A fan blade strain monitoring method based on optical fiber sensing

Abstract

The application discloses a kind of fan blade strain monitoring methods based on optical fiber sensing, it is related to fan blade monitoring technical field, comprising: the vibration signal in the process of acquisition blade operation is carried out denoising and time-frequency analysis processing to vibration signal;Calculate a variety of time-frequency parameters, and utilize wavelet packet transform to carry out frequency band division to vibration signal, extract frequency index to obtain signal characteristic parameter;Establish the multivariate linear regression model between blade strain and signal characteristic parameter, and combined with stress analysis theory, construct health assessment model, output health assessment result;Combined with signal characteristic parameter and health assessment result, according to credibility analysis algorithm and load subcoefficient, the residual service life of blade is evaluated, to carry out the real-time monitoring of strain state.The application is laid out optical fiber sensor along axial and ring direction in the target monitoring area of fan blade, combined with the measurement principle of optical time domain reflection technology, improves the spatial resolution and measurement accuracy of strain monitoring.

Description

Technical Field

[0001] This invention relates to the field of wind turbine blade monitoring technology, and more specifically, to a method for monitoring wind turbine blade strain based on fiber optic sensing. Background Technology

[0002] As an indispensable part of clean energy, wind power is playing an increasingly important role in the global energy structure. Wind turbine blades, as one of the core components of a wind power system, directly affect the performance, safety, and reliability of the entire system. However, because wind turbine blades are exposed to harsh working environments for extended periods, they must withstand complex aerodynamic forces, gravity, and environmental factors (such as strong winds, dust storms, and low temperatures), making them highly susceptible to fatigue damage and crack propagation.

[0003] Currently, while there are various methods for monitoring wind turbine blades, they also have many limitations. Traditional strain gauge-based monitoring methods require extensive wiring, which not only increases installation and maintenance costs but is also susceptible to interference in complex environments, resulting in poor stability. Machine vision monitoring, while providing some information about the blade surface, is easily affected by lighting conditions and weather, and struggles to capture internal blade strain. Although non-destructive testing technologies such as acoustic emission and ultrasonic testing can identify blade defects to some extent, these technologies cannot achieve real-time and long-term monitoring of blade strain.

[0004] As wind power generation moves towards larger scale and deeper offshore applications, the size of wind turbine blades is gradually increasing. Traditional monitoring methods are proving inadequate in terms of accuracy and coverage when applied to these large blades. Furthermore, existing monitoring methods often fail to adequately consider multiple factors such as material fatigue characteristics and load randomness when assessing the remaining service life of wind turbine blades, resulting in insufficient accuracy in the prediction results.

[0005] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0006] In view of the problems in related technologies, this invention proposes a method for monitoring the strain of wind turbine blades based on fiber optic sensing, so as to overcome the above-mentioned technical problems existing in the existing related technologies.

[0007] Therefore, the specific technical solution adopted by the present invention is as follows:

[0008] A method for monitoring the strain of wind turbine blades based on fiber optic sensing, the method comprising:

[0009] S1. Collect vibration signals during blade operation and perform noise reduction and time-frequency analysis on the vibration signals;

[0010] S2. Based on the processing results, calculate various time-frequency parameters, and use wavelet packet transform to divide the vibration signal into frequency bands, extract frequency indices to obtain signal characteristic parameters;

[0011] S3. Establish a multiple linear regression model between blade strain and signal characteristic parameters, and combine it with stress analysis theory to construct a health assessment model and output the health assessment results.

[0012] S4. Combining signal characteristic parameters and health assessment results, and based on the reliability analysis algorithm and load partial factor, assess the remaining service life of the blades to enable real-time monitoring of strain status.

[0013] Furthermore, the vibration signals during blade operation are collected, and the vibration signals are processed by noise reduction and time-frequency analysis, including:

[0014] S11. By using fiber optic sensors and local acceleration data sensors deployed along the axial and circumferential directions, the acceleration data of blade vibration is collected synchronously, and a multi-dimensional early warning mechanism is constructed by combining digital twin technology.

[0015] S12. Based on the constructed multi-dimensional early warning mechanism, the vibration signals output by the fiber optic sensor and the acceleration data sensor are collected, and the vibration signals are denoised using the wavelet filtering algorithm.

[0016] S13. Based on the denoised oscillation signal, perform time-frequency analysis on the denoised vibration signal using the moving window fast Fourier transform algorithm.

[0017] Furthermore, by simultaneously acquiring acceleration data of blade vibration through fiber optic sensors deployed along the axial and circumferential directions and local acceleration data sensors, and combining this with digital twin technology to construct a multi-dimensional early warning mechanism, including:

[0018] S111. In the target monitoring area of ​​the wind turbine blade, fiber optic sensors based on optical time domain reflection technology are deployed along the axial and circumferential directions to collect stress distribution data of the blade during operation.

[0019] S112. Using deployed fiber optic sensors to emit pulsed light, and through the backscattering Rayleigh effect, to sense local strain changes caused by blade vibration and identify vibration events.

[0020] S113. Based on optical time-domain reflectometry, measure the round-trip time difference of the optical signal caused by the vibration event, determine the spatial location of the vibration event, and deploy an acceleration data sensor at the corresponding spatial location to synchronously collect the acceleration data of the blade vibration.

[0021] S114. A three-dimensional dynamic simulation model is established using digital twin technology. Combined with Hooke's law, stress data and acceleration data are converted into a three-dimensional stress field. The real-time deformation response of the virtual blade under load is executed through dynamic driving equations.

[0022] S115. Based on the real-time deformation response results, calculate the damage distribution of each element, and construct a multi-dimensional early warning mechanism by combining the stress threshold and the root mean square value of the acceleration data.

[0023] Furthermore, by utilizing deployed fiber optic sensors to emit pulsed light and sensing local strain changes caused by blade vibration through the backscattering Rayleigh effect, vibration events can be identified, including:

[0024] S1121. According to the monitoring requirements, the optical fiber is divided into several segments along its length.

[0025] S1122. Based on the partitioning results, the fiber optic sensor emits pulsed light. During the propagation of the pulsed light, backscattered Rayleigh light is generated and superimposed in each interval before returning to the incident end along the fiber to generate a detection signal.

[0026] S1123. Analyze the generated detection signal, analyze the amplitude vector sum and phase vector sum of the backscattered Rayleigh light in each interval, and combine the amplitude and phase values ​​of the backscattered Rayleigh light in each interval to calculate the amplitude sum of the scattered light intensity in each interval and obtain the light source signal.

[0027] S1124. Post-process the light source signal, combine the incident light power at the corresponding position with the fiber transmission loss, and extract the power change caused by the disturbance to identify vibration events.

[0028] Furthermore, the expression for the change in power is:

[0029]

[0030] In the formula, p B The power at point B, a distance d from the fiber optic incident section, represents the power at that point; x represents the positional variable at the distance from the fiber optic incident section; p o denoted by initial power; d represents the distance from the incident fiber segment; a0(x) represents the transmission loss of the incident light.

[0031] Furthermore, based on the constructed multi-dimensional early warning mechanism, vibration signals output from fiber optic sensors and accelerometers are collected, and wavelet filtering algorithms are used to denoise the vibration signals, including:

[0032] S121. Based on the backscattering Rayleigh effect, the fiber optic sensor acquires the fiber strain field changes caused by blade vibration by emitting pulsed light and detecting reflected signals, obtains the corresponding optical signals, and converts the optical signals into electrical signals.

[0033] S122. Based on the change in charge generated by the internal piezoelectric effect, the local acceleration data sensor collects the acceleration data generated by the blade vibration and converts the acceleration data into an electrical signal.

[0034] S123. Combine the electrical signal converted from optical signal with the acceleration data converted into an electrical signal to establish a vibration signal. Select the wavelet basis function and the number of decomposition layers to perform multi-layer decomposition on the vibration signal to obtain the smooth component and the oscillation component.

[0035] S124. The wavelet coefficients in the oscillation component are processed by a preset threshold. Wavelet coefficients smaller than the preset threshold are set to zero to remove noise interference. The processed wavelet coefficients are then reconstructed to obtain the denoised vibration signal.

[0036] Furthermore, the time-frequency parameters include: mean, root mean square value, standard deviation, frequency domain energy, centroid frequency, and average frequency.

[0037] Furthermore, a multiple linear regression model between blade strain and signal characteristic parameters is established, and combined with stress analysis theory, a health assessment model is constructed. The output health assessment results include:

[0038] S31. Based on the acquired signal characteristic parameters, establish a multiple linear regression model between blade strain and signal characteristic parameters;

[0039] S32. Combining stress analysis theory with Hooke's law and tensor analysis algorithm, obtain the stress components of the blade in different directions;

[0040] S33. Based on the stress components in each direction, calculate the equivalent stress at each location, identify the stress concentration area, set the stress threshold, and assess the health status of the blade by combining the changing trend of signal characteristic parameters over time.

[0041] S34. Based on the health status assessment results, construct a health assessment model based on support vector machine, input the real-time signal feature parameters into the health assessment model, and output the current leaf health assessment results.

[0042] Furthermore, by combining signal characteristic parameters and health assessment results, and based on reliability analysis algorithms and load partial factors, the remaining service life of the blades is evaluated to enable real-time monitoring of strain status, including:

[0043] S41. Analyze the relationship between signal characteristic parameters and health assessment results, calculate the unit damage based on the damage accumulation theory, and calculate the remaining strength of the blade by combining the number of load cycles experienced by the blade.

[0044] S42. Based on the remaining strength of the blade, the reliability analysis algorithm is used to calculate the reliability of the blade in the current state, and combined with the load partial factor under different working conditions, the reliability under multiple working conditions is obtained.

[0045] S43. By combining the reliability under multiple operating conditions with the number of load cycles that have occurred, assess the remaining service life of the blade under the current operating condition.

[0046] Furthermore, the formula for calculating credibility is:

[0047]

[0048] In the formula, R represents the reliability of the blade; σ y The standard deviation of blade strength is represented by y; y represents a variable defined based on blade strength and working stress; μ y This represents the mean of y.

[0049] The beneficial effects of this invention are as follows:

[0050] 1. This invention enables full-coverage sensing of strain in all directions of the wind turbine blade by deploying fiber optic sensors along the axial and circumferential directions in the target monitoring area of ​​the blade. Combined with the measurement principle of optical time domain reflectance technology, it improves the spatial resolution and measurement accuracy of strain monitoring, and can more sensitively capture minute strain changes, thereby obtaining more comprehensive and accurate strain data.

[0051] 2. This invention effectively removes noise interference from vibration data through wavelet filtering and combines it with time-frequency analysis technology to convert the signal to the time-frequency domain, so that the energy distribution characteristics of the signal at different times and frequencies can be clearly displayed. This not only lays a high-quality data foundation for the accurate extraction of subsequent strain characteristic parameters, but also improves the accuracy and stability of the overall data analysis.

[0052] 3. This invention calculates various signal characteristic parameters from both the time and frequency domains and introduces wavelet packet transform to achieve fine decomposition of vibration signals. It can comprehensively and accurately capture the vibration characteristics of wind turbine blades and accurately reflect their strain state, providing solid data support and judgment basis for in-depth analysis of blade structural behavior and assessment of their health status. Attached Figure Description

[0053] 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.

[0054] Figure 1This is a flowchart of a wind turbine blade strain monitoring method based on fiber optic sensing according to an embodiment of the present invention;

[0055] Figure 2 This is a schematic flowchart of a wind turbine blade strain monitoring method based on fiber optic sensing according to an embodiment of the present invention. Detailed Implementation

[0056] 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.

[0057] According to an embodiment of the present invention, a method for monitoring the strain of wind turbine blades based on fiber optic sensing is provided.

[0058] 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 method for monitoring the strain of wind turbine blades based on fiber optic sensing includes:

[0059] S1. Collect vibration signals during blade operation and perform noise reduction and time-frequency analysis on the vibration signals.

[0060] In this optional embodiment, collecting vibration signals during blade operation and performing noise reduction and time-frequency analysis on the vibration signals includes:

[0061] S11. By using fiber optic sensors and local acceleration data sensors deployed along the axial and circumferential directions, the acceleration data of blade vibration is collected synchronously, and a multi-dimensional early warning mechanism is constructed by combining digital twin technology.

[0062] In this optional embodiment, the acceleration data of blade vibration is synchronously collected by fiber optic sensors and local acceleration data sensors deployed along the axial and circumferential directions, and a multi-dimensional early warning mechanism is constructed by combining digital twin technology, including:

[0063] S111. In the target monitoring area of ​​the wind turbine blade, fiber optic sensors based on optical time domain reflection technology are deployed along the axial and circumferential directions to collect stress distribution data of the blade during operation.

[0064] S112. By using deployed fiber optic sensors to emit pulsed light and sensing the local strain changes caused by blade vibration through the backscattering Rayleigh effect, vibration events can be identified.

[0065] In this optional embodiment, fiber optic sensors are used to emit pulsed light, and the local strain changes caused by blade vibration are sensed through the backscattering Rayleigh effect. Vibration events are identified, including:

[0066] S1121. According to the monitoring requirements, the optical fiber is divided into several segments along its length.

[0067] S1122. Based on the partitioning results, the fiber optic sensor emits pulsed light. During the propagation of the pulsed light, backscattered Rayleigh light is generated and superimposed in each interval before returning to the incident end along the fiber to generate a detection signal.

[0068] S1123. Analyze the generated detection signal, analyze the amplitude vector sum and phase vector sum of the backscattered Rayleigh light in each interval, and combine the amplitude and phase values ​​of the backscattered Rayleigh light in each interval to calculate the amplitude sum of the scattered light intensity in each interval and obtain the light source signal.

[0069] S1124. Post-process the light source signal, combine the incident light power at the corresponding position with the fiber transmission loss, and extract the power change caused by the disturbance to identify vibration events.

[0070] In this optional embodiment, the expression for the power change is:

[0071]

[0072] In the formula, p B The power at point B, a distance d from the fiber optic incident section, represents the power at that point; x represents the positional variable at the distance from the fiber optic incident section; p o denoted by initial power; d represents the distance from the incident fiber segment; a0(x) represents the transmission loss of the incident light.

[0073] S113. Based on optical time-domain reflectometry, measure the round-trip time difference of the optical signal caused by the vibration event to determine the spatial location of the vibration event, and deploy an acceleration data sensor at the corresponding spatial location to synchronously collect the acceleration data of the blade vibration.

[0074] S114. A three-dimensional dynamic simulation model is established using digital twin technology. Combined with Hooke's law, stress data and acceleration data are converted into a three-dimensional stress field. The real-time deformation response of the virtual blade under load is executed through dynamic driving equations.

[0075] S115. Based on the real-time deformation response results, calculate the damage distribution of each element, and construct a multi-dimensional early warning mechanism by combining the stress threshold and the root mean square value of the acceleration data.

[0076] It should be further explained that, assuming the total length of the optical fiber is L, the entire optical fiber is divided into N segments according to the required precision, and the interval length is... The backscattered Rayleigh rays generated by the pulsed light after N scatterings will all superimpose at discrete points and return to the point of incidence; let A k and Let a be the sum of the amplitude vectors and the sum of the phase vectors of the M backscattered Rayleigh rays in the k-th segment. i and Ω i Let be the amplitude and phase values ​​of the i-th backscattered Rayleigh beam within ΔL, respectively. Then, the sum of the scattered light intensity amplitudes of the k-th fiber segment is expressed as:

[0077]

[0078] In the formula, A k Let represent the sum of the amplitude vectors of the backscattered Rayleigh light in the k-th segment; j represents the imaginary unit; The sum of the phase vectors of the backscattered Rayleigh light in segment k represents the sum of the phase vectors of the backscattered Rayleigh light; M represents the number of backscattered Rayleigh lights; i represents the index value; a i Ω represents the amplitude of the i-th backscattered Rayleigh beam; i This represents the phase value of the i-th backscattered Rayleigh beam.

[0079] Regarding the method of converting light intensity, let the power at point B, which is a distance d from the incident section of the optical fiber, be p. B If (x), then the power generated by the disturbance at that location in a short time is expressed as:

[0080]

[0081] In the formula, p B (x) represents the power at point B, which is a distance d from the fiber optic incident section; x represents the positional variable of the fiber optic incident section; p o denoted by initial power; d represents the distance from the incident fiber segment; a0(x) represents the transmission loss of the incident light.

[0082] In order to achieve precise localization of vibration events, based on the principle of optical time domain reflectometer (OTDR), the distance from the event occurrence point to the incident segment can be obtained by measuring the incident and reflected light information:

[0083]

[0084] In the formula, d represents the distance from the incident section of the optical fiber; c represents the speed of light in a vacuum; τ represents the time difference; and n represents the refractive index.

[0085] Simultaneously, combined with a local acceleration sensor, when the sensor is subjected to vibration, the internal cantilever beam vibrates, causing a change in internal resistance, which in turn causes a change in the charge Q. According to Newton's second law F=ma, Q is proportional to a, i.e., Q=e*F=e*ma (e is the piezoelectric constant). Finally, the charge is converted into a voltage signal by a charge amplifier, and a voltage value proportional to the acceleration is output, thereby obtaining the vibration acceleration information of the blade.

[0086] Next, digital twin technology was introduced to establish a three-dimensional dynamic simulation model, constructing a deep fusion driving system between fiber optic sensing data and the virtual model; by developing a cross-scale mapping algorithm for strain-stress-modality, the axial strain ε monitored by the fiber optic cable was mapped... x Circumferential strain ε y The data were converted into a three-dimensional stress field using a derivative formula of Hooke's Law. The derivative formula of Hooke's Law is as follows:

[0087]

[0088] In the formula, σ ij The stress tensor at a point on the blade (i,j = 1, 2, 3, representing three mutually perpendicular directions); E represents the elastic modulus of the material; v represents the Poisson's ratio of the material (describing the ratio of transverse strain to longitudinal strain); ε kk The sum of strain in three directions (ε) 11 +ε 22 +ε 33 );δ ij Represents the Kronecker symbol; ε ij This represents the total strain at a point on the blade.

[0089] Real-time deformation simulation of virtual blades is achieved through dynamic driving equations, and element damage distribution is dynamically calculated using element damage formulas. Furthermore, a multi-dimensional early warning mechanism is constructed based on the VonMises stress threshold and the root mean square value of vibration acceleration. The dynamic driving equations are as follows:

[0090]

[0091] In the formula, M represents the blade mass matrix (related to mass distribution); The modal coordinates represent the second derivative with respect to time; C represents the damping matrix (characterizing vibration energy loss); F(t) represents the first derivative of the modal coordinates with respect to time; F(t) represents the external load (such as aerodynamic load, gravity); F sens (t) represents the load equivalent to the fiber optic sensing data (calculated by the integral formula of the equivalent load of sensing data).

[0092] The formula for the integral of the equivalent load is:

[0093] F sens (t)=∫ Ω B T σ(ε(t))dΩ;

[0094] In the formula, F sens (t) represents the load equivalent to the fiber optic sensing data (calculated by the integral formula); Ω represents the structural space integral domain covered by the fiber optic monitoring; B represents the element strain-displacement matrix; T represents the matrix transpose; σ(ε(t)) represents the stress field calculated from the strain ε(t) through the constitutive relation.

[0095] The unit damage formula is:

[0096]

[0097] In the formula, D e The value represents the element damage (reflecting the degree of damage); k represents the element number in the finite element model of the blade structure; i represents the index value of the load cycle, used to mark the i-th load cycle experienced by the blade; n i N represents the weight coefficient or quantity of the i-th unit (to be determined based on the specific modeling logic); i σ represents the fatigue life at the stress level of the i-th element; max,i This represents the maximum stress value borne by the k-th element in the i-th load cycle.

[0098] The VonMises stress threshold is:

[0099]

[0100] In the formula, σ VM σ1, σ2, and σ3 represent the Von Mises equivalent stress (also known as the equivalent yield stress); σ1, σ2, and σ3 represent the normal stresses in three orthogonal directions (in the shear-free state) obtained after the stress tensor is decomposed by eigenvalues.

[0101] Furthermore, when the multi-dimensional early warning mechanism detects abnormal vibration signals in a certain area (such as the root mean square value of acceleration exceeding a set threshold), the system will mark the signal in that area as "high priority" and enable a more stringent signal processing strategy (such as reducing noise tolerance and refining frequency discrimination) to ensure that key details in the abnormal signal are preserved and to prevent effective information from being filtered out. When the early warning mechanism further triggers the "stress exceeding standard" state, the model will focus on analyzing the changing trend of signal characteristic parameters in this state (such as frequency domain energy mutation, frequency shift, etc.) and dynamically adjust the parameter weights in the evaluation model (such as increasing the proportion of stress concentration area's contribution to damage), thereby enhancing the sensitivity and accuracy of health status assessment.

[0102] S12. Based on the constructed multi-dimensional early warning mechanism, the vibration signals output by the fiber optic sensor and the acceleration data sensor are collected, and the vibration signals are denoised using the wavelet filtering algorithm.

[0103] In this optional embodiment, based on the constructed multi-dimensional early warning mechanism, vibration signals output by fiber optic sensors and acceleration data sensors are collected, and wavelet filtering algorithms are used to denoise the vibration signals, including:

[0104] S121. Based on the backscattering Rayleigh effect, the fiber optic sensor acquires the fiber strain field changes caused by blade vibration by emitting pulsed light and detecting reflected signals, obtains the corresponding optical signals, and converts the optical signals into electrical signals.

[0105] S122. Based on the change in charge generated by the internal piezoelectric effect, the local acceleration data sensor collects the acceleration data generated by the blade vibration and converts the acceleration data into an electrical signal.

[0106] S123. Combine the electrical signal converted from optical signal with the acceleration data converted into an electrical signal to establish a vibration signal. Select the wavelet basis function and the number of decomposition layers to perform multi-layer decomposition on the vibration signal to obtain the smooth component and the oscillation component.

[0107] S124. The wavelet coefficients in the oscillation component are processed by a preset threshold. Wavelet coefficients smaller than the preset threshold are set to zero to remove noise interference. The processed wavelet coefficients are then reconstructed to obtain the denoised vibration signal.

[0108] S13. Based on the denoised oscillation signal, perform time-frequency analysis on the denoised vibration signal using the moving window fast Fourier transform algorithm.

[0109] It should be noted that the data on the change in backscattered Rayleigh light signal caused by the change in the refractive index of the optical fiber due to blade vibration is collected using the following formula:

[0110]

[0111] In the formula, p B (x) represents the power at point B, which is a distance d from the fiber optic incident section; x represents the positional variable of the fiber optic incident section; p o denoted by initial power; d represents the distance to the incident fiber segment; a0(x) represents the transmission loss of the incident light; the collected optical signal is converted into vibration-related electrical signal data.

[0112] The local accelerometer is based on the principle of Q = e * F (where Q represents the amount of charge generated internally, e represents the piezoelectric constant, F = ma, m represents the mass, and a represents the acceleration) to collect the acceleration data generated by the vibration of the blade and convert it into an electrical signal output.

[0113] First, wavelet filtering is used to denoise the data. The formula is as follows:

[0114]

[0115] In the formula, W f (a,b) represent wavelet transform coefficients; f(t) represents the vibration signal; denoted by ; a represents the conjugate of the wavelet basis functions; b represents the scale parameter; and t represents time.

[0116] After noise removal, to more clearly obtain the signal characteristics, the following formula is used to perform time-frequency analysis on the denoised signal; STFT is calculated using a moving windowed fast Fourier transform to obtain a series of spectrums of the signal, which can reflect the change of the signal spectrum over time. The formula is as follows:

[0117]

[0118] In the formula, STFT represents the moving windowed fast Fourier transform; j represents the imaginary unit; w represents the angular frequency; m represents the translation parameter of the window function w(tm); s(t) represents the vibration signal; and w(tm) represents the window function.

[0119] S2. Based on the processing results, calculate various time-frequency parameters, and use wavelet packet transform to divide the vibration signal into frequency bands, extracting frequency indices to obtain signal characteristic parameters.

[0120] In this optional embodiment, the time-frequency parameters include: mean, root mean square value, standard deviation, frequency domain energy, centroid frequency, and average frequency.

[0121] It should be further explained that, from a time-domain perspective, various parameters reflecting the signal amplitude and variation characteristics are calculated; the mean, as an indicator describing the stable components of the signal, is calculated using the following formula: Where, x i Let N represent the amplitude of the signal at the i-th sampling point, and N represent the total number of sampling points; the root mean square (RMS) value is used to measure the signal strength, and its formula is: The larger this value, the higher the overall energy of the signal, reflecting the severity of the wind turbine blade vibration; the standard deviation describes the degree of dispersion of the signal from the mean, and the formula is: The larger the standard deviation, the greater the signal fluctuation, which may indicate that the strain state of the blade is unstable.

[0122] In the frequency domain, the time-domain signal is converted to the frequency domain using the Fast Fourier Transform (FFT), and then the frequency domain energy is calculated; the formula for calculating the frequency domain energy is as follows: Where X(k) represents the frequency domain component after FFT transformation, this parameter reflects the energy distribution of the signal at different frequency components, which helps to analyze the frequency range of concentrated blade vibration energy; in addition, it determines the maximum energy frequency f. max That is, f max =argmax k |X(k)| represents the frequency point where the energy is most concentrated in the signal.

[0123] Considering the complexity of wind turbine blade vibration signals, this invention introduces wavelet packet transform to perform a more refined decomposition of the signal; assuming the signal is decomposed into 2... n Each frequency band has a bandwidth of [missing information]. (f s (Indicates the sampling frequency), the frequency band of the m-th wavelet packet is:

[0124]

[0125] In the formula, fr m The frequency band of the m-th wavelet packet is represented; m represents the index value; f s Indicates the sampling frequency; 2 n+1 This indicates the number of frequency bands into which the signal is decomposed.

[0126] To further analyze the signal characteristics, the centroid frequency S1 and average frequency S2 of the signal are calculated; the centroid frequency describes the frequency of the signal with larger components in the spectrum, and its calculation formula is as follows: Where P(k) represents the corresponding power spectral density value, f k This represents the frequency amplitude at the corresponding point, reflecting the centroid of the signal energy distribution on the frequency axis; the average frequency, as the average value of the power spectrum, is calculated using the following formula: This parameter reflects the average level of the signal frequency from another perspective.

[0127] S3. Establish a multiple linear regression model between blade strain and signal characteristic parameters, and combine it with stress analysis theory to construct a health assessment model and output the health assessment results.

[0128] In this optional embodiment, a multiple linear regression model is established between blade strain and signal characteristic parameters. Combined with stress analysis theory, a health assessment model is constructed, and the output health assessment results include:

[0129] S31. Based on the acquired signal characteristic parameters, establish a multiple linear regression model between blade strain and signal characteristic parameters;

[0130] S32. Combining stress analysis theory with Hooke's law and tensor analysis algorithm, obtain the stress components of the blade in different directions;

[0131] S33. Based on the stress components in each direction, calculate the equivalent stress at each location, identify the stress concentration area, set the stress threshold, and assess the health status of the blade by combining the changing trend of signal characteristic parameters over time.

[0132] S34. Based on the health status assessment results, construct a health assessment model based on support vector machine, input the real-time signal feature parameters into the health assessment model, and output the current leaf health assessment results.

[0133] It should be noted that, firstly, a correlation model between strain and characteristic parameters is established; considering that in actual situations, the strain of wind turbine blades may be affected by a combination of factors, a multiple linear regression model is used to describe the relationship between strain and each characteristic parameter.

[0134] To more accurately assess the overall strain state of the blade, stress analysis is introduced. According to Hooke's law, within the elastic range, stress and strain have a linear relationship σ = Eε (σ represents stress, and E represents the elastic modulus of the material). For the complex structure of wind turbine blades, considering the stress distribution in different directions, tensor analysis is used to describe the stress state. Let the stress tensor at a certain point on the blade be σ. ij (i,j=1,2,3, representing three mutually perpendicular directions), by transforming and calculating the extracted strain characteristic parameters, the stress components of the point in different directions can be obtained.

[0135] To assess the overall stress on the blade, the concept of Von Mises equivalent stress is introduced. For the plane stress state (simplified model, which can be extended to three dimensions as needed), the formula for calculating Von Mises stress is:

[0136]

[0137] In the formula, Von Mises stress; σ 11 σ 22 σ 33 Let σ represent the normal stress components at a point on the blade in three mutually perpendicular directions (usually defined as the x, y, and z axes), where σ 11 Corresponding to the normal stress in the x-direction, σ 22 Corresponding to the normal stress in the y-direction, σ 33 Corresponding normal stress in the z-direction; σ 12 σ 23 σ 13 Let σ represent the shear stress components at a point on the blade in different directional planes, where σ 12 The corresponding shear stress in the xy plane, σ 23 The corresponding shear stress in the yz plane, σ 13Corresponding to the shear stress in the xz plane; under the plane stress assumption σ 33 =σ 23 =σ 13 =0; By calculating the VonMises stress at different locations, the areas of stress concentration on the blade can be determined, and these areas are often prone to damage.

[0138] When assessing blade health, a stress threshold σ is set. threshold When the calculated VonMises stress When the threshold is exceeded, it indicates that the part of the blade is under high stress and there is a potential risk of damage. At the same time, the health status of the blade is further judged by combining the changing trend of characteristic parameters over time.

[0139] S4. Combining signal characteristic parameters and health assessment results, and based on the reliability analysis algorithm and load partial factor, assess the remaining service life of the blades to enable real-time monitoring of strain status.

[0140] In this optional embodiment, the remaining service life of the blade is assessed by combining signal characteristic parameters and health assessment results, based on a reliability analysis algorithm and load partial factors, in order to perform real-time monitoring of strain status, including:

[0141] S41. Analyze the relationship between signal characteristic parameters and health assessment results, calculate the unit damage based on the damage accumulation theory, and calculate the remaining strength of the blade by combining the number of load cycles experienced by the blade.

[0142] S42. Based on the remaining strength of the blade, the reliability analysis algorithm is used to calculate the reliability of the blade in the current state, and combined with the load partial factor under different working conditions, the reliability under multiple working conditions is obtained.

[0143] In this optional embodiment, the formula for calculating credibility is:

[0144]

[0145] In the formula, R represents the reliability of the blade; σ y The standard deviation of blade strength is represented by y; y represents a variable defined based on blade strength and working stress; μ y This represents the mean of y.

[0146] S43. By combining the reliability under multiple operating conditions with the number of load cycles that have occurred, assess the remaining service life of the blade under the current operating condition.

[0147] It should be further explained that, according to Miner's damage accumulation theory, this theory posits that when a material is subjected to cyclic loading, its internal damage gradually accumulates. When the accumulated damage reaches a certain level, the material will experience fatigue failure. Let N be the number of load cycles borne by the wind turbine blade, D be the unit damage caused by each load cycle, and x be the initial strength of the blade. o Then, after N cycles, the remaining strength x of the blade N Represented as:

[0148] x N = x0(1-ND);

[0149] In the formula, x N This represents the remaining strength of the blade after N cycles; x o N represents the initial strength of the blade; D represents the number of load cycles the wind turbine blades bear; and D represents the unit damage caused by each load cycle.

[0150] In practical applications, determining the unit damage quantity D is crucial. By analyzing the extracted strain characteristic parameters and the calculated stress state, the unit damage quantity D can be approximately expressed as:

[0151]

[0152] In the formula, D represents the unit damage caused by each load cycle; N f It represents the fatigue life of a material under a specific stress amplitude.

[0153] Introducing the concept of credibility (reliability), we assume that the strength of the blade follows a normal distribution. The working stress follows a normal distribution Where μ s and σ s These are the mean and standard deviation of the intensity, respectively, μ. l and σ l Let be the mean and standard deviation of the working stress, respectively; define the variable y = x. s -x l (x s For intensity, x l (For the working stress), according to the reliability analysis algorithm (reliability principle), the reliability R of the blade is expressed as:

[0154]

[0155] In the formula, R represents the reliability of the blade; σ y The standard deviation of leaf strength is represented by y; y represents the standard deviation of leaf strength (σ). s ) and working stress (σ l ) Defined variables; μ y The mean of y is represented; where, μ y =μ s -μ l μ s μ represents the mean value of the blade strength. l This represents the average stress on the blade.

[0156] Considering that different types of loads have different degrees of impact on blade damage, a load partial factor γ is introduced. i The reliability R1 under multiple operating conditions is expressed as:

[0157]

[0158] In the formula, R1 represents the reliability under multiple operating conditions; σ y1 The standard deviation of variable y1 is represented by μ; y represents a variable defined based on blade strength and working stress; μ represents the standard deviation of variable y1. y1 Let represent the mean of variable y1; where, y1 represents the variables involved in the reliability calculation formula under multiple operating conditions; n represents the actual number of load cycles borne by the wind turbine blades; i represents the index value; γ i N represents the weight coefficient of the i-th influencing factor; i D represents the number of load cycles a material can withstand before fatigue failure (i.e., fatigue life); i x represents the unit damage caused by each load cycle under the i-th working condition or stress level; si This represents a characteristic value of the blade strength under the i-th operating condition; x li σ represents a characteristic value of the blade's working stress under the i-th operating condition; si σ represents the standard deviation of blade strength under the i-th operating condition; li μ represents the standard deviation of the blade's working stress under the i-th operating condition. si μ represents the mean blade strength under the i-th operating condition; li This represents the average working stress of the blade under the i-th working condition.

[0159] The lifespan N corresponding to the blade's reliability in its current state is calculated using the method described above. Given the number of load cycles N1 already experienced, the remaining lifespan T of the wind turbine blade can be calculated. remaining for:

[0160]

[0161] In the formula, T remaining N represents the remaining service life of the wind turbine blades; N represents the number of load cycles; N1 represents the number of load cycles that have been experienced so far. This represents the average number of load cycles per unit time.

[0162] By comprehensively considering the strain state, stress level, material fatigue characteristics, and load randomness of the blades, this method can accurately predict the remaining service life of wind turbine blades, providing a scientific basis for wind turbine maintenance and replacement and ensuring the safe and stable operation of the wind turbine.

[0163] like Figure 2 As shown in the illustration, in a specific embodiment, a method for monitoring the strain of wind turbine blades based on fiber optic sensing includes:

[0164] (1) Construct a fiber optic sensing and monitoring system;

[0165] (2) Acquiring and preprocessing vibration data;

[0166] (3) Extract strain characteristic parameters;

[0167] (4) Analyze the strain status and assess the health status;

[0168] (5) Predict the remaining useful life.

[0169] In the fiber optic sensing and monitoring system described in (1), fiber optic sensors are deployed along the axial and circumferential directions of the blades in key areas of the wind turbine blades, such as the root, middle, and tip of the blades, where stress concentration is likely to occur. The axial sensors are used to monitor axial strain, and the circumferential sensors are used to monitor circumferential strain, thereby comprehensively acquiring strain information of the blades in different directions. A fiber optic sensor based on the optical time domain reflection (OTDR) principle is selected. Its working principle is based on the backscattering of incident light in the fiber. When the fiber is subjected to external forces, such as the strain caused by the vibration of the wind turbine blades, the refractive index of the fiber will change, which in turn will change the intensity and phase of the backscattered Rayleigh light. According to the characteristics of the backscattered Rayleigh light, its intensity is inversely proportional to the fourth power of the wavelength. Using this relationship, the strain of the blades can be sensed by monitoring the change in the intensity of the backscattered Rayleigh light.

[0170] In practical applications, to improve monitoring accuracy, improved OTDR technology, such as phase-sensitive optical time-domain reflectometry, is adopted. This technology, by detecting and analyzing the phase of backscattered Rayleigh light, can more sensitively monitor minute strain changes in optical fibers. Its basic principle is to process the interferometric optical signal, converting phase changes into light intensity changes for measurement. Let the total length of the optical fiber be L, and based on the system's accuracy, the entire fiber is divided into N segments, with interval lengths... The backscattered Rayleigh rays generated by the pulsed light after N scatterings will all superimpose at discrete points and return to the point of incidence; let A k and Let a be the sum of the amplitude vectors and the sum of the phase vectors of the M backscattered Rayleigh rays in the k-th segment. i and Ω iLet be the amplitude and phase values ​​of the i-th backscattered Rayleigh beam within ΔL, respectively. Then, the sum of the scattered light intensity amplitudes of the k-th fiber segment is expressed as:

[0171] Regarding the method for converting light intensity, the system receives the light source signal and performs post-processing such as bandpass and lowpass filtering, as well as signal amplification, to perform calculations; let the power at point B, which is d away from the incident section of the optical fiber, be p. B If (x), then the power generated by the disturbance at that location in a short time is expressed as:

[0172] To achieve precise localization of vibration events, based on the principle of optical time domain reflectometer (OTDR), the distance from the event occurrence point to the incident segment can be obtained by measuring the incident and reflected light information.

[0173] Simultaneously, by combining local acceleration sensors, such as IEPE piezoelectric acceleration sensors, the working principle is based on the mutual conversion between force and acceleration. When the sensor is subjected to vibration, the internal cantilever beam vibrates, causing a change in internal resistance, which in turn causes a change in the amount of charge Q. According to Newton's second law F=ma, Q is proportional to a, that is, Q=e*F=e*ma (e is the piezoelectric constant). Finally, the charge is converted into a voltage signal by a charge amplifier, and a voltage value proportional to the acceleration is output, thereby obtaining the vibration acceleration information of the blade.

[0174] Next, digital twin technology was introduced to establish a three-dimensional dynamic simulation model, constructing a deep fusion driving system between fiber optic sensing data and the virtual model, breaking through the limitations of traditional monitoring based solely on the analysis of a single physical quantity; by developing a cross-scale mapping algorithm for strain-stress-modality, the axial strain ε monitored by the fiber optic cable was mapped to the desired values. x Circumferential strain ε y The data were converted into a three-dimensional stress field using a derivative formula of Hooke's Law. The derivative formula of Hooke's Law is as follows: This invention combines Miner's damage accumulation theory with finite element fatigue analysis to achieve real-time deformation simulation of virtual blades through dynamic driving equations. The formula dynamically calculates the damage distribution of the unit, while also being based on the VonMises stress threshold. A multi-dimensional early warning mechanism is constructed by combining the root mean square value of vibration acceleration, enabling the virtual model to not only reproduce the actual mechanical behavior of the blade, but also to predict potential risk areas through parameter sensitivity analysis.

[0175] In the process of collecting and preprocessing vibration data in (2), after the fiber optic sensing and monitoring system is constructed in (1), the system begins to continuously collect vibration data of the wind turbine blades during operation. The distributed fiber optic sensor, based on the principle of backscattering of light by Rayleigh, collects the change data of backscattered light signal caused by the change in the refractive index of the fiber optic due to blade vibration. The formula is as follows:

[0176] The local accelerometer is based on the principle of Q = e * F (where Q represents the amount of charge generated internally, e represents the piezoelectric constant, F = ma, m represents the mass, and a represents the acceleration) to collect the acceleration data generated by the vibration of the blade and convert it into an electrical signal output.

[0177] The raw data collected often contains various noises, which can affect the accuracy of subsequent analysis. Therefore, preprocessing is necessary. First, wavelet filtering is used to denoise the data. The formula is: By selecting appropriate wavelet basis functions and decomposition levels, the acquired vibration signal is decomposed into multiple layers, dividing the signal into low-frequency (smoothing components) and high-frequency (oscillatory components). The low-frequency components mainly contain the main characteristic information of the signal, while the high-frequency components contain noise and detail information. By setting a threshold, the wavelet coefficients of the high-frequency components are processed, and wavelet coefficients smaller than the threshold are set to zero, thereby removing the noise component. The processed wavelet coefficients are then reconstructed to obtain the denoised signal.

[0178] After noise removal, to more clearly obtain the signal characteristics, the following formula is used to perform time-frequency analysis on the denoised signal; STFT is calculated using a moving windowed fast Fourier transform to obtain a series of spectrums of the signal, which can reflect the change of the signal spectrum over time. The formula is as follows: This transformation converts the vibration signal in the time domain to the time-frequency domain, allowing us to obtain the energy distribution of the signal at different times and frequencies. This enables us to observe the characteristics of the signal more intuitively and provides a better data foundation for the subsequent accurate extraction of strain characteristic parameters.

[0179] In the process of extracting strain characteristic parameters in (3), after completing the preprocessing of the collected vibration data in (2), based on the signal after denoising and time-frequency analysis, further characteristic parameters that can accurately characterize the strain state of the wind turbine blades are extracted.

[0180] From a time-domain perspective, various parameters reflecting the amplitude and variation characteristics of the signal are calculated; the mean, as an indicator describing the stable components of the signal, is calculated using the following formula: Where, x i Let N represent the amplitude of the signal at the i-th sampling point, and N represent the total number of sampling points; the root mean square (RMS) value is used to measure the signal strength, and its formula is: The larger this value, the higher the overall energy of the signal, reflecting the severity of the wind turbine blade vibration; the standard deviation describes the degree of dispersion of the signal from the mean, and the formula is: The larger the standard deviation, the greater the signal fluctuation, which may indicate that the strain state of the blade is unstable.

[0181] In the frequency domain, the time-domain signal is converted to the frequency domain using the Fast Fourier Transform (FFT), and then the frequency domain energy is calculated; the formula for calculating the frequency domain energy is as follows: Where X(k) represents the frequency domain component after FFT transformation, this parameter reflects the energy distribution of the signal at different frequency components, which helps to analyze the frequency range of concentrated blade vibration energy; in addition, it determines the maximum energy frequency f. max That is, f max =argmax k |X(k)| represents the frequency point where the energy is most concentrated in the signal.

[0182] Considering the complexity of wind turbine blade vibration signals, this invention introduces wavelet packet transform to perform a more refined decomposition of the signal; assuming the signal is decomposed into 2... n Each frequency band has a bandwidth of [missing information]. (f s (Indicates the sampling frequency), the frequency band of the m-th wavelet packet is:

[0183] To further analyze the signal characteristics, the centroid frequency S1 and average frequency S2 of the signal are calculated; the centroid frequency describes the frequency of the signal with larger components in the spectrum, and its calculation formula is as follows: Where P(k) represents the corresponding power spectral density value, f k This represents the frequency amplitude at the corresponding point, reflecting the centroid of the signal energy distribution on the frequency axis; the average frequency, as the average value of the power spectrum, is calculated using the following formula: This parameter reflects the average level of the signal frequency from another perspective.

[0184] In section (4), which analyzes the strain state and assesses the health status, after extracting the strain characteristic parameters in section (3), these parameters are used to analyze the strain state of the wind turbine blades in depth and assess their health status. Based on the extracted time-domain and frequency-domain characteristic parameters, a correlation model between strain and characteristic parameters is first established. Considering that in actual situations, the strain of wind turbine blades may be affected by a combination of factors, a multiple linear regression model is used to describe the relationship between strain and each characteristic parameter. Let the strain value be ε, and the multiple characteristic parameters be x1, x2, ..., x n (e.g., mean, root mean square value, frequency domain energy, etc.), establish the regression equation ε=β0+β1x1+β2x2+…++β n x n+ε, where β0,β1,…,β n ε is the regression coefficient, and ε is the error term. The regression coefficients are solved by the least squares method so that the model can best fit the relationship between strain and characteristic parameters.

[0185] To more accurately assess the overall strain state of the blade, stress analysis is introduced. Referring to relevant mechanics theories, and based on Hooke's law, within the elastic range, stress and strain have a linear relationship σ = Eε (where σ represents stress and E represents the material's elastic modulus). For the complex structure of wind turbine blades, considering the stress distribution in different directions, tensor analysis is used to describe the stress state. Let the stress tensor at a certain point on the blade be σ. ij (i,j=1,2,3, representing three mutually perpendicular directions), by transforming and calculating the extracted strain characteristic parameters, the stress components of the point in different directions can be obtained.

[0186] To assess the overall stress on the blade, the concept of Von Mises equivalent stress is introduced. For a plane stress state (simplified model, which can be extended to three dimensions as needed), the formula for calculating Von Mises stress is: (Under the plane stress setting), by calculating the Von Mises stress at different locations, the areas of stress concentration on the blade can be determined. These areas are often the parts that are prone to damage.

[0187] When assessing blade health, a stress threshold σ is set. threshold When the calculated VonMises stress When the threshold is exceeded, it indicates that the part of the blade is under high stress and there is a potential risk of damage. At the same time, the health status of the blade is further judged by combining the changing trend of characteristic parameters over time.

[0188] Furthermore, by using machine learning algorithms to learn and train on a large amount of historical monitoring data, a health assessment model can be established. The Support Vector Machine (SVM) algorithm can be used to effectively classify the data on healthy and faulty states by finding an optimal classification hyperplane. During the training process, the extracted feature parameters are used as inputs, and the corresponding leaf health status (healthy or faulty) is used as outputs. The parameters of the SVM are solved by an optimization algorithm so that it can accurately identify the health status of the leaves. For new monitoring data, the extracted feature parameters are input into the trained SVM model, and the model will output the health assessment result of the leaves, such as healthy, sub-healthy, or faulty states.

[0189] In the (5) prediction of remaining service life, based on the analysis of the strain state and health status assessment of the wind turbine blades completed in (4), the remaining service life of the wind turbine blades is predicted by combining material fatigue theory and probability statistics methods.

[0190] First, based on Miner's damage accumulation theory, this theory posits that when a material is subjected to cyclic loading, internal damage gradually accumulates, and when the accumulated damage reaches a certain level, the material will experience fatigue failure. Let N be the number of load cycles borne by the wind turbine blade, D be the unit damage caused by each load cycle, and x be the initial strength of the blade. o Then, after N cycles, the remaining strength x of the blade N Represented as: x N = x0(1-ND).

[0191] In practical applications, determining the unit damage amount D is crucial. This is achieved by analyzing the strain characteristic parameters extracted in (3) and the stress state calculated in (4), and by combining the fatigue characteristic curves of the blade material (such as the SN curve, which describes the fatigue life of the material under different stress levels) to estimate D; based on the stress amplitude σ... a Find the corresponding fatigue life N on the SN curve. f Then the unit damage amount D can be approximately expressed as:

[0192] Considering the randomness and uncertainty of the loads on wind turbine blades during actual operation, a probabilistic statistical method is used to more accurately predict the remaining service life; it is assumed that the blade strength follows a normal distribution. The working stress follows a normal distribution Where μ s and σ s These are the mean and standard deviation of the intensity, respectively, μ. l and σ l Let be the mean and standard deviation of the working stress, respectively; define the variable y = x. s -x l (x s For intensity, x l (where the stress is the working stress), according to the reliability principle, the reliability R of the blade is expressed as: in, μ y =μ s -μ l By statistically analyzing historical monitoring data, the mean and standard deviation of strength and working stress can be estimated. As the blades are used, their strength gradually decreases, and according to Miner's damage accumulation theory, the mean and standard deviation of strength will also change. Let's assume that after N1 load cycles, the mean strength becomes μ. s1 =(1-N1D)μ sWith the standard deviation remaining unchanged (this assumption can be made in simplified models, but can be corrected according to more complex models in practice), the N value corresponding to the confidence level R (i.e. the blade life, expressed in terms of load cycles) can be obtained by solving the above integral equation using numerical calculation methods.

[0193] Furthermore, considering that different types of loads have different degrees of impact on blade damage, a load partial factor γ is introduced. i For wind turbine blades, common load types include aerodynamic loads and gravity loads. Each load has a corresponding partial factor. Therefore, the reliability R1 under multiple operating conditions is expressed as:

[0194] The lifespan N corresponding to the blade's reliability in its current state is calculated using the method described above. Given the number of load cycles N1 already experienced, the remaining lifespan T of the wind turbine blade can be calculated. remaining for:

[0195] By comprehensively considering the strain state, stress level, material fatigue characteristics, and load randomness of the blades, this method can accurately predict the remaining service life of wind turbine blades, providing a scientific basis for wind turbine maintenance and replacement and ensuring the safe and stable operation of the wind turbine.

[0196] 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 method for monitoring the strain of wind turbine blades based on fiber optic sensing, characterized in that, The method includes: S1. Collect vibration signals during blade operation and perform noise reduction and time-frequency analysis on the vibration signals; S2. Based on the processing results, calculate various time-frequency parameters, and use wavelet packet transform to divide the vibration signal into frequency bands, extract frequency indices to obtain signal characteristic parameters; S3. Establish a multiple linear regression model between blade strain and signal characteristic parameters, and combine it with stress analysis theory to construct a health assessment model and output the health assessment results. S4. Combining signal characteristic parameters and health assessment results, and based on the reliability analysis algorithm and load partial factor, assess the remaining service life of the blades to enable real-time monitoring of strain status; The process of collecting vibration signals during blade operation and performing noise reduction and time-frequency analysis on the vibration signals includes: S11. By using fiber optic sensors and local acceleration data sensors deployed along the axial and circumferential directions, the acceleration data of blade vibration is collected synchronously, and a multi-dimensional early warning mechanism is constructed by combining digital twin technology. S12. Based on the constructed multi-dimensional early warning mechanism, the vibration signals output by the fiber optic sensor and the local acceleration data sensor are collected, and the vibration signals are denoised using the wavelet filtering algorithm. S13. Based on the denoised vibration signal, perform time-frequency analysis on the denoised vibration signal using the moving windowed fast Fourier transform algorithm; The process involves establishing a multiple linear regression model between blade strain and signal characteristic parameters, and combining this with stress analysis theory to construct a health assessment model. The output health assessment results include: S31. Based on the acquired signal characteristic parameters, establish a multiple linear regression model between blade strain and signal characteristic parameters; S32. Combining stress analysis theory with Hooke's law and tensor analysis algorithm, obtain the stress components of the blade in different directions; S33. Based on the stress components in each direction, calculate the equivalent stress at each location, identify the stress concentration area, set the stress threshold, and assess the health status of the blade by combining the changing trend of signal characteristic parameters over time. S34. Based on the health status assessment results, construct a health assessment model based on support vector machine, input the real-time signal feature parameters into the health assessment model, and output the current leaf health assessment result; The process of combining signal characteristic parameters and health assessment results, and using a reliability analysis algorithm and load partial factors to assess the remaining service life of the blades for real-time monitoring of strain status includes: S41. Analyze the relationship between signal characteristic parameters and health assessment results, calculate the unit damage based on the damage accumulation theory, and calculate the remaining strength of the blade by combining the number of load cycles experienced by the blade. S42. Based on the remaining strength of the blade, the reliability analysis algorithm is used to calculate the reliability of the blade in the current state, and combined with the load partial factor under different working conditions, the reliability under multiple working conditions is obtained. S43. By combining the reliability under multiple operating conditions with the number of load cycles that have occurred, assess the remaining service life of the blade under the current operating condition.

2. The method for monitoring wind turbine blade strain based on fiber optic sensing according to claim 1, characterized in that, The method of synchronously acquiring acceleration data of blade vibration through fiber optic sensors and local acceleration data sensors deployed along the axial and circumferential directions, and constructing a multi-dimensional early warning mechanism by combining digital twin technology, includes: S111. In the target monitoring area of ​​the wind turbine blade, fiber optic sensors based on optical time domain reflection technology are deployed along the axial and circumferential directions to collect stress distribution data of the blade during operation. S112. Using deployed fiber optic sensors to emit pulsed light, and through the backscattering Rayleigh effect, to sense local strain changes caused by blade vibration and identify vibration events. S113. Based on optical time-domain reflectometry, measure the round-trip time difference of the optical signal caused by the vibration event, determine the spatial location of the vibration event, and deploy local acceleration data sensors at the corresponding spatial locations to synchronously collect the acceleration data of the blade vibration. S114. A three-dimensional dynamic simulation model is established using digital twin technology. Combined with Hooke's law, stress data and acceleration data are converted into a three-dimensional stress field. The real-time deformation response of the virtual blade under load is executed through dynamic driving equations. S115. Based on the real-time deformation response results, calculate the damage distribution of each element, and construct a multi-dimensional early warning mechanism by combining the stress threshold and the root mean square value of the acceleration data.

3. The method for monitoring wind turbine blade strain based on fiber optic sensing according to claim 2, characterized in that, The method of using deployed fiber optic sensors to emit pulsed light and sensing local strain changes caused by blade vibration through the backscattering Rayleigh effect to identify vibration events includes: S1121. According to the monitoring requirements, the optical fiber is divided into several segments along its length. S1122. Based on the partitioning results, the fiber optic sensor emits pulsed light. During the propagation of the pulsed light, backscattered Rayleigh light is generated and superimposed in each interval before returning to the incident end along the fiber to generate a detection signal. S1123. Analyze the generated detection signal, analyze the amplitude vector sum and phase vector sum of the backscattered Rayleigh light in each interval, and combine the amplitude and phase values ​​of the backscattered Rayleigh light in each interval to calculate the amplitude sum of the scattered light intensity in each interval and obtain the light source signal. S1124. Post-process the light source signal, combine the incident light power at the corresponding position with the fiber transmission loss, and extract the power change caused by the disturbance to identify vibration events.

4. The method for monitoring wind turbine blade strain based on fiber optic sensing according to claim 3, characterized in that, The expression for the power change is: ； In the formula, p B The power at point B, a distance d from the fiber optic incident section, represents the power at that point; x represents the positional variable at the distance from the fiber optic incident section; p o denoted by initial power; d represents the distance from the incident fiber segment; a0(x) represents the transmission loss of the incident light.

5. The method for monitoring wind turbine blade strain based on fiber optic sensing according to claim 4, characterized in that, The constructed multi-dimensional early warning mechanism collects vibration signals output from fiber optic sensors and local acceleration data sensors, and uses wavelet filtering algorithms to denoise the vibration signals, including: S121. Based on the backscattering Rayleigh effect, the fiber optic sensor acquires the fiber strain field changes caused by blade vibration by emitting pulsed light and detecting reflected signals, obtains the corresponding optical signals, and converts the optical signals into electrical signals. S122. Based on the change in charge generated by the internal piezoelectric effect, the local acceleration data sensor collects the acceleration data generated by the blade vibration and converts the acceleration data into an electrical signal. S123. Combining the electrical signal converted from optical signal and the electrical signal converted from acceleration data, a vibration signal is established. The wavelet basis function and the number of decomposition layers are selected to perform multi-layer decomposition on the vibration signal to obtain the smooth component and the oscillation component. S124. The wavelet coefficients in the oscillation component are processed by a preset threshold. Wavelet coefficients smaller than the preset threshold are set to zero to remove noise interference. The processed wavelet coefficients are then reconstructed to obtain the denoised vibration signal.

6. The method for monitoring wind turbine blade strain based on fiber optic sensing according to claim 1, characterized in that, The time-frequency parameters include: mean, root mean square value, standard deviation, frequency domain energy, centroid frequency, and average frequency.

7. The method for monitoring wind turbine blade strain based on fiber optic sensing according to claim 1, characterized in that, The formula for calculating the credibility is: ； In the formula, R represents the reliability of the blade; σ y The standard deviation of blade strength is represented by y; y represents a variable defined based on blade strength and working stress; μ y This represents the mean of y.

Images