Wind turbine generator blade clearance value prediction method

Abstract

The invention relates to the technical field of blade clearance value prediction, and discloses a wind turbine generator blade clearance value prediction method, which comprises the following steps: constructing a blade clearance value prediction model based on historical environment data; establishing a plurality of adjustment periods, obtaining environment data according to a preset adjustment time node, and generating a clearance value prediction curve in a single adjustment period according to the environment data and the blade clearance value prediction model; setting a plurality of monitoring time nodes in a single adjusting period, obtaining the actual value of each monitoring point, and judging whether to correct the clearance value prediction curve or not according to the deviation degree between the actual value and the clearance value prediction curve. By continuously monitoring an actual value and comparing the actual value with a predicted value, the deviation of a prediction model can be found in time, and the model is adjusted accordingly, so that the accuracy of long-term prediction is improved, and the reliability of the system is enhanced; and production interruption and maintenance cost caused by sudden faults are reduced, so that economic benefits are brought.

Description

Technical Field

[0001] The invention relates to the technical field of blade clearance value prediction, and in particular to a method for predicting the blade clearance value of a wind turbine generator set. Background Art

[0002] The clearance of a wind turbine refers to the vertical distance between the tip of the wind turbine blade and the tower. This distance is critical to the safe operation of the wind turbine, because if the clearance is insufficient, the blades may hit the tower during rotation, causing serious damage or even casualties. As the size of wind turbines increases, especially the length of blades, it becomes particularly important to ensure sufficient clearance. In recent years, with the advancement of sensor technology, computer vision, machine learning, and data analysis methods, the prediction method of the clearance value of wind turbine blades has also been continuously developed. For example, the dynamic digital twin method of wind turbine blades based on real-time vision can analyze images through computer vision technology, calculate the clearance value between the blades and the tower, and realize real-time monitoring and early warning.

[0003] In severe climate environments, the collected clearance values ​​may have low credibility, and if protection control is based on these data, it may lead to unnecessary economic losses. Therefore, it is necessary to develop methods to identify and exclude these low-credibility data. Summary of the invention

[0004] The purpose of the present invention is to conduct blade deformation tests and establish a prediction model through a temperature gradient value and a wind gradient mapping model; predict the deformation distance of the blade under temperature changes and wind changes; obtain the blade clearance value in real time, and generate a preliminary prediction value based on the prediction result; and obtain the final blade clearance value prediction value through deviation value correction.

[0005] In order to achieve the above object, the present invention provides a method for predicting the clearance value of blades of a wind turbine generator set, comprising:

[0006] Construct a blade clearance value prediction model based on historical environmental data;

[0007] Establish multiple adjustment cycles, obtain environmental data according to preset adjustment time nodes, and generate a clearance value prediction curve within a single adjustment cycle based on the environmental data and the blade clearance value prediction model;

[0008] Multiple monitoring time nodes are set in a single adjustment cycle to obtain the actual value of each monitoring point, and whether to correct the headroom value prediction curve is determined based on the deviation between the actual value and the headroom value prediction curve.

[0009] In some embodiments of the present invention, the construction of a blade clearance value prediction model based on historical environmental data includes:

[0010] Conduct simulation tests based on historical environmental data and establish a blade deformation prediction model;

[0011] The blade tip deformation prediction value is generated through the blade deformation prediction model and the current environmental data, including:

[0012] Obtain historical meteorological data of the area where the wind turbine is located, and record the range of meteorological changes in the current area, including temperature change data and wind speed change data;

[0013] Based on the obtained temperature change data, determine the temperature change range of the current area, and divide the temperature within the change range into intervals, each interval representing a gradient value;

[0014] Based on the acquired wind force change data, determine the wind force change range in the current area, and develop a wind force gradient change mapping model to convert the wind force vector data into scalar force data acting on the blades;

[0015] Based on the temperature gradient value and wind gradient change mapping model, blade deformation test is carried out.

[0016] According to the results of blade deformation test, a blade deformation prediction model is constructed;

[0017] Generate a blade tip deformation prediction value through a blade deformation prediction model;

[0018] Based on the predicted value of the blade tip deformation degree, a predicted value of the wind turbine blade clearance value is generated.

[0019] In some embodiments of the present invention, the blade deformation test includes:

[0020] Get the temperature gradient test data set T, T = [t1, t2…t i …t n ];

[0021] Among them, t i represents the gradient value of the i-th temperature interval; n represents the number of intervals;

[0022] Get wind test dataset

[0023] represents the wind vector in the ith case n represents the total number of wind vector situations;

[0024] Through experiments, a wind mapping model is established to convert the wind vector Convert to wind scalar f i ;

[0025] Generate a scalar force test data set F acting on the blade, F = [f1,f2…fi …f n ];

[0026] where f i represents the wind vector in the ith case The corresponding scalar force data acting on the blade, n represents the total number of wind force value data;

[0027] Obtain the deformation distance Li of the blade tip in the i-th case;

[0028] Generate a blade deformation model:

[0029] Li=K1*t i +K2*f i ;

[0030] Where Li represents the deformation distance of the blade tip in the i-th case, K1 represents the influence coefficient of temperature on blade stiffness, and K2 represents the influence coefficient of wind force on blade stiffness;

[0031] The values ​​of K1 and K2 are calculated by the control variable method.

[0032] In some embodiments of the present invention, the calculation of the values ​​of K1 and K2 by the control variable method includes:

[0033] Calculate the influence coefficient K1 of temperature on blade stiffness:

[0034] (1) Obtain the initial stiffness value E0 and the initial temperature value T0, adjust the temperature value to T1, and obtain the current stiffness value E1; at this time, calculate the influence coefficient K1 of temperature on blade stiffness;

[0035]

[0036] Calculate the influence coefficient K2 of wind force on blade stiffness:

[0037] (2) Obtaining the initial stiffness value E0 and the initial wind force value F0, adjusting the wind force value to F1, and obtaining the current stiffness value E1; at this time, calculating the influence coefficient K2 of the wind force on the blade stiffness;

[0038]

[0039] In some embodiments of the present invention, generating the predicted value of blade tip deformation includes:

[0040] Based on remote sensing data, generate temperature change prediction data PT, PT = [pt1, pt2…pt t …pt T ];

[0041] Wind force change prediction data PW,PW=[pw1,pw2…pwt …pw T ];

[0042] Generate a scalar force data set PF, PF = [pf1, pf2…pf t …pf T ];

[0043] Among them, pt t represents the predicted temperature value at time t, pw t represents the wind force forecast value at time t, pf t represents the predicted value of the scalar force at time t, and T is the total number of predicted moments;

[0044] Calculate the deformation distance Lx of the blade tip based on the blade deformation prediction model;

[0045] Lx=K1*pt t +K2*pf t .

[0046] In some embodiments of the present invention, generating a blade clearance value prediction value includes:

[0047] Get the actual value of the blade clearance value and generate the actual value data set L of the blade clearance value, L = [L1, L2…L t …L T ];

[0048] Among them, L1 represents the actual value of the blade clearance at the current moment, L t It represents the actual value of the blade clearance at time t, and T represents the time length;

[0049] Based on the deformation distance Lx of the blade tip, the predicted change value Ly of the clearance value is calculated;

[0050] Ly = Lx*sinB;

[0051] Wherein, B is the current pitch angle of the wind turbine;

[0052] Generate the predicted value Lo of the first-level headroom value at time t;

[0053] Lo=L1-Ly

[0054] =L1-Lx*sinB

[0055] =L1-sinB*(K1*pt t +K2*pf t );

[0056] The first-level clearance value prediction value Lo is corrected to generate the blade clearance value prediction value Lz, Lz = [Lz1, Lz2…Lz t …Lz T ];

[0057] Among them, Lz t It represents the predicted value of the blade clearance value at time t, and T represents the time length.

[0058] In some embodiments of the present invention, the correction of the first-level headroom value prediction value Lo includes:

[0059] Based on the actual value L of the current blade clearance value at time t t With the first-level headroom value prediction value Lo, establish the correction model θ(t);

[0060] θ(t)=L t -Lo

[0061] =L t -(L1-sinB*(K1*pt t +K2*pf t ))

[0062] Based on the correction model θ(t), a correction strategy is formulated to correct the first-level headroom value prediction value Lo.

[0063] In some embodiments of the present invention, the formulating of a correction strategy based on the correction model θ(t) includes:

[0064] Group the values ​​of θ(t) into equally spaced intervals;

[0065] Count the number of data points in each interval, then draw a histogram with the interval as the horizontal axis and the data or frequency as the vertical axis; form correction instructions by observing the shape of the histogram.

[0066] In some embodiments of the present invention, the forming of the correction instruction includes:

[0067] If the histogram is close to a normal distribution, set a sliding window containing m time nodes and obtain the value range of θ(t) in the sliding window, θ(t)min≤θ(t)≤θ(t)max;

[0068] Among them, θ(t)min represents the minimum value of θ(t) in the sliding window, and θ(t)max represents the maximum value of θ(t) in the sliding window;

[0069] θ(t)max is used as the first deviation value C1 for predicting the clearance value of the wind turbine generator set;

[0070] Lz=Lo+C1;

[0071] If the histogram is skewed or multimodal, the data acquisition device should be calibrated to determine whether there is any setting error in the data acquisition device.

[0072] Compared with the prior art, the method for predicting the clearance value of wind turbine blades in the embodiment of the present invention has the following beneficial effects:

[0073] The constructed blade deformation prediction model can predict the deformation of blades under different environmental conditions based on real-time meteorological data, which not only helps to improve the operating efficiency of wind turbines, but also effectively avoids safety risks caused by excessive blade deformation.

[0074] By accurately quantifying the impact of temperature and wind force on the deformation of wind turbine blades, key data support is provided for building a deformation prediction model.

[0075] By calculating the factors affecting the deformation of the blade tip and quantifying the effects of temperature and wind force on blade stiffness, key parameters are provided for predicting the deformation of wind turbine blades under different environmental conditions, thereby optimizing the operation and maintenance strategies of wind turbines.

[0076] By extracting the characteristic information of blade temperature and wind data, calculating the environmental change evaluation value, considering the impact of environmental factors on wind turbines, generating a more accurate prediction value of wind turbine blade clearance value, and improving the accuracy of the prediction

[0077] By constructing a headroom value prediction curve, deviations in the prediction model can be discovered in a timely manner and the model can be adjusted accordingly to improve the accuracy of long-term predictions, enhance system reliability, and identify trends that may lead to equipment damage or safety accidents, so that preventive measures can be taken to reduce the occurrence of accidents.

[0078] Through sophisticated data analysis and multi-level deviation correction, the accuracy of wind turbine blade clearance value prediction is effectively improved, which has significant technical advantages for the operation and maintenance management of wind power generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 The present invention provides a flow chart of a method for predicting the clearance value of wind turbine blades. DETAILED DESCRIPTION

[0080] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0081] In the description of the present invention, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0082] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0083] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0084] like Figure 1 As shown, an embodiment of the present invention discloses a method for predicting the clearance value of blades of a wind turbine generator set, comprising:

[0085] Construct a blade clearance value prediction model based on historical environmental data;

[0086] Establish multiple adjustment cycles, obtain environmental data according to preset adjustment time nodes, and generate a clearance value prediction curve within a single adjustment cycle based on the environmental data and the blade clearance value prediction model;

[0087] Multiple monitoring time nodes are set in a single adjustment cycle to obtain the actual value of each monitoring point, and whether to correct the headroom value prediction curve is determined based on the deviation between the actual value and the headroom value prediction curve.

[0088] In order to prevent risks caused by insufficient clearance, in this embodiment, adjustment strategies can be generated based on the predicted clearance value of the wind turbine blades. These strategies include adjusting the operating parameters of the wind turbine, such as the pitch angle, to increase the distance between the blades and the tower to avoid collision. For example, the pitch angle can be automatically adjusted when it is predicted that the clearance value will drop to a critical level, thereby increasing the tower clearance and ensuring the safety of the unit in extreme wind conditions.

[0089] Embodiment 1:

[0090] The blade clearance value prediction model is constructed based on historical environmental data, including:

[0091] Conduct simulation tests based on historical environmental data and establish a blade deformation prediction model;

[0092] The blade tip deformation prediction value is generated through the blade deformation prediction model and the current environmental data, including:

[0093] Obtain historical meteorological data of the area where the wind turbine is located, and record the range of meteorological changes in the current area, including temperature change data and wind speed change data;

[0094] Based on the obtained temperature change data, determine the temperature change range of the current area, and divide the temperature within the change range into intervals, each interval representing a gradient value;

[0095] Based on the acquired wind force change data, determine the wind force change range in the current area, and develop a wind force gradient change mapping model to convert the wind force vector data into scalar force data acting on the blades;

[0096] Based on the temperature gradient value and wind gradient change mapping model, blade deformation test is carried out.

[0097] According to the results of blade deformation test, a blade deformation prediction model is constructed;

[0098] Generate a blade tip deformation prediction value through a blade deformation prediction model;

[0099] Based on the predicted value of the blade tip deformation degree, a predicted value of the blade clearance value is determined, and based on the predicted value of the blade clearance value, a clearance value prediction curve within a single adjustment cycle is generated.

[0100] In this embodiment:

[0101] Get historical weather data:

[0102] First, obtain historical meteorological data, including temperature change data and wind speed change data, from the meteorological station or relevant meteorological database in the area where the wind turbine is located. These data should cover a sufficiently long time range to reflect changes in different seasons and weather conditions.

[0103] Temperature change interval division:

[0104] Based on historical temperature data, the range of temperature changes is determined and subdivided into multiple temperature gradient intervals. Each interval represents a specific gradient value, for example, a temperature increase or decrease of 5°C is a gradient. This can quantify the impact of temperature changes on blade deformation.

[0105] Wind gradient mapping model establishment:

[0106] The wind force variation data is used to determine the wind force variation range. Then a wind force gradient variation mapping model is established to convert the wind force vector data into scalar force data perpendicular to the blade direction. The key to this step is to convert the wind force vector into the magnitude of the force that directly affects the blade.

[0107] Blade deformation test:

[0108] Conduct blade deformation tests in a laboratory or wind farm. During the test, simulate the environment under different meteorological conditions based on the temperature gradient value and wind gradient mapping model to observe the degree of blade deformation. The test data should include the blade deformation under different temperature and wind conditions.

[0109] Building a predictive model:

[0110] According to the results of the blade deformation test, a mathematical model is established or a machine learning algorithm is used to construct a blade deformation prediction model. The input of the model should include meteorological parameters such as temperature and wind force, and the output is the predicted value of the blade deformation degree. During the model construction process, training and verification are required to optimize the model parameters to ensure the accuracy of the prediction.

[0111] Model validation and optimization:

[0112] The model was validated using an independent historical data set to check the difference between the predicted value and the actual blade deformation and evaluate the prediction accuracy of the model. The model parameters were adjusted according to the validation results and necessary optimization was performed.

[0113] The constructed blade deformation prediction model can predict the deformation of blades under different environmental conditions based on real-time meteorological data, providing an important decision-making basis for the operation of wind turbines. This not only helps to improve the operating efficiency of wind turbines, but also effectively avoids safety risks caused by excessive blade deformation.

[0114] Embodiment 2:

[0115] The blade deformation test comprises:

[0116] Get the temperature gradient test data set T, T = [t1, t2…t i …t n ];

[0117] Among them, t i represents the gradient value of the i-th temperature interval; n represents the number of intervals;

[0118] Get wind test dataset

[0119] represents the wind vector in the ith case n represents the total number of wind vector situations;

[0120] Through experiments, a wind mapping model is established to convert the wind vector Convert to wind scalar f i ;

[0121] Generate a scalar force test data set F acting on the blade, F = [f1,f2…f i …f n ];

[0122] where f i represents the wind vector in the ith case The corresponding scalar force data acting on the blade, n represents the total number of wind force value data;

[0123] Obtain the deformation distance Li of the blade tip in the i-th case;

[0124] Generate a blade deformation model:

[0125] Li=K1*t i +K2*f i ;

[0126] Where Li represents the deformation distance of the blade tip in the i-th case, K1 represents the influence coefficient of temperature on blade stiffness, and K2 represents the influence coefficient of wind force on blade stiffness;

[0127] The values ​​of K1 and K2 are calculated by the control variable method.

[0128] In this embodiment:

[0129] For each wind condition, the angle α between the wind vector and the normal vector of the wind turbine blade is calculated.

[0130] Convert wind vector to scalar force data set F

[0131] Based on the angle α between the wind force dataset W and the normal vector, the scalar force dataset F perpendicular to the blade direction is generated using the vector decomposition principle. Each fi is the projection value of the wind force vector in the direction of the blade normal vector.

[0132] Arrange test points and record deformation

[0133] Monitoring points are arranged at the blade tip to record the blade deformation distance L under different temperature and wind conditions.

[0134] Simulation test and calculation of influencing factors

[0135] Through simulation experiments, the effects of temperature and wind force on blade tip deformation are analyzed.

[0136] Calculate the influence coefficient: determine the influence coefficient K1 of temperature on blade stiffness and the influence coefficient k2 of wind force on blade stiffness.

[0137] Calculate deformation distance L1

[0138] The deformation distance L1 of the blade tip is calculated using the influence coefficients K1 and k2, combined with the temperature gradient value ti and the wind force value fi.

[0139] The formula is: L1=K1*ti+k2*fi.

[0140] The calculation of the angle α should be based on the geometric structure of the blade and the wind direction, using trigonometric functions or vector operations.

[0141] The generation of the scalar force data set F is to decompose the wind force vector in a specific direction to ensure the accuracy of the calculation.

[0142] The monitoring of deformation distance L requires the use of high-precision sensors to ensure the reliability of the data.

[0143] The determination of the influence coefficients K1 and k2 may require multiple experiments and data analysis.

[0144] Embodiment 3:

[0145] The calculation of the values ​​of K1 and K2 by the control variable method includes:

[0146] Calculate the influence coefficient K1 of temperature on blade stiffness:

[0147] (1) Obtain the initial stiffness value E0 and the initial temperature value T0, adjust the temperature value to T1, and obtain the current stiffness value E1; at this time, calculate the influence coefficient K1 of temperature on blade stiffness;

[0148]

[0149] Calculate the influence coefficient K2 of wind force on blade stiffness:

[0150] (2) Obtaining the initial stiffness value E0 and the initial wind force value F0, adjusting the wind force value to F1, and obtaining the current stiffness value E1; at this time, calculating the influence coefficient K2 of the wind force on the blade stiffness;

[0151]

[0152] In this example:

[0153] Data accuracy: Ensure accurate E0 and E1 measurements under different conditions to improve the reliability of K1 and k2 calculations.

[0154] Temperature and wind control: When adjusting the temperature and wind speed, ensure the control accuracy during the change process and avoid interference from other factors.

[0155] Verification of multiple sets of data: To improve the accuracy of the calculation results, it is recommended to repeat the experiment under different temperature and wind conditions, calculate multiple sets of K1 and k2, and then take the average value or perform statistical analysis.

[0156] Through the above method, the influence of temperature and wind force on blade stiffness can be quantified, providing key parameters for predicting the deformation of wind turbine blades under different environmental conditions, thereby optimizing the operation and maintenance strategy of wind turbines.

[0157] Embodiment 4:

[0158] The generating of the blade tip deformation prediction value includes:

[0159] Based on remote sensing data, generate temperature change prediction data PT, PT = [pt1, pt2…pt t …pt T ];

[0160] Wind force change prediction data PW,PW=[pw1,pw2…pw t …pw T ];

[0161] Generate a scalar force data set PF, PF = [pf1, pf2…pf t …pf T ];

[0162] Among them, pt t represents the predicted temperature value at time t, pw t represents the wind force forecast value at time t, pf t represents the predicted value of the scalar force at time t, and T is the total number of predicted moments;

[0163] Calculate the deformation distance Lx of the blade tip based on the blade deformation prediction model;

[0164] Lx=K1*pt t +K2*pf t .

[0165] In this example:

[0166] Remote sensing data processing: Ensure the quality and update frequency of remote sensing data to improve the accuracy of temperature and wind forecasts.

[0167] Model parameter optimization: Regularly calibrate the influence coefficients K1 and k2 to ensure the model prediction accuracy.

[0168] Data synchronization: ensure the time consistency of temperature, wind and scalar force prediction data to avoid errors caused by data time misalignment.

[0169] Changes in environmental conditions: If there are abnormal fluctuations in environmental conditions during the forecast period, the model parameters should be adjusted in time to improve the adaptability and accuracy of the forecast.

[0170] Data fusion: Integrate multiple data sources (such as ground observation stations, satellite remote sensing, etc.) to enhance the reliability and completeness of predictions.

[0171] Embodiment 5:

[0172] The generating of the blade clearance value prediction value includes:

[0173] Get the actual value of the blade clearance value and generate the actual value data set L of the blade clearance value, L = [L1, L2…L t …L T ];

[0174] Among them, L1 represents the actual value of the blade clearance at the current moment, L t It represents the actual value of the blade clearance at time t, and T represents the time length;

[0175] Based on the deformation distance Lx of the blade tip, the predicted change value Ly of the clearance value is calculated;

[0176] Ly = Lx*sinB;

[0177] Wherein, B is the current pitch angle of the wind turbine;

[0178] Generate the predicted value Lo of the first-level headroom value at time t;

[0179] Lo=L1-Ly

[0180] =L1-Lx*sinB

[0181] =L1-sinB*(K1*pt t +K2*pf t );

[0182] The first-level clearance value prediction value Lo is corrected to generate the blade clearance value prediction value Lz, Lz = [Lz1, Lz2…Lz t …Lz T ];

[0183] Among them, Lz t It represents the predicted value of the blade clearance value at time t, and T represents the time length.

[0184] In this embodiment:

[0185] Get the actual value data set of the blade clearance value

[0186] Data collection

[0187] The clearance value monitoring equipment installed on the wind turbine, such as a laser rangefinder, is used to continuously collect blade clearance value data.

[0188] Assuming that the monitoring device collects data at a certain time interval (such as 1 minute), after a period of time T (assuming T=60 minutes), a series of actual values ​​of the blade clearance values ​​are obtained.

[0189] For example, at t = 1 minute, the actual blade clearance value L1 = 10 meters (this is just an assumed value); at t = 2 minutes, L2 = 9.8 meters, etc., resulting in a data set L = [L1, L2, ..., Lt, ..., LT].

[0190] The tip deformation distance is determined based on the tip deformation distance calculation clearance value prediction change value, and the tip deformation distance L2 of the blade is obtained through the sensor or structural analysis model of the wind turbine.

[0191] Embodiment 6:

[0192] The correction of the first-level headroom value prediction value Lo includes:

[0193] Based on the actual value L of the current blade clearance value at time t t With the first-level headroom value prediction value Lo, establish the correction model θ(t);

[0194] θ(t)=L t -Lo

[0195] =L t -(L1-sinB*(K1*pt t +K2*pf t ))

[0196] Based on the correction model θ(t), a correction strategy is formulated to correct the first-level headroom value prediction value Lo.

[0197] In this embodiment:

[0198] In the scenario of wind turbine blade clearance value prediction, the correction of the first-level clearance value prediction value Lo is aimed at improving the accuracy of the prediction and ensuring the safe operation of the wind turbine. By establishing a correction model θ(t) and adjusting it according to the difference between the actual blade clearance value Lt and the predicted value Lo, the prediction error can be reduced and the potential collision risk caused by inaccurate prediction can be avoided. The implementation of this correction technology can bring the following technical effects:

[0199] Improve forecast accuracy: Through real-time correction, the forecast value can be made closer to the actual value and the forecast deviation can be reduced.

[0200] Enhanced safety: Accurate clearance prediction helps to take timely protective measures to prevent blade-tower collisions.

[0201] Optimize maintenance planning: More accurate forecasts can help operations teams plan maintenance work more effectively and reduce unplanned downtime.

[0202] Improve economic benefits: By reducing maintenance costs and production losses caused by forecasting errors, the economic benefits of the entire wind turbine can be improved.

[0203] Quantitative indicators of technical effects may include the percentage reduction of prediction errors, the reduction of safety accident rates, the amount of maintenance cost savings, etc. In actual cases, the technical effect can be evaluated by comparing the prediction data before and after the correction with the actual operation data.

[0204] Embodiment 7:

[0205] The formulation of the correction strategy based on the correction model θ(t) includes:

[0206] Group the values ​​of θ(t) into equally spaced intervals;

[0207] Count the number of data points in each interval, then draw a histogram with the interval as the horizontal axis and the data or frequency as the vertical axis; form correction instructions by observing the shape of the histogram.

[0208] Embodiment 8:

[0209] The forming of the correction instruction includes:

[0210] If the histogram is close to a normal distribution, set a sliding window containing m time nodes and obtain the value range of θ(t) in the sliding window, θ(t)min≤θ(t)≤θ(t)max;

[0211] Among them, θ(t)min represents the minimum value of θ(t) in the sliding window, and θ(t)max represents the maximum value of θ(t) in the sliding window;

[0212] θ(t)max is used as the first deviation value C1 for predicting the clearance value of the wind turbine generator set;

[0213] Lz=Lo+C1;

[0214] If the histogram is skewed or multimodal, the data acquisition device should be calibrated to determine whether there is any setting error in the data acquisition device.

[0215] In this embodiment, a series of values ​​of the correction model θ(t) are calculated through the previous steps, and a histogram is drawn to find that it is approximately normally distributed.

[0216] For example, we have 100 values ​​of θ(t) corresponding to different time points t, with a time interval of 1 minute, so the total duration is 100 minutes.

[0217] Setting the sliding window

[0218] The sliding window is set to contain m=10 time nodes.

[0219] Starting from t=1 to t=10 is the first sliding window, and the value range of θ(t) in this window is calculated.

[0220] Assume that within this window, θ(t)min=-0.5 (the unit is assumed to be meter), θ(t)max=0.8 meter.

[0221] Calculate the first deviation value C1 and correct the predicted value

[0222] θ(t)max is used as the first deviation value C1 for predicting the clearance value of the wind turbine generator set, that is, C1=0.8 meters.

[0223] Assuming that the predicted value Lo of the first-level clearance value calculated previously is 5 meters at a certain moment, according to the correction formula LZ=Lo+C1, we get LZ=5+0.8=5.8 meters.

[0224] As time goes by, the sliding window keeps moving, for example, moving to the window from t=11 to t=20, recalculating θ(t)min and θ(t)max, and updating the value of C1, thereby continuously correcting the value of LZ.

[0225] Based on the value of θ(t), a histogram is plotted and found to be skewed (eg, right-skewed) or multimodal.

[0226] Data acquisition equipment calibration check

[0227] First, conduct a comprehensive inspection of the data acquisition equipment.

[0228] For example, check the temperature sensor to see if it is faulty or incorrectly set. If the measurement range of the temperature sensor is set incorrectly, the collected temperature data may be inaccurate, which in turn affects the prediction of the blade clearance value.

[0229] Then check the wind measurement equipment to see if it is calibrated correctly. If the calibration coefficient of the wind measurement equipment is wrong, the collected wind data will be biased, affecting the wind-related calculations in the entire prediction model, such as blade deformation calculations.

[0230] Assume that during the inspection process, it is found that the zero point of the temperature sensor is set incorrectly. After recalibrating it to the correct zero point, recollect data and calculate the value of θ(t), and draw the histogram again to check whether its shape has improved. If the histogram shape is still abnormal, continue to check other factors that may affect data collection, such as whether the frequency of data collection is appropriate and whether the sensor is installed in the correct position.

[0231] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention. These improvements and substitutions should also be regarded as the scope of protection of the present invention.

Claims

1. A method for predicting the clearance value of wind turbine blades, characterized in that: include: Construct a blade clearance value prediction model based on historical environmental data; Establish multiple adjustment cycles, obtain environmental data according to preset adjustment time nodes, and generate a clearance value prediction curve within a single adjustment cycle based on the environmental data and the blade clearance value prediction model; Multiple monitoring time nodes are set in a single adjustment cycle to obtain the actual value of each monitoring point, and whether to correct the headroom value prediction curve is determined based on the deviation between the actual value and the headroom value prediction curve.

2. The wind turbine blade clearance value prediction method according to claim 1, characterized in that: The blade clearance value prediction model is constructed based on historical environmental data, including: Conduct simulation tests based on historical environmental data and establish a blade deformation prediction model; The blade tip deformation prediction value is generated through the blade deformation prediction model and the current environmental data, including: Obtain historical meteorological data of the area where the wind turbine is located, and record the range of meteorological changes in the current area, including temperature change data and wind speed change data; Based on the obtained temperature change data, determine the temperature change range of the current area, and divide the temperature within the change range into intervals, each interval representing a gradient value; Based on the acquired wind force change data, determine the wind force change range in the current area, and develop a wind force gradient change mapping model to convert the wind force vector data into scalar force data acting on the blades; Based on the temperature gradient value and wind gradient change mapping model, blade deformation test is carried out; According to the results of blade deformation test, a blade deformation prediction model is constructed; Generate a blade tip deformation prediction value through a blade deformation prediction model; Based on the predicted value of the blade tip deformation degree, the predicted value of the blade clearance value is calculated to generate a clearance value prediction curve.

3. The wind turbine blade clearance value prediction method according to claim 2, characterized in that: The blade deformation test comprises: Get the temperature gradient test data set T, T = [t1, t2…t i …t n ]; Among them, t i represents the gradient value of the i-th temperature interval; n represents the number of intervals; Get the wind test dataset W, represents the wind vector in the ith case n represents the total number of wind vector situations; Through experiments, a wind mapping model is established to convert the wind vector Convert to wind scalar f i ; Generate a scalar force test data set F acting on the blade, F = [f1,f2…f i …f n ]; where f i represents the wind vector in the ith case The corresponding scalar force data acting on the blade, n represents the total number of wind force value data; Obtain the deformation distance Li of the blade tip in the i-th case; Generate a blade deformation model: Li=K1*t i +K2*f i ; Where Li represents the deformation distance of the blade tip in the i-th case, K1 represents the influence coefficient of temperature on blade stiffness, and K2 represents the influence coefficient of wind force on blade stiffness; The values ​​of K1 and K2 are calculated by the control variable method.

4. The wind turbine blade clearance value prediction method according to claim 3, characterized in that: The calculation of the values ​​of K1 and K2 by the control variable method includes: Calculate the influence coefficient K1 of temperature on blade stiffness: (1) Obtain the initial stiffness value E0 and the initial temperature value T0, adjust the temperature value to T1, and obtain the current stiffness value E1; at this time, calculate the influence coefficient K1 of temperature on blade stiffness; Calculate the influence coefficient K2 of wind force on blade stiffness: (2) Obtaining the initial stiffness value E0 and the initial wind force value F0, adjusting the wind force value to F1, and obtaining the current stiffness value E1; at this time, calculating the influence coefficient K2 of the wind force on the blade stiffness; 5. The method for predicting the wind turbine blade clearance value according to claim 4, characterized in that: The generating of the blade tip deformation prediction value includes: Based on remote sensing data, generate temperature change prediction data PT, PT = [pt1, pt2…pt t …pt T ]; Wind force change prediction data PW,PW=[pw1,pw2…pw t …pw T ]; Generate a scalar force data set PF, PF = [pf1, pf2…pf t …pf T ]; Among them, pt t represents the predicted temperature value at time t, pw t represents the wind force forecast value at time t, pf t represents the predicted value of the scalar force at time t, and T is the total number of predicted moments; Calculate the deformation distance Lx of the blade tip based on the blade deformation prediction model; Lx=K1*pt t +K2*pf t 。 6. The method for predicting the wind turbine blade clearance value according to claim 5, characterized in that: The generating of the blade clearance value prediction value includes: Get the actual value of the blade clearance value and generate the actual value data set L of the blade clearance value, L = [L1, L2…L t …L T ]; Among them, L1 represents the actual value of the blade clearance at the current moment, L t It represents the actual value of the blade clearance at time t, and T represents the time length; Based on the deformation distance Lx of the blade tip, the predicted change value Ly of the clearance value is calculated; Ly = Lx*sinB; Among them, B is the current pitch angle of the wind turbine; Generate the predicted value Lo of the first-level headroom value at time t; Lo=L1-Ly =L1-Lx*sinB =L1-sinB*(K1*pt t +K2*pf t ); The first-level clearance value prediction value Lo is corrected to generate the blade clearance value prediction value Lz, Lz = [Lz1, Lz2…Lz t …Lz T ]; Among them, Lz t It represents the predicted value of the blade clearance value at time t, and T represents the time length.

7. The method for predicting the wind turbine blade clearance value according to claim 6, characterized in that: The correction of the first-level headroom value prediction value Lo includes: Based on the actual value L of the current blade clearance value at time t t With the first-level headroom value prediction value Lo, establish the correction model θ(t); θ(t)=L t -The =L t -(L1-sinB*(K1*pt t +K2*pf t )) Based on the correction model θ(t), a correction strategy is formulated to correct the first-level headroom value prediction value Lo.

8. The method for predicting the wind turbine blade clearance value according to claim 7, characterized in that: The formulation of the correction strategy based on the correction model θ(t) includes: Group the values ​​of θ(t) into equally spaced intervals; Count the number of data points in each interval, then draw a histogram with the interval as the horizontal axis and the data or frequency as the vertical axis; form correction instructions by observing the shape of the histogram.

9. The method for predicting the wind turbine blade clearance value according to claim 8, characterized in that: The forming of the correction instruction includes: If the histogram is close to a normal distribution, set a sliding window containing m time nodes and obtain the value range of θ(t) in the sliding window, θ(t)min≤θ(t)≤θ(t)max; Among them, θ(t)min represents the minimum value of θ(t) in the sliding window, and θ(t)max represents the maximum value of θ(t) in the sliding window; θ(t)max is used as the first deviation value C1 for predicting the clearance value of the wind turbine generator set; Lz=Lo+C1; If the histogram is skewed or multimodal, the data acquisition device should be calibrated to determine whether there is any setting error in the data acquisition device.

Images