A real-time safety monitoring method for fan blades in harsh environments

Abstract

The present invention discloses a method for real-time safety monitoring of wind turbine blades in harsh environments. The method first divides the deployment area, then divides the windward and leeward surfaces into grids, and deploys optical fiber sensors at the nodes of the grid according to the deployment density of the corresponding area. The actual strain at each monitoring point is calculated and the strain curve of the blade surface is reconstructed. The presence of load is determined by analyzing the changes in the strain curve image. When load occurs, emergency pitch control measures are first implemented. The blade's natural frequency is then calculated to determine whether damage has occurred. The location of the sudden load is determined based on strain curve analysis, and the extent of small-scale damage is determined by analyzing changes in low-order natural frequencies. The location of large-scale damage is determined by comparing the synchronization of high-order natural frequencies obtained by adjacent optical fiber sensors. The present invention satisfies the requirements for real-time status monitoring of wind turbine blades in harsh operating environments, allowing for timely, accurate, and rapid responses, thereby improving maintenance efficiency and the lifespan of wind turbine blades.

Description

Technical Field

[0001] The present invention relates to the technical field of fan blade monitoring, and in particular to a method for real-time safety monitoring of fan blades in a harsh environment. Background Art

[0002] The operating environment of wind turbines is often very harsh, and the blades are subjected to various loads during operation. Fatigue loads refer to loads that may cause fatigue damage under the action of cyclical reciprocating stress on the blades. The main sources include: wind loads, gravity loads, centrifugal forces, variable speed and loads, temperature changes, uneven vibrations, icing, heavy rainfall, etc. Due to the cyclic action of these fatigue loads, cracks will appear at the root or joints of the blades, especially in the connection area between the blades and the main shaft. Large amounts of wind and sand will then enter the cracks, accelerating the cracking of the blade surface, which will then evolve into fractures, causing huge economic losses. Currently, for monitoring wind turbine blades, fiber Bragg grating sensing technology is the monitoring method that is least affected by environmental factors and can effectively monitor the strain changes on the blade surface. However, in actual applications, there is a lack of targeted damage judgment methods, which makes it impossible to detect damage in time and take reasonable maintenance measures. The maintenance efficiency is low, resulting in a short service life of wind turbine blades operating in harsh environments, causing economic losses, and even safety hazards. Summary of the Invention

[0003] Purpose of the invention: In order to overcome the deficiencies in the prior art, the present invention provides a real-time monitoring method for the safety of wind turbine blades in harsh environments, which satisfies the real-time status monitoring of wind turbine blades in harsh operating environments, so as to make correct and rapid responses in a timely manner, thereby improving maintenance efficiency and the life of wind turbine blades.

[0004] Technical solution: To achieve the above-mentioned purpose, the present invention provides a method for real-time monitoring of wind turbine blade safety in a harsh environment, comprising the following steps:

[0005] Step I: Build a scale model of the wind turbine blade to be monitored and generate equivalent strain cloud maps of its windward and leeward sides to divide the layout area of ​​the fiber optic sensors. The layout density of fiber optic sensors in the same area is consistent.

[0006] Step II: Divide the windward and leeward sides of the blade model into grids, and deploy optical fiber sensors at the nodes of the grids according to the density of the corresponding areas.

[0007] Step III: By collecting strain signals from multiple optical fiber sensors, the actual strain at each monitoring point is calculated and the surface strain curve of the blade is reconstructed. By analyzing the changes in the strain curve image, it is determined whether a sudden load or a continuous load occurs;

[0008] Step IV: After determining whether a sudden load or a continuous load has occurred, emergency pitch control measures are first taken based on the location, magnitude, and direction of the strain. The natural frequency of the blade is then calculated using the strain signals measured by each of the optical fiber sensors. The calculated natural frequencies of the optical fiber sensors are compared with the natural frequencies of undamaged blades to determine whether the blade is damaged.

[0009] Step V: When determining that small-scale damage has occurred, first determine the location of the sudden load based on the strain curve analysis, then determine the extent of the small-scale damage by analyzing the changes in the low-order natural frequencies, and plan and execute small-scale damage maintenance events based on the sudden load location and damage extent; when determining that large-scale damage has occurred, determine the location of the large-scale damage by comparing the synchronization of the high-order natural frequencies obtained by adjacent optical fiber sensors, and plan and execute large-scale damage maintenance events based on the location of the large-scale damage.

[0010] Furthermore, the optical fiber sensor is provided with three sensing units, which are respectively used to measure the linear strain in three directions of its layout position, and the grid is composed of grid lines in two mutually orthogonal directions, and the grid lines in the two directions are respectively arranged along the axis of the rotating shaft of the wind turbine blade and perpendicular to the axis; when arranging a single optical fiber sensor, first determine the layout node position on the grid, take the node as the coordinate origin O, and the grid line perpendicular to the axis of the rotating shaft as the X-axis, and make the intersection of the measurement directions of the three sensing units coincide with the coordinate origin O, and then make one of the sensing units be arranged along the X-axis direction, with the counterclockwise direction as the positive direction, and then arrange the other two sensing units in the positive direction at an angle a.

[0011] Furthermore, in step III, reconstructing the strain curve of the blade surface includes constructing strain curves of the windward side and the leeward side of the blade. When drawing the strain curve of the windward side or the leeward side, the actual strain value corresponding to each point is first calculated based on the measured value of each optical fiber sensor according to its own stress transfer rate. Then, the strain curve is drawn using the actual strain values ​​obtained by multiple sensors along the axis of the blade shaft and located on the same straight line, thereby obtaining strain curves of multiple strip areas arranged parallel to the X-axis direction on the windward side and the leeward side, respectively.

[0012] Furthermore, in step III, the reconstructed strain curve is a curve of the dynamic change of stress along the axis of the wind turbine over time. When at a certain moment, the strain value within a certain range from the root of the blade increases and then returns to the normal strain value within a certain period of time, it is judged that a sudden load has occurred; when at a certain moment, the strain value within a certain range from the root of the blade increases and then remains higher than the normal strain value within a certain period of time, it is judged that a continuous load has occurred.

[0013] Furthermore, in step V, when a sudden load causes small-scale damage, the strip area involved in the damage is first determined based on the strain curve analysis. The strain increase monitoring point farthest from the blade root in each strip area is the load approach point. Multiple load approach points are connected to form a boundary line, and adjacent monitoring points on the side of each load approach point away from the blade root are connected to form another boundary line. The area between the two boundary lines is the area where small-scale damage occurs.

[0014] Furthermore, in step IV-V, if the high-order natural frequencies measured by adjacent optical fiber sensors are not synchronized, it is determined that a large-scale damage exists, and the damage location is located between the two sensors whose high-order natural frequencies are not synchronized.

[0015] Furthermore, assuming that the strain value measured by the optical fiber sensor is M and the actual strain value is N, then,

[0016]

[0017] in, is the average strain transfer rate: ;

[0018]

[0019]

[0020] Where, is the average strain transfer rate of the rubber encapsulated optical fiber, is the average strain transfer rate between the fan and the rubber substrate; k and k1 are the correlation coefficients of the material and longitudinal section size of the sensing unit, L is the half-length of the bonding between the optical fiber and the rubber substrate, and L1 is the half-length of the bonding between the rubber substrate and the fan blade.

[0021] Furthermore, the layout angle between the three sensor units is a=45°, and the measurement values ​​of the three sensor units are , then the measured strain value M, including the maximum principal strain and minimum principal strain :

[0022]

[0023]

[0024] The actual strain value N includes the actual maximum principal strain and the actual minimum principal strain , reconstructing the surface strain curve of the wind turbine blade using the actual maximum principal strain.

[0025] Beneficial effects: The present invention provides a real-time monitoring method for the safety of wind turbine blades in harsh environments. The fiber optic Bragg grating sensing technology is used to reconstruct the strain curves of the windward and leeward sides of the blades in different regions. According to the analysis of the curve images, corresponding pitch control measures are taken for sudden strain and continuous strain to reduce the risk of damage or slow down the rapid expansion of damage. The location and extent of the damage are determined by combining the calculation and analysis of the natural frequency, which facilitates the rapid execution of responsive maintenance plans, improves maintenance efficiency, effectively ensures the operation safety of wind turbines in harsh environments, and reduces maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A block diagram of the real-time monitoring method for wind turbine blade safety according to the present invention;

[0027] Figure 2 The equivalent strain cloud diagrams of the windward and leeward sides of a fan blade model according to an embodiment are shown;

[0028] Figure 3 This is a schematic diagram of the layout of a single optical fiber sensor relative to a monitoring point according to an embodiment of the present invention;

[0029] Figure 4 This is a schematic diagram of the division of small-scale damage location areas according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] The present invention will be further described below with reference to the accompanying drawings.

[0031] As attached Figure 1-4 The method for real-time monitoring of wind turbine blade safety in a harsh environment comprises the following steps:

[0032] Step I: Build a proportional model of the wind turbine blade to be monitored and generate the equivalent strain cloud diagrams of its windward and leeward sides to divide the layout area of ​​the optical fiber sensors. The layout density of the optical fiber sensors in the same area is consistent, and the distance between the sensors on both sides of the boundary line of adjacent areas is half of the sum of the distances between the sensors in the two areas. Figure 2 Figure 2 shows the equivalent strain cloud diagrams of the windward and leeward sides of a wind turbine blade. It can be seen that the windward side is primarily tensile, while the leeward side is primarily compressive. The strain values ​​near the root and middle of the blade are higher than those from the rear end to the tip, and the strain values ​​at the leading edge are higher than those at the trailing edge. In other words, the tensile strain at the leading edge in the windward side diagram is significantly higher than the tensile strain at the trailing edge, while the compressive strain at the leading edge in the leeward side diagram is significantly higher than the compressive strain at the trailing edge. Areas with higher strain are more susceptible to damage, so a higher density of deployment can be used in areas with relatively high strain to more accurately locate damage.

[0033] Step II, mesh the windward and leeward sides of the blade model. The mesh is composed of grid lines in two mutually orthogonal directions, and the grid lines in the two directions are respectively arranged along the axis of the shaft of the wind turbine blade and perpendicular to the axis; fiber optic sensors are arranged on the nodes of the grid according to the layout density of the corresponding area. Ensure that multiple sensors are aligned in the axial direction and the distances perpendicular to the axial direction are regular, which is convenient for subsequent data processing. When generating the equivalent strain cloud map, when the model is input into the software, finite element mesh division is required. The mesh division of this step can follow the finite element mesh division, so that the wind separation line of the area just passes through multiple grid nodes, which is more conducive to the layout of sensors.

[0034] In step III, the strain signals of multiple optical fiber sensors are collected to calculate the actual strain at each monitoring point and reconstruct the surface strain curve of the blade. The changes in the strain curve image are analyzed to determine whether a sudden load or a continuous load occurs.

[0035] Reconstructing the blade surface strain curve includes constructing strain curves for the windward and leeward sides of the blade. When drawing the strain curve for the windward or leeward side, the actual strain value corresponding to each point is first calculated based on the measured values ​​of each optical fiber sensor according to its own stress transfer rate. The strain curve is then drawn using the actual strain values ​​obtained by multiple sensors located along the axis of the blade shaft and on the same straight line. This results in strain curves for multiple strip areas arranged parallel to the X-axis on the windward and leeward sides, respectively. The strip areas are symmetrically distributed about the corresponding multi-sensor line, and adjacent strip areas may overlap. Preferably, as many sensors as possible are deployed so that the widths of the multiple strip areas along the X-axis are consistent.

[0036] Assuming that the strain value measured by the optical fiber sensor is M and the actual strain value is N, then,

[0037]

[0038] in, is the average strain transfer rate: ;

[0039]

[0040]

[0041] Where, is the average strain transfer rate of the rubber encapsulated optical fiber, is the average strain transfer rate between the fan and the rubber substrate; L is the half-length of the bonding between the optical fiber and the rubber substrate, and L1 is the half-length of the bonding between the rubber substrate and the fan blade; k and k1 are the correlation coefficients of the material and longitudinal section size of the sensing unit:

[0042]

[0043]

[0044] Where, E f is the elastic modulus of the core layer; r f , r1, r2, r R They are the radius of the core layer, coating layer, first glue layer, and conductive layer, that is, the distance from the outermost layer of each layer to point O; G1, G2, G R are the shear moduli of the coating layer, the first glue layer, and the conductive layer respectively. E3 is the elastic modulus of the second glue layer; h R , h3 is the distance from the outermost side of the conductive layer and the second glue layer to point Q; G3 is the shear modulus of the second glue layer.

[0045] Among them, the core layer and the coating layer are the structural layers that constitute the fiber grating part, the first glue layer is the structural layer formed by the adhesive between the optical fiber and the rubber substrate, the conductive layer is the part of the rubber substrate located between the optical fiber and the fan blade, and the second glue layer is the structural layer formed by the adhesive between the rubber substrate and the fan blade. Therefore, it can be seen that k and k1 can be determined according to the selected sensor structure and the properties of its material. L and L1 are also determined after the sensor is deployed. Therefore, after the sensor enters the working state, k, k1, L and L1 can all be regarded as known fixed values, and then the average strain transfer rate between the fan and the rubber substrate can be calculated respectively. and the average strain transfer rate of the rubber encapsulated optical fiber , and then the average strain transfer rate corresponding to each sensor can be calculated, so that the calculated strain value is closer to the actual strain value, and the strain curve drawn has more reference value for state judgment, making the monitoring results more effective and reliable.

[0046] The fiber optic sensor is provided with three sensing units, each used to measure the linear strain in the three directions of its arrangement position. When a single fiber optic sensor is arranged, the arrangement node position on the grid is first determined, with the node as the coordinate origin O and the grid line perpendicular to the axis of the rotating shaft as the X-axis. The intersection of the measurement directions of the three sensing units is coincident with the coordinate origin O. Then, one of the sensing units is arranged along the X-axis, with the counterclockwise direction as the positive direction, and the other two sensing units are arranged in the positive direction at an angle a. By arranging three sensing units at each monitoring point, the linear strain values ​​in the three directions of the point can be detected simultaneously. The magnitude and direction of the principal strain value at the point can be calculated using the strain values ​​in the three directions. This can provide effective adjustment parameters for pitch change events and simplify the complex stress changes on the surface of the wind turbine blades, facilitating the generation of strain curve images.

[0047] In order to facilitate calculation and improve the response efficiency of the system, the layout angle between the three sensor units is set to a=45°, and the measurement values ​​of the three sensor units are , then the strain value M measured by calculation includes the maximum principal strain and minimum principal strain :

[0048]

[0049]

[0050] The actual strain value N includes the actual maximum principal strain and the actual minimum principal strain .

[0051] The surface strain curve of the wind turbine blade is reconstructed using the actual maximum principal strain. After analysis, it is found that the direction of the maximum principal stress on the blade surface is basically parallel to the blade axis, and the direction of the minimum principal strain is basically perpendicular to the blade axis. In addition, the strain curve required to be drawn in this scheme mainly depicts the real-time stress change curve along the blade axis in a single strip area. Therefore, the calculated actual maximum principal strain is used to reconstruct the surface strain curve of the wind turbine blade.

[0052] Since the root of the wind turbine blade is fixedly connected to the central rotating base, when a certain point on the blade surface is subjected to a load, the part between that point and the root may be subjected to different degrees of tensile stress and compressive stress, while the strain generated from that point to the tip of the blade is smaller. Based on this, the position of the end load can be roughly determined; when at a certain moment, the strain value within a certain range from the root of the blade increases and then returns to the normal strain value within a certain period of time, it is judged that a sudden load has occurred, such as a bird strike, a momentary strong wind, etc.; when at a certain moment, the strain value within a certain range from the root of the blade increases and remains higher than the normal strain value for a certain period of time, it is judged that a continuous load has occurred, such as flying debris entangled in the blade, ice and snow covering the blade surface, etc.

[0053] Step IV: After determining whether a sudden load or a continuous load occurs, take emergency pitch control measures based on the location, magnitude, and direction of the strain. The determination of the location of the strain and the calculation of its magnitude are as shown above. The direction of the strain is :

[0054]

[0055] When the sensor senses a large strain on the blade surface, in order to avoid possible damage to the blade, the force-bearing area can be adjusted by adjusting the angle of the blade. For example, when the wind direction and wind force are unstable, a sudden change in wind force may cause the strain on the blade surface to increase. By adjusting the pitch, the wind angle of the windward side can be reduced, thereby reducing the effect of the wind on the blade surface and avoiding damage caused by excessive strain on the blade surface. This adjustment can be dynamically corrected according to real-time monitoring feedback to ensure the stable working state of the blade.

[0056] Because sudden changes in wind force require a certain period of sustained load to damage the blades, timely pitch adjustment can effectively avoid this. However, if a lightning strike occurs, the high temperature causes the blade surface to expand, resulting in stress concentration. This damage is immediate and irreversible. Therefore, in addition to emergency pitch adjustment to prevent further damage, it is also necessary to quickly locate the damage location and determine the extent of the damage so that maintenance can be carried out as soon as possible to minimize losses. Therefore, this solution calculates the natural frequency of the blade using the strain signal measured by each fiber optic sensor. By comparing the natural frequency calculation results of each fiber optic sensor with the natural frequency of an undamaged blade, it is determined whether the blade is damaged. First, the center wavelength value of each fiber optic sensor is collected, and the strain signal of each fiber optic sensor is calculated using the center wavelength signal. Then, the vibration response of the blade is analyzed based on the strain signal to calculate the blade natural frequency under the corresponding situation. This solution uses MATLAB software to perform a fast Fourier transform on the strain data, obtaining a frequency domain spectrum of the blade. This clearly captures the first three natural frequencies, and analysis shows that the second-order natural frequency is relatively low, while the first and third-order natural frequencies are more pronounced. Therefore, in practical applications, the first and third natural frequencies can be analyzed. Because the natural frequencies are unaffected by the location of the excitation point, damage caused by loads at any position on the wind turbine blade can be determined based on this. When the natural frequencies measured by each sensor decrease synchronously, with a more pronounced decrease in the third-order natural frequency, this indicates the presence of small-scale damage.

[0057] Step V: Small-scale damage may be caused by impact or lightning strike. The damage point roughly coincides with the load application position. Therefore, when determining that small-scale damage has occurred, first determine the strip area involved in the damage based on the strain curve analysis. The strain increase monitoring point farthest from the blade root in each strip area is the load approach point, as shown in the attached figure. Figure 4As shown in the figure, it is a distribution map of monitoring points in a certain area of ​​the blade. If there is an irregular black damage area as shown in the figure, its right side is close to the root side of the blade, and the multiple sensor layout points on the right side are load approach points. Connecting multiple load approach points forms a boundary line, and connecting adjacent monitoring points on the side of each load approach point away from the root of the blade forms another boundary line. The area between the two boundary lines is the area where small-scale damage occurs, such as the gray covered area in the figure, that is, the sudden load position. Based on this, a more certain damage range can be obtained, which is convenient for maintenance workers to quickly locate the small-scale damage position, thereby improving maintenance efficiency; and then by analyzing the changes in low-order natural frequencies, the damage degree of small-scale damage is judged. The more obvious the decrease in the first-order natural frequency, the higher the damage degree. Small-scale damage maintenance events are planned and executed based on the sudden load position and damage degree.

[0058] When an assembly is damaged on a large scale, such as a blade breakage, the blades on both sides of the damaged area can no longer be completely treated as a whole, and there will be asynchrony in the high-order natural frequencies. Therefore, if the high-order natural frequencies measured by adjacent fiber optic sensors are not synchronized, it is determined that large-scale damage exists, and the damage location is located between the two sensors with asynchronous high-order natural frequencies. Large-scale damage maintenance events are planned and executed based on the location of the large-scale damage.

[0059] The reconstruction of the blade surface strain curve and the calculation of the natural frequency are both based on data processing of the strain monitoring values, so the two can be performed simultaneously. The calculation results of the natural frequency can also be used to control the pitch event, to avoid resonance when the wind vibration frequency equals the natural frequency of the wind turbine blade when the wind speed changes. This prevents the vibration of the wind turbine blade from being intensified, accelerating damage or even fracture, and effectively maintaining the operational safety of the wind turbine.

[0060] The above are all methods for judging damage within a single surface. The windward and leeward sides of the blade are calculated and analyzed separately. After the damage location and degree results are obtained respectively, the results of the windward and leeward sides are combined to analyze the damage status of the entire blade.

[0061] The above description is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the above principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A real-time safety monitoring method for wind turbine blades in harsh environments, characterized in that: The following steps are involved: Step I: Build a scale model of the wind turbine blade to be monitored and generate equivalent strain cloud maps of its windward and leeward sides to divide the layout area of ​​the fiber optic sensors. The layout density of fiber optic sensors in the same area is consistent. Step II: Divide the windward and leeward sides of the blade model into grids, and deploy optical fiber sensors at the nodes of the grids according to the density of the corresponding areas. Step III: By collecting strain signals from multiple optical fiber sensors, the actual strain at each monitoring point is calculated and the surface strain curve of the blade is reconstructed. By analyzing the changes in the strain curve image, it is determined whether a sudden load or a continuous load has occurred. The reconstructed strain curve is a curve of the dynamic change of stress along the axis of the wind turbine over time. The strain curve is drawn using the actual strain values ​​obtained by multiple sensors located on the same straight line along the axis of the blade shaft, thereby obtaining strain curves of multiple strip areas arranged parallel to the X-axis on the windward side and the leeward side respectively. Step IV: After determining whether a sudden load or a continuous load has occurred, emergency pitch control measures are first taken based on the location, magnitude, and direction of the strain. The natural frequency of the blade is then calculated using the strain signals measured by each of the optical fiber sensors. The calculated natural frequencies of the optical fiber sensors are compared with the natural frequencies of undamaged blades to determine whether the blade is damaged. Step V: When determining that small-scale damage has occurred, first determine the location of the sudden load based on strain curve analysis, then determine the extent of the small-scale damage by analyzing the changes in low-order natural frequencies. Plan and execute small-scale damage maintenance events based on the sudden load location and damage extent. When a sudden load causes small-scale damage, the strip area involved in the damage is first determined based on the strain curve analysis. The strain increase monitoring point farthest from the blade root in each strip area is the load approach point. Multiple load approach points are connected to form a boundary line. The adjacent monitoring points on the side of each load approach point away from the blade root are connected to form another boundary line. The area between the two boundary lines is the area where small-scale damage occurs. When large-scale damage occurs, the location of the large-scale damage is determined by comparing the synchronization of the high-order natural frequencies obtained by adjacent optical fiber sensors. Large-scale damage maintenance events are planned and executed based on the location of the large-scale damage.

2. The method for real-time monitoring of wind turbine blade safety in a harsh environment according to claim 1, characterized in that: The optical fiber sensor is provided with three sensing units, each for measuring linear strain in three directions of its arrangement position. The grid is composed of grid lines in two mutually orthogonal directions, and the grid lines in the two directions are respectively arranged along the axis of the rotating shaft of the fan blade and perpendicular to the axis. When laying out a single optical fiber sensor, first determine the position of its layout node on the grid, take the node as the coordinate origin O, and the grid line perpendicular to the axis of the rotating shaft as the X-axis. The intersection of the measurement directions of the three sensing units coincides with the coordinate origin O. Then, one of the sensing units is laid out along the X-axis, with the counterclockwise direction as the positive direction, and the other two sensing units are laid out in the positive direction at an angle a.

3. The method for real-time monitoring of wind turbine blade safety in a harsh environment according to claim 2, characterized in that: In step III, reconstructing the strain curve of the blade surface includes constructing the strain curves of the windward and leeward sides of the blade. When drawing the strain curve of the windward or leeward side, the actual strain value corresponding to each point is first calculated based on the measured value of each optical fiber sensor according to its own stress transfer rate. Then, the strain curve is drawn using the actual strain values ​​obtained by multiple sensors along the axis of the blade shaft and located on the same straight line, thereby obtaining the strain curves of multiple strip areas arranged parallel to the X-axis direction on the windward and leeward sides respectively.

4. The method for real-time monitoring of wind turbine blade safety in a harsh environment according to claim 3, characterized in that: In step III, when at a certain moment, the strain value within a certain range from the blade root increases and then returns to the normal strain value within a certain period of time, it is judged that a sudden load has occurred; when at a certain moment, the strain value within a certain range from the blade root increases and then remains higher than the normal strain value within a certain period of time, it is judged that a continuous load has occurred.

5. The method for real-time monitoring of wind turbine blade safety in a harsh environment according to claim 4, characterized in that: In step IV-V, if the high-order natural frequencies measured by adjacent optical fiber sensors are not synchronized, it is determined that a large-scale damage exists, and the damage location is located between the two sensors whose high-order natural frequencies are not synchronized.

6. The method for real-time safety monitoring of wind turbine blades in harsh environments according to claim 5, characterized in that: Assuming that the strain value measured by the optical fiber sensor is M and the actual strain value is N, then, ； in, is the average strain transfer rate: ; ； ； Where, is the average strain transfer rate of the rubber encapsulated optical fiber, is the average strain transfer rate between the fan and the rubber substrate; k and k1 are the correlation coefficients of the material and longitudinal section size of the sensing unit, L is the half-length of the bonding between the optical fiber and the rubber substrate, and L1 is the half-length of the bonding between the rubber substrate and the fan blade.

7. The method for real-time monitoring of wind turbine blade safety in a harsh environment according to claim 6, characterized in that: The layout angle between the three sensor units is a=45°, and the measurement values ​​of the three sensor units are , then the measured strain value M, including the maximum principal strain and minimum principal strain : ； ； The actual strain value N includes the actual maximum principal strain and the actual minimum principal strain , reconstructing the surface strain curve of the wind turbine blade using the actual maximum principal strain.

Images