Method and system for obtaining yaw deviation of wind turbine generator set

Abstract

The present application relates to a method and system for obtaining the yaw deviation of a wind turbine generator set, which belongs to the field of machine learning technology. The method for obtaining the yaw deviation of a wind turbine generator set includes obtaining an optimal data set containing multiple groups of working condition data above the rated wind speed; obtaining wind-containing power; dividing each group of working condition data above the rated wind speed into a corresponding wind deviation interval; obtaining a power mean fitting function; determining the optimal wind deviation interval [x1, x2], and obtaining a yaw deviation value Yawmiss; the yaw deviation acquisition system of the wind turbine generator set includes an optimal data set acquisition module, a wind-containing power acquisition module, a wind deviation interval division module, a power mean fitting function acquisition module, and a yaw deviation acquisition module. Compared with the related art, the present application has the effect of improving the problem of errors in the measured yaw deviation results.

Description

Technical Field

[0001] The present application relates to the field of machine learning technology, and in particular to a method and system for obtaining yaw deviation of a wind turbine generator set. Background Art

[0002] Wind turbines include components such as rotors, generators, and weather vanes. The rotors contain blades, hubs, and reinforcements. Wind turbines have functions such as blades rotating under wind power to generate electricity and generator heads rotating. During the initial installation of wind turbines, due to installation errors, the wind vane being located at the tail of the nacelle, and the measurement results being affected by factors such as the rotor, nacelle, improper maintenance, and measurement and signal links, large wind turbines will have varying degrees of wind deviation during operation. Wind deviation will result in a loss in the output performance of the unit, which will affect the power generation of the wind turbine and cause economic losses.

[0003] In order to correct the wind deviation, manual calibration or installation of laser radar measurement can be used to evaluate the wind situation of the entire unit, but both require a lot of manpower and material resources and are costly. In recent years, a variety of yaw deviation calculation methods based on operating data have been proposed. These calculation methods are based on statistical comparison of power in different wind speed ranges and different yaw angles, and comprehensive analysis of various wind speed ranges to find the highest point of power with respect to the yaw position, which is considered to be the optimal wind angle.

[0004] With respect to the above-mentioned related technologies, the inventor believes that the above methods do not take into account that wind speed measurement will be affected by a variety of factors. Under some wind turbine models, wind speed measurement is affected by the impeller and will be offset at different wind angles. When wind speed measurement is offset, the wind speed interval division will be distorted, so that the highest point of power at different wind angles will also be distorted, resulting in errors in the measured yaw deviation results. Summary of the invention

[0005] In order to improve the problem of errors in the measured yaw deviation results, the present application provides a method and system for obtaining the yaw deviation of a wind turbine generator set.

[0006] In a first aspect, the present application provides a method for obtaining the yaw deviation of a wind turbine generator set, which adopts the following technical solution:

[0007] A method for obtaining a yaw deviation of a wind turbine generator set, the method comprising:

[0008] Based on multiple groups of historical operating data of wind turbines, combined with normalization method and automatic clustering method, an optimal data set containing multiple groups of operating data above rated wind speed is obtained, each group of historical operating data includes but is not limited to the yaw angle to wind, wind direction, power, pitch position, speed and unit status of the wind turbine at the same time;

[0009] According to the optimization data set, multiple pitch positions are obtained, and the wind speed V is obtained according to the pitch positions, and the wind power P is obtained according to the wind speed V. wind ,

[0010]

[0011] Where ρ is the air density, V is the wind speed, and S is the impeller swept area;

[0012] The normal power generation interval angle is divided to obtain multiple wind deviation intervals. The yaw wind angle of each group of working condition data above the rated wind speed is used as the interval division standard, and each group of working condition data above the rated wind speed is divided into the corresponding wind deviation interval;

[0013] Based on wind power P wind , get the equivalent power mean of each wind deviation interval, fit all equivalent power means to get the power mean fitting function; and,

[0014] The continuous derivative result is obtained by continuously derivation of the power mean fitting function. Combining the optimal power position determination rule and the continuous derivative result, the optimal wind deviation interval [x1, x2] is determined, and the yaw deviation value Yawmiss is obtained.

[0015] Yawmiss=(x1+x2) / 2 (2),

[0016] Among them, x1 represents the starting value of the interval angle of the optimal wind deviation interval, and x2 represents the ending value of the interval angle of the optimal wind deviation interval.

[0017] By adopting the above technical scheme, an optimal data set containing multiple groups of operating data above the rated wind speed is extracted from the historical operating data through an unsupervised learning method, and multiple wind-containing powers of the optimal data set are obtained through the wind-containing power calculation formula. At the same time, through the yaw wind angle in the optimal data set, each group of operating data above the rated wind speed in the optimal data set is divided into a corresponding wind deviation interval, and then the equivalent power mean of each wind deviation interval obtained based on the wind-containing power is fitted to obtain a power mean fitting function, and the power mean fitting function is continuously differentiated and the optimal power position is determined to obtain the optimal wind deviation, so as to obtain the yaw deviation, without dividing the wind speed interval or judging the wind speed. To a certain extent, the influence of wind speed measurement factors is avoided, which helps to improve the problem of errors in the measured yaw deviation results.

[0018] Optionally, the specific method of extracting the optimization data set includes:

[0019] Acquire multiple groups of historical operating condition data of the wind turbine generator set, each group of the historical operating condition data includes but is not limited to the yaw angle to wind, wind direction, power, pitch position, speed and unit status of the wind turbine generator set;

[0020] Each set of historical operating condition data is processed using the normalization function Fone to obtain data set One;

[0021] Using a Gaussian mixture model to automatically perform clustering calculations on data set One, dividing data set One into multiple new data sets; and,

[0022] Based on multiple new data sets and combined with extraction methods, an optimal data set is obtained.

[0023] By adopting the above technical solution, the normalization function Fone is used to simplify the calculation of each group of historical operating condition data to obtain the data set One, and then the data set One is quantified through the Gaussian mixture model, and the automatic clustering calculation is performed. This unsupervised machine learning method is divided into multiple new data sets, and then combined with the extraction method, the optimal data set is obtained. The whole process is completed automatically, and the optimal data set can be obtained quickly, which helps to improve the problem of errors in the measured yaw deviation results.

[0024] Optionally, the extraction method comprises:

[0025] The above-rated index calculation method is used to obtain the above-rated index of each data element in each new data set;

[0026] Based on the rated above index of each data element, obtain the index mean of each new data set; and,

[0027] Compare the index means of each new data set, and the one with the largest index mean is the optimal data set.

[0028] By adopting the above technical solution and with the help of the rated-above index calculation method, the rated-above index of each data element in each new data set can be quickly obtained, and the index mean of each new data set can be obtained based on the rated-above index, and finally the optimal data set with the largest index mean can be obtained, which facilitates the rapid acquisition of optimal data.

[0029] Optionally, the specific method for obtaining the optimal wind deviation interval [x1, x2] includes:

[0030] Continuously derive the power mean fitting function;

[0031] According to the continuous derivation results and the optimal power position determination rule, the optimal power of the power mean fitting function is obtained; and,

[0032] According to the wind deviation interval where the optimal power is located, the optimal wind deviation interval [x1, x2] is obtained.

[0033] By adopting the above technical solution, after continuously differentiating the power mean fitting function, multiple extreme points are obtained, and then the optimal power of the power mean fitting function can be determined by using the optimal power position determination rule, so that the optimal wind deviation interval where the optimal power is located can be obtained.

[0034] Optionally, the optimal power position determination rule includes that, among multiple extreme value points of the continuous derivative results, the extreme value closest to zero is the optimal power.

[0035] By adopting the above technical solution, the zero value point is the turning point of the increase or decrease of the power mean value fitting function, so the point closest to the zero value is the optimal power point.

[0036] Optionally, the optimal power position determination rule also includes: among the multiple extreme points of the continuous derivative results, the extreme values ​​on both sides of the zero value are median values, and the one with the smallest absolute value among the medians is the optimal power.

[0037] By adopting the above technical solution, the zero point is the turning point of the increase or decrease of the power mean fitting function, so the point with the smallest absolute value among the medians is the optimal power point.

[0038] In a second aspect, the present application provides a yaw deviation acquisition system for a wind turbine generator set, which adopts the following technical solution:

[0039] A yaw deviation acquisition system for a wind turbine generator set, the system comprising:

[0040] An optimization data set acquisition module is used to obtain an optimization data set containing multiple sets of operating data above the rated wind speed based on multiple sets of historical operating data of the wind turbine generator set, in combination with a normalization method and an automatic clustering method, wherein each set of the historical operating data includes but is not limited to the yaw angle to wind, wind direction, power, pitch position, speed and unit status of the wind turbine generator set at the same time;

[0041] The wind power acquisition module is used to obtain multiple pitch positions according to the optimization data set, and obtain the wind speed V according to the pitch positions, and obtain the wind power P according to the wind speed V. wind ,

[0042]

[0043] Where ρ is the air density, V is the wind speed, and S is the impeller swept area;

[0044] The wind deviation interval division module is used to divide the normal power generation interval angle to obtain multiple wind deviation intervals, and the yaw wind angle of each group of working condition data above the rated wind speed is used as the interval division standard, and each group of working condition data above the rated wind speed is divided into the corresponding wind deviation interval;

[0045] The power mean fitting function acquisition module is used to obtain the wind power P wind , get the equivalent power mean of each wind deviation interval, fit all equivalent power means to get the power mean fitting function; and,

[0046] The yaw deviation acquisition module is used to continuously derive the power mean fitting function to obtain the continuous derivative result, combine the optimal power position determination rule and the continuous derivative result, determine the optimal wind deviation interval [x1, x2], and obtain the yaw deviation value Yawmiss.

[0047] Yawmiss=(x1+x2) / 2 (2),

[0048] Among them, x1 represents the starting value of the interval angle of the optimal wind deviation interval, and x2 represents the ending value of the interval angle of the optimal wind deviation interval.

[0049] By adopting the above technical scheme, the optimal data set acquisition module extracts the optimal data set containing multiple groups of working condition data above the rated wind speed from the historical working condition data through an unsupervised learning method, the wind-containing power acquisition module obtains multiple wind-containing powers of the optimal data set through the wind-containing power calculation formula, the wind deviation interval division module divides each group of working condition data above the rated wind speed in the optimal data set into the corresponding wind deviation interval through the yaw wind angle in the optimal data set, the power mean fitting function acquisition module fits the equivalent power mean of each wind deviation interval obtained based on the wind-containing power to obtain the power mean fitting function, the yaw deviation acquisition module continuously derives the power mean fitting function and determines the optimal power position to obtain the optimal wind deviation, so as to obtain the yaw deviation, without dividing the wind speed interval or judging the wind speed, thus avoiding the influence of wind speed measurement factors to a certain extent and greatly improving the problem of error in the measured yaw deviation result.

[0050] In a third aspect, the present application provides a smart terminal, which adopts the following technical solution:

[0051] An intelligent terminal comprises a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and executed by any method in the first aspect.

[0052] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution:

[0053] A computer-readable storage medium stores a computer program that can be loaded by a processor and execute any method according to the first aspect of claim 1. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1This is a first flow chart of the method for obtaining the yaw deviation of a wind turbine generator set according to an embodiment of the present application.

[0055] Figure 2 This is a second flow chart of the method for obtaining the yaw deviation of a wind turbine generator set according to an embodiment of the present application.

[0056] Figure 3 This is a third flow chart of the method for obtaining the yaw deviation of a wind turbine generator set according to an embodiment of the present application.

[0057] Figure 4 This is a fourth flow chart of the method for obtaining the yaw deviation of a wind turbine generator set according to an embodiment of the present application.

[0058] Figure 5 It is a structural block diagram of the yaw deviation acquisition system of the wind turbine generator set in an embodiment of the present application. DETAILED DESCRIPTION

[0059] The following is combined with Figure 1-5 This application is described in further detail.

[0060] The embodiment of the present application discloses a method for obtaining the yaw deviation of a wind turbine generator set. Figure 1 and Figure 2 , the acquisition method includes the following steps:

[0061] 101. Based on multiple sets of historical operating data of wind turbines, combined with normalization method and automatic clustering method, an optimal data set containing multiple sets of operating data above rated wind speed is obtained.

[0062] Each set of historical operating condition data includes, but is not limited to, variables such as the yaw angle to wind, wind direction, power, pitch position, speed, and unit status of the wind turbine generator set at the same time. The speed here refers to the impeller speed. For example, it may also include ambient temperature, generator speed, etc.

[0063] Normalization is a way to simplify calculations, which is to transform a dimensional expression into a dimensionless expression, thus becoming a scalar.

[0064] Automatic clustering method is a typical unsupervised machine learning method. Clustering attempts to divide the samples in the data set into several usually non-overlapping subsets, each of which becomes a cluster. Through such division, each cluster may correspond to some potential concepts.

[0065] It should be noted that a set of operating data above the rated wind speed is a data element of the optimization data set. The process of extracting operating data above the rated wind speed is also the process of automatic operating condition identification. The automatic operating condition identification adopts a common automatic operating condition identification method, such as identification based on power and unit status, which will not be further elaborated in this embodiment.

[0066] 102. According to the optimization data set, multiple pitch positions are obtained, and the wind speed V is obtained according to the pitch positions, and the wind power P is obtained according to the wind speed V. wind .

[0067] in,

[0068]

[0069] And ρ is the air density, V is the wind speed, and S is the impeller swept area. Among them, the impeller swept area can be calculated using the circle area formula based on the parameters of the impeller diameter, and the air density can be obtained by measurement.

[0070] The power value of the selected working condition data above the rated wind speed is near the rated power, and the power is controlled within a stable range through the pitch control. According to the operating principle of the variable pitch wind turbine, at the same power level, the greater the power, the larger the pitch position is required to maintain a stable power output. Therefore, the wind power P can be calculated based on the pitch position data. wind size.

[0071] It should be noted that there are multiple data elements about working conditions above the rated wind speed in the optimization data set, and there are multiple pitch positions in the optimization data set, so the wind power P wind The corresponding relationship between wind speed and pitch position can be obtained according to the design parameters of the actual wind turbine generator set, or can be obtained by fitting a large amount of historical operation data, which will not be further elaborated in this embodiment.

[0072] 103. The normal power generation interval angle is divided to obtain multiple wind deviation intervals. The yaw wind angle of each group of operating data above the rated wind speed is used as the interval division standard, and each group of operating data above the rated wind speed is divided into the corresponding wind deviation interval.

[0073] The normal power generation interval angle is the windward interval angle when the wind turbine is working normally, which is generally -30° to +30°. In this embodiment, each 1° angle is used as an interval interval, so 60 windward deviation intervals are obtained. When the size of the yaw windward angle of each set of working condition data above the rated wind speed is within a certain windward deviation interval, the set of working condition data above the rated wind speed is classified into this windward deviation interval.

[0074] It should be noted that step 101 may be performed first, then step 102, and finally step 103; or step 101 may be performed first, then step 103, and finally step 102.

[0075] 104. Based on wind power P wind , get the equivalent power mean of each wind deviation interval, fit all equivalent power means, and get the power mean fitting function.

[0076] There are multiple working conditions above the rated wind speed in each wind deviation interval, so there are multiple pitch angles, that is, multiple wind power P wind , multiple winds with power P wind The average value is added to obtain the equivalent power mean value for each wind deviation interval.

[0077] 105. The power mean fitting function is continuously differentiated to obtain a continuous differentiation result. The optimal power position determination rule and the continuous differentiation result are combined to determine the optimal wind deviation interval [x1, x2], and the yaw deviation value Yawmiss is obtained.

[0078] in,

[0079] Yawmiss=(x1+x2) / 2 (2),

[0080] Among them, x1 represents the starting value of the interval angle of the optimal wind deviation interval, and x2 represents the ending value of the interval angle of the optimal wind deviation interval. For example, the optimal wind deviation interval may be [-30°, -29°], at this time, the yaw deviation value Yawmiss = -29.5°. The process of obtaining the yaw deviation value Yawmiss according to the optimal wind deviation interval [x1, x2] is also the process of estimating according to the optimal wind deviation interval.

[0081] It can be concluded from step 104 and step 105 that step 104 and step 105 are a process of searching for an optimum based on equivalent power.

[0082] In the above implementation, an optimal data set containing multiple groups of working condition data above the rated wind speed is extracted from the historical working condition data by an unsupervised learning method, and multiple wind power P of the optimal data set are obtained by the wind power calculation formula. wind At the same time, through the yaw angle of the optimal data set, each group of working condition data above the rated wind speed in the optimal data set is divided into the corresponding wind deviation interval, and then the wind power P is calculated based on the wind power P. windThe equivalent power mean of each wind deviation interval is fitted to obtain the power mean fitting function, and the power mean fitting function is continuously differentiated and the optimal power position is determined to obtain the optimal wind deviation, so as to obtain the yaw deviation. There is no need to divide the wind speed interval or judge the wind speed. To a certain extent, the influence of wind speed measurement factors is avoided, which helps to improve the problem of error in the measured yaw deviation results.

[0083] Reference Figure 2 and Figure 3 ,The specific steps of extracting the optimization data set are as follows:

[0084] 201. Obtain multiple sets of historical operating data of wind turbine generator sets.

[0085] Each set of historical operating data includes but is not limited to variables such as the yaw angle to wind, power, pitch position, speed and unit status of the wind turbine. The historical operating data is obtained from the SCADA system, that is, the variables are the time series data in the SCADA data record. The wind direction is measured by the wind vane, the power is obtained by the generator grid-connected power, the pitch position is obtained according to the pitch angle, the speed is obtained by the speed sensor, the unit status includes shutdown, grid connection and failure, etc., and the yaw angle to wind is obtained by the yaw encoder. In this embodiment, a set of historical operating data is obtained every 10 seconds with a sampling interval of 10 seconds. Each set of historical operating data is composed of a vector X t ,

[0086] X t =(wind speed , wind direction , power, pitch, speed, state) (3).

[0087] At each time t, there is a vector X t Produced, and together form a X t is a data set A with elements. At this time, each set of historical operating condition data is a data element of data set A. speed Represents the wind speed V, which is obtained by the variable pitch position. direction Represents wind direction, power represents power, pitch represents yaw angle to wind, speed represents rotation speed, and state represents unit status.

[0088] 202. Use normalization function Fone to process each set of historical operating condition data to obtain data set One.

[0089] That is, the normalization function Fone is used to normalize each data element in the data set A to obtain a data set One, that is,

[0090] One=Fone(A) (4),

[0091] Among them, Fone performs the following calculation on each element X:

[0092]

[0093] That is, the variables such as wind deviation angle, wind direction, power, pitch position, speed and unit status in each set of data above rated wind speed conditions are proportionally calculated according to the physical range and dimension of the variable words, so as to transform the value of each variable to a quantitative level. rated Indicates rated power.

[0094] 203. Use the Gaussian mixture model to automatically perform clustering calculation on the data set One, and divide the data set One into multiple new data sets.

[0095] Among them, the Gaussian mixture model uses the Gaussian probability density function, that is, the normal distribution curve, to accurately quantify things. It is a model that decomposes things into several Gaussian probability density functions (normal distribution curves).

[0096] The specific method of automatic clustering calculation adopts the DBSCAN method which has been widely verified by academia and engineering. DBSCAN is a relatively representative density-based clustering algorithm, which defines a cluster as the largest set of density-connected points, can divide areas with sufficiently high density into clusters, and can find clusters of any shape in noisy spatial databases.

[0097] In this implementation, the Python-based open source code library sklearn.cluster.DBSCAN is introduced. The specific calculation code is as follows:

[0098] db=DBSCAN(eps=0.3, min_samples=10).fit(One)

[0099] labels = db.labels_

[0100] The labels is a set of integers, which corresponds one-to-one to each element in the dataset One, indicating which clustering result the element belongs to.

[0101] According to labels, we divide the data set One into several different new data sets: O1, O2, …, On, where n represents the number of new data sets.

[0102] 204. The above-rated index calculation method is used to obtain the above-rated index of each data element in each new data set.

[0103] The calculation method of the above-rated index is as follows:

[0104] Indexrated(Xone)=(1if wind_speed>0.8else 0)+(1if power>power_ratedelse 0)+(1if pitch>0else 0)) (6).

[0105] According to formula (6), the above-rated index can be quickly calculated.

[0106] 205. Based on the rated index of each data element, the index mean of each new data set is obtained.

[0107] 206. Compare the index means of each new data set, and the one with the largest index mean is the optimal data set.

[0108] In the implementation method of extracting the optimal data set, the historical operating data are quickly processed with the help of the normalization function Fone, the Gaussian mixture model, the automatic clustering calculation method and the above-rated index calculation method, and multiple groups of optimal data sets of operating data above the rated wind speed are obtained. There is no need to divide the wind speed interval. The obtained optimal data set avoids the influence of wind speed measurement factors to a certain extent, which helps to improve the problem of error in the measured yaw deviation.

[0109] Reference Figure 2 and Figure 4 , as an implementation of a specific method for obtaining the optimal wind deviation interval [x1, x2], includes the following steps:

[0110] 301. Continuously derive the power mean fitting function.

[0111] Continuous derivation is a common mathematical calculation method, which will not be further elaborated in this implementation. The number of continuous derivations can be adjusted according to the actual power mean fitting function.

[0112] 302. According to the continuous derivation result and the optimal power position determination rule, the optimal power of the power mean fitting function is obtained.

[0113] The result of continuous derivation is the extreme point obtained.

[0114] 303. According to the wind deviation interval where the optimal power is located, obtain the optimal wind deviation interval [x1, x2].

[0115] In the implementation of the above-mentioned specific method for obtaining the optimal wind deviation interval [x1, x2], after continuously differentiating the power mean fitting function, a continuous differentiation result, i.e., multiple extreme points, is obtained. Then, the optimal power of the power mean fitting function can be determined by using the optimal power position determination rule, thereby obtaining the optimal wind deviation interval [x1, x2] where the optimal power is located.

[0116] As an implementation of the optimal power position determination rule, the optimal power position determination rule includes that among multiple extreme value points of the continuous derivation results, the extreme value closest to zero is the optimal power.

[0117] As an implementation of the optimal power position determination rule, the optimal power position determination rule also includes taking the median of the extreme values ​​on both sides of the zero value among multiple extreme value points of the continuous derivation results, and the one with the smallest absolute value among the medians is the optimal power.

[0118] It should be noted that the continuous derivative results may contain multiple zero values. In this case, the median implementation method is selected, and the zero value point is the turning point of the increase or decrease of the power mean fitting function.

[0119] This embodiment also discloses a yaw deviation acquisition system for a wind turbine generator set, referring to Figure 5 ,The yaw deviation acquisition system of the wind turbine generator set includes,

[0120] The optimal data set acquisition module 401 is used to obtain an optimal data set containing multiple sets of operating data above the rated wind speed based on multiple sets of historical operating data of the wind turbine generator set, in combination with a normalization method and an automatic clustering method, wherein each set of historical operating data includes but is not limited to the yaw angle to wind, wind direction, power, pitch position, speed and unit status of the wind turbine generator set at the same time;

[0121] The wind power acquisition module 402 is used to obtain multiple pitch positions according to the optimization data set, and obtain the wind speed V according to the pitch positions, and obtain the wind power P according to the wind speed V. wind .

[0122] in,

[0123]

[0124] Where ρ is the air density, V is the wind speed, and S is the impeller swept area;

[0125] The wind deviation interval division module 403 is used to divide the normal power generation interval angle to obtain multiple wind deviation intervals, and use the yaw wind angle of each group of working condition data above the rated wind speed as the interval division standard, and divide each group of working condition data above the rated wind speed into the corresponding wind deviation interval;

[0126] The power mean fitting function acquisition module 404 is used to obtain the power mean value based on the wind power P wind , get the equivalent power mean of each wind deviation interval, fit all equivalent power means to get the power mean fitting function; and,

[0127] The yaw deviation acquisition module 405 is used to continuously derive the power mean fitting function to obtain a continuous derivative result, combine the optimal power position determination rule and the continuous derivative result, determine the optimal wind deviation interval [x1, x2], and obtain the yaw deviation value Yawmiss.

[0128] in,

[0129] Yawmiss=(x1+x2) / 2 (2),

[0130] Among them, x1 represents the starting value of the interval angle of the optimal wind deviation interval, and x2 represents the ending value of the interval angle of the optimal wind deviation interval.

[0131] In the implementation of the above-mentioned yaw deviation acquisition system for wind turbines, the optimal data set acquisition module 401 extracts an optimal data set containing multiple groups of operating data above the rated wind speed from the historical operating data by an unsupervised learning method, and the wind power acquisition module 402 obtains multiple wind power P of the optimal data set by a wind power calculation formula. wind The wind deviation interval division module divides each set of working condition data above the rated wind speed in the optimization data set into the corresponding wind deviation interval according to the yaw wind angle in the optimization data set, and the power mean fitting function acquisition module 404 obtains the wind deviation interval based on the wind power P wind The equivalent power mean of each wind deviation interval is fitted to obtain a power mean fitting function. The yaw deviation acquisition module 405 continuously derives the power mean fitting function and determines the optimal power position to obtain the optimal wind deviation to obtain the yaw deviation value Yawmiss. There is no need to divide the wind speed interval or judge the wind speed. To a certain extent, the influence of wind speed measurement factors is avoided, and the problem of error in the measured yaw deviation result is greatly improved.

[0132] The yaw deviation acquisition system of a wind turbine provided in the present application can be applied to a large wind turbine with an active yaw system. By collecting the operating data of the wind turbine at a rated wind speed or higher, the yaw deviation value of the wind turbine is estimated by using the calculation method provided by the yaw deviation acquisition method. Thus, the impeller's angle to the wind can be corrected by calculating the yaw deviation value of the wind turbine, thereby effectively improving the power generation of the wind turbine.

[0133] This embodiment also discloses an intelligent terminal, including a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor to execute any method in a method for obtaining a yaw deviation of a wind turbine generator set.

[0134] This embodiment further discloses a computer-readable storage medium storing a computer program that can be loaded by a processor and executed, for example, any method in a method for acquiring a yaw deviation of a wind turbine generator set.

[0135] The implementation principle of the yaw deviation acquisition method, system and readable storage medium of a wind turbine generator set in the embodiment of the present application is as follows: the optimal data set acquisition module 401 extracts the optimal data set containing multiple groups of working condition data above the rated wind speed from the historical working condition data by an unsupervised learning method, and the wind power acquisition module 402 obtains multiple wind power P of the optimal data set by the wind power calculation formula. wind The wind deviation interval division module 403 divides each set of working condition data above the rated wind speed in the optimization data set into the corresponding wind deviation interval according to the yaw wind angle in the optimization data set, and the power mean fitting function acquisition module 404 obtains the wind deviation interval based on the wind power P wind The equivalent power mean of each wind deviation interval is fitted to obtain a power mean fitting function. The yaw deviation acquisition module 405 continuously derives the power mean fitting function and determines the optimal power position to obtain the optimal wind deviation to obtain the yaw deviation.

[0136] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for obtaining yaw deviation of a wind turbine generator set, characterized in that: The acquisition method includes: Based on multiple groups of historical operating data of wind turbines, combined with normalization method and automatic clustering method, an optimal data set containing multiple groups of operating data above rated wind speed is obtained, each group of historical operating data includes but is not limited to the yaw angle to wind, wind direction, power, pitch position, speed and unit status of the wind turbine at the same time; According to the optimization data set, multiple pitch positions are obtained, and the wind speed V is obtained according to the pitch positions, and the wind power P is obtained according to the wind speed V. wind , Where ρ is the air density, V is the wind speed, and S is the impeller swept area; The normal power generation interval angle is divided to obtain multiple wind deviation intervals. The yaw wind angle of each group of working condition data above the rated wind speed is used as the interval division standard, and each group of working condition data above the rated wind speed is divided into the corresponding wind deviation interval; Based on wind power P wind , get the equivalent power mean of each wind deviation interval, fit all equivalent power means to get the power mean fitting function; and, The power mean fitting function is continuously differentiated to obtain the continuous derivative result. Combining the optimal power position determination rule and the continuous derivative result, the optimal wind deviation interval [x1, x2] is determined, and the yaw deviation value Yawmiss is obtained. Yawmiss=(x1+x2) / 2 (2), Among them, x1 represents the starting value of the interval angle of the optimal wind deviation interval, and x2 represents the ending value of the interval angle of the optimal wind deviation interval.

2. The method for obtaining the yaw deviation of a wind turbine generator set according to claim 1, characterized in that: The specific method of extracting the optimization data set includes: Acquire multiple groups of historical operating condition data of the wind turbine generator set, each group of the historical operating condition data includes but is not limited to the yaw angle to wind, wind direction, power, pitch position, speed and unit status of the wind turbine generator set; Each set of historical operating condition data is processed using the normalization function Fone to obtain data set One; Using a Gaussian mixture model to automatically perform clustering calculations on data set One, dividing data set One into multiple new data sets; and, Based on multiple new data sets and combined with extraction methods, an optimal data set is obtained.

3. The method for obtaining the yaw deviation of a wind turbine generator set according to claim 2, characterized in that: The extraction method comprises: The above-rated index calculation method is used to obtain the above-rated index of each data element in each new data set; Based on the rated above index of each data element, the index mean of each new data set is obtained; as well as, Compare the index means of each new data set, and the one with the largest index mean is the optimal data set.

4. The method for obtaining the yaw deviation of a wind turbine generator set according to claim 1, characterized in that: The specific method for obtaining the optimal wind deviation interval [x1, x2] includes: Continuously derive the power mean fitting function; According to the continuous derivation result and the optimal power position determination rule, the optimal power of the power mean fitting function is obtained; and According to the wind deviation interval where the optimal power is located, the optimal wind deviation interval [x1, x2] is obtained.

5. A method for obtaining yaw deviation of a wind turbine generator set according to claim 1 or 4, characterized in that: The optimal power position determination rule includes that, among the multiple extreme value points of the continuous derivation results, the extreme value closest to zero is the optimal power.

6. A method for obtaining yaw deviation of a wind turbine generator set according to claim 1 or 4, characterized in that: The optimal power position determination rule includes: among the multiple extreme value points of the continuous derivative results, the extreme values ​​on both sides of the zero value are respectively taken as the median, and the one with the smallest absolute value among the medians is the optimal power.

7. A yaw deviation acquisition system for a wind turbine generator set, characterized in that: The system comprises, An optimization data set acquisition module (401) is used to obtain an optimization data set containing multiple sets of operating data above the rated wind speed based on multiple sets of historical operating data of the wind turbine generator set, in combination with a normalization method and an automatic clustering method, wherein each set of the historical operating data includes but is not limited to the yaw angle to wind, wind direction, power, pitch position, rotation speed and unit status of the wind turbine generator set at the same time; The wind power acquisition module (402) is used to obtain multiple pitch positions according to the optimization data set, and obtain the wind speed V according to the pitch positions, and obtain the wind power P according to the wind speed V. wind , Where ρ is the air density, V is the wind speed, and S is the impeller swept area; A wind deviation interval division module (403) is used to divide the normal power generation interval angle to obtain multiple wind deviation intervals, using the yaw wind angle of each group of working condition data above the rated wind speed as the division standard, and dividing each group of working condition data above the rated wind speed into a corresponding wind deviation interval; The power mean fitting function acquisition module (404) is used to obtain the wind power P based on the wind power P wind , get the equivalent power mean of each wind deviation interval, fit all equivalent power means to get the power mean fitting function; and, The yaw deviation acquisition module (405) is used to continuously derive the power mean fitting function to obtain a continuous derivative result, combine the optimal power position determination rule and the continuous derivative result, determine the optimal wind deviation interval [x1, x2], and obtain the yaw deviation value Yawmiss, Yawmiss=(x1+x2) / 2 (2), Among them, x1 represents the starting value of the interval angle of the optimal wind deviation interval, and x2 represents the ending value of the interval angle of the optimal wind deviation interval.

8. An intelligent terminal, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program which can be loaded by the processor and executed by the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the method according to any one of claims 1 to 6.

Images