Wireless passive surface acoustic wave bolt stress monitoring system and method

Abstract

The invention discloses a wireless passive surface acoustic wave bolt stress monitoring system and method, and particularly relates to the technical field of bolt stress monitoring. Sensors are installed at key parts of a wind turbine generator, the sensors are excited to generate surface acoustic wave signals, propagation characteristics of waves are monitored, changes of the surface acoustic wave signals are collected and analyzed in real time, parameters reflecting the stress state of a machine body are obtained, the parameters comprise the surface roughness and the nonlinear stress-strain relation of materials, and the stress state of the machine body is obtained by analyzing the parameters. The measurement precision value of the sensor is calculated, the monitoring precision of the sensor is judged, and an early warning signal is generated to prompt maintenance personnel when the precision is low; and according to an analysis result and historical data, the working frequency and gain of the sensor are dynamically adjusted, and long-term stability correction is performed, so that the long-term stability and measurement precision of the sensor are effectively improved, and the defect that the performance of the wind turbine generator sensor is influenced by long-term mechanical vibration, repeated stress change and temperature fluctuation in the prior art is overcome.

Description

Technical Field

[0001] The invention relates to the technical field of wave bolt stress monitoring, and in particular to a wireless passive surface acoustic wave bolt stress monitoring system and method. Background Art

[0002] Wireless passive SAW bolt stress monitoring refers to a method of monitoring the stress state of bolts under wireless and passive conditions by using SAW technology. Surface acoustic wave is an elastic wave that propagates on the surface of a solid and can be used to detect stress changes in materials. When a bolt is subjected to external force or changes, the stress and deformation on the surface will cause changes in the propagation characteristics of the SAW. By installing SAW sensors on bolts and combining them with wireless technology (such as RFID), real-time stress monitoring can be achieved without batteries or external power supplies, thereby improving the safety and reliability of engineering structures.

[0003] The prior art has the following deficiencies: During use, wind turbines are subject to long-term mechanical vibration, repeated stress changes, temperature changes and other environmental factors, which may affect the long-term stability of surface acoustic wave sensors. Especially when the sensor needs to work in an extreme vibration and impact environment, it may cause wear or loosening of the sensor element, thereby affecting its measurement accuracy. In addition, long-term fatigue loads and stress changes may cause tiny cracks or loosening at the junction of the sensor and the structure, resulting in signal distortion or failure. Summary of the invention

[0004] The object of the present invention is to provide a wireless passive surface acoustic wave bolt stress monitoring system and method to solve the deficiencies in the background technology.

[0005] In order to achieve the above object, the present invention provides the following technical solution: a wireless passive surface acoustic wave bolt stress monitoring method, comprising the following steps: S1: Install several wireless passive surface acoustic wave sensors at key locations of the wind turbine, including the main bearing, gearbox, generator, blade root, blade bearing and tower connection; S2: exciting the surface acoustic wave sensor to generate a surface acoustic wave signal, and monitoring the propagation characteristics of the surface acoustic wave, including propagation speed, amplitude and phase, through the surface acoustic wave sensor; S3: collecting the surface acoustic wave signal in real time, and obtaining the surface roughness of the fuselage reflecting the stress state of the fuselage and the nonlinear stress-strain relationship of the material according to the change of the surface acoustic wave signal; S4: Analyze the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material, calculate the measurement accuracy value of the wind turbine monitoring sensor through the machine learning model, and judge the sensor monitoring status of the wind turbine fuselage according to the accuracy value; S5: According to the calculation results and historical data, the working state of the surface acoustic wave sensor is adjusted to perform dynamic correction, wherein the correction step includes adjusting the working frequency and gain of the sensor to ensure the long-term stability of the sensor's measurement accuracy.

[0006] Preferably, a wireless passive surface acoustic wave sensor is used to collect echo signals from key parts of the wind turbine generator set, and multiple features related to surface roughness are extracted from the echo signals, wherein the multiple features include the amplitude, phase and propagation velocity of the echo signals, and the body surface roughness anomaly index is calculated by a covariance matrix analysis method, specifically: The echo signals of the key parts of the wind turbine are collected by using surface acoustic wave sensors. Multiple features related to surface roughness are extracted from the echo signals. These features are organized into a data matrix. Each row represents the features of a moment or a measurement point, and each column represents different characteristic variables. The eigenvalues ​​and eigenvectors of the covariance matrix are calculated. The eigenvalues ​​represent the variance of each principal component. The eigenvalues ​​and eigenvectors are obtained through eigendecomposition. The principal components are selected according to the size of the eigenvalues. After the principal components are selected, the data are projected onto the principal components to obtain a new characteristic matrix: the surface roughness anomaly index of the body is calculated, and the expression is: ;in, represents the standard deviation of the ith principal component, p is the number of principal components selected, m is the number of features in the matrix, It is the abnormality index of the fuselage surface roughness.

[0007] Preferably, in S3, a nonlinear effect drift index is generated after comparing the propagation velocity data under different stresses, the propagation velocity data is recorded at different measurement points and time points, and a method for obtaining the nonlinear effect drift index is: collecting the propagation velocity data under different stresses, selecting a nonlinear model, and setting the relationship between the propagation velocity and the stress to be a power law relationship: ; v represents the propagation velocity, E represents the stress, a is a constant, and b is the power exponent. The least squares method is used to estimate the model parameters a and b by minimizing the error between the model output and the actual data. The nonlinear effect drift index is calculated by fitting the parameters. The expression is: in, represents the propagation speed at the i-th moment, is the predicted value obtained by fitting the model, w is the total number of data points, and QER is the nonlinear effect drift index.

[0008] Preferably, in S4, the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material are analyzed and calculated to obtain the measurement accuracy value of the wind turbine monitoring sensor, specifically: The fuselage surface roughness anomaly index and nonlinear effect drift index are converted into comprehensive feature vectors, which are used as input of the machine learning model. The machine learning model is trained, and the measurement accuracy value of the wind turbine monitoring sensor is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0009] Preferably, the sensor monitoring status of the fuselage part is judged according to the accuracy value, specifically: the obtained measurement accuracy value of the wind turbine monitoring sensor is compared with a preset measurement accuracy reference threshold; if the measurement accuracy value of the wind turbine monitoring sensor is greater than or equal to the preset measurement accuracy reference threshold, it means that the measurement accuracy of the wind turbine monitoring sensor is high, and no warning signal is generated at this time; if the measurement accuracy value of the wind turbine monitoring sensor is less than the preset measurement accuracy reference threshold, it means that the measurement accuracy of the wind turbine monitoring sensor is low, and a warning signal is generated at this time to remind the unit maintenance personnel to check or replace the sensor.

[0010] Preferably, in S5, the working state of the surface acoustic wave sensor is adjusted according to the analysis results and historical data, and dynamic correction is performed, and the correction step includes adjusting the working frequency and gain of the sensor, specifically: The measurement accuracy and measurement error of wind turbine monitoring sensors are used as input items of fuzzy logic and divided into different fuzzy sets. The sensor's operating frequency and gain adjustment results are used as output items of fuzzy logic and divided into different fuzzy sets; Formulate fuzzy rules to describe the impact of wind turbine monitoring sensor measurement accuracy and historical data on sensor operating frequency and gain; Perform fuzzy reasoning according to fuzzy rules and adjust the sensor operating frequency and gain.

[0011] The present invention also provides a wireless passive surface acoustic wave bolt stress monitoring system, comprising a sensor configuration module, a signal excitation module, a data processing module, a measurement accuracy calculation module and an adaptive adjustment module; Sensor configuration module: installing at least one wireless passive surface acoustic wave sensor at key parts of the wind turbine, including the main bearing, gearbox, generator, blade root, blade bearing and tower connection; Signal excitation module: excites the surface acoustic wave sensor to generate a surface acoustic wave signal, and monitors the propagation characteristics of the surface acoustic wave through the sensor, wherein the propagation characteristics include propagation speed, amplitude and phase; Data processing module: collects the surface acoustic wave signal in real time, and obtains parameters reflecting the stress state of the fuselage according to the change of the surface acoustic wave signal, wherein the parameters include the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material; Measurement accuracy calculation module: Analyze and calculate the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material to obtain the measurement accuracy value of the wind turbine monitoring sensor, and judge the sensor monitoring status of the fuselage part according to the accuracy value; Adaptive adjustment module: according to the analysis results and historical data, the working state of the surface acoustic wave sensor is adjusted to perform dynamic correction. The correction step includes adjusting the working frequency and gain of the sensor to ensure the long-term stability of the sensor's measurement accuracy.

[0012] In the above technical solution, the technical effects and advantages provided by the present invention are: 1. The present invention can effectively solve the problems of wear, looseness, microcracks, etc. that may occur in sensors of wind turbines under complex environmental conditions such as long-term mechanical vibration, repeated stress changes and temperature fluctuations. By real-time acquisition of surface acoustic wave signals and analysis of their propagation characteristics, combined with the abnormality index of the fuselage surface roughness and the nonlinear effect drift index, the system can accurately evaluate the measurement accuracy of the sensor. Based on the machine learning model, the system can intelligently predict the long-term stability of the sensor and generate early warning signals, prompting the unit maintenance personnel to take timely measures to ensure the reliability and security of the monitoring data.

[0013] 2. The present invention also introduces a dynamic correction mechanism, which controls the operating frequency and gain of the sensor through fuzzy logic, and combines historical data and real-time analysis results to ensure the stability and accuracy of the sensor after long-term use. This method not only improves the anti-interference ability of the sensor, but also maintains its efficient measurement performance under extreme conditions, thereby providing a more accurate, stable and long-term solution for wind turbine stress monitoring, greatly improving the safety and reliability of wind turbines. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0015] Figure 1 The figure is a flow chart of the method of the present invention.

[0016] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example 1, please refer to Figure 1 As shown, the wireless passive surface acoustic wave bolt stress monitoring method described in this embodiment includes the following steps: S1: Install at least one wireless passive surface acoustic wave sensor at key parts of the wind turbine, including the main bearing, gearbox, generator, blade root, blade bearing and tower connection; S2: Exciting a surface acoustic wave sensor to generate a surface acoustic wave signal, and monitoring the propagation characteristics of the surface acoustic wave through the sensor, wherein the propagation characteristics include propagation speed, amplitude and phase; S3: collecting the surface acoustic wave signal in real time, and obtaining parameters reflecting the stress state of the fuselage according to changes in the surface acoustic wave signal, wherein the parameters include the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material; S4: Analyze and calculate the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material to obtain the measurement accuracy value of the wind turbine monitoring sensor, and judge the monitoring status of the sensor at the fuselage according to the accuracy value, and generate an early warning signal to prompt the maintenance personnel of the unit; S5: According to the analysis results and historical data, the working state of the surface acoustic wave sensor is adjusted to perform dynamic correction, wherein the correction step includes adjusting the working frequency and gain of the sensor to ensure the long-term stability of the sensor's measurement accuracy.

[0019] The wireless passive surface acoustic wave (SAW) sensor is a high-precision sensor based on the surface acoustic wave effect and has the following characteristics: Wireless transmission: no need to lay out complex wired connections, suitable for complex environments such as wind turbines; Passive power supply: No external power supply is required, relying on RF signals for excitation and data reading; High precision and high stability: can be used to measure temperature, stress, vibration, pressure and other parameters; Adapt to harsh environment: able to work stably in environments with high humidity and large temperature changes.

[0020] Wind turbines contain multiple core components. In order to achieve efficient monitoring, SAW sensors need to be installed in the following key locations: Main bearing: used to monitor bearing temperature and vibration to prevent bearing overheating or damage; Gearbox: installed on the input shaft, output shaft and gear meshing parts of the gearbox to detect the temperature, vibration and lubrication status of the gearbox to prevent gearbox failure; Generator: monitor the temperature and vibration of the stator and rotor to ensure the safe operation of the generator; Blade root and blade bearing: detect blade stress changes, prevent blade fatigue damage, and improve the reliability of the wind turbine; Key locations of the tower structure: such as the tower connection, monitor structural stress and vibration to prevent structural fatigue damage.

[0021] The installation methods of SAW sensors are generally as follows: Adhesive installation: suitable for parts such as blade roots where stress needs to be measured; Embedded installation: suitable for internal components such as bearings and gearboxes, fixed by reserved installation positions; Surface mount: fixed on the metal surface using high-temperature weather-resistant adhesives or welding to monitor temperature and vibration changes.

[0022] After installation, a wireless data acquisition terminal is required, which is usually composed of a wireless RF antenna and a signal demodulator, responsible for reading and transmitting the sensor's signals. The data can be processed through edge computing devices or cloud monitoring platforms to achieve remote real-time monitoring and early warning.

[0023] Surface acoustic wave sensors use the propagation characteristics of surface waves to sense changes in the surrounding medium, especially changes in the stress state of the material surface. The excitation of surface acoustic waves is generally carried out in the following ways: Excitation signal source: Piezoelectric materials are usually used as excitation sources to convert electrical signals into mechanical waves. These piezoelectric elements (for example, PZT materials) can vibrate under the action of electrical signals, thereby exciting surface acoustic waves. The frequency, amplitude and waveform of the excitation signal are usually determined by the design and measurement requirements of the sensor. By precisely controlling the excitation signal, it is possible to ensure that the surface acoustic wave propagates within the operating frequency band of the sensor, ensuring the clarity and accuracy of the signal.

[0024] Excitation method: Short-duration pulse signals, sine waves or linear frequency modulation signals (chirp) are usually used as excitation waveforms. These signals can adjust the frequency, amplitude and duration according to specific monitoring requirements to stimulate surface acoustic waves to propagate on the sensor surface. Pulse signals are usually used for time domain analysis, while sine waves or frequency modulation signals are used for frequency domain analysis.

[0025] Surface acoustic wave types: Commonly used surface acoustic wave types include Rayleigh waves and Einhope waves. Rayleigh waves are the most common type of surface waves. They propagate along the surface of the material, and the trajectory of the wave particles is elliptical, which is suitable for monitoring changes in surface conditions. Different types of surface acoustic waves have different adaptability to different stress fields and sensor designs.

[0026] The propagation characteristics of surface acoustic waves reflect the state changes of the material surface, including stress, temperature changes, structural defects, etc. By monitoring the propagation speed, amplitude and phase of surface acoustic waves, stress changes and the health status of the structure can be effectively evaluated.

[0027] Propagation speed: The propagation speed of surface acoustic waves is affected by the elastic modulus, density and surface stress of the material. Stress changes (such as stretching, compression, bending, etc.) will cause changes in the elastic properties of the material, thereby affecting the propagation speed of the surface acoustic wave. Therefore, monitoring the changes in propagation speed can directly reflect the stress state of various parts of the wind turbine, especially the local stress concentration.

[0028] Relationship between stress and propagation velocity: In a tensile state, the stiffness of the material increases, and the propagation velocity of the surface acoustic wave increases; in a compressed state, the propagation velocity decreases. By comparing the propagation velocities at different locations or at different time points, changes in stress distribution can be evaluated and possible locations of fatigue damage can be identified.

[0029] Amplitude: Amplitude refers to the amplitude of the surface acoustic wave signal, and its changes can also reflect changes in the surface state of the material. Stress changes may cause changes in the surface roughness of the material, which in turn affects the attenuation and amplitude of the wave. With the accumulation of fatigue damage, defects such as microcracks, corrosion or fatigue pits may appear on the surface of the material, which will cause the attenuation of the surface acoustic wave to increase during propagation and reduce the amplitude.

[0030] The relationship between wave amplitude and material defects: When cracks or tiny physical damage appear on the surface of a material, the wave amplitude will usually be significantly weakened, or the signal will attenuate. Therefore, by analyzing the change in the wave amplitude of the surface acoustic wave, it is possible to infer fatigue damage or tiny cracks on the surface of the material.

[0031] Phase: Phase is an important parameter that describes the state of wave propagation. It is closely related to the wave propagation speed and amplitude. The change of phase reflects the relative delay or advance experienced by the wave during propagation. Phase change may be caused by the stress state of the material, temperature change or surface morphology change.

[0032] Relationship between phase and structural health: Since materials will deform slightly when subjected to external loads, this deformation will cause the propagation path of the surface acoustic wave to change, thereby changing the phase of the wave. If the phase changes abnormally, it may mean that the structure has undergone slight deformation, cracks or other damage. Therefore, by monitoring the phase change, important information about the health status of the structure can be provided.

[0033] Changes in the roughness of the fuselage surface are usually related to changes in the surface microstructure (such as cracks, corrosion, wear, etc.). Surface acoustic wave signals can propagate through the surface and are significantly affected by changes in the surface state, thereby reflecting changes in roughness.

[0034] Relationship between fuselage surface roughness and surface acoustic wave: Relationship between surface roughness and amplitude: When the surface roughness increases, the tiny irregularities on the surface (such as bumps or depressions) will cause the surface acoustic wave to scatter and attenuate during propagation, thus affecting the amplitude of the wave. A rougher surface will cause the wave to decay faster, so changes in the amplitude can directly reflect changes in surface roughness.

[0035] Relationship between surface roughness and phase: In addition to the amplitude, a rough surface may also cause changes in the wave propagation speed. Especially on surfaces with different roughness, the propagation speed of sound waves will be affected by tiny changes in surface morphology, which in turn leads to changes in the signal phase.

[0036] The process of obtaining surface roughness: The surface acoustic wave signal is stimulated by the sensor, and the echo signal is monitored in real time. The waveform information monitored will include the scattering and attenuation effects caused by the surface roughness. Use appropriate algorithms (such as time domain analysis, frequency domain analysis, etc.) to analyze the amplitude of the echo signal. The attenuation rate of the amplitude is proportional to the surface roughness. By comparing with the benchmark data, the size of the surface roughness can be estimated. By analyzing the propagation speed and phase change of the surface acoustic wave, the influence of the surface morphology on the propagation characteristics can be further confirmed, thereby inferring the surface roughness.

[0037] Based on the above analysis methods, mathematical models and machine learning algorithms can be used to fit the collected signals to quantitatively analyze the changes in surface roughness. For example, methods such as autoregressive moving average (ARMA) model, wavelet transform or Fourier transform can be used to extract features from time series signals and calculate surface roughness in combination with empirical models.

[0038] After analyzing the amplitude of the echo signal, the fuselage surface roughness anomaly index is generated. The method for obtaining the fuselage surface roughness anomaly index is as follows: The surface acoustic wave sensor is used to collect echo signals from key parts of the wind turbine. The echo signal can include multiple features (such as propagation speed, amplitude, phase, etc.). The echo signal is processed by denoising and filtering to eliminate environmental interference and noise.

[0039] Extract multiple features related to surface roughness from the echo signal. Common features include: Amplitude: reflects the influence of surface roughness. Rough surface will cause the amplitude to attenuate. Propagation Velocity: The propagation velocity of surface acoustic waves is affected by the surface state. The increase of roughness will change the propagation velocity. Phase Shift: Slight deformation, cracks or changes in roughness of the surface will affect the propagation phase of the wave. Organize these features into a data matrix, where each row represents the features of a moment or a measurement point, and each column represents different feature variables.

[0040] Assuming that m features (such as amplitude, propagation velocity, phase, etc.) are extracted from n measurement points, the data matrix is: in, Represents the mth eigenvalue of the nth measurement point. Standardized data PCA requires data to have the same dimension, so the data needs to be standardized. Each feature column is subtracted from its mean and divided by the standard deviation so that the mean of each feature is 0 and the standard deviation is 1.

[0041] The standardized data is used to calculate the covariance matrix. The covariance matrix describes the relationship between different features and reflects the linear dependence of the features: ; Where C is the covariance matrix, Z is the standardized data matrix, and T is the matrix transpose; Calculate the eigenvalues ​​and eigenvectors of the covariance matrix C. The eigenvalues ​​represent the variance of each principal component, and the eigenvectors represent the importance of each direction in the data. Through eigendecomposition, we get the eigenvalues and the eigenvector ; According to the size of the eigenvalue, select the principal components, that is, the first few principal components whose cumulative variance reaches 90%~95%. For example, if the eigenvalue of the kth principal component is large, it contributes more to the original data. Usually the first p principal components are selected. After selecting the principal components, project the data onto these principal components to obtain a new feature matrix: ;in, is a matrix containing the first p principal component eigenvectors, and Y is the feature matrix after dimensionality reduction.

[0042] The calculation of the body surface roughness anomaly index is as follows: ;in, represents the standard deviation of the ith principal component, p is the number of principal components selected, is the abnormal index of the fuselage surface roughness. The larger the abnormal index, the more significant the change in surface roughness.

[0043] The nonlinear stress-strain relationship of a material describes the nonlinear change between stress and strain of a material under the action of an external force. In practical applications, the nonlinear behavior of a material is particularly significant when it is affected by external loads (such as fatigue loads, impact, etc.).

[0044] Relationship between nonlinear stress-strain relationship and surface acoustic waves: Influence of nonlinear effect on propagation velocity: When a material experiences a large stress, the elastic modulus of the material will change as the stress increases. This change not only affects the stiffness of the material, but also directly affects the propagation velocity of the surface acoustic wave. The relationship between the propagation velocity of the surface acoustic wave and stress-strain shows a nonlinear characteristic, that is, as the stress increases, the change in the propagation velocity will also change nonlinearly.

[0045] Influence of nonlinear effects on wave amplitude: When a material undergoes a large strain, it may cause the formation of micro cracks or local deformation of the surface, thereby affecting the wave propagation process and causing nonlinear changes in the attenuation rate of the wave amplitude.

[0046] The process of obtaining the nonlinear stress-strain relationship: The propagation characteristics of the surface acoustic wave under different loads are obtained by real-time acquisition of the surface acoustic wave signal and combining the stress change and the strain state of the material. According to the change of the propagation velocity of the surface acoustic wave under different stress states, a relationship model between stress and propagation velocity is established. By comparing the propagation velocity data under different stresses, the nonlinear stress-strain relationship of the material under different loads can be obtained. The amplitude and phase can also reflect the nonlinear stress-strain behavior of the material. The nonlinear attenuation of the amplitude and the change of the phase can provide valuable information about the nonlinear strain state of the material.

[0047] By performing regression analysis on data such as propagation velocity, amplitude, and phase under different stress states, a nonlinear stress-strain relationship model can be established. Common modeling methods include using polynomial regression, neural network models, finite element simulation (FEM), and other methods. These models can capture the nonlinear behavior of materials under different stresses and fit them with experimental data. The accuracy of the model is verified through experimental tests. During the experiment, different loads can be applied and signals can be collected through surface acoustic wave sensors. The consistency of the experimental results with the model prediction results can be compared to further optimize the stress-strain relationship model.

[0048] In order to accurately extract the parameters of the surface roughness and nonlinear stress-strain relationship of the fuselage, data analysis is the key link. Common data processing methods include: time domain and frequency domain analysis: Process the signal in the time domain and frequency domain to extract the delay, frequency component and amplitude information of the waveform. This information reflects the changes in the surface and stress state of the material. Pattern recognition algorithm: Through machine learning and pattern recognition algorithms, such as support vector machines (SVM), decision trees, neural networks, etc., features are automatically extracted from the collected signals to predict the roughness and stress-strain relationship. Signal filtering and denoising: Apply filtering technology to remove noise and improve the accuracy and reliability of the signal. Common filtering methods include Kalman filtering, wavelet transform, etc.

[0049] After comparing the propagation velocity data under different stresses, the nonlinear effect drift index is generated. The method for obtaining the nonlinear effect drift index is as follows: collect the propagation velocity data under different stresses and obtain a set of ,in is stress, is the propagation velocity at the wth moment, select the nonlinear model, and set the relationship between propagation velocity and stress to a power law relationship: ; v represents the propagation velocity, E represents the stress, and a is a constant that represents the proportional relationship between stress E and propagation velocity v. When stress E is equal to 1, the propagation velocity v is equal to the value of a. b is the power exponent that determines the degree of influence of stress E on the change of propagation velocity v. The least squares method is used to estimate the model parameters a and b by minimizing the error between the model output and the actual data. The nonlinear effect drift index is calculated by fitting the parameters. The expression is: in, represents the propagation speed at the i-th moment, is the predicted value obtained by fitting the model, w is the total number of data points, and QER is the nonlinear effect drift index.

[0050] S4: Analyze and calculate the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material to obtain the measurement accuracy value of the wind turbine monitoring sensor, and judge the sensor monitoring status of the fuselage part based on the accuracy value, and generate an early warning signal to prompt the unit maintenance personnel.

[0051] The fuselage surface roughness anomaly index and nonlinear effect drift index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts the measurement accuracy value label of the wind turbine monitoring sensor with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of the prediction errors of the measurement accuracy value labels of all wind turbine monitoring sensors as the training target. The machine learning model is trained until the sum of the prediction errors converges, and the model training is stopped. The measurement accuracy value of the wind turbine monitoring sensor is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0052] The method for obtaining the measurement accuracy value of the wind turbine monitoring sensor is as follows: from the comprehensive feature vector training data of the trained machine learning model, the corresponding function expression is obtained: ; In the formula, is the output function of the model, is the abnormal roughness index of the fuselage surface, QER is the nonlinear effect drift index, The measurement accuracy value of the wind turbine monitoring sensor.

[0053] The obtained measurement accuracy value of the wind turbine monitoring sensor is compared with the pre-set measurement accuracy reference threshold. If the measurement accuracy value of the wind turbine monitoring sensor is greater than or equal to the pre-set measurement accuracy reference threshold, it means that the measurement accuracy of the wind turbine monitoring sensor is high, and no warning signal is generated at this time; if the measurement accuracy value of the wind turbine monitoring sensor is less than the pre-set measurement accuracy reference threshold, it means that the measurement accuracy of the wind turbine monitoring sensor is low, and a warning signal is generated at this time to remind the unit maintenance personnel to check or replace the sensor.

[0054] S5: According to the analysis results and historical data, the working state of the surface acoustic wave sensor is adjusted to perform dynamic correction, wherein the correction step includes adjusting the working frequency and gain of the sensor to ensure the long-term stability of the sensor's measurement accuracy.

[0055] In order to ensure the stability of the measurement accuracy of the surface acoustic wave sensor during long-term use, dynamic correction is performed through a fuzzy logic control system in combination with the analysis results and historical data. The specific steps are as follows: Prepare input data items based on real-time monitoring data and historical records. Input items include: Measurement accuracy value: The current measurement accuracy value of the wind turbine monitoring sensor predicted by the polynomial regression model represents the accuracy of the sensor in the current working state.

[0056] Historical data: including historical measurement accuracy values, historical sensor operating frequencies, and gain settings. These data help determine the long-term stability changes of the sensor.

[0057] These data will serve as input to the fuzzy logic system and influence the decision of sensor calibration.

[0058] The fuzzy logic system is a control method based on fuzzy sets and fuzzy reasoning rules, which simulates the human decision-making process through fuzzy rules and membership functions. In this system, both the input and output items are fuzzy quantities, and the corresponding output values ​​are obtained through reasoning.

[0059] Input: Fuzzification of measurement accuracy values: According to the measurement accuracy values, they are divided into fuzzy sets such as "high accuracy", "medium accuracy", and "low accuracy", and membership functions are given (for example, triangle, trapezoid, etc. fuzzy membership functions). Historical data: The measurement error of the sensor in the historical data also needs to be fuzzified. The measurement error is the difference between the actual measurement value of the sensor and the standard measurement value.

[0060] Output: Operating frequency adjustment: Dynamically adjust the operating frequency of the SAW sensor based on the current measurement accuracy and historical data. Usually, changes in the operating frequency affect the propagation speed, amplitude, and phase characteristics, thereby improving the measurement accuracy. Gain adjustment: Gain control determines the degree of amplification of the sensor to the signal. If the measurement accuracy is low, it may be necessary to adjust the gain to improve the sensor's responsiveness and sensitivity to the signal.

[0061] Design a series of fuzzy rules and use fuzzy reasoning to decide how to adjust the operating frequency and gain of the sensor. For example: Rule 1: If the measurement accuracy is "low accuracy" and the sensor measurement error in the historical data continues to increase, increase the gain and adjust the frequency to medium and high frequencies to improve the sensor's responsiveness.

[0062] Rule 2: If the measurement accuracy is "high accuracy" and the sensor measurement error is stable, maintain the current gain and keep the frequency unchanged to avoid over-adjustment.

[0063] Rule 3: If the measurement accuracy is "medium accuracy" and the sensor measurement error is small, increase the gain appropriately and adjust the frequency according to the current environmental conditions (for example, the frequency may need to be reduced at high temperatures to reduce the impact of noise).

[0064] Through the fuzzy reasoning process, the membership of each input item is combined to generate the corresponding output. For example: The input terms "low precision" and "performance degradation" may lead to "higher frequency" and "higher gain" as outputs through rules.

[0065] The frequency and gain adjustment results of the fuzzy inference output need to be converted into actual control values ​​through a "defuzzification" method (such as the centroid method and the maximum membership method).

[0066] According to the output results obtained by fuzzy reasoning, adjust the operating frequency and gain of the sensor: Working frequency adjustment: According to the output value, adjust the working frequency of the sensor to adapt to the current environmental factors such as stress and temperature of the fuselage. Gain adjustment: Adjust the gain of the sensor through the electronic control interface to improve the signal response and ensure the measurement accuracy.

[0067] Once the frequency and gain adjustments are complete, restart the sensor and monitor it to assess whether its measurement accuracy has reached the expected target. If it still has not reached the expected accuracy, continue to perform dynamic corrections based on new measurement results until the measurement accuracy stabilizes.

[0068] Through continuous monitoring and feedback mechanisms, the system can adjust the working state of the sensor in real time. As new historical data accumulates, the fuzzy logic rule base may self-optimize and adjust more appropriate control strategies to further improve the long-term stability and measurement accuracy of the sensor.

[0069] Example 2, please refer to Figure 2 As shown, a wireless passive surface acoustic wave bolt stress monitoring system described in this embodiment includes a sensor configuration module, a signal excitation module, a data processing module, a measurement accuracy calculation module and an adaptive adjustment module; Sensor configuration module: installing at least one wireless passive surface acoustic wave sensor at key parts of the wind turbine, including the main bearing, gearbox, generator, blade root, blade bearing and tower connection; Signal excitation module: excites the surface acoustic wave sensor to generate a surface acoustic wave signal, and monitors the propagation characteristics of the surface acoustic wave through the sensor, wherein the propagation characteristics include propagation speed, amplitude and phase; Data processing module: collects the surface acoustic wave signal in real time, and obtains parameters reflecting the stress state of the fuselage according to the change of the surface acoustic wave signal, wherein the parameters include the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material; Measurement accuracy calculation module: Analyze and calculate the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material to obtain the measurement accuracy value of the wind turbine monitoring sensor, and judge the monitoring status of the sensor in the fuselage according to the accuracy value, and generate an early warning signal to prompt the maintenance personnel of the unit; Adaptive adjustment module: according to the analysis results and historical data, the working state of the surface acoustic wave sensor is adjusted to perform dynamic correction. The correction step includes adjusting the working frequency and gain of the sensor to ensure the long-term stability of the sensor's measurement accuracy.

[0070] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0071] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0072] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0073] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A wireless passive surface acoustic wave bolt stress monitoring method, characterized in that: The following steps are involved: S1: Install several wireless passive surface acoustic wave sensors at key locations of the wind turbine, including the main bearing, gearbox, generator, blade root, blade bearing and tower connection; S2: exciting the surface acoustic wave sensor to generate a surface acoustic wave signal, and monitoring the propagation characteristics of the surface acoustic wave, including propagation speed, amplitude and phase, through the surface acoustic wave sensor; S3: collecting the surface acoustic wave signal in real time, and obtaining the surface roughness of the fuselage reflecting the stress state of the fuselage and the nonlinear stress-strain relationship of the material according to the change of the surface acoustic wave signal; S4: Analyze the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material, calculate the measurement accuracy value of the wind turbine monitoring sensor through the machine learning model, and judge the sensor monitoring status of the wind turbine fuselage according to the accuracy value; S5: According to the calculation results and historical data, the working state of the surface acoustic wave sensor is adjusted to perform dynamic correction, wherein the correction step includes adjusting the working frequency and gain of the sensor to ensure the long-term stability of the sensor's measurement accuracy.

2. A wireless passive surface acoustic wave bolt stress monitoring method according to claim 1, characterized in that: Wireless passive surface acoustic wave sensors are used to collect echo signals from key parts of wind turbines, and multiple features related to surface roughness are extracted from the echo signals, including the amplitude, phase and propagation velocity of the echo signals. The surface roughness anomaly index of the wind turbine body is calculated by the covariance matrix analysis method, which is specifically: The echo signals of the key parts of the wind turbine are collected by using surface acoustic wave sensors. Multiple features related to surface roughness are extracted from the echo signals. These features are organized into a data matrix. Each row represents the features of a moment or a measurement point, and each column represents different characteristic variables. The eigenvalues ​​and eigenvectors of the covariance matrix are calculated. The eigenvalues ​​represent the variance of each principal component. The eigenvalues ​​and eigenvectors are obtained through eigendecomposition. The principal components are selected according to the size of the eigenvalues. After the principal components are selected, the data are projected onto the principal components to obtain a new characteristic matrix: the surface roughness anomaly index of the body is calculated, and the expression is: ;in, represents the standard deviation of the ith principal component, p is the number of principal components selected, m is the number of features in the matrix, It is the abnormality index of the fuselage surface roughness.

3. A wireless passive surface acoustic wave bolt stress monitoring method according to claim 2, characterized in that: In S3, the nonlinear effect drift index is generated after comparing the propagation velocity data under different stresses. The propagation velocity data is recorded at different measurement points and time points. The nonlinear effect drift index is obtained by collecting the propagation velocity data under different stresses, selecting a nonlinear model, and setting the relationship between the propagation velocity and the stress to be a power law relationship: ; v represents the propagation velocity, E represents the stress, a is a constant, and b is the power exponent. The least squares method is used to estimate the model parameters a and b by minimizing the error between the model output and the actual data. The nonlinear effect drift index is calculated by fitting the parameters. The expression is: in, represents the propagation speed at the i-th moment, is the predicted value obtained by fitting the model, w is the total number of data points, and QER is the nonlinear effect drift index.

4. A wireless passive surface acoustic wave bolt stress monitoring method according to claim 3, characterized in that: In S4, the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material are analyzed and calculated to obtain the measurement accuracy value of the wind turbine monitoring sensor, which is specifically: The fuselage surface roughness anomaly index and nonlinear effect drift index are converted into comprehensive feature vectors, which are used as input of the machine learning model. The machine learning model is trained, and the measurement accuracy value of the wind turbine monitoring sensor is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

5. A wireless passive surface acoustic wave bolt stress monitoring method according to claim 4, characterized in that: According to the accuracy value, the sensor monitoring status of the fuselage is judged, specifically: the obtained measurement accuracy value of the wind turbine monitoring sensor is compared with a preset measurement accuracy reference threshold; if the measurement accuracy value of the wind turbine monitoring sensor is greater than or equal to the preset measurement accuracy reference threshold, it means that the measurement accuracy of the wind turbine monitoring sensor is high, and no warning signal is generated at this time; if the measurement accuracy value of the wind turbine monitoring sensor is less than the preset measurement accuracy reference threshold, it means that the measurement accuracy of the wind turbine monitoring sensor is low, and a warning signal is generated at this time to remind the unit maintenance personnel to check or replace the sensor.

6. A wireless passive surface acoustic wave bolt stress monitoring method according to claim 5, characterized in that: In S5, according to the analysis results and historical data, the working state of the surface acoustic wave sensor is adjusted to perform dynamic correction, and the correction step includes adjusting the working frequency and gain of the sensor, specifically: The measurement accuracy and measurement error of wind turbine monitoring sensors are used as input items of fuzzy logic and divided into different fuzzy sets. The sensor's operating frequency and gain adjustment results are used as output items of fuzzy logic and divided into different fuzzy sets; Formulate fuzzy rules to describe the impact of wind turbine monitoring sensor measurement accuracy and historical data on sensor operating frequency and gain; Perform fuzzy reasoning according to fuzzy rules and adjust the sensor operating frequency and gain.

7. A wireless passive surface acoustic wave bolt stress monitoring system, used to implement a wireless passive surface acoustic wave bolt stress monitoring method according to any one of claims 1 to 6, characterized in that: It includes a sensor configuration module, a signal excitation module, a data processing module, a measurement accuracy calculation module and an adaptive adjustment module; Sensor configuration module: installing at least one wireless passive surface acoustic wave sensor at key parts of the wind turbine, including the main bearing, gearbox, generator, blade root, blade bearing and tower connection; Signal excitation module: excites the surface acoustic wave sensor to generate a surface acoustic wave signal, and monitors the propagation characteristics of the surface acoustic wave through the sensor, wherein the propagation characteristics include propagation speed, amplitude and phase; Data processing module: collects the surface acoustic wave signal in real time, and obtains parameters reflecting the stress state of the fuselage according to the change of the surface acoustic wave signal, wherein the parameters include the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material; Measurement accuracy calculation module: Analyze and calculate the surface roughness of the fuselage and the nonlinear stress-strain relationship of the material to obtain the measurement accuracy value of the wind turbine monitoring sensor, and judge the sensor monitoring status of the fuselage part according to the accuracy value; Adaptive adjustment module: according to the analysis results and historical data, the working state of the surface acoustic wave sensor is adjusted to perform dynamic correction. The correction step includes adjusting the working frequency and gain of the sensor to ensure the long-term stability of the sensor's measurement accuracy.

Images