A method and device for predicting the risk of wind turbine blade fracture

Abstract

The present application provides a method and device for predicting the risk of wind turbine blade fracture, which is as follows: typhoon meteorological data and blade real-time status data are acquired, and these two types of data are pre-processed to obtain de-anomaly data and smoothed data. Based on the de-anomaly data, the turbulence intensity and wind direction change rate are calculated to construct a meteorological risk factor; based on the smoothed data, the vibration energy and strain energy density are calculated to construct a blade status risk factor. Then, an optimization algorithm is used to assign weight coefficients to these two risk factors, and a comprehensive risk factor is generated through weight fusion. According to the numerical value of the comprehensive risk factor, the seven-level risk threshold range is matched, and the corresponding warning signal and operation and maintenance operation are triggered. This method quantifies meteorological and blade status risks, optimizes weight allocation and comprehensively evaluates through multi-dimensional data acquisition and processing and feature engineering, and can accurately predict fracture risks, provide early warnings and guide operation and maintenance, thereby reducing losses and costs.

Description

Technical Field

[0001] The present application relates to the field of wind turbine blade safety, and in particular to a method and device for predicting the risk of wind turbine blade fracture. Background Art

[0002] In the field of wind power generation, wind turbine blades are key components that capture wind energy and convert it into electricity. Their stability and safety are crucial to the operation of the entire wind power generation system. However, typhoons, as extreme meteorological phenomena, pose a serious threat to wind turbine blades with their strong winds and complex and changing airflow characteristics. During a typhoon, wind turbine blades are not only subjected to enormous aerodynamic loads, but also to the combined effects of alternating stresses and complex environmental factors, which significantly increase the risk of blade fracture.

[0003] Traditional blade fracture risk assessment methods mainly rely on regular manual inspections and monitoring based on simple equipment operating parameters. However, these methods have significant limitations. Manual inspections are not only time-consuming and labor-intensive, but are also limited by the experience of inspectors and the inspection environment, making it difficult to comprehensively and accurately detect potential safety hazards in blades. Especially before a typhoon arrives, due to time constraints and harsh environments, the efficiency and accuracy of manual inspections are even more difficult to guarantee. At the same time, monitoring methods based on simple operating parameters, such as only monitoring wind speed and wind direction, although they can provide certain information, cannot fully consider the dynamic relationship between complex meteorological conditions during a typhoon and the actual operating status of the blades. This static monitoring method has significant limitations and lags in predicting blade fracture risks, and cannot accurately and timely reflect the true risk status of blades in typhoons. Summary of the Invention

[0004] The purpose of this application is to overcome the defects in the above-mentioned prior art and provide a method and device for predicting the risk of wind turbine blade fracture.

[0005] This application also provides a method for predicting the risk of wind turbine blade fracture, comprising:

[0006] Obtain typhoon meteorological data and real-time status data of blades;

[0007] Performing preprocessing operations on the meteorological data and the real-time status data respectively, and generating abnormality-removed data and smoothed data respectively;

[0008] Calculating turbulence intensity and wind direction change rate based on the de-anomaly data, and calculating vibration energy and strain energy density based on the smoothed data;

[0009] constructing a meteorological risk factor according to the turbulence intensity and the wind direction change rate, and constructing a blade state risk factor according to the vibration energy and the strain energy density;

[0010] Allocating weight coefficients of the meteorological risk factor and the leaf state risk factor by an optimization algorithm;

[0011] fusing the meteorological risk factor and the leaf state risk factor using the weight coefficient to generate a comprehensive risk factor;

[0012] According to the numerical value of the comprehensive risk factor matching the seven-level risk threshold range, the corresponding level of early warning signal and operation and maintenance operation is triggered.

[0013] Optionally, the preprocessing operation includes:

[0014] Outlier removal is based on statistical criteria within a sliding window;

[0015] The smoothing process uses an adaptive filtering algorithm.

[0016] Optionally, the step of extracting the strain energy density includes:

[0017] The strain energy density is obtained based on the elastic properties of the blade material and real-time deformation data.

[0018] Optionally, the meteorological risk factor and the leaf state risk factor are fused using the weight coefficient to generate a comprehensive risk factor, including:

[0019] The weight coefficient is dynamically allocated according to the historical frequency of typhoons and the operating years of the wind turbine.

[0020] Optionally, the seven risk level classification thresholds include:

[0021] Typhoon landing critical value, blade resonance critical value and material deformation limit value.

[0022] The present application provides a device for predicting the risk of wind turbine blade fracture, comprising:

[0023] Acquisition module, which obtains typhoon meteorological data and real-time status data of blades;

[0024] A preprocessing module performs preprocessing operations on the meteorological data and the real-time status data, respectively, to generate abnormality-free data and smoothed data;

[0025] a calculation module for calculating turbulence intensity and wind direction change rate based on the de-anomalyed data, and calculating vibration energy and strain energy density based on the smoothed data;

[0026] a factor module, constructing a meteorological risk factor according to the turbulence intensity and the wind direction change rate, and constructing a blade state risk factor according to the vibration energy and the strain energy density;

[0027] an allocation module, which allocates weight coefficients of the meteorological risk factor and the leaf state risk factor through an optimization algorithm;

[0028] A weight module, which fuses the meteorological risk factor and the leaf state risk factor using the weight coefficient to generate a comprehensive risk factor;

[0029] The early warning module matches the seven-level risk threshold range according to the numerical value of the comprehensive risk factor, and triggers the corresponding level of early warning signal and operation and maintenance operation.

[0030] Optionally, the preprocessing operation includes:

[0031] Outlier removal is based on statistical criteria within a sliding window;

[0032] The smoothing process uses an adaptive filtering algorithm.

[0033] Optionally, the step of extracting the strain energy density includes:

[0034] The strain energy density is obtained based on the elastic properties of the blade material and real-time deformation data.

[0035] Optionally, the meteorological risk factor and the leaf state risk factor are fused using the weight coefficient to generate a comprehensive risk factor, including:

[0036] The weight coefficient is dynamically allocated according to the historical frequency of typhoons and the operating years of the wind turbine.

[0037] Optionally, the seven risk level classification thresholds include:

[0038] Typhoon landing critical value, blade resonance critical value and material deformation limit value.

[0039] The beneficial effects of this application are:

[0040] The present application also provides a method for predicting the risk of wind turbine blade fracture, including: obtaining typhoon meteorological data and real-time status data of blades; performing preprocessing operations on the meteorological data and the real-time status data respectively, and generating de-anomaly data and smoothed data accordingly; calculating turbulence intensity and wind direction change rate based on the de-anomaly data, and calculating vibration energy and strain energy density based on the smoothed data; constructing a meteorological risk factor according to the turbulence intensity and the wind direction change rate, and constructing a blade status risk factor according to the vibration energy and the strain energy density; allocating weight coefficients of the meteorological risk factor and the blade status risk factor through an optimization algorithm; fusing the meteorological risk factor and the blade status risk factor using the weight coefficient to generate a comprehensive risk factor; matching the seven-level risk threshold range according to the numerical value of the comprehensive risk factor, triggering the corresponding level of warning signal and operation and maintenance operation. The present application uses multi-dimensional data collection and processing, combined with feature engineering to construct a risk prediction model, quantify meteorological and blade status risks, optimize weight allocation and comprehensive evaluation, to achieve accurate fracture risk prediction, early warning and guidance of operation and maintenance, and reduce damage and cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a schematic diagram of the wind turbine blade fracture risk prediction process in this application. DETAILED DESCRIPTION

[0042] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it is understood that various forms of implementing the present disclosure are not limited by the embodiments set forth herein. Rather, the embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0043] The present application provides a method for predicting the risk of wind turbine blade fracture, comprising: a meteorological monitoring terminal, a blade status monitoring terminal, a data processing center, and a risk prediction model.

[0044] Among them, the meteorological monitoring terminal and the blade status monitoring terminal establish communication connections with the data processing center respectively, and the risk prediction model is integrated into the data processing center.

[0045] The meteorological monitoring terminal integrates a variety of high-precision sensors, enabling comprehensive, real-time monitoring of typhoon-related meteorological data. It continuously collects this data and transmits it to the data processing center via a stable communication link, providing critical meteorological information support for the entire risk prediction system.

[0046] The blade condition monitoring terminal is equipped with a series of specialized monitoring devices to collect real-time data on the blade's status. These sensors work together to convert the collected physical quantities into electrical or digital signals, which are then pre-processed and promptly transmitted to the data processing center. This allows the system to monitor the blade's operating status in real time, providing internal status information for assessing blade health and fracture risk.

[0047] After receiving data from the meteorological monitoring terminal and the blade condition monitoring terminal, the data processing center first performs data preprocessing. Next, it conducts feature engineering and, based on a pre-set calculation formula, derives a comprehensive risk factor, R. Finally, based on the value of R, it assigns a risk level and sends risk warning information to operations and maintenance personnel.

[0048] The data processing center is the core hub of the entire risk prediction system, responsible for data processing, analytical calculations, and decision support. It transforms raw data from various terminals into valuable information. Through complex calculations and analysis, it assesses the risk of blade breakage in typhoons and provides intuitive risk levels and early warning information to operators, helping them take timely measures to ensure the safe operation of wind turbines.

[0049] The risk prediction model, built within the data processing center, is key to intelligently predicting blade fracture risk. By studying and analyzing extensive historical data and combining physical principles with mathematical algorithms, it quantifies the impact of meteorological conditions and blade status on blade fracture risk. By continuously optimizing model parameters, the accuracy and reliability of predictions can be improved.

[0050] As attached Figure 1 The present application provides a method for predicting the risk of wind turbine blade fracture, which specifically includes:

[0051] S101, obtaining typhoon meteorological data and real-time status data of blades;

[0052] The meteorological monitoring terminal is used to collect typhoon-related meteorological data in real time, and the blade status monitoring terminal is used to collect real-time status data of the blades.

[0053] The meteorological data include: current wind speed V, maximum wind speed V at the typhoon center max , average wind speed V a , wind direction θ, air pressure P, typhoon intensity level I, typhoon moving speed V m , gust coefficient G, horizontal wind shear S h , vertical wind shear S v And the real-time distance d between the typhoon center and the wind turbine.

[0054] The real-time status data includes: blade vibration frequency f, vibration amplitude A, blade equivalent mass m, blade material elastic modulus E, blade strain ε, strain change rate ζ, blade temperature T, blade temperature change rate T c , blade surface damage area ratio S and blade surface roughness R a .

[0055] The blade vibration frequency and amplitude are measured using high-precision accelerometers installed at the blade root, mid-blade, and tip, with signal processing algorithms used to obtain accurate data. Blade strain is collected using strain gauges placed at key locations on the blade and transmitted in real time via a data acquisition card. Blade temperature is monitored using multiple distributed temperature sensors, combined with a temperature compensation algorithm to improve measurement accuracy. The percentage of blade surface damage is determined by analyzing the blade using image recognition and deep learning algorithms using a drone equipped with a high-definition camera and taking 360° images of the blade. Blade surface roughness is measured using a non-contact optical measuring instrument. Blade rotational speed is indirectly measured using the generator's speed sensor.

[0056] S102, respectively perform pre-processing operations on the meteorological data and the real-time status data, and generate abnormal data and smoothed data accordingly.

[0057] The collected meteorological data and blade status data were preprocessed, and the 3σ criterion based on statistical principles was used to remove outliers, and the sliding average filter algorithm was used to smooth the data with large fluctuations.

[0058] S103, calculating turbulence intensity and wind direction change rate based on the de-abnormalized data, and calculating vibration energy and strain energy density based on the smoothed data;

[0059] Perform feature engineering based on preprocessed data.

[0060] The feature engineering includes calculating the turbulence intensity of wind speed:

[0061]

[0062] Where N is the number of wind speed data samples, V i Represents the i-th wind speed data, V a Indicates the average wind speed.

[0063] In reality, wind speed is not stable but fluctuates.

[0064] It is used to measure the degree of dispersion of wind speed around the average value, which can reflect the amplitude of wind speed fluctuation. Dividing it by the average wind speed, the turbulence intensity can be used to standardize the intensity of wind speed fluctuation. vThe larger the value, the more severe the wind speed fluctuation.

[0065] Severely fluctuating wind speeds can subject blades to unstable aerodynamic loads, causing additional vibration and fatigue stress. v When the wind speed is high, the blades experience frequent changes in magnitude and direction of wind force at different times during rotation, which accelerates blade wear and fatigue and increases the risk of blade breakage. It is an important indicator for evaluating the stability of the blade operating environment. The feature engineering includes calculating the rate of change of wind direction:

[0066]

[0067] Among them, θ i+1 and θ i They represent the wind direction data at adjacent moments, and △t represents the time interval.

[0068] Changes in wind direction will change the direction of the aerodynamic loads on the blades.

[0069] By calculating the difference in wind direction between adjacent moments and dividing it by the time interval, we can determine the change in wind direction per unit time, or the wind direction change rate. This quantifies the speed of wind direction change. Frequent wind direction changes result in a larger Δθ value, which means increased unstable loads on the blades.

[0070] For example: when the wind direction changes rapidly, the blades need to constantly adjust their posture to adapt to the new wind direction, which will cause the blade structure to bear greater torsional and bending stresses, easily causing damage to the internal structure of the blades, thereby increasing the risk of blade breakage. It is one of the key parameters for evaluating the impact of meteorological conditions on blade stress.

[0071] The feature engineering involves calculating the energy index of the blade vibration:

[0072]

[0073] Where m represents the equivalent mass of the blade, f represents the blade vibration frequency, and A represents the vibration amplitude.

[0074] During blade vibration, mass m, vibration frequency f and vibration amplitude A are key factors affecting vibration energy. This formula can comprehensively consider these factors and accurately calculate the energy possessed by blade vibration. The vibration energy is proportional to the square of these parameters, reflecting the degree of their important influence on energy. Ev reflects the energy of the blade during vibration. The higher the vibration energy, the more violent the blade vibration. Excessive vibration energy will accelerate the fatigue damage of the blade and make the blade more prone to fatigue cracks. Over time, these cracks may expand and eventually lead to blade breakage. Therefore, Ev is an important indicator to measure the impact of the blade's own vibration state on its structural health; the feature engineering includes calculating the blade strain energy density:

[0075]

[0076] Where E represents the elastic modulus of the blade material and ε represents the blade strain.

[0077] Strain energy density is the energy stored per unit volume of material when it is deformed.

[0078] For blade materials, the elastic modulus E reflects the material's ability to resist elastic deformation, U ε It can measure the energy stored in the blade during the deformation process. When the blade is strained by external force, U ε Larger values ​​indicate more energy stored within the blade. Excessive strain energy accumulation can cause plastic deformation or even damage to the blade material. This parameter is a crucial parameter for assessing the blade's stress deformation state and potential damage risk, reflecting the impact of changes in the blade material's internal state on its safety.

[0079] S104, constructing a meteorological risk factor according to the turbulence intensity and the wind direction change rate, and constructing a blade state risk factor according to the vibration energy and the strain energy density;

[0080] Constructing an analysis and prediction model: The data processing center constructs a risk prediction model integrating meteorological factors and blade status factors, and defines the meteorological risk factor R m and leaf status risk factor R s .

[0081] The meteorological risk factors are:

[0082]

[0083] in, , V max Indicates the maximum wind speed at the center of the typhoon, V def Indicates the maximum wind speed the blade is designed to withstand.

[0084] , I v Turbulence intensity indicating wind speed; , d represents the real-time distance between the typhoon center and the wind turbine;

[0085]

[0086] P represents air pressure, and P0 represents the reference value of air pressure under normal meteorological conditions.

[0087] , S v Indicates vertical wind shear. , S h Indicates horizontal wind shear. , G represents the gust coefficient.

[0088] , V m Indicates the moving speed of the typhoon.

[0089] , I represents the intensity level of the typhoon. 10 =△θ, △θ represents the rate of change of wind direction, α1 to α 10 Represents the weight coefficient.

[0090] And through the particle swarm optimization algorithm combined with historical data to optimize the training, meet the .

[0091] After multiple training verifications, we determined that α1=0.18, α2=0.13, α3=0.09, α4=0.07, α5=0.11, α6=0.09, α7=0.09, α8=0.04, α9=0.09, and α 10 =0.11.

[0092] The meteorological risk factor is a quantitative indicator of blade risk that integrates multiple meteorological factors.

[0093] Each Xi represents a quantitative relationship between a meteorological factor and blade breakage risk, and α is a weight coefficient, determined through a particle swarm optimization algorithm combined with historical data training. This linear weighted summation approach is based on the combined influence of multiple factors. Considering the varying contributions of different meteorological factors to blade breakage risk, the weight coefficient adjusts the relative importance of each factor.

[0094] Rm comprehensively reflects the risk level of typhoon meteorological conditions to blades.

[0095] For example: X1 reflects the relative relationship between the maximum wind speed of a typhoon and the design bearing capacity of the blade. The closer the wind speed is to or exceeds the design value, the higher the risk.

[0096] IV, S v 、S h , △θ, etc. reflect the impact of the unstable characteristics of wind on the blades.

[0097] X3 represents the impact of the distance from the typhoon center on the risk. The closer the distance, the greater the risk.

[0098] X4 reflects the effects of air pressure changes.

[0099] G reflects the effect of gusts.

[0100] X8 measures the impact of a typhoon's moving speed.

[0101] I reflects the impact of typhoon intensity level.

[0102] By combining these factors, R m It can comprehensively assess the contribution of meteorological conditions to the risk of blade breakage.

[0103] The leaf condition risk factors:

[0104]

[0105] in, , f represents the vibration frequency of the blade, f0 represents the vibration frequency of the blade during normal operation, and fmax is the maximum vibration frequency allowed by the blade.

[0106] , A represents the vibration amplitude, A max Indicates the safety threshold of blade vibration amplitude.

[0107] ,ζ represents the strain change rate, ε max represents the ultimate strain of the blade.

[0108] , U ε represents the blade strain energy density, U ε-max represents the strain energy density of the blade at the ultimate strain.

[0109] , , T represents the blade temperature, Tc represents the blade temperature change rate, T0 represents the normal operating temperature of the blade, T max Indicates the maximum temperature the blade can withstand.

[0110] , S represents the percentage of damaged area on the leaf surface. , R a Indicates the surface roughness of the blade, R a-max Indicates the maximum allowable value of blade surface roughness.

[0111] , E v Energy index of blade vibration, E v-max It represents the energy index of the blade under the maximum vibration condition allowed by the design; β1 to β9 represent weight coefficients, which are optimized and determined by genetic algorithm to meet .

[0112] After multiple training verifications, it was determined that β1=0.13, β2=0.1, β3=0.13, β4=0.11, β5=0.09, β6=0.09, β7=0.16, β8=0.16, and β9=0.03.

[0113] The leaf condition risk factor is a quantitative indicator of risk that integrates multiple leaf condition factors.

[0114] Each Y j Represents a quantitative relationship between leaf condition and risk. β is a weighting coefficient, determined through optimization using a genetic algorithm combined with expert experience. Similarly, a linear weighted summation approach is used to adjust the weighting coefficients, taking into account the varying impacts of different leaf condition factors on leaf fracture risk. Significance: Rs comprehensively reflects the impact of leaf condition on fracture risk.

[0115] For example, Y1 and Y2 reflect the degree to which the blade vibration frequency and amplitude deviate from the normal range, respectively; greater deviations indicate a higher risk. Y3 reflects the relationship between the strain change rate and the ultimate strain, reflecting the degree of danger of blade strain changes. Y4 measures the relationship between strain energy density and the ultimate value. Y5 and Y6 consider the impact of temperature and its changes on the blade. S reflects the damage to the blade surface. Y8 reflects the impact of blade surface roughness. Y9 represents the relationship between blade vibration energy and maximum allowable energy. Combining these factors, Rs comprehensively assesses the contribution of the blade's own condition to fracture risk.

[0116] S105, allocating weight coefficients of the meteorological risk factor and the leaf state risk factor by using an optimization algorithm;

[0117] Based on the risk prediction model constructed by the data processing center, a comprehensive risk factor R is established.

[0118] S6. Using the weight coefficient to fuse the meteorological risk factor and the leaf state risk factor to generate a comprehensive risk factor;

[0119] The combined risk factors are:

[0120]

[0121] Among them, γ represents the comprehensive weight of meteorological risk factor and blade status risk factor, and its value range is between 0 and 1. In areas where typhoons occur frequently and wind turbines have been in operation for a long time, γ increases; in areas where typhoons have less impact and the equipment is newer, γ decreases.

[0122] The risk prediction model divides risk levels according to the value of the comprehensive risk factor R and sets seven risk levels.

[0123] The seven risk levels are specifically: when R<0.05, it is classified as the first risk level; when 0.05<=R<0.15, it is classified as the second risk level; when 0.15<=R<0.3, it is classified as the third risk level; when 0.3<=R<0.45, it is classified as the fourth risk level; when 0.45<=R<0.6, it is classified as the fifth risk level; when 0.6<=R<0.8, it is classified as the sixth risk level; when R>0.8, it is classified as the seventh risk level.

[0124] S107: Match the seven-level risk threshold range according to the numerical value of the comprehensive risk factor, and trigger the corresponding level of early warning signal and operation and maintenance operation.

[0125] Issue corresponding warning signals based on the risk level risk prediction model.

[0126] The corresponding early warning signals include:

[0127] When the prediction is the first risk level, it is determined to be a safe state and the first warning signal is issued. This signal indicates that the blades will basically not be under obvious risk threats during the typhoon. The data processing center only performs routine data recording and system operation status monitoring, and arranges equipment inspection and maintenance work according to the normal operation and maintenance plan.

[0128] When the forecast is the second risk level, it is determined to be an extremely low risk state and a second warning signal is issued. This signal indicates that the current risk is low, but it is necessary to start paying close attention to typhoon dynamics and changes in blade status. The system automatically pushes information on typhoon path and intensity changes to operation and maintenance personnel, reminding them to prepare basic detection tools and be ready to conduct more detailed inspections of the equipment at any time.

[0129] When the prediction is the third risk level, it is determined to be a low-risk state and a third early warning signal is issued. This signal indicates that the blades are facing certain risks. The operation and maintenance personnel need to conduct a comprehensive inspection of the key parts of the wind turbine to ensure that all components are firmly connected and operating normally. At the same time, check the operation of the meteorological monitoring terminal and the blade status monitoring terminal to ensure the accuracy and stability of data collection.

[0130] When the fourth risk level is predicted, it is determined to be a medium-low risk state and a fourth warning signal is issued, which indicates that the risk has increased. In addition to the low-risk level inspection work, non-destructive testing of the blade surface is also required to detect whether there are potential defects or damage inside the blade. Based on the test results, a more accurate assessment of the safety status of the blade can be made.

[0131] When the prediction is the fifth risk level, it is determined to be a medium-risk state and the fifth warning signal is issued, which indicates that the blades are in a medium-risk state. On the basis of continuing to strengthen the above-mentioned inspections and preparations, appropriate load reduction operations are carried out on wind turbines according to the expected path and intensity changes of the typhoon to reduce the load on the blades. At the same time, the frequency of meteorological monitoring and blade status monitoring is adjusted to increase the density of data collection.

[0132] When the prediction is the sixth risk level, it is determined to be a medium-to-high risk state and the sixth warning signal is issued. This signal indicates that the blades are facing a greater risk of breakage. The emergency plan is immediately activated, the wind turbine is shut down urgently, and the blade position is locked to prevent it from rotating freely in strong winds. At the same time, a warning area is set up around the equipment, and unauthorized personnel are prohibited from approaching to ensure personnel safety.

[0133] When the prediction is the seventh risk level, it is determined to be a high-risk state and the seventh early warning signal is issued. This signal indicates that the blades are at an extremely high risk of fracture and may fracture at any time. In addition to implementing all measures for the medium and high risk levels, personnel and equipment within a certain range of the surrounding area must be quickly evacuated to ensure the safety of personnel and property, and the wind turbine must be fully isolated for safety to prevent broken blades from causing further damage to surrounding facilities.

[0134] The present application provides a device for predicting the risk of wind turbine blade fracture, comprising:

[0135] Acquisition module, which obtains typhoon meteorological data and real-time status data of blades;

[0136] A preprocessing module performs preprocessing operations on the meteorological data and the real-time status data, respectively, to generate abnormality-free data and smoothed data;

[0137] a calculation module for calculating turbulence intensity and wind direction change rate based on the de-anomalyed data, and calculating vibration energy and strain energy density based on the smoothed data;

[0138] a factor module, constructing a meteorological risk factor according to the turbulence intensity and the wind direction change rate, and constructing a blade state risk factor according to the vibration energy and the strain energy density;

[0139] an allocation module, which allocates weight coefficients of the meteorological risk factor and the leaf state risk factor through an optimization algorithm;

[0140] A weight module, which fuses the meteorological risk factor and the leaf state risk factor using the weight coefficient to generate a comprehensive risk factor;

[0141] The early warning module matches the seven-level risk threshold range according to the numerical value of the comprehensive risk factor, and triggers the corresponding level of early warning signal and operation and maintenance operation.

[0142] Optionally, the preprocessing operation includes:

[0143] Outlier removal is based on statistical criteria within a sliding window;

[0144] The smoothing process uses an adaptive filtering algorithm.

[0145] Optionally, the step of extracting the strain energy density includes:

[0146] The strain energy density is obtained based on the elastic properties of the blade material and real-time deformation data.

[0147] Optionally, the meteorological risk factor and the leaf state risk factor are fused using the weight coefficient to generate a comprehensive risk factor, including:

[0148] The weight coefficient is dynamically allocated according to the historical frequency of typhoons and the operating years of the wind turbine.

[0149] Optionally, the seven risk level classification thresholds include:

[0150] Typhoon landing critical value, blade resonance critical value and material deformation limit value.

[0151] The above description of the embodiments is intended to facilitate understanding and application of this application by those skilled in the art. It will be readily apparent to those skilled in the art that various modifications to the above embodiments can be made, and the general principles described herein can be applied to other embodiments without requiring creative effort. Therefore, this application is not limited to the above embodiments. Any improvements or modifications made to this application by those skilled in the art based on the disclosure of this application should fall within the scope of protection of this application.

Claims

1. A method for predicting the risk of wind turbine blade fracture, characterized in that: include: Obtain typhoon meteorological data and real-time status data of blades; Performing preprocessing operations on the meteorological data and the real-time status data respectively, and generating abnormality-removed data and smoothed data respectively; Calculating turbulence intensity and wind direction change rate based on the de-anomaly data, and calculating vibration energy and strain energy density based on the smoothed data; constructing a meteorological risk factor according to the turbulence intensity and the wind direction change rate, and constructing a blade state risk factor according to the vibration energy and the strain energy density; Allocating weight coefficients of the meteorological risk factor and the leaf state risk factor by an optimization algorithm; fusing the meteorological risk factor and the leaf state risk factor using the weight coefficient to generate a comprehensive risk factor; According to the numerical value of the comprehensive risk factor matching the seven-level risk threshold range, the corresponding level of early warning signal and operation and maintenance operation is triggered.

2. A wind turbine blade fracture risk prediction method according to claim 1, characterized in that: The pre-processing operation includes: Outlier removal is based on statistical criteria within a sliding window; The smoothing process uses an adaptive filtering algorithm.

3. A wind turbine blade fracture risk prediction method according to claim 1, characterized in that: The step of extracting the strain energy density comprises: The strain energy density is obtained based on the elastic properties of the blade material and real-time deformation data.

4. A wind turbine blade fracture risk prediction method according to claim 1, characterized in that: The meteorological risk factor and the leaf state risk factor are fused using the weight coefficient to generate a comprehensive risk factor, including: The weight coefficient is dynamically allocated according to the historical frequency of typhoons and the operating years of the wind turbine.

5. A wind turbine blade fracture risk prediction method according to claim 1, characterized in that: The thresholds for the seven risk levels include: Typhoon landing critical value, blade resonance critical value and material deformation limit value.

6. A wind turbine blade fracture risk prediction device, characterized in that: include: Acquisition module, which obtains typhoon meteorological data and real-time status data of blades; A preprocessing module performs preprocessing operations on the meteorological data and the real-time status data, respectively, to generate abnormality-free data and smoothed data; a calculation module for calculating turbulence intensity and wind direction change rate based on the de-anomalyed data, and calculating vibration energy and strain energy density based on the smoothed data; a factor module, constructing a meteorological risk factor according to the turbulence intensity and the wind direction change rate, and constructing a blade state risk factor according to the vibration energy and the strain energy density; an allocation module, which allocates weight coefficients of the meteorological risk factor and the leaf state risk factor through an optimization algorithm; A weight module, which fuses the meteorological risk factor and the leaf state risk factor using the weight coefficient to generate a comprehensive risk factor; The early warning module matches the seven-level risk threshold range according to the numerical value of the comprehensive risk factor, and triggers the corresponding level of early warning signal and operation and maintenance operation.

7. A wind turbine blade fracture risk prediction device according to claim 6, characterized in that: The pre-processing operation includes: Outlier removal is based on statistical criteria within a sliding window; The smoothing process uses an adaptive filtering algorithm.

8. The wind turbine blade fracture risk prediction device according to claim 6, characterized in that: The step of extracting the strain energy density comprises: The strain energy density is obtained based on the elastic properties of the blade material and real-time deformation data.

9. The wind turbine blade fracture risk prediction device according to claim 6, characterized in that: The meteorological risk factor and the leaf state risk factor are fused using the weight coefficient to generate a comprehensive risk factor, including: The weight coefficient is dynamically allocated according to the historical frequency of typhoons and the operating years of the wind turbine.

10. The wind turbine blade fracture risk prediction device according to claim 6, characterized in that: The seven risk level classification thresholds include: Typhoon landing critical value, blade resonance critical value and material deformation limit value.

Images