A wind turbine blade strain built-in sensing system based on distributed fiber grating array

Abstract

The application discloses a wind power blade strain built-in sensing system based on a distributed fiber grating array, relates to the technical field of wind turbine blade detection, and comprises a distributed fiber grating array, a packaging module, a fiber signal processor and a data analysis unit. The distributed fiber grating array is used for capturing strain signals inside a blade; the packaging module is used for fixing the distributed fiber grating array and protecting the same, so that the strain signals can be accurately captured when the blade is deformed under stress; the fiber signal processor is used for receiving and processing the strain signals transmitted by the fiber grating array; the data analysis unit is used for analyzing the processed strain signals by using a multi-modal data fusion algorithm and calculating strain information of the wind power blade; and the wireless transmission module is used for wirelessly transmitting the strain signals to a monitoring center. The application can accurately detect internal strain changes of a wind turbine blade, analyze the internal health state of the blade, adapt to complex and changeable environmental conditions, and meet the demand of modern wind power generation technology.

Description

Technical Field

[0001] This invention relates to the field of wind turbine blade testing technology, and more specifically, to a wind turbine blade strain-integrated sensing system based on a distributed fiber optic grating array. Background Technology

[0002] With the increasing demand for renewable energy, wind power, as a clean energy source, has seen its technology and applications widely developed and promoted. Wind turbine blades are one of the core components of wind turbines, and their performance and safe operation have a decisive impact on the efficiency and stability of the entire wind turbine generator set.

[0003] During the operation of wind turbines, wind turbine blades are subjected to complex mechanical loads, including wind pressure, gravity, and centrifugal force. These loads cause strain in the blades, which in turn affects their structural integrity and service life. Long-term or excessive strain may lead to blade damage or fatigue, and may even cause the entire wind turbine generator set to fail. Therefore, real-time and accurate monitoring of the strain status of wind turbine blades is crucial to ensuring the efficient and safe operation of wind turbine generator sets.

[0004] Existing wind turbine blade strain monitoring technologies mainly include strain gauges and vibration sensors; however, these technologies have certain limitations. For example, strain gauges are complex to install and maintain, while vibration sensors may not provide sufficient strain details. Furthermore, these traditional sensors are typically sensitive to environmental conditions; changes in temperature and humidity can affect their measurement accuracy. Additionally, most current sensors are mounted on the surface of the wind turbine blades, making them ineffective at detecting early internal faults and susceptible to environmental interference, resulting in short lifespans and insufficient reliability. Especially in offshore wind farm environments, sensors face even harsher operating conditions, such as high humidity, high salt spray, and extreme temperature variations, which further reduce the lifespan and measurement reliability of traditional sensors.

[0005] Furthermore, with the increasing complexity and size of wind turbine blades, single-point measurements can no longer meet the needs of comprehensive monitoring. Existing technologies struggle to simultaneously monitor the strain state at different locations on the blade, lacking a comprehensive understanding of the overall strain distribution, resulting in low early fault identification rates and delayed warnings. Simultaneously, existing data processing methods often ignore the dynamic impact of environmental factors on measurement results and lack effective noise processing and multi-source data fusion mechanisms, reducing the accuracy and reliability of strain monitoring.

[0006] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0007] To address the problems in related technologies, this invention proposes a wind turbine blade strain built-in sensing system based on a distributed fiber optic grating array. This system has the advantages of high accuracy, real-time monitoring, strong anti-interference capability, and low maintenance cost, thereby solving the problems of traditional external wind turbine blade sensors being susceptible to environmental interference, having insufficient reliability of the packaging structure, and low signal processing accuracy in the prior art.

[0008] Therefore, the specific technical solution adopted by the present invention is as follows:

[0009] A wind turbine blade strain-integrated sensing system based on a distributed fiber Bragg grating array, comprising:

[0010] A distributed fiber Bragg grating array is placed on the inside of the wind turbine blade to capture strain signals inside the blade.

[0011] The encapsulation module, including a flexible substrate and an adhesive, is disposed on the top inner side of the wind turbine blade to fix and protect the distributed grating fiber array, so as to ensure accurate capture of strain signals when the blade is subjected to stress and deformation.

[0012] An optical fiber signal processor, connected to a distributed fiber Bragg grating array, is used to receive and process strain signals transmitted by the fiber Bragg grating array.

[0013] The data analysis unit, connected to the fiber optic signal processor, is used to analyze the processed strain signal using a multi-modal data fusion algorithm and to calculate the strain information of the wind turbine blades.

[0014] The wireless transmission module, connected to the data analysis unit, is used to wirelessly transmit strain signals to the monitoring center.

[0015] Specifically, the wind turbine blade strain built-in sensing system based on a distributed fiber Bragg grating array according to the present invention includes:

[0016] Distributed fiber Bragg grating array: Based on the finite element stress analysis results of the blade, grating nodes are arranged in a zigzag pattern in the high strain gradient region (blade root, leading edge, airfoil transition region), with a grating density ≥ 3 gratings / m. 2 The low-strain zones (blade tip and trailing edge) are sparsely distributed, with a density of ≤1 zone / m². 2 The overall number of sensors is reduced by 20% to 30%.

[0017] The encapsulation module includes a flexible polyimide substrate, adhesive, etc., consisting of a flexible silicone buffer layer (0.5–1 mm thick, Shore A30 hardness) to reduce installation stress; carbon fiber reinforced polymer (CFRP, 2–3 mm thick, fiber volume fraction ≥60%) to resist salt spray corrosion and UV aging; and a polyurethane waterproof coating (0.1 mm thick, contact angle ≥110°) to meet IP68 protection standards. The fiber direction of the CFRP layer is consistent with the main load-bearing direction of the blade, with a 0° layup ratio ≥70%.

[0018] Fiber Optic Signal Processor: Employs a tunable laser (TLS) and a high-speed FPGA demodulation unit, integrating a Gauss-Newton iterative algorithm to fit the peak value of the reflection spectrum in real time.

[0019] Data Analysis Unit: Connected to the fiber optic signal processor, it analyzes the processed signals to obtain the strain information of the wind turbine blades, integrates the strain data, the blade SCADA system (wind speed, pitch angle) and vibration spectrum, and eliminates noise through Kalman filtering.

[0020] Wireless transmission module: Connects to the data analysis unit and is used to wirelessly transmit strain information to the monitoring center.

[0021] Specifically, the working principle of this invention is as follows:

[0022] When wind turbine blades deform under wind force, the fiber Bragg grating array distributed on the blades also deforms, causing a change in the center wavelength of the gratings. This change is transmitted to a data analysis unit via a fiber optic signal processor, which processes these signals using specialized algorithms to accurately calculate the strain of the blade. Through a wireless transmission module, this data can be sent in real-time to a remote monitoring center, enabling real-time monitoring of the wind turbine blade's strain status.

[0023] Furthermore, when the data analysis unit analyzes the processed strain signal using a multimodal data fusion algorithm and calculates the strain information of the wind turbine blade, it includes:

[0024] S1. Based on the strain data acquired by fiber Bragg grating, time synchronization and unit standardization are performed, and the maximum strain value and strain gradient are extracted to obtain the first mode feature.

[0025] S2. Process the wind speed and pitch angle data using interpolation methods to ensure time alignment with the strain data, and extract the wind speed change rate and pitch angle adjustment frequency to obtain the second mode characteristics;

[0026] S3. Use a filtering algorithm to process the vibration spectrum data, extract the vibration frequency and amplitude as indirect representations of the dynamic load, and obtain the third mode characteristics;

[0027] S4. Based on the multimodal data fusion algorithm, integrate the first modal features, the second modal features, and the third modal features to obtain the fused strain information of the wind turbine blade;

[0028] S5. A noise covariance matrix update strategy based on environmental changes is implemented, the filtering algorithm is dynamically optimized, and the fused strain information of wind turbine blades is denoised.

[0029] Furthermore, based on the multimodal data fusion algorithm, the first modal features, second modal features, and third modal features are integrated to obtain the fused strain information of the wind turbine blade, including:

[0030] S41. Construct a state vector based on the first modal features, the second modal features, and the third modal features;

[0031] S42. Based on the nonlinear relationship between wind speed, temperature and strain, construct the strain information weighting coefficient, and set the vibration signal weighting coefficient based on the rate of change of wind speed and vibration frequency.

[0032] S43. Calculate the temperature compensation coefficient based on real-time temperature changes, and establish the vibration frequency compensation coefficient by analyzing the changes in vibration modes and introducing the coupling effect of temperature and load.

[0033] S44. Using the strain information weighting coefficient, vibration signal weighting coefficient, temperature compensation coefficient and vibration frequency compensation coefficient, the modal data in the state vector are weighted and fused to obtain the preliminary fusion result of the wind turbine blade.

[0034] S45. Based on the coupling relationship between strain and vibration, and temperature and vibration in the wind turbine blade, the preliminary fusion result of the wind turbine blade is optimized by introducing the first cross-coupling coefficient and the second cross-coupling coefficient, and the fusion strain information of the wind turbine blade is obtained.

[0035] Furthermore, based on the coupling relationship between strain and vibration, and temperature and vibration in the wind turbine blade, the preliminary fusion results of the wind turbine blade are optimized by introducing a first cross-coupling coefficient and a second cross-coupling coefficient, resulting in the following fused strain information of the wind turbine blade:

[0036] S451. Based on the product between the normalized strain value and the normalized vibration frequency value, and combined with empirical data, calculate the first cross-coupling coefficient to characterize the dynamic coupling effect between strain and vibration.

[0037] S452. Calculate the second cross-coupling coefficient by using the product between the temperature difference and the normalized value of the vibration frequency, combined with an adjustment factor, to characterize the effect of temperature change on vibration characteristics.

[0038] S453. By weighting and integrating the first cross-coupling coefficient and the second cross-coupling coefficient of each modal data in the state vector, the preliminary fusion result of the wind turbine blade is nonlinearly corrected to obtain the fusion strain information of the wind turbine blade.

[0039] Furthermore, the expression for the fused strain information of the wind turbine blade is as follows:

[0040]

[0041] In the formula, X fusion w(t) represents the fused strain information of the wind turbine blade; w1(t) represents the strain information weighting coefficient; w2(t) represents the vibration signal weighting coefficient; w3(t) represents the temperature compensation coefficient; w4(t) represents the vibration frequency compensation coefficient; w5(t) represents the first cross-coupling coefficient; w6(t) represents the second cross-coupling coefficient; ε represents the strain, v represents the wind speed, β represents the blade pitch angle, and f represents the wind speed. vib The frequency is the vibration frequency.

[0042] Furthermore, the distributed fiber grating array is conformally bonded to the inner curved surface of the blade through a polymer film. The distributed fiber grating array includes several fiber gratings etched on a single optical fiber to cover different center wavelength bands. The reflectivity of the fiber gratings is controlled within a preset range to avoid crosstalk between adjacent grating signals. Based on the finite element stress simulation results of the blade, the distributed fiber grating array adopts a dense grating arrangement in the high strain sensitive area and a sparse arrangement in the low strain area to optimize the sensor layout and reduce the overall number.

[0043] Furthermore, the encapsulation module includes a flexible substrate with a thickness not exceeding a preset value, and the surface of the flexible substrate is coated with an adhesive layer of a specific thickness. The encapsulation module uses a vacuum bagging process to simultaneously cure the adhesive layer and the blade layup within a preset pressure range and at a preset temperature. The cured guide groove is filled with a flexible sealing material of a preset hardness to ensure that the aerodynamic performance parameters of the wind turbine blade do not change beyond a preset threshold after the sensor is installed.

[0044] Furthermore, the fiber optic signal processor includes a wavelength demodulation device for converting the wavelength variation of the fiber Bragg grating into an electrical signal. When receiving and processing the strain signal transmitted from the fiber Bragg grating array, the fiber optic signal processor includes: using a broadband light source with a transmission bandwidth capable of covering the wavelength range of all fiber Bragg gratings in the distributed fiber Bragg grating array; introducing the optical signal into the distributed fiber Bragg grating array via a fiber coupler, causing each fiber Bragg grating to reflect a specific wavelength, while the remaining wavelengths are transmitted to subsequent gratings or terminating absorbers; converting the wavelength variation into an intensity variation using the edge linear region of a filter, so that the intensity of the reflected light after passing through the filter is linearly related to the wavelength shift; and converting the voltage signal into a wavelength shift using the filter slope based on the linear relationship.

[0045] Furthermore, based on the noise covariance matrix update strategy for environmental changes, the filtering algorithm is dynamically optimized, and the denoising process for the fused strain information of wind turbine blades includes:

[0046] S51. Based on the fused strain information of wind turbine blades, establish state equations and observation equations; wherein, the state equations include state vectors, state transition matrices, control input matrices and process noise, and the observation equations include observed values, observation matrices and observation noise;

[0047] S52. Based on the changes in the operating environment of wind turbine blades, the weights of historical data and current data are balanced using a forgetting factor to dynamically update the process noise covariance matrix and the observation noise covariance matrix.

[0048] S53. Based on the updated noise covariance matrix, perform the prediction and update process of the filtering algorithm to denoise the fused strain data.

[0049] Furthermore, based on the updated noise covariance matrix, the prediction and update process of the filtering algorithm is performed to denoise the fused strain data, including:

[0050] S531. In the prediction stage, the prior error covariance matrix is ​​calculated based on the state transition matrix and the process noise covariance matrix.

[0051] S532. Calculate the Kalman gain based on the prior error covariance matrix and the observation noise covariance matrix;

[0052] S533. During the update phase, the state estimate is corrected using Kalman gain and observations, and the filtered state vector is output to obtain the fused strain data of the wind turbine blade after denoising, so as to monitor the strain state of the wind turbine blade in real time.

[0053] The beneficial effects of this invention are as follows:

[0054] (1) This invention proposes a sensing system consisting of a carbon fiber reinforced polymer encapsulation and an internally integrated distributed fiber optic grating sensor. Addressing the problems of traditional external wind turbine blade sensors being susceptible to environmental interference, having insufficient encapsulation reliability, and low signal processing accuracy, this invention innovatively employs a flexible polyimide substrate. Through a multi-layer structure, mechanical protection and temperature stress compensation are achieved, effectively extending the sensor's lifespan. Based on the non-uniform fiber optic grating array layout derived from blade finite element stress analysis, the sensor layout is optimized, reducing the overall number of sensors by 20%–30% and improving monitoring efficiency. The system integrates a multi-modal data fusion algorithm and dynamic wavelength demodulation technology, combined with a Kalman filter noise reduction method dynamically optimized based on environmental factors, to achieve high-precision early warning of cracks. Simultaneously, a low-power dual-mode wireless transmission module is used to ensure stable long-distance communication in high-humidity marine environments.

[0055] (2) This invention develops a novel wind turbine blade strain monitoring technology that can provide high-precision and high-reliability strain monitoring, can detect the strain changes inside the wind turbine blade in real time and accurately, effectively analyze the internal health status of the blade, and can adapt to complex and ever-changing environmental conditions to meet the needs of modern wind power generation technology.

[0056] (3) The present invention provides a sensing system that can be built into the blade structure, has distributed measurement capabilities, strong environmental adaptability and long life, and is equipped with advanced data fusion and processing algorithms, which can realize comprehensive and real-time monitoring of the health status of wind turbine blades.

[0057] (4) The sensor of the present invention has the advantages of high precision, real-time monitoring, strong anti-interference ability and low maintenance cost. It can detect early cracks in the wind turbine blades in advance and is suitable for various types of wind turbine units. In particular, it is of great significance for improving the operating efficiency and safety of wind turbine units in large offshore wind farms operating in harsh environments.

[0058] (5) This invention achieves high-precision detection of internal strain of wind turbine blades through distributed fiber Bragg grating array technology, which can accurately capture strain changes in various parts of the blade. In particular, a dense arrangement strategy is adopted in the high strain gradient region to ensure the monitoring accuracy of key areas. The system uses a wireless transmission module to realize real-time data transmission and monitoring, so that wind farm operation and maintenance personnel can keep track of the blade status at any time. Fiber Bragg grating arrays have the characteristics of resisting electromagnetic interference, which are suitable for the complex electromagnetic environment of wind turbine units and will not be affected by external interference such as lightning strikes. At the same time, the optical fiber material has excellent stability and durability, which greatly reduces the maintenance cost of the system and is suitable for long-term deployment. More importantly, this invention can effectively identify early internal faults of wind turbine blades by analyzing the internal strain change law, providing a scientific basis for preventive maintenance and significantly improving the operational safety and service life of wind turbine units. Attached Figure Description

[0059] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1 This is a schematic diagram of a wind turbine blade strain built-in sensing system based on a distributed fiber optic grating array according to an embodiment of the present invention.

[0061] Figure 2This is a schematic diagram of a wind turbine blade strain built-in sensing system based on a distributed fiber optic grating array according to an embodiment of the present invention.

[0062] Figure 3 This is a schematic diagram of the encapsulation module in a wind turbine blade strain built-in sensing system based on a distributed fiber grating array according to an embodiment of the present invention.

[0063] Figure 4 This is a flowchart illustrating a data analysis unit in a wind turbine blade strain-embedded sensing system based on a distributed fiber grating array, according to an embodiment of the present invention. Detailed Implementation

[0064] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0065] According to an embodiment of the present invention, a wind turbine blade strain built-in sensing system based on a distributed fiber Bragg grating array is provided.

[0066] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 As shown, according to an embodiment of the present invention, a wind turbine blade strain-integrated sensing system based on a distributed fiber Bragg grating array is provided. This wind turbine blade strain-integrated sensing system based on a distributed fiber Bragg grating array includes:

[0067] A distributed fiber Bragg grating array is placed on the inside of the wind turbine blade to capture strain signals inside the blade.

[0068] The encapsulation module, including a flexible substrate and an adhesive, is disposed on the top inner side of the wind turbine blade to fix and protect the distributed grating fiber array, so as to ensure accurate capture of strain signals when the blade is subjected to stress and deformation.

[0069] An optical fiber signal processor, connected to a distributed fiber Bragg grating array, is used to receive and process strain signals transmitted by the fiber Bragg grating array.

[0070] The data analysis unit, connected to the fiber optic signal processor, is used to analyze the processed strain signal using a multi-modal data fusion algorithm and to calculate the strain information of the wind turbine blades.

[0071] The wireless transmission module, connected to the data analysis unit, is used to wirelessly transmit strain signals to the monitoring center.

[0072] In one embodiment, when the data analysis unit analyzes the processed strain signal using a multimodal data fusion algorithm and calculates the strain information of the wind turbine blade, it includes:

[0073] S1. Based on the strain data acquired by fiber Bragg grating, time synchronization and unit standardization are performed, and the maximum strain value and strain gradient are extracted to obtain the first mode feature.

[0074] S2. Process the wind speed and pitch angle data using interpolation methods to ensure time alignment with the strain data, and extract the wind speed change rate and pitch angle adjustment frequency to obtain the second mode characteristics;

[0075] S3. Use a filtering algorithm to process the vibration spectrum data, extract the vibration frequency and amplitude as indirect representations of the dynamic load, and obtain the third mode characteristics;

[0076] S4. Based on the multimodal data fusion algorithm, integrate the first modal features, the second modal features, and the third modal features to obtain the fused strain information of the wind turbine blade;

[0077] S5. A noise covariance matrix update strategy based on environmental changes is implemented, the filtering algorithm is dynamically optimized, and the fused strain information of wind turbine blades is denoised.

[0078] In one embodiment, based on a multimodal data fusion algorithm, the first modal features, the second modal features, and the third modal features are integrated to obtain the fused strain information of the wind turbine blade, including:

[0079] S41. Construct a state vector based on the first modal features, the second modal features, and the third modal features;

[0080] S42. Based on the nonlinear relationship between wind speed, temperature and strain, construct the strain information weighting coefficient, and set the vibration signal weighting coefficient based on the rate of change of wind speed and vibration frequency.

[0081] S43. Calculate the temperature compensation coefficient based on real-time temperature changes, and establish the vibration frequency compensation coefficient by analyzing the changes in vibration modes and introducing the coupling effect of temperature and load.

[0082] S44. Using the strain information weighting coefficient, vibration signal weighting coefficient, temperature compensation coefficient and vibration frequency compensation coefficient, the modal data in the state vector are weighted and fused to obtain the preliminary fusion result of the wind turbine blade.

[0083] S45. Based on the coupling relationship between strain and vibration, and temperature and vibration in the wind turbine blade, the preliminary fusion result of the wind turbine blade is optimized by introducing the first cross-coupling coefficient and the second cross-coupling coefficient, and the fusion strain information of the wind turbine blade is obtained.

[0084] In one embodiment, based on the coupling relationship between strain and vibration, and temperature and vibration in the wind turbine blade, the initial fusion result of the wind turbine blade is optimized by introducing a first cross-coupling coefficient and a second cross-coupling coefficient, and the fused strain information of the wind turbine blade is obtained, including:

[0085] S451. Based on the product between the normalized strain value and the normalized vibration frequency value, and combined with empirical data, calculate the first cross-coupling coefficient to characterize the dynamic coupling effect between strain and vibration.

[0086] S452. Calculate the second cross-coupling coefficient by using the product between the temperature difference and the normalized value of the vibration frequency, combined with an adjustment factor, to characterize the effect of temperature change on vibration characteristics.

[0087] S453. By weighting and integrating the first cross-coupling coefficient and the second cross-coupling coefficient of each modal data in the state vector, the preliminary fusion result of the wind turbine blade is nonlinearly corrected to obtain the fusion strain information of the wind turbine blade.

[0088] Specifically, this invention integrates a multimodal data fusion algorithm, including strain data, blade SCADA system (wind speed, blade pitch angle), and vibration spectrum. For example... Figure 4 As shown, data preprocessing is first performed. The strain data acquired by the fiber Bragg grating is synchronized in time and standardized in units. SCADA data (wind speed, pitch angle) is interpolated to ensure alignment with the strain data. Vibration spectrum data is filtered to extract characteristic frequencies. Features such as maximum strain value and strain gradient are extracted from the strain data. Features such as wind speed change rate and pitch angle adjustment frequency are extracted from the SCADA data. The dominant frequency (vibration frequency) and amplitude are extracted from the vibration spectrum as indirect representations of the dynamic load. A state vector X = [ε, v, β, f] is constructed. vib ], where ε is strain, v is wind speed, β is pitch angle, and f vib Let be the vibration frequency. A weighted fusion method is used to initially integrate multimodal data. The weights are determined based on the contribution of each modal data point to the blade strain, yielding the preliminary fusion result for the wind turbine blade, expressed as:

[0089] X fusion (t)=w1(t)ε+w2(t)v+w3(t)β+w4(t)f vib ;

[0090] In the formula, w1(t), w2(t), w3(t), and w4(t) are the strain information weighting coefficient, vibration signal weighting coefficient, temperature compensation coefficient, and vibration frequency compensation coefficient, respectively.

[0091] Specifically, the dynamic adjustment of the weighting coefficients in this invention is determined based on the stress characteristics of the blades under different operating conditions. w1(t) is the strain information weighting coefficient, which, based on a machine learning regression method, describes the nonlinear relationship between wind speed, temperature, and strain, and is expressed as:

[0092] w1(t)=f1(v(t),T(t),ε(t));

[0093] In the formula, v(t) is the wind speed; T(t) is the temperature; ε(t) is the strain; and w1(t) is the dynamic adjustment value calculated by wind speed, temperature and the strain response of the blade.

[0094] Specifically, w2(t) is the vibration signal weighting coefficient, which reflects the dynamic response of the blade during operation; the adjustment of w2(t) is based on the rate of change of wind speed and vibration frequency, and the expression is:

[0095]

[0096] Δv(t) is the rate of change of wind speed at time t, α' and β' are dynamic adjustment factors determined based on previous operational data, and Δf vib (t) represents the rate of change of vibration at time t.

[0097] w3(t) is the temperature compensation coefficient: temperature is a key factor affecting sensor performance, and the adjustment of w3(t) depends on real-time temperature changes. The expression is:

[0098] w3(t) = 1 + γ(T(t) - T0);

[0099] In the formula, γ is the temperature sensitivity coefficient, T0 is the reference temperature (in this embodiment, the standard temperature of the sensor or the laboratory ambient temperature), and T(t) is the current ambient temperature.

[0100] w4(t) is the vibration frequency compensation coefficient: The calculation of the vibration frequency compensation coefficient considers the coupling effect between the blade's vibration mode and factors such as temperature and load. The value of w4 is determined by analyzing the changes in the vibration mode, and the coupling effect of temperature and load is introduced for adjustment. δ and θ are adjustment factors, adjusted according to real-time operating data.

[0101]

[0102] Furthermore, considering the coupling relationship between parameters, by introducing the first and second cross terms of strain, vibration, and temperature, the coupling effect between parameters can be captured more accurately. Extending the formula, the expression for the fused strain information of the wind turbine blade is obtained as follows:

[0103]

[0104] In the formula, X fusion w(t) represents the fused strain information of the wind turbine blade; w1(t) represents the strain information weighting coefficient; w2(t) represents the vibration signal weighting coefficient; w3(t) represents the temperature compensation coefficient; w4(t) represents the vibration frequency compensation coefficient; w5(t) represents the first cross-coupling coefficient; w6(t) represents the second cross-coupling coefficient; ε represents the strain, v represents the wind speed, β represents the blade pitch angle, and f represents the wind speed. vib The frequency is the vibration frequency.

[0105] Specifically, w5(t) and w6(t) are the first and second cross-coupling coefficients, respectively, used to describe the coupling relationship between strain and vibration, and temperature and vibration; the expression for the first cross-coupling coefficient is:

[0106]

[0107] In the formula, ε(t) represents the strain at the current moment. max (t) represents the historical maximum strain, f vib (t) is the vibration frequency at time t, f vib,max (t) represents the historical maximum vibration frequency, and α5 represents empirical data obtained after actual operation.

[0108] Specifically, the expression for the second cross-coupling coefficient is:

[0109]

[0110] In the formula, T max (t) represents the historical maximum temperature, and α6 is an adjustment factor determined based on operational data for a certain period.

[0111] In one embodiment, the distributed fiber grating array is conformally bonded to the inner curved surface of the blade via a polymer film. The distributed fiber grating array includes several fiber gratings etched on a single optical fiber to cover different center wavelength bands.

[0112] The reflectivity of fiber optic gratings is controlled within a preset range to avoid crosstalk between adjacent grating signals;

[0113] Based on the finite element stress simulation results of the blade, the distributed fiber grating array adopts a dense grating arrangement in the high strain sensitive region and a sparse arrangement in the low strain region to optimize the sensor layout and reduce the overall number.

[0114] Specifically, the present invention inscribes multiple fiber gratings (FBGs) on a single optical fiber, and the grating spacing is dynamically adjusted according to the blade strain gradient (e.g., spacing ≤10cm in the root region and ≤30cm in the tip region); the grating reflectivity is controlled at 1% to 5% to avoid crosstalk between adjacent grating signals and ensure wavelength demodulation accuracy ≤5pm.

[0115] Specifically, this invention supports wavelength division multiplexing (WDM) technology, where a single optical fiber can accommodate 50 to 100 gratings, covering different center wavelength bands.

[0116] Specifically, based on the finite element stress simulation results of the blade, this invention employs a dense grating arrangement in high strain sensitive areas (such as the blade root, leading edge, and airfoil transition area) and a sparse arrangement in low strain areas, reducing the number of sensors by 20% to 30%. The grating array is conformally bonded to the inner curved surface of the blade through a polyimide film, with a bonding error ≤0.1mm, avoiding the introduction of additional stress during installation.

[0117] In one embodiment, the encapsulation module includes a flexible substrate with a thickness not exceeding a preset value, and the surface of the flexible substrate is coated with an adhesive layer of a specific thickness.

[0118] Specifically, the encapsulation module uses a vacuum bagging process to simultaneously cure the adhesive layer and blade layup within a preset pressure range and at a preset temperature. The cured guide groove is filled with a flexible sealing material of preset hardness to ensure that the aerodynamic performance parameters of the wind turbine blade do not change beyond a preset threshold after the sensor is installed.

[0119] Specifically, such as Figure 2 As shown, the laser emits a stable optical signal, which is initially amplified by a laser amplifier (SOA), and the amplification process is precisely controlled by a microcontroller through a driver module. The amplifier (EDFA) further enhances the intensity of the optical signal. A circulator is used to guide the amplified optical signal into a distributed fiber optic grating (FBG) array deployed inside the wind turbine blade. The reflected signal of a specific wavelength is then guided back by the circulator to a photoelectric converter (PD) for photoelectric conversion. The resulting electrical signal is digitized by a digital converter (ADC), acquired by the microcontroller, and transmitted to a host computer (PC).

[0120] Specifically, such as Figure 3 As shown, this invention embeds a flexible polyimide substrate (thickness ≤0.2mm) encapsulating a fiber Bragg grating array into a wind turbine blade as an encapsulation module. An epoxy resin adhesive layer with a thickness of 50-100μm is coated on the substrate surface as an adhesive. A vacuum bagging process is used to simultaneously cure the adhesive layer and the blade layup. The curing pressure is 0.5-1MPa and the temperature is 120℃. After curing, the guide groove is filled with a flexible silicone sealant with a Shore hardness of A30-50. After installation, wind tunnel testing verifies that the change in blade lift-to-drag ratio is ≤1%.

[0121] In one embodiment, the fiber optic signal processor includes a wavelength demodulation device for converting wavelength variations of a fiber optic grating into electrical signals.

[0122] The fiber optic signal processor, when receiving and processing strain signals transmitted by a fiber Bragg grating array, includes:

[0123] Based on a broadband light source, the transmission bandwidth can cover the optical signal of all fiber grating wavelength ranges in the distributed fiber grating array;

[0124] Optical signals are introduced into a distributed fiber optic array via fiber couplers, so that each fiber optic grating reflects a specific wavelength, while the remaining wavelengths are transmitted to subsequent gratings or terminal absorbers.

[0125] By utilizing the linear region at the edge of the filter, wavelength changes are converted into light intensity changes, so that the light intensity of the reflected light after passing through the filter is linearly related to the wavelength shift.

[0126] Based on the linear relationship, the voltage signal is converted into a wavelength offset by the filter slope.

[0127] Specifically, the fiber optic signal processor includes a wavelength demodulation unit for converting wavelength variations of the fiber Bragg grating (FBG) into electrical signals. A broadband light source is used to emit optical signals covering the FBG's reflection band; the light source bandwidth must cover the wavelength range of all FBGs. The optical signals enter the distributed FBG array through fiber optic couplers. Each FBG reflects its specific wavelength λ. B The remaining wavelengths are transmitted to subsequent gratings or terminal absorbers. The wavelength change is converted into an intensity change using the linear region at the filter's edge. After passing through the filter, the intensity I of the reflected light is linearly related to the wavelength shift Δλ, expressed as:

[0128] I = k·Δλ + I0;

[0129] In the formula, k is the filter slope, I is the current signal; I0 is the initial current value; Δλ is the wavelength offset;

[0130] Specifically, Δλ can be converted into a current signal I by calibrating the coefficient k (filter slope).

[0131] Specifically, the wireless transmission module supports a variety of wireless communication protocols, including but not limited to Wi-Fi, Bluetooth, or mobile communication networks.

[0132] Specifically, the sensor of the present invention has waterproof and dustproof functions, and is suitable for harsh environments such as offshore wind farms.

[0133] In one embodiment, the noise covariance matrix update strategy based on environmental changes, the dynamic optimization of the filtering algorithm, and the denoising processing of the fused strain information of the wind turbine blades include:

[0134] S51. Based on the fused strain information of wind turbine blades, establish state equations and observation equations; wherein, the state equations include state vectors, state transition matrices, control input matrices and process noise, and the observation equations include observed values, observation matrices and observation noise;

[0135] S52. Based on the changes in the operating environment of wind turbine blades, the weights of historical data and current data are balanced using a forgetting factor to dynamically update the process noise covariance matrix and the observation noise covariance matrix.

[0136] S53. Based on the updated noise covariance matrix, perform the prediction and update process of the filtering algorithm to denoise the fused strain data.

[0137] In one embodiment, the denoising process of the fused strain data, based on the updated noise covariance matrix, involves performing a prediction and update process for the filtering algorithm, including:

[0138] S531. In the prediction stage, the prior error covariance matrix is ​​calculated based on the state transition matrix and the process noise covariance matrix.

[0139] S532. Calculate the Kalman gain based on the prior error covariance matrix and the observation noise covariance matrix;

[0140] S533. During the update phase, the state estimate is corrected using Kalman gain and observations, and the filtered state vector is output to obtain the fused strain data of the wind turbine blade after denoising, so as to monitor the strain state of the wind turbine blade in real time.

[0141] Specifically, due to the randomness of wind loads and mechanical vibrations in the operating environment of wind turbine blades, the measurement noise and process noise covariance matrices of sensors may change over time. High humidity, temperature fluctuations, and blade vibrations in offshore winds can cause dynamic drift in noise characteristics. Traditional Kalman filtering, which uses fixed noise covariance matrices Q and R, cannot adapt to such changes. Therefore, this invention proposes an adaptive noise covariance matrix update strategy, forming a Kalman filtering method based on dynamic optimization of environmental factors, to denoise the fused data.

[0142] Specifically, the state equation is:

[0143] X k =AX k-1 +Bu k +w k ;

[0144] In the formula, X k Let u be the state vector at time k, A be the state transition matrix, B be the control input matrix, and u be the state vector at time k. k For external inputs (such as changes in wind speed), w k This is process noise;

[0145] The observation equation is:

[0146] Z k =HX k +v k ;

[0147] In the formula, Z k For the observed value, HX k Let v be the observation matrix. k The observed noise. The process noise covariance matrix is ​​Q. k , representing the noise of the system process. The observation noise covariance matrix R k , which represents the uncertainty of noise in the observed data.

[0148] Specifically, the operating environment of wind turbine blades (such as wind speed and temperature) affects the performance of sensors, and consequently, the noise characteristics. Therefore, this invention designs a noise covariance matrix update strategy based on environmental changes, expressed as:

[0149]

[0150] In the formula, α1 and β1 are forgetting factors, with values ​​ranging from (0,1), used to balance the weights of historical data and current data; K k For Kalman gain; y k To measure the residual, P k | k-1 and P k | k These are the prior and posterior error covariance matrices, respectively; H k This is the measurement matrix.

[0151] Specifically, in the filtering process, the prediction stage:

[0152]

[0153] Update phase formula:

[0154] K k =P k|k-1 H T HP k|k-1 H T +R) -1 ;

[0155]

[0156] P k|k =(IK k H)P k|k-1 ;

[0157] In the formula, K k Let P be the Kalman gain, H be the error covariance matrix, H be the measurement matrix describing the linear relationship between the state variables and the measured values, and R be the measurement noise covariance matrix. The output is the prior state estimate of the current state; the output is the filtered state vector. Obtain the denoised strain data.

[0158] Specifically, the wind turbine blade strain sensor of the present invention can detect the early development of wind turbine blade faults by monitoring the strain state of wind turbine blades in real time with high precision, which can significantly improve the operating efficiency and safety of wind turbine generator sets.

[0159] To facilitate understanding of the above technical solutions of the present invention, specific embodiments are described below:

[0160] Example 1

[0161] To verify the feasibility of this invention, a certain type of wind turbine blade was selected as the test object. First, the fiber grating (FBG) arrangement was determined: based on the finite element stress analysis results of the blade, grating nodes were arranged in a zigzag pattern in the high strain gradient region (blade root, leading edge, and airfoil transition region), with a grating density ≥ 3 gratings / m. 2 The low-strain zones (blade tip and trailing edge) are sparsely distributed, with a density of ≤1 zone / m². 2 The overall number of sensors is reduced by 20% to 30%, while ensuring that the critical stress areas of the entire blade are covered. The layout density of the distributed fiber grating array is positively correlated with the high strain gradient region (strain change rate ≥ 0.1 με / mm) in the finite element analysis of the blade.

[0162] The distributed fiber Bragg grating (FBG) encapsulation employs a three-layer structure design: the encapsulation mold consists of carbon fiber reinforced polymer (CFRP), a flexible silicone buffer layer, and a polyurethane waterproof layer, divided into an inner layer, a middle layer, and an outer layer. The upper layer features a small groove aligned with the fiber direction in the center, facilitating fiber fixation and effectively protecting the fiber from breakage during encapsulation. The fiber direction of the CFRP layer is consistent with the main load-bearing direction of the blade, with over 70% of the layer being laid up at 0°.

[0163] The signal processing and analysis system consists of a fiber optic signal processor and a data analysis unit. The fiber optic grating is connected to the fiber optic signal processor, which employs a tunable laser and an FPGA high-speed demodulation unit. A high-sensitivity wavelength demodulation device accurately converts the change in the center wavelength of the fiber optic grating into an electrical signal. The data analysis unit receives these electrical signals, calculates the blade strain using a multi-mode data fusion algorithm, and effectively eliminates the influence of environmental noise using Kalman filtering technology.

[0164] In terms of wireless data transmission, the data analysis unit is connected to the wireless transmission module, which supports multiple communication protocols and can send the monitored data to the remote monitoring center in real time, thereby realizing real-time monitoring of the strain status of wind turbine blades.

[0165] During the system testing and optimization phase, comprehensive tests were conducted in both actual wind turbine blades and simulated environments. By adjusting the fiber Bragg grating arrangement density and signal processing parameters, the system's accuracy and response speed were optimized to ensure accurate reflection of the wind turbine blade's strain state under various operating conditions. Test results show that the system can capture minute strain changes of 0.1 με with a response time of less than 10 ms.

[0166] Example 2

[0167] Building upon Example 1, the system's environmental adaptability and data processing capabilities have been further enhanced. To adapt to harsh environments such as offshore wind farms, the sensor has undergone reinforced waterproofing and dustproofing treatment. The encapsulation employs a double-layer waterproof design, with an outer fluorocarbon coating and an inner modified silicone rubber layer, providing a temperature resistance range of -40℃ to 85℃. Simultaneously, corrosion-resistant and high / low temperature-resistant materials are selected to ensure the sensor's stability and reliability in extreme environments, meeting IP68 protection requirements.

[0168] In terms of data processing, dedicated data processing software was developed, which can not only display blade strain data in real time, but also perform trend analysis based on historical data to provide timely warnings of potential blade damage or failure. The software adopts an adaptive weight adjustment mechanism to dynamically optimize the multimodal data fusion algorithm based on the operating status, further improving the accuracy of strain monitoring.

[0169] To verify the system's versatility, sensor compatibility tests were conducted on wind turbine generator sets of different brands and models. The results show that the invention is widely applicable to various wind turbine generator sets and has good compatibility and scalability.

[0170] Through the above embodiments, the wind turbine blade strain sensor of the present invention can operate stably under various operating conditions and provide accurate and real-time strain monitoring data, which is of great significance for improving the operating efficiency and safety of wind turbine units.

[0171] In summary, based on the above-mentioned technical solution of this invention, a sensing system composed of a carbon fiber reinforced polymer encapsulation and an internally integrated distributed fiber Bragg grating sensor is proposed. Addressing the problems of traditional external wind turbine blade sensors being susceptible to environmental interference, having insufficient reliability of the encapsulation structure, and low signal processing accuracy, this system innovatively employs a flexible polyimide substrate. Through a multi-layer structure, it achieves mechanical protection and temperature stress compensation, effectively extending the sensor's lifespan. Based on the non-uniform fiber Bragg grating array layout derived from blade finite element stress analysis, the sensor layout is optimized, reducing the overall number of sensors by 20%–30% and improving monitoring efficiency. The system integrates a multi-modal data fusion algorithm and dynamic wavelength demodulation technology, combined with a Kalman filter noise reduction method dynamically optimized based on environmental factors, to achieve high-precision early warning of cracks. Simultaneously, a low-power dual-mode wireless transmission module is employed to ensure stable long-distance communication in high-humidity marine environments.

[0172] This invention develops a novel wind turbine blade strain monitoring technology that provides high-precision and high-reliability strain monitoring. It can accurately detect strain changes inside the wind turbine blade in real time, effectively analyze the internal health status of the blade, and adapt to complex and changing environmental conditions to meet the needs of modern wind power generation technology.

[0173] This invention provides a sensing system that can be built into the blade structure, has distributed measurement capabilities, strong environmental adaptability and long lifespan, and is equipped with advanced data fusion and processing algorithms, enabling comprehensive and real-time monitoring of the health status of wind turbine blades.

[0174] The sensor of this invention has the advantages of high precision, real-time monitoring, strong anti-interference ability and low maintenance cost. It can detect early cracks in the wind turbine blades and is suitable for various types of wind turbine units. In particular, it is of great significance for strain monitoring of wind turbine blades in large offshore wind farms operating in harsh environments, and can improve the operating efficiency and safety of wind turbine units.

[0175] This invention achieves high-precision detection of internal strain in wind turbine blades using distributed fiber Bragg grating array (FBG) technology. It accurately captures strain changes in various parts of the blade, employing a dense array strategy, particularly in high strain gradient regions, to ensure monitoring accuracy in critical areas. The system utilizes a wireless transmission module for real-time data transmission and monitoring, enabling wind farm maintenance personnel to monitor blade status at any time. FBG arrays inherently possess electromagnetic interference resistance, making them suitable for the complex electromagnetic environment of wind turbines and unaffected by external interference such as lightning strikes. Furthermore, the excellent stability and durability of optical fiber significantly reduce system maintenance costs, making it suitable for long-term deployment. More importantly, this invention can effectively identify early internal faults in wind turbine blades by analyzing internal strain change patterns, providing a scientific basis for preventative maintenance and significantly improving the operational safety and service life of wind turbines.

[0176] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A wind turbine blade strain-embedded sensing system based on a distributed fiber optic grating array, characterized in that, include: A distributed fiber Bragg grating array is placed on the inside of the wind turbine blade to capture strain signals inside the blade. The encapsulation module, including a flexible substrate and an adhesive, is disposed on the top inner side of the wind turbine blade to fix and protect the distributed grating fiber array, so as to ensure accurate capture of strain signals when the blade is subjected to stress and deformation. An optical fiber signal processor, connected to a distributed fiber Bragg grating array, is used to receive and process strain signals transmitted by the fiber Bragg grating array. The data analysis unit, connected to the fiber optic signal processor, is used to analyze the processed strain signal using a multi-modal data fusion algorithm and to calculate the strain information of the wind turbine blades. The wireless transmission module, connected to the data analysis unit, is used to wirelessly transmit strain signals to the monitoring center. The data analysis unit, when analyzing the processed strain signal using a multimodal data fusion algorithm and calculating the strain information of the wind turbine blades, includes: S1. Based on the strain data acquired by fiber Bragg grating, time synchronization and unit standardization are performed, and the maximum strain value and strain gradient are extracted to obtain the first mode feature. S2. Process the wind speed and pitch angle data using interpolation methods to ensure time alignment with the strain data, and extract the wind speed change rate and pitch angle adjustment frequency to obtain the second mode characteristics; S3. Use a filtering algorithm to process the vibration spectrum data, extract the vibration frequency and amplitude as indirect representations of the dynamic load, and obtain the third mode characteristics; S4. Based on the multimodal data fusion algorithm, integrate the first modal features, second modal features, and third modal features to obtain the fused strain information of the wind turbine blade. Specifically, this includes: S41. Constructing a state vector based on the first modal features, second modal features, and third modal features; S42. Constructing strain information weighting coefficients based on the nonlinear relationship between wind speed, temperature, and strain, and setting vibration signal weighting coefficients based on the rate of change of wind speed and vibration frequency; S43. Calculating the temperature compensation coefficient based on real-time temperature changes, and establishing a vibration frequency compensation coefficient by analyzing the changes in vibration modes and introducing the coupling effect of temperature and load; S44. Using the strain information weighting coefficient, vibration signal weighting coefficient, temperature compensation coefficient, and vibration frequency compensation coefficient, weighted fusion of each modal data in the state vector is performed to obtain the preliminary fusion result of the wind turbine blade; S45. Based on the coupling relationship between strain and vibration, and temperature and vibration in the wind turbine blade, optimizing the preliminary fusion result of the wind turbine blade by introducing the first cross-coupling coefficient and the second cross-coupling coefficient to obtain the fused strain information of the wind turbine blade. S5. A noise covariance matrix update strategy based on environmental changes is implemented, the filtering algorithm is dynamically optimized, and the fused strain information of wind turbine blades is denoised.

2. The wind turbine blade strain-embedded sensing system based on a distributed fiber optic grating array according to claim 1, characterized in that, Based on the coupling relationship between strain and vibration, and temperature and vibration in wind turbine blades, the preliminary fusion results of wind turbine blades are optimized by introducing a first cross-coupling coefficient and a second cross-coupling coefficient, resulting in the following fused strain information of the wind turbine blades: S451. Based on the product between the normalized strain value and the normalized vibration frequency value, and combined with empirical data, calculate the first cross-coupling coefficient to characterize the dynamic coupling effect between strain and vibration. S452. Calculate the second cross-coupling coefficient by using the product between the temperature difference and the normalized value of the vibration frequency, combined with an adjustment factor, to characterize the effect of temperature change on vibration characteristics. S453. By weighting and integrating the first cross-coupling coefficient and the second cross-coupling coefficient of each modal data in the state vector, the preliminary fusion result of the wind turbine blade is nonlinearly corrected to obtain the fusion strain information of the wind turbine blade.

3. The wind turbine blade strain-embedded sensing system based on a distributed fiber optic grating array according to claim 2, characterized in that, The expression for the fused strain information of the wind turbine blade is: ； In the formula, X fusion w(t) represents the fused strain information of the wind turbine blade; w1(t) represents the strain information weighting coefficient; w2(t) represents the vibration signal weighting coefficient; w3(t) represents the temperature compensation coefficient; w4(t) represents the vibration frequency compensation coefficient; w5(t) represents the first cross-coupling coefficient; w6(t) represents the second cross-coupling coefficient; ε represents the strain, v represents the wind speed, β represents the blade pitch angle, and f represents the wind speed. vib The frequency is the vibration frequency.

4. The wind turbine blade strain-embedded sensing system based on a distributed fiber optic grating array according to claim 1, characterized in that, The distributed fiber grating array is conformally bonded to the inner curved surface of the blade through a polymer film. The distributed fiber grating array includes several fiber gratings etched on a single optical fiber to cover different center wavelength bands. The reflectivity of the fiber optic grating is controlled within a preset range to avoid crosstalk between adjacent grating signals; The distributed fiber grating array is based on the finite element stress simulation results of the blade. It adopts a dense grating arrangement in the high strain sensitive region and a sparse arrangement in the low strain region to optimize the sensor layout and reduce the overall number.

5. The wind turbine blade strain-embedded sensing system based on a distributed fiber optic grating array according to claim 1, characterized in that, The packaging module includes a flexible substrate with a thickness not exceeding a preset value, and the surface of the flexible substrate is coated with an adhesive layer of a specific thickness; The encapsulation module uses a vacuum bagging process to simultaneously cure the adhesive layer and blade layup within a preset pressure range and at a preset temperature. The cured guide groove is filled with a flexible sealing material of preset hardness to ensure that the aerodynamic performance parameters of the wind turbine blade do not change beyond a preset threshold after the sensor is installed.

6. The wind turbine blade strain-embedded sensing system based on a distributed fiber optic grating array according to claim 1, characterized in that, The fiber optic signal processor includes a wavelength demodulation device for converting wavelength changes of fiber Bragg gratings into electrical signals. The fiber optic signal processor, when receiving and processing the strain signal transmitted by the fiber Bragg grating array, includes: Based on a broadband light source, the transmission bandwidth can cover the optical signal of all fiber grating wavelength ranges in the distributed fiber grating array; Optical signals are introduced into a distributed fiber optic array via fiber couplers, so that each fiber optic grating reflects a specific wavelength, while the remaining wavelengths are transmitted to subsequent gratings or terminal absorbers. By utilizing the linear region at the edge of the filter, wavelength changes are converted into light intensity changes, so that the light intensity of the reflected light after passing through the filter is linearly related to the wavelength shift. Based on the linear relationship, the voltage signal is converted into a wavelength offset by the filter slope.

7. The wind turbine blade strain-embedded sensing system based on a distributed fiber optic grating array according to claim 1, characterized in that, The noise covariance matrix update strategy based on environmental changes, the dynamic optimization of the filtering algorithm, and the denoising processing of the fused strain information of the wind turbine blades include: S51. Based on the fused strain information of wind turbine blades, establish state equations and observation equations; wherein, the state equations include state vectors, state transition matrices, control input matrices, and process noise, and the observation equations include observed values, observation matrices, and observation noise; S52. Based on the changes in the operating environment of wind turbine blades, the weights of historical data and current data are balanced using a forgetting factor to dynamically update the process noise covariance matrix and the observation noise covariance matrix. S53. Based on the updated noise covariance matrix, perform the prediction and update process of the filtering algorithm to denoise the fused strain data.

8. The wind turbine blade strain-embedded sensing system based on a distributed fiber optic grating array according to claim 7, characterized in that, The process of performing prediction and update of the filtering algorithm based on the updated noise covariance matrix to denoise the fused strain data includes: S531. In the prediction stage, the prior error covariance matrix is ​​calculated based on the state transition matrix and the process noise covariance matrix. S532. Calculate the Kalman gain based on the prior error covariance matrix and the observation noise covariance matrix; S533. During the update phase, the state estimate is corrected using Kalman gain and observations, and the filtered state vector is output to obtain the fused strain data of the wind turbine blade after denoising, so as to monitor the strain state of the wind turbine blade in real time.

Images