Wind turbine blade fatigue life prediction platform and prediction method

Abstract

The invention provides a wind turbine blade fatigue life prediction platform and method, and relates to the technical field of fatigue life prediction.The wind turbine blade fatigue life prediction method comprises the steps that firstly, a geometric model and an experimental test piece are constructed according to design parameters, the limit fatigue cycle number under different stresses is obtained through a fatigue performance experiment, and an S-N curve is drawn and used for calibrating the geometric model; secondly, stress values of the wind turbine blade in historical operation are collected, and the current damage condition is analyzed so as to preliminarily evaluate whether the residual fatigue life exists or not; a fault prediction model is constructed by using historical operation data, so that the fault probability is predicted. And performing fault prediction on the operation data at the current moment through the fault prediction model, and analyzing the influence of the environmental data on the fatigue life. And finally, calculating the residual fatigue life by comprehensively analyzing the damage condition, the fault probability and the environmental influence, carrying out state judgment, and sending out safety or danger early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of fatigue life prediction, and in particular to a wind turbine blade fatigue life prediction platform and prediction method. Background Art

[0002] With the rapid development of renewable energy, wind energy has been widely used as an important clean energy source. However, as a key component in wind power systems that withstand wind loads and environmental influences, the prediction and management of fatigue life of wind turbine blades has become increasingly important. Traditional wind turbine blade fatigue analysis methods mostly rely on static tests and empirical formulas, lacking real-time monitoring and dynamic evaluation, making it difficult to accurately predict the fatigue state of blades in actual operation, which may lead to failures and safety hazards, affecting the stability and economic benefits of wind power systems.

[0003] In recent years, with the development of computer technology and the application of advanced tools such as finite element analysis and computational fluid dynamics, blade design and fatigue analysis have become more precise. However, how to combine existing technologies to achieve real-time monitoring and comprehensive evaluation of multiple variables in wind turbine blades during actual operation, especially in dealing with complex wind speed changes and fatigue behavior under different environmental conditions, remains a challenge.

[0004] As a crucial component of wind power systems, wind turbine blades are exposed to complex conditions over long periods of time and are susceptible to fatigue damage. Traditional fatigue analysis methods typically rely on static testing and empirical formulas. In practical applications, these methods often fail to fully account for the impact of dynamic load variations, environmental factors, and material properties on blade fatigue behavior. Consequently, the lack of real-time monitoring and dynamic assessment methods leads to significant uncertainty in fatigue life predictions for wind turbine blades, potentially leading to unexpected failures and downtime, impacting the economic benefits and safe operation of wind farms. Summary of the Invention

[0005] The purpose of the present invention is to provide a wind turbine blade fatigue life prediction platform and prediction method to solve the problems raised in the above background technology.

[0006] A wind turbine blade fatigue life prediction platform, comprising:

[0007] The finite element analysis module is used to construct the geometric model and experimental specimens of the wind turbine blades based on the design parameters, conduct fatigue performance tests on the experimental specimens to obtain the limit number of fatigue cycles under different stresses, draw the SN curve, perform finite element analysis on the geometric model, and calibrate the geometric model based on the SN curve;

[0008] The fatigue damage analysis module is used to collect the stress values ​​of wind turbine blades during historical operation, obtain the number of fatigue cycles of wind turbine blades under the stress values ​​during historical operation through finite element analysis, and analyze the damage status at the current moment based on the number of fatigue cycles;

[0009] The fault probability model building module is used to obtain the historical operating data of wind turbine blades and assign corresponding fault probabilities through expert scoring. The module uses the historical operating data of wind turbine blades as input and the corresponding fault probabilities as labels to build a fault prediction model based on the machine learning model.

[0010] A real-time fault analysis module collects the current operating data of the wind turbine blades and the environmental data of their location, inputs the operating data into the trained fault prediction model, obtains the current fault probability, and analyzes the impact of the current environmental data on the fatigue life of the wind turbine blades;

[0011] The fatigue life prediction model is used to analyze the remaining fatigue life of the wind turbine blades at the current moment based on the damage status of the wind turbine blades, the failure probability and the impact of environmental data on the fatigue life of the wind turbine blades at the current moment, and to determine whether to issue an early warning for the remaining fatigue life at the current moment.

[0012] Furthermore, performing finite element analysis on the collective model specifically includes:

[0013] Several wind turbine blade specimens were made according to the design parameters of the wind turbine blades. Fatigue performance tests were carried out on the wind turbine blade specimens. The blade root, blade tip and blade middle were set as observation points. Different stress levels were applied to the observation points of each experimental model. The limit fatigue cycle number and stress value of the observation point at each stress level were recorded. The limit fatigue cycle number and stress value were fitted to obtain the SN curve. The fitted SN curve equation is expressed as:

[0014] log 10 (N f )=ab·log 10 (σ)

[0015] Among them, N f is the number of ultimate fatigue cycles, σ is the stress value, a and b are fitting parameters respectively;

[0016] The design parameters include geometric data and material properties of the wind turbine blade, the geometric data including the length, width, thickness, cross-sectional shape of the wind turbine blade, and curvature and camber of the wind turbine blade, the material properties including the material used for the wind turbine blade and mechanical properties, the mechanical properties including elastic modulus, Poisson's ratio, shear modulus, tensile strength and compressive strength; creating a geometric model of the blade based on the geometric data of the wind turbine blade, inputting material properties, defining the boundary conditions of the root of the wind turbine blade as fixed, determining the grid type, applying loads to observation points of the geometric model, performing fatigue analysis and recording the number of fatigue cycles at the observation points output by the finite element analysis, comparing the results output by the finite element analysis with the SN curve, calculating the absolute difference between the fatigue life in the SN curve of each observation point and the limit fatigue cycle number output by the finite element analysis, and considering that the geometric model calibration is completed when the absolute difference is less than the error threshold;

[0017] The calculation formula for converting the limit fatigue cycle number into the initial fatigue life is:

[0018]

[0019] Among them, E life is the initial fatigue life, where the initial fatigue life is in hours, N f is the limit fatigue cycle number, and f is the operating frequency.

[0020] Furthermore, analyzing the damage condition of the wind turbine blade at the current moment specifically includes:

[0021] The stress values ​​of the wind turbine blades in the historical operation are collected. The stress values ​​carried in each historical operation are used as input and the finite element analysis is performed on the geometric model. The output number of fatigue cycles is obtained and the damage degree in each historical operation is calculated. The calculation formula is:

[0022]

[0023] Where D(i) is the damage degree under the i-th stress level in the historical operation, N(i) is the number of cycles under the i-th stress level in the historical operation, and N f (i) is the limit fatigue cycle number in the SN curve corresponding to the stress value of the wind turbine blade under the i-th stress level, and i is the index of different stress levels;

[0024] The damage degree under each stress level is accumulated, and the formula is as follows:

[0025]

[0026] Where D is the total damage degree at the current moment, n is the type of stress level, and i∈[1,n];

[0027] If D<1, it indicates that the wind turbine blade still has remaining fatigue life. If D>1, the stress conditions in the historical operation have caused fatigue damage to the wind turbine blade. The fatigue life of the wind turbine blade is determined to be 0, and a fatigue life danger warning is issued.

[0028] Furthermore, constructing the fault prediction model specifically includes:

[0029] The historical operation data includes stress, vibration and blade temperature. The mean of the stress, vibration and blade temperature data at the observation point of the wind turbine blade is used as the stress, vibration and blade temperature data of the wind turbine blade. The historical operation data is analyzed by an expert scoring method to assign corresponding failure probabilities. The historical operation data and the corresponding failure probabilities are integrated into a sample set, and the sample set is divided into a training set and a test set with a division ratio of 7:3. The historical operation data is used as input and trained based on the machine learning algorithm model. The corresponding failure probability is used as a label, and the mean square error is selected as the loss function. When the loss function is lower than the set threshold, it indicates that the model training is completed. The threshold can be set to 0.01. After the training is completed, the historical operation data in the test set is input into the trained fault prediction model, and the mean absolute error and determination coefficient indicators are used to evaluate the model performance.

[0030] Furthermore, the analysis of the current operating data and environmental data specifically includes:

[0031] Obtain the stress, vibration, and blade temperature of the observation point at the current moment, calculate the mean value as the stress, vibration, and blade temperature of the wind turbine blade at the current moment, obtain environmental data at the location of the wind turbine blade through the weather station, input the operating data into the trained fault prediction model, and obtain the fault probability predicted by the model;

[0032] Analyze the impact of environmental data on the fatigue life of wind turbine blades, and calculate the environmental impact index using the following formula:

[0033]

[0034] Among them, FEI is the environmental impact index, T en 、H c 、W c are temperature, humidity and wind speed data respectively, T r 、H r 、W r are the suitable values ​​of temperature, humidity and wind speed respectively, and β1, β2 and β3 are the influence coefficients of each item respectively.

[0035] Furthermore, analyzing the remaining fatigue life of the wind turbine blade at the current moment specifically includes:

[0036] Comprehensively analyze the damage condition, failure probability and environmental data of the wind turbine blades at the current moment on the fatigue life of the wind turbine blades to obtain the remaining fatigue life at the current moment. The calculation formula is:

[0037]

[0038] Among them, R life is the remaining fatigue life, E life is the initial fatigue life, D is the total damage degree, FEI is the environmental impact index, CFP is the failure probability, λ1, λ2, and λ3 are the weight coefficients of the corresponding items, 0<λ2<λ3<λ1<1;

[0039] The remaining fatigue life at the current moment is judged. If the remaining fatigue life at the current moment is higher than the life threshold, it is judged that the wind turbine blades at the current moment are in a safe state, and the wind turbine blades continue to be monitored. If the remaining fatigue life at the current moment is lower than the life threshold, it is judged that the wind turbine blades at the current moment are in a fatigue damage state, and a fatigue life danger warning is issued.

[0040] The present invention further provides a method for predicting the fatigue life of a wind turbine blade. The method is executed by the above-mentioned wind turbine blade fatigue life prediction platform, and the specific steps include:

[0041] Step 1: Construct a geometric model and experimental specimen of the wind turbine blade based on the design parameters. Perform fatigue performance tests on the experimental specimen to obtain the limit number of fatigue cycles under different stresses, draw the SN curve, perform finite element analysis on the geometric model, and calibrate the geometric model based on the SN curve.

[0042] Step 2: Collect the stress values ​​of the wind turbine blades during historical operation, obtain the number of fatigue cycles of the wind turbine blades under the stress values ​​during historical operation through finite element analysis, and analyze the damage status at the current moment based on the number of fatigue cycles;

[0043] Step 3: Obtain historical operating data of wind turbine blades and assign corresponding failure probabilities using an expert scoring method. Using the historical operating data of wind turbine blades as input and the corresponding failure probabilities as labels, a fault prediction model is constructed based on a machine learning model.

[0044] Step 4: Collect the current operating data of the wind turbine blades and the environmental data of their location. Input the operating data into the trained fault prediction model to obtain the current failure probability and analyze the impact of the current environmental data on the fatigue life of the wind turbine blades.

[0045] Step 5: Based on the damage condition of the wind turbine blades at the current moment, the failure probability and the impact of environmental data on the fatigue life of the wind turbine blades, the remaining fatigue life of the wind turbine blades at the current moment is analyzed, and it is determined whether to issue an early warning for the remaining fatigue life at the current moment.

[0046] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0047] This solution combines finite element analysis, fatigue damage analysis, and machine learning models to assess the fatigue state and remaining life of wind turbine blades in real time. This platform not only systematically analyzes design and material parameters to optimize blade design, but also intelligently predicts failure probability by learning from historical data. This approach significantly improves wind turbine monitoring accuracy and response speed, effectively reducing the risk of failure due to fatigue damage, providing reliable assurance for the safe operation of wind power generation, and thus enhancing the overall economic benefits and sustainability of wind power generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Schematic diagram of the platform structure of the present invention;

[0049] Figure 2 Schematic diagram of the overall method flow of the present invention;

[0050] Figure 3 This is a relationship diagram between temperature and environmental impact index of the present invention;

[0051] Figure 4 This is a relationship diagram between humidity and environmental impact index of the present invention;

[0052] Figure 5 This is a relationship diagram between wind speed and environmental impact index of the present invention. DETAILED DESCRIPTION

[0053] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to specific embodiments.

[0054] It should be noted that, unless otherwise defined, the technical or scientific terms used in the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0055] Example:

[0056] See also Figure 1 The present invention provides a wind turbine blade fatigue life prediction platform, comprising:

[0057] The finite element analysis module is used to construct the geometric model and experimental specimens of the wind turbine blades based on the design parameters, conduct fatigue performance tests on the experimental specimens to obtain the limit number of fatigue cycles under different stresses, draw the SN curve, perform finite element analysis on the geometric model, and calibrate the geometric model based on the SN curve;

[0058] In this embodiment, performing finite element analysis on the collective model specifically includes:

[0059] Several wind turbine blade specimens were made according to the design parameters of the wind turbine blades. Fatigue performance tests were carried out on the wind turbine blade specimens. The blade root, blade tip and blade middle were set as observation points. Different stress levels were applied to the observation points of each experimental model. The limit fatigue cycle number and stress value of the observation point at each stress level were recorded. The limit fatigue cycle number and stress value were fitted to obtain the SN curve. The fitted SN curve equation is expressed as:

[0060] log 10 (N f )=ab·log 10 (σ)

[0061] Among them, N f is the limit fatigue cycle number, where the limit fatigue cycle number refers to the maximum fatigue cycle number that the wind turbine blade can withstand under the corresponding stress value, σ is the stress value, a and b are fitting parameters, and the least square method can be used for fitting;

[0062] The design parameters include geometric data and material properties of the wind turbine blade, the geometric data including the length, width, thickness, cross-sectional shape of the wind turbine blade (such as NACA series airfoil) and curvature and camber of the wind turbine blade, the material properties including the material used for the wind turbine blade (such as glass fiber, carbon fiber, composite material, etc.) and mechanical properties, the mechanical properties including elastic modulus, Poisson's ratio, shear modulus, tensile strength and compressive strength; creating a geometric model of the blade based on the geometric data of the wind turbine blade, inputting material properties, defining the boundary condition of the root of the wind turbine blade as fixed, determining the grid type, applying loads to the observation points of the geometric model, performing fatigue analysis and recording the number of fatigue cycles at the observation points output by the finite element analysis, comparing the results output by the finite element analysis with the SN curve, calculating the absolute difference between the fatigue life in the SN curve of each observation point and the limit fatigue cycle number output by the finite element analysis, and considering that the geometric model calibration is completed when the absolute difference is less than the error threshold;

[0063] The calculation formula for converting the limit fatigue cycle number into the initial fatigue life is:

[0064]

[0065] Among them, E life is the initial fatigue life, where the initial fatigue life is in hours, N f is the limit fatigue cycle number, f is the operating frequency, here refers to the rotation frequency of the wind turbine.

[0066] By constructing the geometric model and experimental specimens of wind turbine blades based on design parameters, this module can ensure the accuracy of the model. The input of detailed geometric data and material properties enables the model to truly reflect the stress conditions of the blades under actual working conditions, thereby improving the reliability of the analysis. The recording and analysis of the number of ultimate fatigue cycles under different stress levels in the module enables researchers to fully understand the fatigue characteristics of materials in practical applications. By setting the root, top and middle of the blade as observation points, the stress response of different parts can be obtained, further enriching the understanding of fatigue performance. By fitting the experimental data, the SN curve model is formed, which provides theoretical support for subsequent fatigue life prediction work. The SN curve not only provides the necessary parameters for fatigue analysis, but also provides a convenient calculation method for practical engineering applications in the form of mathematical models.

[0067] The fatigue damage analysis module is used to collect the stress values ​​of wind turbine blades during historical operation, obtain the number of fatigue cycles of wind turbine blades under the stress values ​​during historical operation through finite element analysis, and analyze the damage status at the current moment based on the number of fatigue cycles;

[0068] In this embodiment, analyzing the damage condition of the wind turbine blade at the current moment specifically includes:

[0069] The stress values ​​of the wind turbine blades in the historical operation are collected. The stress values ​​carried in each historical operation are used as input and the finite element analysis is performed on the geometric model. The output number of fatigue cycles is obtained and the damage degree in each historical operation is calculated. The calculation formula is:

[0070]

[0071] Where D(i) is the damage degree under the i-th stress level in the historical operation, N(i) is the number of cycles under the i-th stress level in the historical operation, and N f (i) is the limit fatigue cycle number in the SN curve corresponding to the stress value of the wind turbine blade under the i-th stress level, and i is the index of different stress levels;

[0072] The damage degree under each stress level is accumulated, and the formula is as follows:

[0073]

[0074] Where D is the total damage degree at the current moment, n is the type of stress level, and i∈[1,n];

[0075] If D<1, it indicates that the wind turbine blade still has remaining fatigue life. If D>1, the stress conditions in the historical operation have caused fatigue damage to the wind turbine blade. The fatigue life of the wind turbine blade is determined to be 0, and a fatigue life danger warning is issued.

[0076] By collecting stress values ​​from historical wind turbine blade operations, this module dynamically reflects the loads experienced by blades in actual operation. This historical data-based analysis method makes damage degree calculation more meaningful and accurately reflects the blade's current fatigue state. The damage degree calculation formula intuitively represents the degree of fatigue damage to the blade under specific stress conditions. This proportional relationship accurately reflects the fatigue performance of the material under different loads. By presetting the total damage degree from historical operations, it is possible to directly determine whether fatigue failure has occurred during historical operation. If so, it indicates that the wind turbine blade has reached the limit of fatigue cycles during historical operation. Fatigue failure is considered to have occurred, and fatigue life under future stresses is no longer determined. Instead, a fatigue life risk warning is issued. This module enables real-time analysis and monitoring of wind turbine blade damage, providing timely safety warnings. If fatigue life is detected to have dropped to a dangerous level, prompt action can be taken, such as shutting down the unit for maintenance, to prevent a major failure, thereby improving wind turbine safety.

[0077] The fault probability model building module is used to obtain the historical operating data of wind turbine blades and assign corresponding fault probabilities through expert scoring. The module uses the historical operating data of wind turbine blades as input and the corresponding fault probabilities as labels to build a fault prediction model based on the machine learning model.

[0078] In this embodiment, constructing the fault prediction model specifically includes:

[0079] Stress sensors and vibration sensors are installed inside the center of the upper layer of the blade to obtain the stress and vibration data of the blade. The historical operation data includes stress, vibration and blade temperature. The mean of the stress, vibration and blade temperature data at the observation point of the wind turbine blade is used as the stress, vibration and blade temperature data of the wind turbine blade. The historical operation data is analyzed by expert scoring method to assign corresponding fault probabilities. The historical operation data and the corresponding fault probabilities are integrated into a sample set, and the sample set is divided into a training set and a test set with a division ratio of 7:3. The historical operation data is used as input and trained based on the machine learning algorithm model. The corresponding fault probability is used as the label, and the mean square error is selected as the loss function. When the loss function is lower than the set threshold, it indicates that the model training is completed. The threshold can be set to 0.01. After the training is completed, the historical operation data in the test set is input into the trained fault prediction model, and the mean absolute error and determination coefficient indicators are used to evaluate the model performance.

[0080] Machine learning algorithms are able to handle complex nonlinear relationships between input features and outputs. In wind turbine blade fault prediction, there may be complex interactions between factors that affect the occurrence of failures (such as stress, vibration, temperature, etc.). Machine learning models can capture these relationships by learning from historical data, thereby improving prediction accuracy. Through machine learning models, the system can continuously optimize and update itself based on new historical operating data. During the long-term operation of wind turbine blades, performance and failure modes may change over time and environmental changes. Machine learning models can quickly adapt to these changes and improve the accuracy of fault prediction. In addition, the model is suitable for processing large-scale, high-dimensional data sets, thereby making more accurate fault predictions.

[0081] Input data includes stress, vibration, and blade temperature, parameters that are closely related to the operating status of wind turbine blades. Stress and vibration directly affect blade fatigue performance and reliability, while temperature can affect the physical properties and fatigue life of the material. Therefore, historical operating data can comprehensively reflect the blade's operating environment and condition. Stress is the force per unit area and is typically caused by forces such as aerodynamics and gravity that act on wind turbine blades during operation. Wind speed and direction fluctuations, blade weight, and centrifugal forces caused by rotation all induce varying degrees of stress in the blades. High stress can lead to material fatigue, causing cracks, deformation, or ultimately failure. Repeated stress cycles (fatigue) accelerate fatigue damage in materials, increasing structural fragility. Therefore, continuous monitoring of stress levels is crucial for predicting blade failure. Vibration is the mechanical vibration of an object caused by external or internal forces. In wind turbine blades, vibration can arise from wind flow, blade imbalance, structural defects, or failures of other mechanical components. Excessive vibration can cause fatigue damage, loosening of joints, or even failure. Especially at the connection between the blade and the hub, vibration may accelerate the formation and expansion of fatigue cracks, leading to structural failure. Therefore, monitoring and analyzing vibration data can help to identify potential faults in a timely manner. Blade temperature refers to the temperature of the wind turbine blade body, which is usually affected by factors such as the external environment, the thermal conductivity characteristics of the blade material, and wind speed. Temperature changes can affect the mechanical properties of the blade material. For example, as the temperature rises, the strength of some materials may decrease. In addition, excessively high temperatures may cause the material to age, deform, or lose elasticity, thereby increasing the risk of failure. By monitoring the blade temperature, its operating condition and potential failure risks can be assessed. Therefore, using the above data as input can effectively determine whether there is a fault in the wind turbine blade, making the constructed failure probability prediction model more accurate.

[0082] The expert scoring method combines failure probability labels with machine learning models to reduce reliance on human experience. The machine learning model uses a data-driven approach to make predictions, reducing the errors and uncertainties associated with human judgment. The expert scoring method can set stress states to three levels: high, medium, and low; vibration states to three levels: normal, abnormal frequency, and severe vibration; and blade temperature states to normal, slightly high, and excessively high. A design scoring scale is then established, using a 1-to-5 rating system. 1 indicates extremely low risk (little to no probability of failure); 2 indicates low risk (occasionally possible failure); 3 indicates medium risk (some probability of failure); 4 indicates high risk (high probability of failure); and 5 indicates extremely high risk (almost certain failure). Experts first determine the corresponding states for stress, vibration, and blade temperature, then assign a design score based on these states. Finally, the stress, vibration, and blade temperature scores are normalized and weighted to obtain the final failure probability. Blade temperature is weighted the highest, stress the medium, and vibration the lowest. The basis for probability judgment is then given. For example, if the failure probability is within 0-0.2, it indicates low risk and the possibility of failure is small. If the failure probability is within 0.2-0.5, it indicates medium risk and attention should be paid to the wind turbine blades. If the failure probability is within 0.5-0.8, it indicates high risk and measures should be taken. If the failure probability is within 0.8-1, it indicates extremely high risk and failure is almost inevitable and needs to be handled immediately.

[0083] The performance of blade materials (such as composites) can significantly degrade in high-temperature environments. Increased temperatures affect material strength, stiffness, and fatigue life, so high temperatures are often considered a key factor in material failure. Temperature fluctuations cause thermal expansion and contraction, leading to thermal stresses. These thermal stresses can exacerbate existing material defects and cause crack propagation. High temperatures often accelerate material aging and corrosion, impacting long-term safety. Therefore, the impact of temperature is direct and significant. Stress is a significant factor in fatigue cracking, especially under repeated loading. High stress levels can easily cause fatigue damage in blade materials. While stress is a significant factor in wind turbine blade operation, its impact depends to some extent on other factors (such as temperature and environmental influences). Therefore, a medium weighting is assigned to reflect its importance while also acknowledging the interplay of other factors. Although vibration can also contribute to fatigue and damage, in many cases its impact can be relatively minor, especially under normal operating conditions. If a wind turbine is well designed and maintained, vibration problems are generally minimal and manageable. The impact of vibration is also related to its frequency and amplitude. Under normal operating conditions, wind turbine designs can generally withstand some vibration. Therefore, a low weighting is assigned to low vibration.

[0084] A real-time fault analysis module collects the current operating data of the wind turbine blades and the environmental data of their location, inputs the operating data into the trained fault prediction model, obtains the current fault probability, and analyzes the impact of the current environmental data on the fatigue life of the wind turbine blades;

[0085] In this embodiment, analyzing the current operating data and environmental data specifically includes:

[0086] Obtain the stress, vibration, and blade temperature of the observation point at the current moment, calculate the mean value as the stress, vibration, and blade temperature of the wind turbine blade at the current moment, obtain environmental data at the location of the wind turbine blade through the weather station, input the operating data into the trained fault prediction model, and obtain the fault probability predicted by the model;

[0087] Analyze the impact of environmental data on the fatigue life of wind turbine blades, and calculate the environmental impact index using the following formula:

[0088]

[0089] Among them, FEI is the environmental impact index, T en 、H c 、W c are temperature, humidity and wind speed data respectively, T r 、H r 、W r are the suitable values ​​of temperature, humidity and wind speed respectively, β1, β2 and β3 are the influence coefficients of each item respectively, 0<β3<β2<β1<1.

[0090] By calculating the FEI, the fatigue life of wind turbine blades under current environmental conditions can be promptly assessed and predicted. The independent variables in this formula are temperature, humidity, and wind speed. Ambient temperature affects the thermal expansion and contraction of materials, potentially leading to stress concentrations at joints. Furthermore, high temperatures can induce thermal aging of materials and accelerate fatigue damage. High humidity can cause moisture accumulation on the blade surface, increasing the risk of corrosion, especially if microcracks or defects are present. Furthermore, humidity fluctuations can affect the mechanical properties of blade materials, particularly composite materials, whose hygroscopicity can lead to performance degradation. Excessively high wind speeds can subject wind turbine blades to extreme loads, increasing stress and vibration. Furthermore, strong winds can cause unexpected vibration or distortion in the blades, increasing the risk of fatigue failure. Furthermore, changes in wind speed can cause unstable aerodynamic loads, leading to overall performance degradation.

[0091] When environmental conditions exceed the optimal value, wind turbine blades may face a higher fatigue risk. For example, high temperature may cause material performance degradation, excessively high or low humidity may affect the physical properties of the material, and extreme wind speeds may cause unbalanced loads and vibrations. FEI is proportional to the deviation from the optimal value. That is, if the actual environmental conditions (T en 、H c 、W c ) deviates from the appropriate value (T r 、H r 、W r ) increases, indicating a stronger negative impact of the environment on blade fatigue life. Therefore, the deviation of environmental conditions from the optimal value is positively correlated with the Environmental Impact Index. If actual environmental conditions are consistent with the optimal value, the Environmental Impact Index approaches 0, indicating no significant environmental impact on wind turbine blade fatigue life.

[0092] The optimum temperature is set at 25°C, the ideal operating temperature for wind turbine blade materials (typically composite materials) and ensuring stable mechanical properties. Temperature has the most significant impact on material properties. High temperatures can lead to decreased material strength and accelerated aging, directly impacting blade fatigue life. Therefore, it is reasonable to assign the highest weight to temperature. The optimum humidity is set at 50%. This level typically strikes a balance between material stability and structural safety. Excessively high or low humidity levels can have some impact on material properties. Humidity has a relatively minor impact on materials, primarily affecting their hygroscopicity and durability. While humidity changes can affect material physical properties, the impact is typically less significant than that of temperature, so it is assigned a medium weight. The optimum wind speed is set at 10 m / s. This is the optimal operating speed for wind turbines, ensuring efficient energy generation without excessive mechanical loads. The impact of wind speed is primarily reflected in mechanical loads and vibration. In a well-designed wind turbine, wind speeds within the normal range do not cause direct damage to the blades, so it is assigned the lowest weight.

[0093] In this embodiment, the optimum temperature value is set to 25°C, the optimum humidity value is set to 50%, the optimum wind speed value is set to 10m / s, the temperature weight coefficient is set to 0.5, the humidity weight coefficient is set to 0.3, and the wind speed weight coefficient is set to 0.1. The environmental data of nine groups of wind turbine blades are collected and analyzed. The specific measurement data are shown in the following table:

[0094]

[0095]

[0096] Reference Figures 3 to 5As can be seen from the data in the table above, as the temperature approaches the suitable value, the environmental impact index decreases. The collected sample temperature data changes from less than the suitable value to greater than the suitable value. Figure 3 It can also be seen that the trend of the environmental impact index is gradually decreasing and then gradually increasing, reflecting that the closer to the temperature suitable value, the smaller the environmental impact index, and the further away from the temperature suitable value, the larger the environmental impact index. Similarly, the collected sample humidity data is from less than the humidity suitable value to greater than the humidity suitable value, from Figure 4 It can also be seen that the trend of the environmental impact index is gradually decreasing and then gradually increasing, reflecting that the closer to the suitable humidity value, the smaller the environmental impact index, and the further away from the suitable humidity value, the larger the environmental impact index. Similarly, the collected sample wind speed data is from less than the suitable wind speed value to greater than the suitable wind speed value, from Figure 5 It can also be seen that the trend of the environmental impact index is gradually decreasing and then gradually increasing, reflecting that the closer to the suitable wind speed value, the smaller the environmental impact index, and the further away from the suitable wind speed value, the larger the environmental impact index.

[0097] The fatigue life prediction model is used to analyze the remaining fatigue life of the wind turbine blades at the current moment based on the damage status of the wind turbine blades, the failure probability and the impact of environmental data on the fatigue life of the wind turbine blades at the current moment, and to determine whether to issue an early warning for the remaining fatigue life at the current moment.

[0098] In this embodiment, analyzing the remaining fatigue life of the wind turbine blade at the current moment specifically includes:

[0099] Comprehensively analyze the damage condition, failure probability and environmental data of the wind turbine blades at the current moment on the fatigue life of the wind turbine blades to obtain the remaining fatigue life at the current moment. The calculation formula is:

[0100]

[0101] Among them, R life is the remaining fatigue life, E life is the initial fatigue life, D is the total damage degree, FEI is the environmental impact index, CFP is the failure probability, λ1, λ2, and λ3 are the weight coefficients of the corresponding items, 0<λ2<λ3<λ1<1;

[0102] The remaining fatigue life at the current moment is judged. If the remaining fatigue life at the current moment is higher than the life threshold, it is judged that the wind turbine blades at the current moment are in a safe state, and the wind turbine blades continue to be monitored. If the remaining fatigue life at the current moment is lower than the life threshold, it is judged that the wind turbine blades at the current moment are in a fatigue damage state, and a fatigue life danger warning is issued.

[0103] Remaining fatigue life refers to the estimated safe operating time or lifespan of a wind turbine blade under current conditions, after experiencing certain damage and environmental impacts. It is a key performance indicator used to determine whether a wind turbine blade requires repair or replacement. Remaining fatigue life can assess the current condition of wind turbine blades, ensuring maintenance before fatigue occurs, reducing the risk of failure, and ensuring the safe operation of wind farms. By analyzing remaining fatigue life, wind farm operators can develop appropriate maintenance plans, optimize resource allocation, and improve economic efficiency. Real-time monitoring and prediction can also identify potential problems in advance, reducing economic losses caused by equipment failure.

[0104] The total damage degree reflects the degree of accumulated damage suffered by the blade during use. The higher the damage degree, the shorter the remaining fatigue life of the blade. The environmental impact index (FEI) indicates the impact of current environmental conditions on the fatigue life of the blade. The more unfavorable the environmental factors (such as high temperature and high humidity), the higher the FEI value, which will further reduce the remaining fatigue life. The failure probability indicates the possibility of equipment failure. The higher the failure probability, the greater the risk, which reduces the remaining fatigue life. The increase in the total damage degree and the failure probability will directly reduce the remaining fatigue life in the form of exponential decay. The increase in the total damage degree is directly inversely correlated with the remaining fatigue life. The higher the damage degree, the shorter the remaining fatigue life. The environmental impact index affects the remaining fatigue life through a logarithmic function. As the environmental impact index increases, the log 10 An increase in the value of (1+λ2·FEI) reduces the remaining fatigue life, demonstrating the inverse effect of environmental factors on blade performance. An increase in the probability of failure also leads to a decrease in the remaining fatigue life, indicating that a greater risk of failure shortens the safe operating time. This term represents the reduction in fatigue life of wind turbine blades due to accumulated damage. As the damage degree D increases, the remaining fatigue life decays exponentially, indicating that the effect of each unit of damage on fatigue life is nonlinear. This exponential decay relationship is very consistent with material fatigue theory, and the accumulation of damage has an increasingly greater impact on the material's load-bearing capacity and life. Therefore, by The form of can accurately reflect the significant impact of the current damage on the remaining fatigue life. (1-log 10The term (1+λ2·FEI) quantifies the impact of environmental factors on the fatigue life of wind turbine blades. FEI, standing for Environmental Impact Index, reflects the effect of changes in environmental conditions such as temperature, humidity, and wind speed on blade performance. Expressing it in a logarithmic form allows the increase in environmental impact to taper off over a wide range, avoiding overly drastic effects. This logarithmic function linearizes the impact of environmental factors, particularly at low FEI values, where the environmental impact on fatigue life is minimal; at higher FEI values, the impact gradually increases. This more accurately reflects the variation in blade fatigue life under varying environmental conditions, providing a smoother adjustment mechanism. This term reflects the impact of failure probability on the remaining fatigue life of wind turbine blades. As the failure probability (CFP) increases, the risk of equipment failure in its current state increases, leading to a decrease in remaining fatigue life. Using an exponential decay formula to express the impact of failure probability on life, similar to the effect of damage severity, emphasizes the importance of failure risk in fatigue life. This formula effectively simulates the dramatic decline in actual equipment reliability and safety under high failure probability, enhancing the model's sensitivity to risk. The initial fatigue life provides a reference baseline, and subsequent correction factors are designed to reflect the impact of various factors in actual operation on this baseline, thereby adjusting the remaining fatigue life to a more realistic value. Multiplying these three correction factors together effectively integrates the combined effects of various influencing factors. The fatigue life of wind turbine blades is not simply a simple linear superposition, but rather the result of the interaction of multiple factors. Using a product formula more realistically reflects this complexity.

[0105] The total damage index is an important indicator for evaluating the current damage status of wind turbine blades. With increasing service life, blades accumulate damage, resulting in a reduction in fatigue life. Therefore, a higher weight for λ1 reflects the direct and significant impact of damage on residual fatigue life. Increased damage generally has a significant impact on structural safety. While environmental factors (such as temperature and humidity) have an important impact on blade performance, their impact is typically less significant than direct structural damage. Therefore, it is reasonable to set the weight for λ2 lower than both λ3 and λ1. The environmental impact index affects residual fatigue life through a logarithmic function, indicating that the impact of environmental factors is gradual rather than linear. The failure probability (CFP) represents the probability of equipment failure. While this factor can also significantly affect the residual fatigue life of a blade, its impact is generally considered a conditional risk assessment. Compared to the damage index, the impact of the failure probability is more random and indirect, so it is reasonable to set it lower than λ1 but higher than λ2. The probability of failure directly impacts safety, but its impact often complements the actual damage level.

[0106] Life thresholds can be set based on historical data, such as setting the average remaining fatigue life of wind turbine blades in historical data as the life threshold. By comparing the remaining fatigue life with the preset life threshold, it is possible to determine in real time whether the wind turbine blades are in a safe operating condition. If the remaining fatigue life is higher than the threshold, the system will prompt the equipment to continue safe operation; otherwise, a fatigue life danger warning will be issued, prompting timely maintenance or replacement. This judgment mechanism ensures the safety of the wind turbine and reduces the risk of accidents caused by equipment failure. By monitoring and predicting the remaining fatigue life in real time, wind power operators can more effectively formulate maintenance and overhaul plans. This predictive method can help identify potential failure points and take proactive measures, thereby reducing unplanned downtime and maintenance costs.

[0107] See also Figure 2 The present invention further provides a method for predicting the fatigue life of a wind turbine blade. The method is executed by the above-mentioned wind turbine blade fatigue life prediction platform, and the specific steps include:

[0108] Step 1: Construct a geometric model and experimental specimen of the wind turbine blade based on the design parameters. Perform fatigue performance tests on the experimental specimen to obtain the limit number of fatigue cycles under different stresses, draw the SN curve, perform finite element analysis on the geometric model, and calibrate the geometric model based on the SN curve.

[0109] Step 2: Collect the stress values ​​of the wind turbine blades during historical operation, obtain the number of fatigue cycles of the wind turbine blades under the stress values ​​during historical operation through finite element analysis, and analyze the damage status at the current moment based on the number of fatigue cycles;

[0110] Step 3: Obtain historical operating data of wind turbine blades and assign corresponding failure probabilities using an expert scoring method. Using the historical operating data of wind turbine blades as input and the corresponding failure probabilities as labels, a fault prediction model is constructed based on a machine learning model.

[0111] Step 4: Collect the current operating data of the wind turbine blades and the environmental data of their location. Input the operating data into the trained fault prediction model to obtain the current failure probability and analyze the impact of the current environmental data on the fatigue life of the wind turbine blades.

[0112] Step 5: Based on the damage condition of the wind turbine blades at the current moment, the failure probability and the impact of environmental data on the fatigue life of the wind turbine blades, the remaining fatigue life of the wind turbine blades at the current moment is analyzed, and it is determined whether to issue an early warning for the remaining fatigue life at the current moment.

[0113] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0114] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled 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 by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software depends on the specific application and design constraints of the technical solution.

[0115] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, and may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment as needed.

[0116] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A wind turbine blade fatigue life prediction platform, characterized by: include: The finite element analysis module is used to construct the geometric model and experimental specimens of the wind turbine blades based on the design parameters, conduct fatigue performance tests on the experimental specimens to obtain the limit number of fatigue cycles under different stresses, draw the SN curve, perform finite element analysis on the geometric model, and calibrate the geometric model based on the SN curve; The fatigue damage analysis module is used to collect the stress values ​​of wind turbine blades during historical operation, obtain the number of fatigue cycles of wind turbine blades under the stress values ​​during historical operation through finite element analysis, and analyze the damage status at the current moment based on the number of fatigue cycles; The fault probability model building module is used to obtain the historical operating data of wind turbine blades and assign corresponding fault probabilities through expert scoring. The module uses the historical operating data of wind turbine blades as input and the corresponding fault probabilities as labels to build a fault prediction model based on the machine learning model. A real-time fault analysis module collects the current operating data of the wind turbine blades and the environmental data of their location, inputs the operating data into the trained fault prediction model, obtains the current fault probability, and analyzes the impact of the current environmental data on the fatigue life of the wind turbine blades; The fatigue life prediction model is used to analyze the remaining fatigue life of the wind turbine blades at the current moment based on the damage status of the wind turbine blades, the failure probability and the impact of environmental data on the fatigue life of the wind turbine blades at the current moment, and to determine whether to issue an early warning for the remaining fatigue life at the current moment.

2. A wind turbine blade fatigue life prediction platform according to claim 1, characterized in that: Performing finite element analysis on the collective model specifically includes: Several wind turbine blade specimens were made according to the design parameters of the wind turbine blades. Fatigue performance tests were carried out on the wind turbine blade specimens. The blade root, blade tip and blade middle were set as observation points. Different stress levels were applied to the observation points of each experimental model. The limit fatigue cycle number and stress value of the observation point at each stress level were recorded. The limit fatigue cycle number and stress value were fitted to obtain the SN curve. The fitted SN curve equation is expressed as: log 10 (N f )=ab·log 10 (s) Among them, N f is the number of ultimate fatigue cycles, σ is the stress value, a and b are fitting parameters respectively; The design parameters include geometric data and material properties of the wind turbine blade, the geometric data including the length, width, thickness, cross-sectional shape of the wind turbine blade, and curvature and camber of the wind turbine blade, the material properties including the material used for the wind turbine blade and mechanical properties, the mechanical properties including elastic modulus, Poisson's ratio, shear modulus, tensile strength and compressive strength; creating a geometric model of the blade based on the geometric data of the wind turbine blade, inputting material properties, defining the boundary conditions of the root of the wind turbine blade as fixed, determining the grid type, applying loads to observation points of the geometric model, performing fatigue analysis and recording the number of fatigue cycles at the observation points output by the finite element analysis, comparing the results output by the finite element analysis with the SN curve, calculating the absolute difference between the fatigue life in the SN curve of each observation point and the limit fatigue cycle number output by the finite element analysis, and considering that the geometric model calibration is completed when the absolute difference is less than the error threshold; The calculation formula for converting the limit fatigue cycle number into the initial fatigue life is: Among them, E life is the initial fatigue life, where the initial fatigue life is in hours, N f is the limit fatigue cycle number, and f is the operating frequency.

3. A wind turbine blade fatigue life prediction platform according to claim 1, characterized in that: Analyzing the damage condition of the wind turbine blade at the current moment specifically includes: The stress values ​​of the wind turbine blades in the historical operation are collected. The stress values ​​carried in each historical operation are used as input and the finite element analysis is performed on the geometric model. The output number of fatigue cycles is obtained and the damage degree in each historical operation is calculated. The calculation formula is: Where D(i) is the damage degree under the i-th stress level in the historical operation, N(i) is the number of cycles under the i-th stress level in the historical operation, and N f (i) is the limit fatigue cycle number in the SN curve corresponding to the stress value of the wind turbine blade under the i-th stress level, and i is the index of different stress levels; The damage degree under each stress level is accumulated, and the formula is as follows: Where D is the total damage degree at the current moment, n is the total number of stress levels, and i∈[1,n]; If D<1, it indicates that the wind turbine blade still has remaining fatigue life. If D>1, the stress conditions in the historical operation have caused fatigue damage to the wind turbine blade. The fatigue life of the wind turbine blade is determined to be 0, and a fatigue life danger warning is issued.

4. A wind turbine blade fatigue life prediction platform according to claim 1, characterized in that: Constructing the fault prediction model specifically includes: The historical operation data includes stress, vibration and blade temperature. The mean of the stress, vibration and blade temperature data at the observation point of the wind turbine blade is used as the stress, vibration and blade temperature data of the wind turbine blade. The historical operation data is analyzed by an expert scoring method to assign corresponding failure probabilities. The historical operation data and the corresponding failure probabilities are integrated into a sample set, and the sample set is divided into a training set and a test set with a division ratio of 7:

3. The historical operation data is used as input and trained based on the machine learning algorithm model. The corresponding failure probability is used as a label, and the mean square error is selected as the loss function. When the loss function is lower than the set threshold, it indicates that the model training is completed. The threshold can be set to 0.

01. After the training is completed, the historical operation data in the test set is input into the trained fault prediction model, and the mean absolute error and determination coefficient indicators are used to evaluate the model performance.

5. The wind turbine blade fatigue life prediction platform according to claim 1, characterized in that: Analyzing the current operating data and environmental data specifically includes: Obtain the stress, vibration, and blade temperature of the observation point at the current moment, calculate the mean value as the stress, vibration, and blade temperature of the wind turbine blade at the current moment, obtain environmental data at the location of the wind turbine blade through the weather station, input the operating data into the trained fault prediction model, and obtain the fault probability predicted by the model; Analyze the impact of environmental data on the fatigue life of wind turbine blades, and calculate the environmental impact index using the following formula: Among them, FEI is the environmental impact index, T en 、H c 、W c are temperature, humidity and wind speed data respectively, T r 、H r 、W r are the suitable values ​​of temperature, humidity and wind speed respectively, and β1, β2 and β3 are the influence coefficients of each item respectively.

6. A wind turbine blade fatigue life prediction platform according to claim 1, characterized in that: Analyzing the remaining fatigue life of the wind turbine blade at the current moment specifically includes: Comprehensively analyze the damage condition, failure probability and environmental data of the wind turbine blades at the current moment on the fatigue life of the wind turbine blades to obtain the remaining fatigue life at the current moment. The calculation formula is: Among them, R life is the remaining fatigue life, E life is the initial fatigue life, D is the total damage degree, FEI is the environmental impact index, CFP is the failure probability, λ1, λ2, and λ3 are the weight coefficients of the corresponding items, 0<λ2<λ3<λ1<1; The remaining fatigue life at the current moment is judged. If the remaining fatigue life at the current moment is higher than the life threshold, it is judged that the wind turbine blades at the current moment are in a safe state, and the wind turbine blades continue to be monitored. If the remaining fatigue life at the current moment is lower than the life threshold, it is judged that the wind turbine blades at the current moment are in a fatigue damage state, and a fatigue life danger warning is issued.

7. A method for predicting fatigue life of wind turbine blades, characterized in that: The method for predicting fatigue life of a wind turbine blade is executed by a wind turbine blade fatigue life prediction platform according to any one of claims 1 to 6, and the specific steps include: Step 1: Construct a geometric model and experimental specimen of the wind turbine blade based on the design parameters. Perform fatigue performance tests on the experimental specimen to obtain the limit number of fatigue cycles under different stresses, draw the SN curve, perform finite element analysis on the geometric model, and calibrate the geometric model based on the SN curve. Step 2: Collect the stress values ​​of the wind turbine blades during historical operation, obtain the number of fatigue cycles of the wind turbine blades under the stress values ​​during historical operation through finite element analysis, and analyze the damage status at the current moment based on the number of fatigue cycles; Step 3: Obtain historical operating data of wind turbine blades and assign corresponding failure probabilities using an expert scoring method. Using the historical operating data of wind turbine blades as input and the corresponding failure probabilities as labels, a fault prediction model is constructed based on a machine learning model. Step 4: Collect the current operating data of the wind turbine blades and the environmental data of their location. Input the operating data into the trained fault prediction model to obtain the current failure probability and analyze the impact of the current environmental data on the fatigue life of the wind turbine blades. Step 5: Based on the damage condition of the wind turbine blades at the current moment, the failure probability and the impact of environmental data on the fatigue life of the wind turbine blades, the remaining fatigue life of the wind turbine blades at the current moment is analyzed, and it is determined whether to issue an early warning for the remaining fatigue life at the current moment.

Images