A wind turbine blade tower-sweeping hard real-time monitoring and control system and method

Abstract

The application discloses a kind of wind power blade sweeps tower hard real-time monitoring and control system and method, system includes integrated perception unit, hard real-time control unit, parallel type safety execution unit, hardware level black box unit and cloud management unit;Method includes S1, multi-sensor synchronous acquisition;S2, sub-microsecond time synchronization;S3, feature level data fusion;S4, multi-frame stable decision;S5, hard real-time emergency stop output;S6, black box record.The application can realize the hardware level emergency stop response within 3ms, reduce the sweep tower accident rate by more than 99%, while solving the problems of high misoperation rate, complex installation, non-compliance and poor reliability of prior art.

Description

Technical Field

[0001] This invention belongs to the field of wind power equipment safety protection technology, and in particular relates to a real-time monitoring and control system and method for wind turbine blade sweeping tower. Background Technology

[0002] With the rapid development of wind power technology, the length of wind turbine blades has increased from 30 meters in the early days to over 120 meters currently, and will exceed 150 meters in the future. The longer the blade, the greater the bending deformation of the blade under the combined effects of factors such as off-center load, extreme wind conditions, and fatigue damage, and the higher the risk of tower sweep. Blade sweep is the most common and most devastating catastrophic accident in wind farms. The direct loss from a single 3MW wind turbine sweep exceeds 5 million yuan (including blade scrapping, tower repair, and downtime losses). More than 200 tower sweep accidents occur nationwide every year, resulting in direct economic losses exceeding 1 billion yuan.

[0003] Existing blade clearance monitoring technologies mainly include: ranging schemes based on single lidar, visual schemes based on camera image recognition, vibration monitoring schemes based on accelerometers, and ranging schemes based on millimeter-wave radar. These existing technologies suffer from the following five fatal flaws: Defect 1: Can only alarm, cannot control - All existing solutions are alarm-type, which can only remind maintenance personnel through sound and light signals, and cannot automatically trigger emergency stop and feathering; Defect 2: Excessive response latency - Ordinary Linux systems have a latency of 200-500ms, PLC systems have a latency of 10-50ms, and the fastest control response time with existing technology is about 100ms, which cannot meet the real-time requirements of tower sweep protection; Defect 3: High false alarm rate - The single sensor solution is greatly affected by the weather. The false alarm rate exceeds 30% under conditions such as night, heavy fog, and sandstorms, which leads to frequent wind turbine shutdowns. Defect 4: Complex and non-compliant installation - Existing solutions require modification of the original wind turbine circuitry or connection to a SCADA system, with an installation cycle of 7 to 30 days; a very small number of solutions that connect to the safety chain use a series connection method, and equipment failure can lead to unexpected wind turbine shutdown. Defect 5: Poor reliability – Distributed sensor solutions have multiple exposed wiring connectors, with a mean time between failures (MTBF) of only about 100,000 hours.

[0004] NB / T 11774-2025, "Technical Specification for Blade Clearance Monitoring System of Wind Turbine Generator" (released in 2025), specifies the response time, false alarm rate and installation compliance requirements for blade clearance monitoring systems. However, existing technical solutions are difficult to meet all the indicators of this standard at the same time.

[0005] Therefore, there is an urgent need for a wind turbine blade sweep protection system and method that can achieve hard real-time automatic control, reliable operation around the clock, 100% compliant installation, and low malfunction rate. Summary of the Invention

[0006] The purpose of this invention is to solve the above-mentioned technical problems existing in the prior art, and to provide a real-time monitoring and control system and method for wind turbine blade sweeping, which can achieve hardware-level emergency stop response within 3ms, reduce the sweeping accident rate by more than 99%, and solve the problems of high malfunction rate, complex installation, non-compliance and poor reliability of the prior art.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A real-time monitoring and control system for wind turbine blade sweeping towers, characterized in that it includes: The integrated sensing unit includes a lidar core chipset, a millimeter-wave radar chip, and an image sensor. The optical centers of the lidar core chipset, the millimeter-wave radar chip, and the image sensor are aligned on the same vertical axis and located in the same plane. The hard real-time control unit runs a customized hard real-time Linux operating system with I / O response latency below a predetermined sub-microsecond threshold, and the control logic exclusively uses one CPU core. The parallel safety actuator includes dual SIL3 level safety relays. The safety relays are connected in parallel with all original emergency stop buttons of the wind turbine. After being connected in parallel, they are connected in series to the main circuit of the wind turbine's hard safety chain. When the parallel safety actuator is powered off, malfunctions, or is damaged, the safety relays remain in a normally closed state, which does not affect the normal operation of the wind turbine. The hardware-level black box unit records all sensor data within a predetermined time window. The data is hardware encrypted and cannot be modified after being written. The cloud management unit communicates with the integrated sensing unit to enable remote monitoring and data analysis.

[0008] Furthermore, the overlap of the field of view of the core LiDAR chipset, millimeter-wave radar chip, and image sensor is no less than 99.9%.

[0009] Furthermore, the hard real-time control unit adopts a three-level multi-sensor fusion decision-making mechanism. The safety relay is triggered to disconnect only when the lidar core chip group, millimeter-wave radar chip and image sensor simultaneously detect that the distance between the blade and the tower is less than the safety threshold and remain stable for multiple consecutive frames.

[0010] Furthermore, the three-level multi-sensor fusion decision-making mechanism is implemented according to the following steps: Level 1 time synchronization: The hardware timer of the main control chip is used to align the data from the three sensors at the sub-microsecond level. Second-level feature fusion: Extract lidar ranging data, millimeter-wave Doppler velocity data and image edge features respectively, and generate the minimum distance estimate between the blade and the tower based on Kalman filter fusion; Third-level decision output: Only when the estimated fusion spacing of the above three sensors is simultaneously lower than the safety threshold and remains stable for 10 consecutive frames, a disconnect command is output to the parallel safety execution unit.

[0011] Furthermore, the hardware-level black box unit has a built-in supercapacitor, which can maintain operation for no less than 30 seconds after a power outage, and the data is encrypted using AES-256 hardware.

[0012] Furthermore, the integrated sensing unit is equipped with an optical panel made of tempered glass, an anti-reflective coating on the lidar window area, a polytetrafluoroethylene (PTFE) wave-transmitting layer integrated into the millimeter-wave radar window area, and an infrared filter attached to the image sensor window area.

[0013] Furthermore, the hard real-time control unit also includes a low-pass filter module to eliminate fluctuations in sensor ranging data caused by tower vibration.

[0014] This invention also discloses a method for real-time monitoring and control of wind turbine blade sweeping, characterized by the following steps: S1. Multi-sensor synchronous acquisition: The distance data between the blade and the tower is synchronously acquired by the lidar core chip group, millimeter-wave radar chip and image sensor integrated on the same PCB board and with their optical centers coaxial and coplanar in the integrated sensing unit. S2, Sub-microsecond time synchronization: The hard real-time control unit uses a hardware timer to align the data from the three sensors at the sub-microsecond timestamp level; S3. Feature-level data fusion: Extract lidar ranging data, millimeter-wave Doppler velocity data and image edge features, fuse them through Kalman filtering, and generate an estimate of the minimum distance between the blade and the tower. S4. Multi-frame stability decision: Determine whether the estimated fusion spacing of the three sensors is lower than the safety threshold and remains stable for multiple consecutive frames; S5, Hard Real-Time Emergency Stop Output: When the conditions in step S4 are met, the hard real-time control unit sends a disconnect command to the parallel safety execution unit within a sub-millisecond delay, triggering the wind turbine hard safety chain to disconnect and execute hard emergency stop and automatic feathering. S6. Black Box Recording: All sensor data within a predetermined time window before and after an event is recorded by a hardware-level black box unit. The data cannot be modified after it is written.

[0015] Furthermore, in step S4, the safety threshold is dynamically adjusted based on real-time wind speed, blade rotation speed, and blade bending deflection.

[0016] Furthermore, in step S5, when the parallel safety actuator itself experiences a power outage, malfunction, or damage, its safety relay remains normally closed, thus not affecting the normal operation of the fan and maintaining the fault-safe characteristics of the method.

[0017] The present invention, by adopting the above-described technical solution, has the following beneficial effects: 1. Safety improved by 30 times: 3ms hard emergency stop response, which is more than 30 times faster than the fastest 100ms response of existing technology, reducing the tower sweeping accident rate by more than 99%; 2. 20 times improved reliability: The fully sealed integrated design achieves a mean time between failures (MTBF) of 2 million hours, which is 20 times better than the existing distributed solutions (approximately 100,000 hours). 3. 240-fold reduction in false alarm rate: Three-sensor AND logic fusion decision-making, with a false alarm rate of less than 0.005 times / year / unit; 4. Installation efficiency increased by 60 times: 20 minutes for a single person to install, no need to modify the original circuit, and no need to stop the machine; 5. 100% Fail-Safe: Parallel connection with a hard safety chain ensures that no system failure will cause the wind turbine to shut down unexpectedly, fully complying with the dual standards of GB / T 19073-2018 and NB / T 11774-2025. Attached Figure Description

[0018] The present invention will be further described below with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the structure of a wind turbine blade sweeping tower hard real-time monitoring and control system according to the present invention; Figure 2 This is a control principle diagram of the present invention; Figure 3 This is a schematic diagram of the internal structure of the integrated sensing unit in this invention; Figure 4 This is a measured waveform of the system emergency stop response time in this invention; Figure 5 This is a flowchart of the real-time monitoring and control method for wind turbine blade sweeping tower in this invention; Figure 6 This is a schematic diagram of the installation of the high-power offshore wind turbine (Example 2) in this invention.

[0019] In the diagram, 1-Integrated sensing unit; 2-Hard real-time control unit; 3-Parallel safety execution unit; 4-Hardware-level black box unit; 5-Cloud management unit; 6-Safety chain; 11-LiDAR core chipset; 12-Millimeter-wave radar chip; 13-Image sensor; 14-Main control chip; 15-Optical panel; 16-Fully sealed metal housing; 17-Aviation connector; K1-Tower base emergency stop button; K2-Nacelle emergency stop button; K3-Wheel hub emergency stop button; K4-Safety relay; SC-Wind turbine hard safety chain; KM-Wind turbine main contactor; YV-Feather solenoid valve. Detailed Implementation

[0020] like Figures 1 to 6 As shown, this invention provides a hard real-time monitoring and control system for wind turbine blade sweeping, comprising an integrated sensing unit 1, including a lidar core chip group 11, a millimeter-wave radar chip 12, and an image sensor 13. The lidar core chip group 11, millimeter-wave radar chip 12, and image sensor 13 are integrated on the same PCB board, with their optical centers aligned on the same vertical axis and located in the same plane. A hard real-time control unit 2 runs a customized hard real-time Linux operating system, with an I / O response latency below a predetermined sub-microsecond threshold, and the control logic exclusively occupies one CPU core. Parallel security... The full-execution unit 3 includes dual-channel SIL3-level safety relays K4. Safety relays K4 are connected in parallel with all original emergency stop buttons on the wind turbine (tower base emergency stop button K1, nacelle emergency stop button K2, hub emergency stop button K3, and safety relays K4). After parallel connection, they are connected in series to the wind turbine's hard safety chain SC main circuit. When the parallel-connected safety execution unit 3 loses power, malfunctions, or is damaged, safety relays K4 remain normally closed, not affecting the normal operation of the wind turbine. The hardware-level black box unit 4 records all sensor data within a predetermined time window. The data is hardware-encrypted and cannot be modified after being written. The cloud management unit 5 communicates with the integrated sensing unit 1 to achieve remote monitoring and data analysis. The system's total emergency stop response time is less than 3ms. Through the above settings, a hardware-level emergency stop response within 3ms is achieved, reducing the tower sweeping accident rate by more than 99%, while simultaneously solving all the problems of high malfunction rate, complex installation, non-compliance, and poor reliability of existing technologies.

[0021] The system installation does not require any modification to the original circuitry of the fan, complies with Clause 5.5 of GB / T 19073-2018, and the installation time for a single person does not exceed 20 minutes.

[0022] As one implementation method, the overlap of the field of view of the LiDAR core chipset 11, the millimeter-wave radar chip 12, and the image sensor is no less than 99.9%, and it is calibrated once at the factory and does not need to be adjusted for life.

[0023] As an implementation method, the hard real-time control unit 2 adopts a three-level multi-sensor fusion decision mechanism. Only when the lidar core chip group 11, the millimeter-wave radar chip 12 and the image sensor 13 simultaneously detect that the distance between the blade and the tower is less than the safety threshold, and remain stable for multiple consecutive frames, will the safety relay K4 be triggered to disconnect.

[0024] Specifically, the three-level multi-sensor fusion decision-making mechanism is implemented according to the following steps: Level 1 time synchronization: The hardware timer of the main control chip is used to align the data from the three sensors at the sub-microsecond level. Second-level feature fusion: Extract lidar ranging data, millimeter-wave Doppler velocity data and image edge features respectively, and generate the minimum distance estimate between the blade and the tower based on Kalman filter fusion; Third-level decision output: Only when the estimated fusion spacing of the above three sensors is simultaneously lower than the safety threshold and remains stable for 10 consecutive frames, a disconnect command is output to the parallel safety execution unit 3.

[0025] In this embodiment, the safety threshold can be dynamically adjusted based on the current wind speed, blade rotation speed, and blade bending deflection.

[0026] As one implementation method, the hardware-level black box unit 4 has a built-in supercapacitor that can continue to work for no less than 30 seconds after power failure, and the data is encrypted using AES-256 hardware.

[0027] As one implementation method, the integrated sensing unit 1 adopts a fully sealed metal housing 16 with a protection level of not less than IP68, and is filled with dry nitrogen gas, with a mean time between failures (MTBF) of not less than 2 million hours.

[0028] As one implementation method, the integrated sensing unit 1 is provided with an optical panel 15, which is made of tempered glass. The lidar window area is coated with a 905nm anti-reflection film, the millimeter-wave radar window area is integrated with a polytetrafluoroethylene wave-transmitting layer, and the image sensor 13 window area is covered with an infrared filter. The center position error of the three windows does not exceed ±0.1mm.

[0029] Specifically, the integrated sensing unit 1 has an electrically heated layer on the back of the optical panel 15, which automatically starts when the ambient temperature is lower than a predetermined low temperature threshold to ensure that the sensor works normally in a low temperature environment.

[0030] As one implementation method, the cloud management unit 5 supports the generation of operation reports, risk analysis reports and accident analysis reports, and supports online management of a large number of devices simultaneously.

[0031] As one implementation method, the hard real-time control unit 2 also includes a low-pass filter module to eliminate fluctuations in sensor ranging data caused by tower vibration, so that the ranging accuracy is not less than ±1cm.

[0032] This invention also discloses a hard real-time monitoring and control method for wind turbine blade sweeping, applied to a system including an integrated sensing unit 1, a hard real-time control unit 2, and a parallel safety execution unit 3, comprising the following steps: S1. Multi-sensor synchronous acquisition: The distance data between the blade and the tower is synchronously acquired by the lidar core chip group 11, the millimeter-wave radar chip 12 and the image sensor 13, which are integrated on the same PCB board and whose optical centers are coaxial and coplanar in the integrated sensing unit 1. S2, Sub-microsecond time synchronization: The hard real-time control unit 2 uses a hardware timer to perform sub-microsecond timestamp alignment on the data from the three sensors; S3. Feature-level data fusion: Extract lidar ranging data, millimeter-wave Doppler velocity data and image edge features, fuse them through Kalman filtering, and generate an estimate of the minimum distance between the blade and the tower. S4. Multi-frame stability decision: Determine whether the estimated fusion spacing of the three sensors is lower than the safety threshold and remains stable for multiple consecutive frames; S5, Hard Real-Time Emergency Stop Output: When the conditions in step S4 are met, the hard real-time control unit 2 sends a disconnection command to the parallel safety execution unit 3 within a sub-millisecond delay, triggering the wind turbine hard safety chain SC to disconnect and execute hard emergency stop and automatic feathering. S6. Black box recording: The hardware-level black box unit 4 records all sensor data within a predetermined time window before and after the event. The data cannot be modified after it is written.

[0033] In step S4, the safety threshold is dynamically adjusted based on real-time wind speed, blade rotation speed, and blade bending deflection.

[0034] In step S5, when the parallel safety actuator 3 experiences a power outage, malfunction, or damage, its safety relay K4 remains normally closed, thus not affecting the normal operation of the fan and maintaining the fault-safe characteristics.

[0035] The present invention also discloses a computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the steps of a method for real-time monitoring and control of wind turbine blade sweeping.

[0036] Experimental verification: To further prove the technical effect of the present invention, the following three sets of comparative experiments were conducted at a wind farm in Inner Mongolia. The experimental environment met the requirements of GB / T 19073-2018 standard, and all experiments were witnessed by a qualified third-party testing agency.

[0037] Experiment 1: Emergency Stop Response Time Test Using a high-precision digital oscilloscope (1GHz bandwidth, 5GS / s sampling rate) and a standard reflective target, the test was repeated 1000 times. The measured results were: minimum 2.1ms, average 2.7ms, maximum 2.9ms, and standard deviation 0.12ms, fully meeting the real-time requirements for tower sweep protection. The test waveform is shown below. Figure 4 As shown.

[0038] Experiment 2: Comparison Experiment of Error Rate Ten prototype units of this invention, along with ten existing dual-sensor devices, operated continuously for 30 days (720 hours) under four operating conditions: clear weather, heavy fog, sandstorms, and nighttime. The proposed solution exhibited no malfunctions during these 30 days, resulting in an annual malfunction rate of less than 0.005 times per year per unit, which is more than 240 times lower than existing dual-sensor solutions.

[0039] Experiment 3: Fail-Safe Comparison Experiment This invention simulates seven fault modes, including power outage, system crash, and relay circuit failure. The parallel connection scheme of this invention ensures the wind turbine operates normally under all seven fault modes, achieving 100% fault safety and reliability; existing series connection schemes cause unexpected wind turbine shutdowns under all seven fault modes.

[0040] Example 1: Installation of a standard 3MW onshore wind turbine like Figure 1 As shown, the wind turbine blade sweeping tower hard real-time monitoring and control system disclosed in this invention includes an integrated sensing unit 1, a hard real-time control unit 2, a parallel safety execution unit 3, a hardware-level black box unit 4, and a cloud management unit 5.

[0041] like Figure 2 As shown, the dual-channel SIL3 level safety relay K4 in the parallel safety execution unit 3 is connected in parallel with the tower base emergency stop button K1, the nacelle emergency stop button K2, and the hub emergency stop button K3, and then connected in series to the wind turbine hard safety chain SC main circuit.

[0042] Normal state: Safety relay K4 outputs a normally closed signal, the fan hard safety chain SC is activated, the fan main contactor KM is energized, and the fan operates normally. Hazardous state: Safety relay K4 is deactivated, the fan hard safety chain SC is deactivated, the fan main contactor KM is de-energized, the feathering solenoid valve YV is energized and actuates, and the fan immediately performs a hard emergency stop and automatic feathering. Fault state: When the equipment of this invention experiences a power outage, malfunction, or damage, safety relay K4 remains normally closed, preventing unexpected fan shutdown (fail-safe design).

[0043] like Figure 3As shown, the integrated sensing unit adopts a fully sealed cylindrical design, with a diameter of 180mm, a height of 280mm, a weight of 5.2kg, and an IP68 protection rating. Internally, it integrates a LiDAR core chipset 11, a millimeter-wave radar chip 12, an image sensor 13, and a main control chip. The optical center patch error of the three sensors does not exceed ±0.05mm, and the field of view overlap is not less than 99.9%.

[0044] To adapt to the low-temperature environment in northern winters, the optical panel 15 has a 5W electric heating layer on the back, which automatically starts when the ambient temperature is below -20℃, ensuring that the sensor works normally at -40℃.

[0045] This invention employs a three-level multi-sensor fusion decision-making mechanism (see...). Figure 5 (corresponding to steps S2 to S4) Level 1 (Time Synchronization, Step S2): The main control chip hardware timer is used to align the three-channel sensor data with sub-microsecond timestamps to eliminate differences in sensor scanning cycles.

[0046] The second stage (feature-level fusion, step S3) extracts lidar ranging data, millimeter-wave radar Doppler velocity data, and image sensor 13 blade edge detection results, and fuses them using the Kalman filter algorithm to generate a real-time estimate of the minimum distance between the blade and the tower, d_est.

[0047] Level 3 (Decision-level fusion, step S4): A disconnect command is issued only when the estimated independent spacing between the LiDAR, millimeter-wave radar, and image sensor 13 is all below the safety threshold (default 3 meters) and remains stable for 10 consecutive frames (100ms), thus controlling the false alarm rate to below 0.005 times / year / unit.

[0048] Installation steps: ① Drill three M8 holes on the top platform of the tower and secure the main unit with expansion bolts; ② Lay a 100m single-core aviation cable to the control cabinet at the tower base; ③ Insert the aviation plug 17 and the 24V power supply and safety chain 6 terminals of the control cabinet into both ends of the cable; ④ Power on and start the machine. A solid green light indicates normal operation. The entire process takes no more than 20 minutes, can be completed by a single person, requires no downtime, and does not require modification of the original circuitry.

[0049] Example 2: Installation of high-power offshore wind turbines (8-16MW) like Figure 6 As shown, for high-power offshore wind turbines (single unit capacity 8-16MW, blade length 90-115m), this invention makes the following targeted improvements to the standard solution: (1) Salt spray resistant reinforced shell: The fully sealed metal shell 16 is made of double-layer titanium alloy in one piece, and the outer layer is coated with an anti-corrosion coating that meets the requirements of ISO 12944-2 C5-M level, ensuring that the corrosion life is not less than 25 years in a high salt spray environment.

[0050] (2) Submarine cable transmission scheme: changed to optical fiber + Category 5e shielded composite cable, with the addition of metal corrugated pipe mechanical protection, the transmission distance is extended to 500m, which meets the flexible cable laying requirements of floating wind turbines.

[0051] (3) Dynamic safety threshold: The safety threshold has been adjusted from a fixed 3 meters to a dynamic calculation based on real-time wind speed and wave load sensor data, with a threshold range of 2.5m to 5m, further reducing the false detection rate.

[0052] (4) Centralized management: The cloud management unit 5 is connected to the offshore control center via submarine optical cable to support the unified management of the entire wind farm units.

[0053] Sea trials have verified that the emergency stop response time of this embodiment is still less than 3ms under typical sea conditions, and the false alarm rate is less than 0.005 times / year / unit.

[0054] Example 3: Implementation Plan for Dynamic Adjustment of Safety Threshold This embodiment adds a dynamic security threshold adjustment function based on embodiment 1.

[0055] Input variables: real-time wind speed V (from the nacelle anemometer, in m / s); blade rotational speed ω (from the pitch controller CAN bus, in rpm); blade bending deflection δ (from the fiber optic strain sensor installed at the blade root, in mm).

[0056] The dynamic threshold calculation formula is: d_threshold = d_base + k1•V + k2•ω + k3•δ, where d_base is the basic safety threshold (default 2.5m), and k1, k2, and k3 are engineering adjustment coefficients, which are provided by the wind turbine manufacturer or calibrated by on-site measurement.

[0057] The threshold is appropriately reduced under low wind speed conditions, automatically expanded under extreme wind conditions, and correspondingly expanded during the blade fatigue aging period, significantly improving the adaptability throughout the entire life cycle.

[0058] The above are merely specific embodiments of the present invention, but the technical features of the present invention are not limited thereto. Any simple changes, equivalent substitutions, or modifications made based on the present invention to solve essentially the same technical problems and achieve essentially the same technical effects are all covered within the protection scope of the present invention.

Claims

1. A real-time monitoring and control system for wind turbine blade sweeping, characterized in that: include, The integrated sensing unit includes a lidar core chipset, a millimeter-wave radar chip, and an image sensor. The optical centers of the lidar core chipset, the millimeter-wave radar chip, and the image sensor are aligned on the same vertical axis and located in the same plane. The hard real-time control unit runs a customized hard real-time Linux operating system with I / O response latency below a predetermined sub-microsecond threshold, and the control logic exclusively uses one CPU core. The parallel safety actuator includes dual SIL3 level safety relays. The safety relays are connected in parallel with all original emergency stop buttons of the wind turbine. After being connected in parallel, they are connected in series to the main circuit of the wind turbine's hard safety chain. When the parallel safety actuator is powered off, malfunctions, or is damaged, the safety relays remain in a normally closed state, which does not affect the normal operation of the wind turbine. The hardware-level black box unit records all sensor data within a predetermined time window. The data is hardware encrypted and cannot be modified after being written. The cloud management unit communicates with the integrated sensing unit to enable remote monitoring and data analysis.

2. The wind turbine blade sweeping tower hard real-time monitoring and control system according to claim 1, characterized in that: The overlap of the field of view of the core LiDAR chipset, the millimeter-wave radar chip, and the image sensor is no less than 99.9%.

3. The wind turbine blade sweeping tower hard real-time monitoring and control system according to claim 1, characterized in that: The hard real-time control unit adopts a three-level multi-sensor fusion decision mechanism. The safety relay is triggered to disconnect only when the lidar core chip group, the millimeter-wave radar chip and the image sensor simultaneously detect that the distance between the blade and the tower is less than the safety threshold and remain stable for multiple consecutive frames.

4. The wind turbine blade sweeping tower hard real-time monitoring and control system according to claim 3, characterized in that: The three-level multi-sensor fusion decision-making mechanism is executed according to the following steps: Level 1 time synchronization: The hardware timer of the main control chip is used to align the data from the three sensors at the sub-microsecond level. Second-level feature fusion: Extract lidar ranging data, millimeter-wave Doppler velocity data and image edge features respectively, and generate the minimum distance estimate between the blade and the tower based on Kalman filter fusion; Third-level decision output: Only when the estimated fusion spacing of the above three sensors is simultaneously lower than the safety threshold and remains stable for 10 consecutive frames, a disconnect command is output to the parallel safety execution unit.

5. The wind turbine blade sweeping tower hard real-time monitoring and control system according to claim 1, characterized in that: The hardware-level black box unit has a built-in supercapacitor and can continue to operate for no less than 30 seconds after a power outage. The data is encrypted using AES-256 hardware.

6. The wind turbine blade sweeping tower hard real-time monitoring and control system according to claim 1, characterized in that: The integrated sensing unit is equipped with an optical panel made of tempered glass. The lidar window area is coated with an anti-reflection film, the millimeter-wave radar window area is integrated with a polytetrafluoroethylene wave-transmitting layer, and the image sensor window area is covered with an infrared filter.

7. The wind turbine blade sweeping tower hard real-time monitoring and control system according to claim 1, characterized in that: The hard real-time control unit also includes a low-pass filter module to eliminate fluctuations in sensor ranging data caused by tower vibration.

8. A method for real-time monitoring and control of wind turbine blade sweeping tower, characterized in that, Includes the following steps: S1. Multi-sensor synchronous acquisition: The distance data between the blade and the tower is synchronously acquired by the lidar core chip group, millimeter-wave radar chip and image sensor integrated on the same PCB board and with their optical centers coaxial and coplanar in the integrated sensing unit. S2, Sub-microsecond time synchronization: The hard real-time control unit uses a hardware timer to align the data from the three sensors at the sub-microsecond timestamp level; S3. Feature-level data fusion: Extract lidar ranging data, millimeter-wave Doppler velocity data and image edge features, fuse them through Kalman filtering, and generate an estimate of the minimum distance between the blade and the tower. S4. Multi-frame stability decision: Determine whether the estimated fusion spacing of the three sensors is lower than the safety threshold and remains stable for multiple consecutive frames; S5, Hard Real-Time Emergency Stop Output: When the conditions in step S4 are met, the hard real-time control unit sends a disconnect command to the parallel safety execution unit within a sub-millisecond delay, triggering the wind turbine hard safety chain to disconnect and execute hard emergency stop and automatic feathering. S6. Black Box Recording: All sensor data within a predetermined time window before and after an event is recorded by a hardware-level black box unit. The data cannot be modified after it is written.

9. A method for real-time monitoring and control of wind turbine blade sweeping tower as described in claim 8, characterized in that: The safety threshold mentioned in step S4 is dynamically adjusted based on real-time wind speed, blade rotation speed, and blade bending deflection.

10. A method for real-time monitoring and control of wind turbine blade sweeping tower as described in claim 8, characterized in that: In step S5, when the parallel safety actuator itself experiences a power outage, malfunction, or damage, its safety relay remains normally closed, thus not affecting the normal operation of the fan. The method maintains its fault-safe characteristics.

Images