliability of the multi-task learning model's output.

[0058] Example 2:

[0059] This embodiment should be understood to include at least all the features of any of the foregoing embodiments, and to further improve upon them;

[0060] This embodiment provides an intelligent UAV identification system for wind turbine blade defects. The system includes an inspection and tracking module, a data acquisition module, a defect identification module, and a data integration module. The inspection and tracking module is used to plan and track the inspection route of the UAV. The data acquisition module is used to acquire image information of the wind turbine blades. The defect identification module is used to analyze and identify defect information of the wind turbine blades. The data integration module is used to integrate the defect information of the wind turbine blades to generate visualization results and maintenance suggestions.

[0061] Furthermore, the data integration module includes a data storage unit, a maintenance suggestion generation unit, a defect cause analysis unit, and a display unit; the data storage unit is used to store image information and defect information of the wind turbine blades; the maintenance suggestion generation unit is used to generate maintenance suggestions for the wind turbine blades; the defect cause analysis unit is used to analyze the cause patterns of wind turbine blade defects; and the display unit is used to visualize and display the content collected and analyzed by the system.

[0062] Furthermore, the wind turbine blade image information stored in the data storage unit includes image information corresponding to each position of the wind turbine blade, and the defect information includes the defect type corresponding to each position of the wind turbine blade and the priority score corresponding to the defect type.

[0063] Furthermore, the maintenance suggestion generation unit generates maintenance suggestions with corresponding priority scores based on the type and priority rating of the defects currently existing in the wind turbine blades, combined with a third-party expert knowledge base, and provides detailed maintenance operation plans.

[0064] Furthermore, such as Figure 3 As shown, the defect cause analysis unit completes the analysis of the cause mode of wind turbine blade defects in the following manner:

[0065] S1: Based on the structural characteristics and functional distribution of the wind turbine blades, they are divided into multiple regions;

[0066] S2: Set a fixed monitoring cycle and obtain defect information at the location of each wind turbine blade area within the monitoring cycle;

[0067] S3: For each region location, calculate the frequency index of each defect type at that region location:

[0068]

[0069] Among them, I k S is the frequency recurrence index of the k-th defect type in this region, which reflects the frequency and severity of the occurrence of this defect type in this region; priority,i γ is the priority score for the i-th occurrence of the k-th defect type, obtained by the defect identification unit during historical data collection; γ is a pre-set adjustment coefficient used to control the influence of defect occurrence time on the frequency repetition index, set through pre-experimentation; t end t is the end time of the monitoring period. total,i Let T be the time point when the k-th defect type occurs for the i-th time. total The total duration of the monitoring period;

[0070] S4: For each region, obtain the frequency index of all corresponding defect types, sort them from high to low, and extract the top L defect types as the key defect types for that region. L can be set according to user needs. The key defect types combine the frequency and severity of defect occurrence, reflecting the multiple defects that have the greatest impact on the reliability of wind turbine operation during the monitoring period.

[0071] S5: For each region location, based on the corresponding key defect type, and combined with the defect cause records in the historical database, match the cause patterns of the same defect type.

[0072] S6: For the matched cause patterns, verify them in conjunction with the operating data and environmental data within the current monitoring period, and filter out cause patterns that do not conform to the current operating conditions or environmental conditions.

[0073] S7: Sort the filtered causal patterns according to the strength of their association with the critical defect type to determine the priority of each causal pattern:

[0074] For a certain filtered causal pattern j:

[0075]

[0076] Among them, R j M is the causal priority coefficient for the j-th causal mode. k,j The correlation strength between the kth defect type and the jth cause mode in the critical defect types is preset based on the degree of matching between the defect cause records and defect types in the historical database.

[0077] S8: For each region, obtain the corresponding cause patterns arranged from largest to smallest according to the cause priority coefficient, and use them as the cause report of the defects in that region.

[0078] Furthermore, part of the running code of the defect cause analysis unit is as follows:

[0079]

[0080]

[0081]

[0082]

[0083] This solution first identifies key defect types and clarifies their impact on wind turbine operational reliability by comprehensively assessing the frequency and severity of defects. Then, it matches and verifies the causal patterns by combining the correlation between key defect types and their causes, and generates a defect cause analysis report according to priority, providing a scientific basis for optimizing maintenance strategies and resource allocation.

[0084] Furthermore, the content that the display unit visually displays to the user specifically includes:

[0085] A 3D display model is constructed based on a geographic information system to show the distribution of defects in each area of ​​each wind turbine in the wind farm; and defect information and image information are inserted into the corresponding positions in the 3D display model.

[0086] It presents maintenance recommendations for each area of ​​each wind turbine in the wind farm;

[0087] List the key defect types and their causes in each area of ​​each wind turbine in the wind farm, and sort them by priority.

[0088] The content disclosed above is only a preferred and feasible embodiment of the present invention, and is not intended to limit the scope of protection of the present invention. Therefore, all equivalent technical changes made based on the content of the present invention specification and drawings are included within the scope of protection of the present invention. Furthermore, the elements therein can be updated as technology develops.

Claims

1. A UAV intelligent identification system for wind turbine blade defects, characterized in that, The system includes an inspection and tracking module, a data acquisition module, a defect identification module, and a data integration module; the inspection and tracking module is used to plan and track the inspection route of the UAV; the data acquisition module is used to acquire image information of the wind turbine blades; The defect identification module is used to analyze and identify defect information of the wind turbine blades; the data integration module is used to integrate the defect information of the wind turbine blades to generate visualization results and maintenance suggestions. The inspection and tracking module includes a route planning unit and a trajectory tracking unit; the route planning unit is used to generate inspection routes based on wind turbine distribution and inspection requirements. The flight tracking unit is used to monitor the flight status of the UAV in real time and control the UAV to travel along the inspection route. The data acquisition module includes an image acquisition unit and a data transmission unit; the image acquisition unit is used to acquire image information of the wind turbine blades, and the data transmission unit is used to transmit the image information to a subsequent processing module on the ground via wireless communication technology. The defect identification module and data integration module are located on the ground. The defect identification module includes an image preprocessing unit and a defect identification unit; the image preprocessing unit is used to preprocess the raw wind turbine blade image information collected by the UAV. The defect identification unit is used to analyze and identify the specific defect type and location of the wind turbine blades from the preprocessed image information of the wind turbine blades. The data integration module includes a data storage unit, a maintenance suggestion generation unit, a defect cause analysis unit, and a display unit; the data storage unit is used to store image information and defect information of the wind turbine blades; the maintenance suggestion generation unit is used to generate maintenance suggestions for the wind turbine blades; The defect cause analysis unit is used to analyze the cause patterns of defects in wind turbine blades. The display unit is used to visually display the content collected and analyzed by the system; The defect identification unit uses a multi-task learning model to classify wind turbine blade defects and prioritize them. The framework of the multi-task learning model includes: Shared feature layer: Used to acquire input image information and extract global features related to wind turbine blade defect classification and priority ranking from the input image information; its form is expressed as follows: ； in, For the extracted global features, The input image information is the wind turbine blade image information after preprocessing by the image preprocessing unit; Model parameters for shared feature layers; These are the operational functions for deep convolutional neural networks; Defect classification branch layer: Used to analyze global features, obtain the specific type of wind turbine blade defect, and output the corresponding category probability; its form is expressed as follows: ； in, The predicted probability distribution for defect types; and For the model parameters of the defect branch classification layer; The activation function for the model; Priority ranking branch layer: Used to assess the severity of defects and output a priority score based on the severity corresponding to the defect type. The priority score represents the urgency of defect repair; a higher score indicates that it needs to be addressed first. Its form is expressed as follows: ； in, Defect priority is scored, with values ​​being continuous. and Prioritize the model parameters of the branch layer; The multi-task learning model employs a joint loss function to optimize the defect classification and prioritization tasks; the joint loss function is expressed as follows: ； in, For the joint loss function; The weights for the defect classification task, The weights of tasks are assigned to prioritize them. The loss function value for the defect classification task. The loss function value for prioritizing tasks; for and satisfy: ； ； The loss function for the defect classification task can be obtained through cross-entropy calculation; the loss function for the priority ranking task can be obtained through mean squared error calculation. The defect identification unit inputs the wind turbine blade image information corresponding to each wind turbine blade position into the multi-task learning model, thereby outputting the specific defect type of the wind turbine blade at that position and the priority score corresponding to the defect type. The defect cause analysis unit completes the analysis of the cause patterns of wind turbine blade defects in the following ways: S1: Based on the structural characteristics and functional distribution of the wind turbine blades, they are divided into multiple regions; S2: Set a fixed monitoring cycle and obtain defect information at the location of each wind turbine blade area within the monitoring cycle; S3: For each region location, calculate the frequency index of each defect type at that region location: ； in, For the first The frequency index of a defect type in the region reflects the frequency and severity of the occurrence of that defect type in the region. For the first Type of defect The priority score at the time of occurrence is obtained through the defect identification unit during the historical data collection process; The pre-set adjustment coefficient is used to control the degree of influence of the defect occurrence time on the frequency recurrence index, and is set through pre-experimentation; This is the end time of the monitoring period. For the first Type of defect The time point when it occurred The total duration of the monitoring period; S4: For each region location, obtain the frequency repetition index of all corresponding defect types, sort them from highest to lowest, and extract the top-ranked regions. The defect type is selected as the key defect type for that area. It can be set according to user needs; the key defect types combine the frequency and severity of defect occurrence, reflecting multiple defects that have the greatest impact on the reliability of wind turbine operation during the monitoring period; S5: For each region location, based on the corresponding key defect type, and combined with the defect cause records in the historical database, match the cause patterns of the same defect type. S6: For the matched cause patterns, verify them in conjunction with the operating data and environmental data within the current monitoring period, and filter out cause patterns that do not conform to the current operating conditions or environmental conditions. S7: Sort the filtered causal patterns according to the strength of their association with the critical defect type to determine the priority of each causal pattern: For a certain filtered cause pattern : ； in, For the first The causal priority coefficient of each causal pattern The first among the critical defect types Type of defect and the first The correlation strength between the various causal patterns is preset based on the degree of matching between defect cause records and defect types in the historical database. S8: For each location, obtain the corresponding cause pattern arranged from largest to smallest according to the cause priority coefficient, and use it as the cause report of the defect in that location.

Images