Lightning protection detection method for blades of wind driven generator of unmanned aerial vehicle

Abstract

The invention discloses an unmanned aerial vehicle wind driven generator blade lightning protection detection method, and belongs to the technical field of wind power generation. The system comprises a multi-mode perception and hidden defect detection module, a digital twin and full life cycle management module, an AI-driven autonomous decision module, an extreme environment adaptation and multi-model compatible module and a full-scene operation and maintenance closed-loop management module. The multi-mode sensing and hidden defect detection module integrates visible light, thermal imaging, ultraviolet, microwave radar and laser radar, and realizes high-precision detection of blade surface lightning stroke traces and internal downlead fracture hidden defects. According to the method, the hidden defect detection rate is improved through a multi-mode perception fusion technology, and the fracture positioning error of the internal downlead is controlled within + / -8cm; the installation angle deviation measurement precision of the lightning arrester is high and is greatly improved compared with a traditional method, the anti-interference design enables the automatic wiping frequency of the ultraviolet lens in the salt fog environment to be reduced, and the dynamic gain adjustment range of the thermal imaging module is expanded under the strong wind condition.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind power generation, and in particular relates to a lightning protection detection method for blades of wind turbines of unmanned aerial vehicles. Background Art

[0002] As the energy transition accelerates, wind power, a core pillar of clean and renewable energy, continues to see its installed capacity rise. By 2025, global wind power capacity will exceed 1.2 trillion watts, with offshore wind power accounting for over 35%. Wind turbine blades, core components for capturing wind energy, are exposed to complex natural environments for extended periods. Their lightning protection performance is directly linked to the reliability of the entire turbine and the economic profitability of the wind farm. Statistics show that lightning strikes account for 18% to 25% of wind turbine failures, with 70% of damage occurring in the blades. A single severe lightning strike can render a blade useless, resulting in significant economic losses.

[0003] Traditional blade lightning protection inspections rely primarily on manual inspections and regular maintenance, resulting in low efficiency, large blind spots, and poor real-time performance. In recent years, drone technology, owing to its flexibility and non-contact detection advantages, has gradually become an important means of blade lightning protection inspection. However, existing drone inspection solutions still face multiple technical bottlenecks: insufficient multi-source data fusion capabilities lead to a high rate of missed detection of hidden defects; sensor signal-to-noise ratio decreases in extreme environments such as strong winds, salt spray, and high temperature differences; inspection strategies rely on manual pre-setting and cannot be dynamically adjusted according to real-time operating conditions; differences in blade structures between different manufacturers lead to incompatible inspection equipment; and the disconnection of inspection data from operations and maintenance systems makes closed-loop management difficult. Summary of the Invention

[0004] The purpose of the present invention is to provide a lightning protection detection method for wind turbine blades of UAV to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solutions: including a multimodal perception and hidden defect detection module, a digital twin and full life cycle management module, an AI-driven autonomous decision-making module, an extreme environment adaptation and multi-model compatibility module, and a full-scenario operation and maintenance closed-loop management module; The multimodal perception and hidden defect detection module integrates visible light, thermal imaging, ultraviolet, microwave radar and lidar to achieve high-precision detection of lightning strike marks on the blade surface and hidden defects such as internal down conductor breakage; the digital twin and the full life cycle management module dynamically integrate detection data, operation data and meteorological data to build a wind turbine lightning protection digital twin; the AI-driven autonomous decision-making module is based on edge computing and reinforcement learning algorithms to identify defect types in real time and autonomously adjust detection strategies; the extreme environment adaptation and multi-model compatibility module adapts to strong wind and salt spray environments through anti-interference design, and modular load configuration supports multi-model detection; the full-scenario operation and maintenance closed-loop management module associates the "lightning risk-equipment" fault model to automatically trigger operation and maintenance work orders; As a further preferred embodiment of this technical solution, the following steps are included: Identify blade faults and construct a 3D model using LiDAR; Real-time synchronization of drone detection data to record the arc path and energy distribution at the moment the lightning receptor strikes lightning; Establish a "lightning strike risk-equipment" failure correlation model; As a further preferred embodiment of the present technical solution: the modal perception and hidden defect detection module detects lightning ablation marks on the blade surface and abnormal temperature rise of electrical connections through visible light and thermal imaging, the ultraviolet imager measures the intensity of local discharge at the down conductor connection, the microwave radar penetrates the composite material to detect the fracture of the internal down conductor, the laser radar constructs a three-dimensional point cloud model of the blade, and measures the installation angle deviation of the lightning receptor; As a further preferred embodiment of this technical solution, the digital twin is dynamically integrated with the full lifecycle management module to synchronize drone detection data, SCADA operation data, and meteorological data in real time to build a digital twin of the wind turbine lightning protection system. The digital twin also simulates lightning strike paths, predicts areas of thermal stress concentration in composite materials, and analyzes historical data showing that the corrosion rate of the unit's lightning receptors is positively correlated with ambient humidity, thereby optimizing the replacement cycle of the lightning receptors. As a further preferred embodiment of this technical solution: a lightweight CNN model is deployed in the AI-driven autonomous decision-making module to identify defect types and classify them in real time; the drone autonomously adjusts the inspection path based on AI feedback and extends the dwell time in key areas; based on the DQN algorithm, the optimal payload combination is autonomously selected during the inspection process; As a further preferred embodiment of the present technical solution: the extreme environment adaptability and multi-model compatibility module automatically switches to anti-interference mode in a strong wind environment, and the frequency of automatic wiping of the UV imager lens is increased to once per minute in a salt fog environment; and through modular payload configuration and dynamic route planning, it is compatible with a variety of direct drive, dual-fed, and semi-direct drive models; As a further preferred embodiment of the present technical solution: the "lightning strike risk-equipment" fault model in the full-scenario operation and maintenance closed-loop management module has input variables including the damage location of the lightning receptor, the electromagnetic field intensity of the thunderstorm cloud, and the operating load of the unit; after the warning, the wind turbine yaw angle is automatically adjusted to reduce the probability of lightning strike, and the backup grounding system is activated at the same time; As a further preferred embodiment of this technical solution: the drone transmits the test data to the ground station in real time. The test engineer completes the data pre-processing on site and outputs a preliminary report through the edge computing unit. The data is synchronized to the wind farm management platform. Combined with the historical test data, a "lightning protection health trend chart" is generated to predict the replacement cycle of the lightning receptor. As a further preferred embodiment of this technical solution: the blade point cloud data and test results collected by the drone are integrated with the BIM model to build a digital twin of the wind turbine lightning protection system. Through the platform, the historical test records of any lightning receptor and the simulated lightning strike path can be viewed in real time; As a further preferred embodiment of this technical solution: analyzing the correlation between the "lightning rod damage location" and the "thunderstorm cloud movement path" in historical detection data; combining the unit operation data of the SCADA system to establish a "lightning strike risk-equipment failure" correlation model to trigger operation and maintenance work orders in advance.

[0006] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention improves the detection rate of hidden defects through multimodal sensing fusion technology, among which the positioning error of internal down conductor fracture is controlled within ±8cm; the measurement accuracy of lightning terminal installation angle deviation is high, which is a significant improvement over traditional methods. The anti-interference design reduces the frequency of automatic wiping of ultraviolet lenses in salt fog environments, and expands the dynamic adjustment range of thermal imaging module gain under strong wind conditions.

[0007] 2. The present invention uses a lightweight CNN model to accurately identify defect types. The DQN algorithm improves the efficiency of detection path optimization, and the residence time in key areas is automatically extended to multiple times of conventional areas. The modular payload configuration supports seamless switching between direct-drive, dual-fed, and semi-direct-drive models. The detection resolution and coverage efficiency are high. The fault model extends the warning lead time to 72 hours, the automatic triggering rate of operation and maintenance work orders is high, and the fault handling response time is shortened. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 This is a structural diagram of a lightning protection detection method for UAV wind turbine blades of the present invention; Figure 2 This is a flow chart of a lightning protection detection method for UAV wind turbine blades according to the present invention. DETAILED DESCRIPTION

[0009] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0010] Example: See also Figure 1 - Figure 2 As shown, the present invention provides a technical solution: including a multimodal perception and hidden defect detection module, a digital twin and full life cycle management module, an AI-driven autonomous decision-making module, an extreme environment adaptation and multi-model compatibility module, and a full-scenario operation and maintenance closed-loop management module; The multimodal perception and hidden defect detection module integrates visible light, thermal imaging, ultraviolet, microwave radar, and lidar to achieve high-precision detection of lightning strike marks on blade surfaces and hidden defects such as internal down conductor breakage. The digital twin and full lifecycle management module dynamically integrate detection data, operational data, and meteorological data to build a digital twin for wind turbine lightning protection. The AI-driven autonomous decision-making module uses edge computing and reinforcement learning algorithms to identify defect types in real time and autonomously adjust detection strategies. The extreme environment adaptation and multi-model compatibility module adapts to strong winds and salt spray environments through anti-interference design, and modular payload configuration supports multi-model detection. The full-scenario operation and maintenance closed-loop management module links the "lightning risk-equipment" fault model to automatically trigger operation and maintenance work orders. In this embodiment, specifically, the following steps are included: Identify blade faults and construct a 3D model using LiDAR; Real-time synchronization of drone detection data to record the arc path and energy distribution at the moment the lightning receptor strikes lightning; Establish a "lightning strike risk-equipment" failure correlation model; In this embodiment, the modal perception and hidden defect detection module uses visible light and thermal imaging to detect lightning ablation marks on the blade surface and abnormal temperature rise of electrical connections. A Zenmuse H20T gimbal camera (20-megapixel wide-angle + 23x zoom) is used. The thermal imaging module parameters are: 640×512 resolution, 0.05°C temperature resolution. The ultraviolet imager measures the intensity of partial discharges at downconductor connections. A microwave radar penetrates composite materials to detect internal downconductor fractures. A lidar constructs a three-dimensional point cloud model of the blade and measures the angle deviation of the lightning receptor installation. Specifically, in this embodiment: dynamic data fusion is performed between the digital twin and the full lifecycle management module, synchronizing drone detection data, SCADA operation data, and meteorological data in real time, including the number of converter overvoltages and the movement path of thunderstorm clouds, to build a digital twin of the wind turbine lightning protection system. Lightning strike paths are simulated to predict areas of concentrated thermal stress in composite materials. Analysis of historical data shows that the corrosion rate of the unit's lightning receptors is positively correlated with ambient humidity. Based on this, the replacement cycle of the lightning receptors is optimized, adjusting it from a fixed three-year cycle to a dynamic two- to four-year cycle. In this embodiment, specifically: a lightweight CNN model is deployed in the AI-driven autonomous decision-making module to identify defect types in real time and classify them into levels: Level I for immediate processing and Level II for planned maintenance; the drone autonomously adjusts its inspection path based on AI feedback, extending its dwell time in key areas; based on the DQN algorithm, it autonomously selects the optimal payload combination during the inspection process, activating a UV imager and microwave radar in high-risk areas; In this embodiment, specifically: the module is adaptable to extreme environments and compatible with multiple models. In strong winds with a wind speed greater than 12m / s, it automatically switches to anti-interference mode, increases the gain of the thermal imaging module to 12dB, enables the lens heating and defogging function, and increases the automatic wiping frequency of the UV imager lens to once per minute in salt fog environments. Through modular payload configuration and dynamic route planning, it is compatible with multiple direct-drive, dual-fed, and semi-direct-drive models. In this embodiment, specifically: the "Lightning Strike Risk - Equipment" fault model in the full-scenario O&M closed-loop management module uses input variables including the location of lightning receptor damage, the electromagnetic field intensity of the thunderstorm cloud, and the unit's operating load. After an early warning, the wind turbine's yaw angle is automatically adjusted to reduce the probability of a lightning strike, while simultaneously activating the backup grounding system. In this embodiment, the drone transmits test data to the ground station in real time. The test engineer uses the edge computing unit to pre-process the data on-site and output a preliminary report. The data is synchronized to the wind farm management platform, and a "lightning protection health trend chart" is generated based on historical test data to predict the replacement cycle of the lightning receptors. In this embodiment, specifically: the blade point cloud data and test results collected by drones are integrated with the BIM model to build a digital twin of the wind turbine lightning protection system. Through the platform, the historical test records of any lightning receptor and the simulated lightning strike path can be viewed in real time; In this embodiment, specifically: the correlation between the "lightning terminal damage location" and the "thunderstorm cloud movement path" in the historical detection data is analyzed; combined with the unit operation data of the SCADA system, a "lightning strike risk-equipment failure" correlation model is established to trigger the operation and maintenance work order in advance.

[0011] Working principle: The drone is equipped with a five-in-one sensing payload. After taking off, it performs a preliminary scan along the preset route. The visible light camera captures the lightning ablation marks on the blade surface at a frame rate of 20fps. The thermal imaging module simultaneously monitors the temperature rise of the electrical connection points. The ultraviolet imager detects the local discharge intensity at the down conductor connection. The microwave radar penetrates the composite material at a frequency of 24GHz to detect the status of the internal down conductor. The lidar constructs a three-dimensional point cloud model of the blade and measures the installation angle deviation of the lightning receptor. Then, the edge computing unit runs the lightweight CNN model in real time and processes the collected data in a hierarchical manner. The detection strategy is then adjusted based on the DQN algorithm: when the local discharge intensity is detected to be greater than 5pC, it automatically switches to the anti-interference mode and prolongs the stay time in the area. The microwave radar priority mode is enabled in strong wind environments and in salt fog environments. Increase the frequency of UV lens wiping to once per minute; at this time, synchronize the detection data, SCADA operation data, and meteorological data to the cloud, simulate the lightning strike path and predict the thermal stress concentration area of ​​the composite material; when the "lightning risk-equipment" fault model predicts a risk value greater than 0.7, automatically adjust the yaw angle of the wind turbine to reduce the probability of lightning strikes, synchronously start the backup grounding system, and trigger an operation and maintenance work order to the nearest maintenance team. The detection engineer completes the data preprocessing on site and outputs a preliminary report containing a "lightning protection health trend chart". The data is synchronized to the wind farm management platform; based on historical detection data, the LSTM algorithm is used to predict the replacement cycle of the lightning rod. The digital twin visualization platform supports real-time viewing of the historical detection records and simulated lightning strike paths of any lightning rod, providing data support for design optimization.

[0012] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations that come within the meaning and range of equivalents of the claims are intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

[0013] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A lightning protection detection method for UAV wind turbine blades, characterized in that: It includes a multimodal perception and hidden defect detection module, a digital twin and full life cycle management module, an AI-driven autonomous decision-making module, an extreme environment adaptation and multi-model compatibility module, and a full-scenario operation and maintenance closed-loop management module; The multimodal perception and hidden defect detection module integrates visible light, thermal imaging, ultraviolet, microwave radar and lidar to achieve high-precision detection of lightning strike marks on the blade surface and hidden defects such as broken internal down conductors; the digital twin and the full life cycle management module dynamically integrate detection data, operation data and meteorological data to construct a wind turbine lightning protection digital twin; the AI-driven autonomous decision-making module is based on edge computing and reinforcement learning algorithms to identify defect types in real time and autonomously adjust detection strategies; the extreme environment adaptability and multi-model compatibility module adapts to strong wind and salt spray environments through anti-interference design, and modular load configuration supports multi-model detection; the full-scenario operation and maintenance closed-loop management module associates the "lightning strike risk-equipment" fault model to automatically trigger operation and maintenance work orders.

2. The lightning protection detection method for UAV wind turbine blades according to claim 1 is characterized in that: The following steps are involved: Identify blade faults and construct a 3D model using LiDAR; Real-time synchronization of drone detection data to record the arc path and energy distribution at the moment the lightning receptor strikes lightning; Establish a "lightning strike risk-equipment" failure correlation model.

3. The lightning protection detection method for UAV wind turbine blades according to claim 2 is characterized in that: The modal perception and hidden defect detection module uses visible light and thermal imaging to detect lightning ablation marks on the blade surface and abnormal temperature rise of electrical connections, an ultraviolet imager to measure the local discharge intensity at the down conductor connection, a microwave radar to penetrate composite materials to detect internal down conductor fractures, and a lidar to construct a three-dimensional point cloud model of the blade and measure the angle deviation of the lightning receptor installation.

4. The lightning protection detection method for UAV wind turbine blades according to claim 3 is characterized in that: The digital twin is dynamically integrated with the full life cycle management module to synchronize drone detection data, SCADA operation data, and meteorological data in real time to build a digital twin of the wind turbine lightning protection system. The digital twin also simulates lightning strike paths, predicts areas of concentrated thermal stress in composite materials, and analyzes historical data showing that the corrosion rate of the unit's lightning receptors is positively correlated with the ambient humidity, thereby optimizing the replacement cycle of the lightning receptors.

5. The lightning protection detection method for UAV wind turbine blades according to claim 4 is characterized in that: The AI-driven autonomous decision-making module deploys a lightweight CNN model to identify and classify defect types in real time. The drone autonomously adjusts its inspection path based on AI feedback, extending its dwell time in key areas. Based on the DQN algorithm, the optimal load combination is autonomously selected during the detection process.

6. The lightning protection detection method for UAV wind turbine blades according to claim 5 is characterized in that: The extreme environment adaptability and multi-model compatibility module automatically switches to anti-interference mode in strong wind environments, and the automatic wiping frequency of the ultraviolet imager lens is increased to once per minute in salt fog environments; and through modular payload configuration and dynamic route planning, it is compatible with a variety of direct-drive, dual-fed, and semi-direct-drive models.

7. The lightning protection detection method for UAV wind turbine blades according to claim 6 is characterized in that: The "lightning strike risk-equipment" fault model in the full-scenario operation and maintenance closed-loop management module uses input variables including the damage location of the lightning receptor, the electromagnetic field intensity of the thunderstorm cloud, and the unit's operating load. After the warning, the wind turbine's yaw angle is automatically adjusted to reduce the probability of lightning strikes, and the backup grounding system is activated at the same time.

8. The lightning protection detection method for UAV wind turbine blades according to claim 7 is characterized in that: The drone transmits the inspection data to the ground station in real time. The inspection engineer completes the data preprocessing on site and outputs a preliminary report through the edge computing unit. The data is synchronized to the wind farm management platform and combined with historical inspection data to generate a "lightning protection health trend chart" to predict the replacement cycle of the lightning rod.

9. The lightning protection detection method for wind turbine blades of a UAV according to claim 8, characterized in that: The blade point cloud data and inspection results collected by drones are integrated with the BIM model to build a digital twin of the wind turbine lightning protection system. Through the platform, the historical inspection records and simulated lightning strike paths of any lightning rod can be viewed in real time.

10. The lightning protection detection method for UAV wind turbine blades according to claim 9, characterized in that: Analyze the correlation between the "lightning terminal damage location" and the "thunderstorm cloud movement path" in historical detection data; combine the unit operation data of the SCADA system to establish a "lightning strike risk-equipment failure" correlation model to trigger operation and maintenance work orders in advance.

Images