A method for calculating yaw-to-wind error of a wind turbine based on goodness of fit

Abstract

The embodiment provides a wind turbine yaw-to-wind error calculation method and system based on goodness of fit, a terminal and a medium, and belongs to the technical field of wind power generation. The embodiment first preprocesses the original data to exclude the interference of variable pitch, shutdown, power limitation and the like on power generation; then performs interval data set division based on wind speed intervals and yaw angles; then uses a hypothesis testing method to check the power distribution of each wind speed interval sub-set, selects a corresponding fitting method through the actual distribution of each sub-set, and determines the yaw wind direction of the maximum power in the interval; and finally comprehensively considers the calculated yaw error angles in each wind speed interval. If the yaw error angles obtained in each wind speed interval have high consistency, and the goodness of the fitting result meets the preset requirement, it is considered that there is a significant yaw error angle in the wind farm, and the yaw error angle is calculated. The present application improves the accuracy of yaw error calculation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wind power generation technology, in particular to a wind turbine yaw-to-wind error calculation method, system, terminal and medium based on goodness of fit. BACKGROUND

[0002] With the increasing emphasis on environmental protection and sustainable development worldwide, the new energy industry, especially the wind power industry, has ushered in an unprecedented development opportunity. As a clean and renewable form of energy, the effective use of wind power is of great significance to reducing carbon emissions and achieving energy structure transformation. However, the development of the wind power industry has not been smooth sailing, and the yaw system, as one of the core components of the wind turbine, directly affects the power generation efficiency of the wind turbine.

[0003] The main function of the yaw system is to adjust the direction of the wind wheel of the wind turbine so that it always points to the wind direction, thereby maximizing the capture of wind energy and improving power generation. However, in actual engineering applications, the yaw system faces many challenges. First of all, the zero deviation problem of the wind vane of the unit cannot be ignored. When the zero degree of the wind vane is no longer aligned with the cabin axis, it will cause the yaw error angle measured by the wind vane to deviate from the true value, thereby affecting the accuracy and reliability of the yaw system.

[0004] Secondly, during the operation of the wind turbine, the high-speed rotation of the impeller will drive the surrounding air to rotate, forming a so-called impeller wake. This phenomenon not only causes interference to the measurement of the wind vane, resulting in a deviation between the measured yaw error and the actual yaw error, but also increases the operating burden of the wind turbine and reduces the power generation efficiency.

[0005] The existence of yaw error undoubtedly wastes the power generation potential of the wind turbine. It not only causes the wind turbine to fail to fully utilize wind energy, but also increases the wear and tear and energy consumption of the unit due to frequent yaw adjustment, further affecting power generation.

[0006] Therefore, how to accurately identify the yaw-to-wind error and improve the accuracy and stability of the yaw system has become a technical problem to be solved in the current wind power industry. SUMMARY

[0007] In view of the deficiencies in the prior art, the present application provides a wind turbine yaw-to-wind error calculation method, system, terminal and medium based on goodness of fit to solve the technical problems existing in the prior art.

[0008] The technical solution adopted by the present application is:

[0009] The first aspect of the present embodiment provides a wind turbine yaw-to-wind error calculation method based on goodness of fit, comprising:

[0010] obtaining SCADA data of a wind turbine generator;

[0011] preprocessing the SCADA data to obtain N wind speed intervals;

[0012] performing K-S normal distribution detection and Jarque-Bera skewness distribution hypothesis detection on power data in each wind speed interval to obtain a wind speed interval subject to K-S normal distribution, a wind speed interval subject to Jarque-Bera skewness distribution, and / or a wind speed interval not conforming to a standard distribution;

[0013] dividing each wind speed interval into a yaw angle interval, and each wind speed interval obtains M sub-intervals;

[0014] calculating average power of the M sub-intervals in each wind speed interval, and each wind speed interval obtains M average power data;

[0015] for the wind speed interval subject to K-S normal distribution, using the number of effective data points in the wind speed interval subject to K-S normal distribution as a weight, performing binomial fitting on the average power and wind direction of each sub-interval of the wind speed interval subject to K-S normal distribution to obtain a quadratic function curve describing the change of power with wind direction;

[0016] if the vertex of the quadratic function curve is within a preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, then the wind speed interval subject to K-S normal distribution exists a yaw error angle, and the abscissa of the vertex of the quadratic function curve is the yaw error angle of the wind speed interval subject to K-S normal distribution;

[0017] for the wind speed interval subject to Jarque-Bera skewness distribution, using the number of effective data points in the wind speed interval subject to Jarque-Bera skewness distribution as a weight, performing cubic fitting on the average power and wind direction of each sub-interval of the wind speed interval subject to Jarque-Bera skewness distribution to obtain a cubic function curve describing the change of power with wind direction;

[0018] if the vertex of the cubic function curve is within a preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, then the wind speed interval subject to Jarque-Bera skewness distribution exists a yaw error angle, and the abscissa of the vertex of the cubic function curve is the yaw error angle of the wind speed interval subject to Jarque-Bera skewness distribution;

[0019] if the number of wind speed intervals with yaw error angles in the wind turbine generator N * exceeds 70% of the total number of wind speed intervals N of the wind turbine generator, and N * the number of wind speed intervals with yaw error angles in the N wind speed intervals exceeds N* If the 90% of the wind speed interval is less than 5°, and the N° is greater than 5°, the yaw error angle of the wind turbine is N°, the N° is the mean value of N * yaw error angle of the wind speed interval.

[0020] Further, the pre-processing of the SCADA data to obtain N wind speed intervals comprises:

[0021] The data in the initial wind turbine SCADA data that does not meet the normal power generation mode, the blade pitch angle is not in the range of [-1°, 1°], the wind turbine active output is less than or equal to 20kW, the wind turbine is not power limited, and the 60s average wind speed is less than the cut-in wind speed and the 60s average wind speed is greater than the rated wind speed are removed to obtain a cleaned data set;

[0022] The wind speed interval division processing is performed on the cleaned data set to obtain N to-be-processed wind speed intervals;

[0023] The wind speed outliers in each to-be-processed wind speed interval are removed to obtain N wind speed intervals.

[0024] Further, the wind speed interval division processing is performed on the cleaned data set to obtain N to-be-processed wind speed intervals, comprising:

[0025] The preset interval length, the cut-in wind speed, and the rated wind speed are obtained;

[0026] The cut-in wind speed is taken as the starting wind speed, the rated wind speed is taken as the cut-off wind speed, and the wind speed interval division processing is performed on the cleaned data set based on the preset interval length to obtain N to-be-processed wind speed intervals.

[0027] Further, the wind speed outliers in each to-be-processed wind speed interval are removed by using the DBSCAN clustering algorithm of the density-based spatial clustering method with noise to obtain N wind speed intervals.

[0028] Further, the yaw angle interval division is performed on each wind speed interval, and each wind speed interval obtains M sub-intervals, comprising:

[0029] The preset yaw angle interval length, the preset starting yaw angle, and the preset cut-off yaw angle of each wind speed interval are obtained;

[0030] The data set in each wind speed interval is subjected to yaw angle interval division based on the preset yaw angle interval length, the preset starting yaw angle, and the preset cut-off yaw angle, and each wind speed interval obtains M sub-intervals.

[0031] Further, the determination coefficient of the fitting result is calculated by using the following formula:

[0032]

[0033] wherein R 2 R2denotes the determination coefficient of the fitting result, y i denotes the average power under the i-th wind direction interval, denotes the fitting value of the average power under the i-th wind direction interval, denotes the average value of the average power in the n wind direction intervals.

[0034] The second aspect of the embodiment provides a yaw-to-wind error calculation system for wind turbines based on goodness of fit, which comprises:

[0035] a SCADA data acquisition module for acquiring SCADA data of a wind turbine;

[0036] a data preprocessing module for preprocessing the SCADA data to obtain N wind speed intervals;

[0037] a distribution detection module for performing K-S normal distribution detection and Jarque-Bera skew distribution hypothesis detection on the power data in each wind speed interval to obtain wind speed intervals conforming to K-S normal distribution, wind speed intervals conforming to Jarque-Bera skew distribution, and / or wind speed intervals not conforming to standard distribution;

[0038] a yaw angle interval division module for dividing each wind speed interval into M subintervals;

[0039] an average power calculation module for calculating the average power of the M subintervals in each wind speed interval, each wind speed interval obtaining M average power data sets;

[0040] a power versus wind direction fitting module for performing binomial fitting on the average power and wind direction of each subinterval of the wind speed intervals conforming to K-S normal distribution using the number of effective data points in the wind speed intervals conforming to K-S normal distribution as the weight, to obtain a quadratic function curve describing the change of power with wind direction;

[0041] if the vertex of the quadratic function curve is within the preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, then the yaw error angle of the wind speed interval conforming to K-S normal distribution exists, and the abscissa of the vertex of the quadratic function curve is the yaw error angle of the wind speed interval conforming to K-S normal distribution;

[0042] For the wind speed interval subject to the Jarque-Bera skew distribution, the effective data points in the wind speed interval subject to the Jarque-Bera skew distribution are used as weights to perform a cubic polynomial fitting on the average power and wind direction of each sub-interval of the wind speed interval subject to the Jarque-Bera skew distribution, so as to obtain a cubic function curve describing the change of the power with the wind direction;

[0043] If the vertex of the cubic function curve is within a preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, then the yaw error angle of the wind speed interval subject to the Jarque-Bera skew distribution exists, and the abscissa of the vertex of the cubic function curve is the yaw error angle of the wind speed interval subject to the Jarque-Bera skew distribution;

[0044] The yaw error angle calculation module: if the number N of wind speed intervals with yaw error angles in the unit * exceeds 70% of the total number N of wind speed intervals in the unit, and N * The number of wind speed intervals with yaw error angles in the range of N°±2° exceeds 90% of N * , and N° is greater than 5°, then the yaw error angle of the wind turbine is N°, and the N° is the average of the yaw error angles of the N * wind speed intervals.

[0045] The third aspect of the embodiment provides a terminal, including a processor, an input device, an output device and a memory, the processor, the input device, the output device and the memory are connected with each other, wherein the memory is used for storing a computer program, the computer program includes program instructions, the processor is configured to call the program instructions, and the method is executed as any one of the first aspect of the embodiment.

[0046] The fourth aspect of the embodiment provides a computer readable storage medium, the computer readable storage medium stores a computer program, the computer program includes program instructions, and the program instructions make the processor execute the method as any one of the first aspect of the embodiment when the processor executes the program instructions.

[0047] From the above technical solutions, the beneficial technical effects of the present application are as follows:

[0048] 1.The application strictly preprocesses SCADA data, eliminates data that does not meet normal power generation conditions, abnormal blade pitch angle, too low power, abnormal wind speed and other non-compliant data, ensures the accuracy of subsequent analysis, and according to the distribution characteristics of wind speed data, respectively uses K-S normal distribution detection and Jarque-Bera skew distribution hypothesis detection, selects appropriate fitting methods (binomial or trinomial fitting) for different distribution characteristics of wind speed intervals, thereby more accurately describing the change relationship of power with wind direction, and by calculating the determination coefficient of the fitting result and setting strict judgment conditions (determination coefficient greater than 0.9, vertex within the preset wind direction range), further improving the reliability of yaw error calculation. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual proportions.

[0050] Figure 1 The method flowchart of the embodiment of the present application is shown in the figure.

[0051] Figure 2 The data preprocessing flowchart of the embodiment of the present application is shown in the figure.

[0052] Figure 3 The cubic fitting curve diagram of the embodiment of the present application is shown in the figure.

[0053] Figure 4 The system schematic diagram of the embodiment of the present application is shown in the figure.

[0054] Figure 5 The terminal schematic diagram of the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0055] The embodiments of the technical solutions of the present application will be described in detail below with reference to the drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application, and therefore only serve as examples, but cannot limit the protection scope of the present application.

[0056] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in the present application should be understood as the usual meanings understood by the skilled in the art to which the present application belongs.

[0057] EMBODIMENT

[0058] The embodiment provides a wind turbine yaw-to-wind error calculation method based on goodness of fit, as shown in the figure, the method comprises the following steps: Figure 1 ​

[0059] acquiring SCADA data of a wind turbine generator system;

[0060] preprocessing the SCADA data to obtain N wind speed intervals;

[0061] performing K-S normal distribution detection and Jarque-Bera skewness distribution hypothesis detection on power data in each wind speed interval to obtain a wind speed interval subject to K-S normal distribution, a wind speed interval subject to Jarque-Bera skewness distribution, and / or a wind speed interval not conforming to a standard distribution;

[0062] dividing each wind speed interval into a yaw angle interval, and each wind speed interval obtains M sub-intervals;

[0063] calculating average power of the M sub-intervals in each wind speed interval, and each wind speed interval obtains M average power data;

[0064] for the wind speed interval subject to K-S normal distribution, using the number of effective data points in the wind speed interval subject to K-S normal distribution as a weight, performing binomial fitting on the average power and wind direction of each sub-interval of the wind speed interval subject to K-S normal distribution to obtain a quadratic function curve describing the change of power with wind direction;

[0065] if the vertex of the quadratic function curve is within a preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, then the wind speed interval subject to K-S normal distribution exists a yaw error angle, and the abscissa of the vertex of the quadratic function curve is the yaw error angle of the wind speed interval subject to K-S normal distribution;

[0066] for the wind speed interval subject to Jarque-Bera skewness distribution, using the number of effective data points in the wind speed interval subject to Jarque-Bera skewness distribution as a weight, performing cubic fitting on the average power and wind direction of each sub-interval of the wind speed interval subject to Jarque-Bera skewness distribution to obtain a cubic function curve describing the change of power with wind direction;

[0067] if the vertex of the cubic function curve is within a preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, then the wind speed interval subject to Jarque-Bera skewness distribution exists a yaw error angle, and the abscissa of the vertex of the cubic function curve is the yaw error angle of the wind speed interval subject to Jarque-Bera skewness distribution;

[0068] if the number of wind speed intervals with yaw error angles in the wind turbine generator system is N * more than 70% of the total number N of wind speed intervals in the wind turbine generator system, and N * the number of wind speed intervals with yaw error angles in the N wind speed intervals is within the range of N°±2° more than 90% of N, and N° is greater than 5°, then the yaw error angle of the wind turbine is N°, which is the average of N * yaw error angles of the N wind speed intervals. *

[0069] In an embodiment, after obtaining the SCADA data, the original data needs to be preprocessed to exclude the interference of other factors on the power generation;

[0070] Table 1 Parameter name and its detailed information

[0071] Parameter name Detail information Working mode Normal power generation mode of wind turbine Blade pitch angle Full-pitch power generation state Active power Active power output state of wind turbine 60s average wind speed Cut-in wind speed and rated wind speed Limited power operation state Wind turbine not limited power

[0072] The SCADA data is preprocessed to obtain N wind speed intervals, as shown in the following formula (1), including: Figure 2

[0073] The data in the initial wind turbine SCADA data that does not meet the normal power generation mode, the blade pitch angle is not in the range of [-1°, 1°], the wind turbine active output is less than or equal to 20kW, the wind turbine is not power limited, and the 60s average wind speed is less than the cut-in wind speed and the 60s average wind speed is greater than the rated wind speed are removed to obtain a cleaned data set.

[0074] The cleaned data set is subjected to wind speed interval division processing to obtain N wind speed intervals. In this embodiment, the preset interval length is 0.5m / s, the cut-in wind speed is the starting wind speed, and the rated wind speed is the cutoff wind speed to divide into N different wind speed intervals (wherein the interval length and the starting and cutoff wind speeds can be corrected according to the actual test feedback results of different wind farms).

[0075] The wind speed outlier data in each wind speed interval is removed to obtain N wind speed intervals. In this embodiment, the DBSCAN clustering algorithm with noise based on density is used to remove the wind speed outlier data in each wind speed interval to obtain N wind speed intervals.

[0076] Specifically, the DBSCAN clustering algorithm is used to remove the wind speed outlier data in each wind speed interval, including: first determining €(neighborhood radius) and MinPts(minimum point number), for each point in each wind speed interval, calculating the number of points in its e-neighborhood. If the number of points is greater than or equal to MinPts, mark the point as a core point and start forming a new cluster. For each core point, find all points in its -neighborhood and add them to the same cluster. Then recursively repeat this process for newly added core points until no new points can be added to the cluster. The remaining points not assigned to any cluster are marked as noise points, then the wind speed data corresponding to the noise points is removed and the remaining data is retained. ​​

[0077] In an embodiment, K-S normal distribution detection and Jarque-Bera skewness distribution hypothesis detection are performed on the power data in each wind speed interval to obtain wind speed intervals that conform to K-S normal distribution, wind speed intervals that conform to Jarque-Bera skewness distribution, and / or wind speed intervals that do not conform to standard distribution.

[0078] Specifically, K-S normal distribution detection and Jarque-Bera skewness distribution hypothesis detection are performed on the power data in each wind speed interval, and the power data is divided into three cases according to whether it conforms to normal distribution, skewness distribution, or neither. Different skew error angle calculation methods are used according to the different cases.

[0079] K-S normal distribution detection is performed on the wind speed interval.

[0080] (1) Calculate the empirical distribution function (EDF):

[0081] 1. Sort the data: Sort the interval power data in ascending order.

[0082] 2. Reassign the sequence number: For the i-th data after sorting (i.e., the i-th smallest power in size order), its sequence number is i.

[0083] 3. Calculate the empirical distribution function EDF:

[0084] F(i) = (i + 1) / n

[0085] Where n is the total number of data points (i.e., the total number of power points under the wind speed interval).

[0086] (2) Calculate the cumulative distribution function (CDF) of the standard normal distribution:

[0087]

[0088] Where, is the error function erf(x);

[0089] Where,

[0090] (3) Calculate fabs(i) = |F(i) - CDF(i)|, and calculate D = max{fabs(i)}. If D < α, the wind speed data in the wind speed interval satisfies the normal distribution, otherwise, it does not, where α is the significance level.

[0091] Jarque-Bera skewness distribution hypothesis detection is performed on the wind speed interval.

[0092] (1) Jarque-Bera statistic

[0093] wherein n is the wind speed data capacity of each wind speed interval, S is the sample skewness, and K is the sample kurtosis; wherein the sample skewness S is one of the key factors for judging whether the distribution is skewed or not;

[0094] (2) If the original hypothesis is rejected, that is, there is evidence that the sample data does not obey the normal distribution. At this time, the value of the skewness S can be further observed to determine whether it is a skewed distribution. Wherein χ 2 represents the chi-square distribution, so represents the significance level α (here 0.05), the chi-square test critical value of degree of freedom = 2, which is a fixed value after the determination of the significance level and the freedom, directly obtained by looking up the χ 2 square distribution table);

[0095] (3) After rejecting the original hypothesis, if S>0, the sample data may be right-skewed distribution, the data is concentrated on the left side, and the probability of extreme value on the right side is relatively high; on the contrary, if S<0, the sample data may be left-skewed distribution, the data is concentrated on the right side, and the probability of extreme value on the left side is relatively high.

[0096] In an embodiment, the wind speed interval is divided into a yaw angle interval, and each wind speed interval obtains M sub-intervals;

[0097] The average power of the M sub-intervals in each wind speed interval is calculated, and each wind speed interval obtains M average power data;

[0098] Specifically, the data set of each wind speed interval is divided into 32 sub-sets with 1 degree as the interval length, -16 degrees as the starting yaw angle, and 16 degrees as the cut-off yaw angle, and the average power and the number of valid data points in each sub-set are calculated.

[0099] In an embodiment, for the wind speed interval obeying K-S normal distribution, the number of valid data points in the wind speed interval obeying K-S normal distribution is used as the weight, and the binomial fitting is performed on the average power and the wind direction of each sub-interval of the wind speed interval obeying K-S normal distribution, to obtain a quadratic function curve describing the change of power with wind direction;

[0100] If the vertex of the quadratic function curve is within the preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, then the yaw error angle of the wind speed interval obeying K-S normal distribution exists, and the abscissa of the vertex of the quadratic function curve is the yaw error angle of the wind speed interval obeying K-S normal distribution;

[0101] Specifically, for the yaw angle maximum power calculation of the wind speed interval subject to normal distribution: if it is verified that the power scatter points of the wind speed interval are subject to normal distribution, it is considered that the power is subject to normal distribution with wind direction in the wind speed interval, that is, it is considered that there is a more obvious vertex of the power with wind direction in the wind speed interval, and the vertex of the normal distribution is the maximum power position in the wind speed interval, and the corresponding wind direction value is the yaw angle when the maximum power is obtained. Since the normal distribution is symmetrically distributed, it is closer to the waveform of the quadratic function, then the average power and the effective data point number in each sub-set are weighted binomial fitted with the effective data point number as the weight, and the determination coefficient (R2_score) of the fitting result is used as an index to determine the goodness of the fitting result:

[0102]

[0103] wherein, R 2 represents the determination coefficient of the fitting result, y i represents the average power in the i-th wind direction interval, represents the fitting value of the average power in the i-th wind direction interval, represents the average value of the average power in the n wind direction intervals, and the range of R2_score is between 0 and 1;

[0104] If the calculated R2_score in the wind speed interval is greater than 0.9, and the vertex of the fitted quadratic curve is in the wind direction range of [-16, 16], it is considered that there is a significant yaw error angle in the wind speed interval subject to K-S normal distribution, and the abscissa (wind direction) of the vertex of the quadratic fitting curve is the yaw error angle of the wind speed interval subject to K-S normal distribution.

[0105] In an embodiment, for the wind speed interval subject to Jarque-Bera skewness distribution, the average power and wind direction of each sub-interval of the wind speed interval subject to Jarque-Bera skewness distribution are fitted with a cubic function curve using the effective data point number in the wind speed interval subject to Jarque-Bera skewness distribution as the weight;

[0106] If the vertex of the cubic function curve is in the preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, the yaw error angle of the wind speed interval subject to Jarque-Bera skewness distribution exists, and the abscissa of the vertex of the cubic function curve is the yaw error angle of the wind speed interval subject to Jarque-Bera skewness distribution;

[0107] Specifically, for the yaw angle maximum power calculation of the wind speed interval subject to skew distribution: if the power scatter points of the wind speed interval are subject to skew distribution after inspection, it is considered that the power is subject to skew distribution with wind direction in the wind speed interval, that is, there is a more obvious vertex of the power with wind direction in the wind speed interval, and the vertex of the distribution function is the maximum power position in the wind speed interval, and the corresponding wind direction value is the yaw angle when the maximum power is obtained. Since the skew distribution is not a symmetric distribution, it is closer to the waveform of a cubic function, then the average power and the number of effective data points in each sub-set are weighted and fitted with a cubic polynomial, and the determination coefficient (R2_score) of the fitting result is used as an index to determine the goodness of the fitting result:

[0108]

[0109] If the calculated R2_score in the wind speed interval is greater than 0.9, and the vertex of the cubic curve is within the range of [-16, 16], it is considered that there is a significant yaw error angle in the wind speed interval, and the abscissa (wind direction) of the vertex of the cubic fitting curve is the yaw error angle of the wind speed interval subject to Jarque-Bera skew distribution. As shown in FIG. 2, the fitting of the average power of the 2# unit in the 3.5-4 wind speed interval is shown. Since the original power scatter points are subject to skew distribution, the model selects cubic fitting, and the abscissa corresponding to the vertex value of the cubic curve is the yaw error angle in the wind speed interval. Figure 3

[0110] In an embodiment, the yaw angle calculation of the wind speed interval not subject to any distribution: if the power of a certain wind speed interval does not significantly subject to normal distribution or skew distribution, it is considered that there is no significant yaw error angle in the wind speed interval.

[0111] In an embodiment, it is judged whether the yaw error angle exists in the unit;

[0112] The number N of wind speed intervals in which the yaw error angle exists in the unit is calculated * ;

[0113] It is judged whether N * is greater than 70% of the total number N of wind speed intervals of the unit;

[0114] If N * > 0.7*N, the mean N° of N * effective yaw error angles is calculated;

[0115] If the number of effective yaw error angles falling within the range of N°±2° is and N°>5°, it is determined that the unit has a significant yaw error angle, and the yaw error angle is N°, and the error range is N°±σ Y , σ​Y the standard deviation of the effective yaw error angle.

[0116] In an embodiment, the calculation of the reliability of the yaw-to-wind error angle is:

[0117] The R2_score of the average fitting result of the effective yaw error angle is taken as the reliability of the yaw-to-wind error angle of the wind turbine.

[0118] In an embodiment, the partial unit yaw error angle, error range and reliability calculated based on the above-described method are shown in Table 2.

[0119] Table 2 is a final result table of partial unit wind yaw error analysis

[0120] Unit number Yaw error angle Error range Reliability 1 8.75 ±3.50 0.91 2 8.17 ±4.80 0.80 3 11.50 ±6.54 0.71

[0121] As shown in Figure 4 the embodiment provides a wind turbine yaw-to-wind error calculation system based on goodness of fit, which comprises:

[0122] The SCADA data acquisition module 101 acquires SCADA data of the wind turbine.

[0123] The data preprocessing module 102 preprocesses the SCADA data to obtain N wind speed intervals.

[0124] The distribution detection module 103 performs K-S normal distribution detection and Jarque-Bera skew distribution hypothesis detection on the power data in each wind speed interval to obtain wind speed intervals conforming to K-S normal distribution, wind speed intervals conforming to Jarque-Bera skew distribution and / or wind speed intervals not conforming to standard distribution.

[0125] The yaw angle interval division module 104 divides the yaw angle interval of each wind speed interval, and each wind speed interval obtains M subintervals.

[0126] The average power calculation module 105 calculates the average power of the M subintervals in each wind speed interval, and each wind speed interval obtains M average power data sets.

[0127] The power versus wind direction fitting module 106 performs binomial fitting on the average power and wind direction of each subinterval of the wind speed interval conforming to K-S normal distribution using the number of effective data points in the wind speed interval conforming to K-S normal distribution as the weight, to obtain a quadratic function curve describing the change of power with wind direction.

[0128] If the vertex of the quadratic function curve is within the preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, then the yaw error angle of the wind speed interval subject to the K-S normal distribution exists, and the abscissa of the vertex of the quadratic function curve is the yaw error angle of the wind speed interval subject to the K-S normal distribution.

[0129] For the wind speed interval subject to the Jarque-Bera skewness distribution, the average power and wind direction of each sub-interval of the wind speed interval subject to the Jarque-Bera skewness distribution are fitted using the number of effective data points in the wind speed interval subject to the Jarque-Bera skewness distribution as the weight, to obtain a cubic function curve describing the change of power with wind direction.

[0130] If the vertex of the cubic function curve is within the preset wind direction range, and the determination coefficient of the fitting result is greater than 0.9, then the yaw error angle of the wind speed interval subject to the Jarque-Bera skewness distribution exists, and the abscissa of the vertex of the cubic function curve is the yaw error angle of the wind speed interval subject to the Jarque-Bera skewness distribution.

[0131] The yaw error angle calculation module 107: if the number N of wind speed intervals with yaw error angles in the unit * exceeds 70% of the total number N of wind speed intervals in the unit, and N * the number of wind speed intervals with yaw error angles in the range of N°±2° exceeds 90% of N * , and N° is greater than 5°, then the yaw error angle of the wind turbine is N°, and the N° is the average of the yaw error angles of the N * wind speed intervals.

[0132] As shown in Figure 5 , the embodiment provides a terminal, including a processor, an input device, an output device and a memory, which are connected with each other, wherein the memory is used to store a computer program, the computer program includes program instructions, the processor is configured to invoke the program instructions, and execute part or all steps described in the wind turbine yaw-to-wind error calculation method based on goodness of fit provided by the embodiment.

[0133] The embodiment further provides a computer readable storage medium, which stores a computer program, the computer program includes program instructions, and the program instructions, when executed by a processor, cause the processor to execute part or all steps described in the wind turbine yaw-to-wind error calculation method based on goodness of fit provided by the embodiment.

[0134] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some or all of the technical features can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and they should be covered in the scope of the claims and the specification of the present application.

Claims

1. A method for calculating the yaw error of a wind turbine based on goodness of fit, characterized in that, include: Obtain SCADA data from wind turbines; The SCADA data is preprocessed to obtain N wind speed ranges; KS normal distribution detection and Jarque-Bera skewed distribution hypothesis detection are performed on the power data in each wind speed range to obtain wind speed ranges that follow KS normal distribution, wind speed ranges that follow Jarque-Bera skewed distribution, and / or wind speed ranges that do not conform to the standard distribution. Each wind speed interval is divided into yaw angle intervals, and each wind speed interval is divided into M sub-intervals. Calculate the average power of M sub-intervals in each wind speed interval, and obtain M average power data for each wind speed interval; For wind speed intervals that follow a KS normal distribution, the number of valid data points in the wind speed intervals that follow a KS normal distribution is used as the weight. The average power and wind direction of each sub-interval of the wind speed intervals that follow a KS normal distribution are fitted with a binomial to obtain a quadratic function curve describing the change of power with wind direction. If the vertex of the quadratic function curve is within the preset wind direction range and the determination coefficient of the fitting result is greater than 0.9, then the yaw error angle of the wind speed range that follows the KS normal distribution is the x-coordinate of the vertex of the quadratic function curve, which is the yaw error angle of the wind speed range that follows the KS normal distribution. For wind speed intervals that follow a Jarque-Bera skewed distribution, the number of valid data points in the wind speed intervals that follow a Jarque-Bera skewed distribution is used as the weight. A trinomial fit is performed on the average power and wind direction of each sub-interval of the wind speed intervals that follow a Jarque-Bera skewed distribution to obtain a cubic function curve describing the change of power with wind direction. If the vertex of the cubic function curve is within the preset wind direction range and the determination coefficient of the fitting result is greater than 0.9, then the yaw error angle of the wind speed range that follows the Jarque-Bera skewed distribution is the x-coordinate of the vertex of the cubic function curve, which is the yaw error angle of the wind speed range that follows the Jarque-Bera skewed distribution. If the number N of wind speed ranges containing yaw error angles in the unit * The number of wind speed ranges N across all units exceeds 70%, and N * Number of wind speed intervals with yaw error angles within the range of N°±2° More than N * If the yaw rate is 90% and N° is greater than 5°, then the yaw error angle of the wind turbine is N°, where N° is N * The average yaw error angle for each wind speed range.

2. The method for calculating the yaw error of a wind turbine based on goodness of fit according to claim 1, characterized in that, The SCADA data is preprocessed to obtain N wind speed ranges, including: The cleaned dataset is obtained by removing data from the initial wind turbine SCADA data that are not in the normal power generation mode, have blade pitch angles outside the range of [-1°, 1°], have active power output less than or equal to 20kW, have no power limit, and have 60s average wind speed less than the cut-in wind speed and 60s average wind speed greater than the rated wind speed. The cleaning dataset is divided into wind speed intervals to obtain N wind speed intervals to be processed; By removing outlier data points from each wind speed range to be processed, N wind speed ranges are obtained.

3. The method for calculating the yaw error of a wind turbine based on goodness of fit according to claim 2, characterized in that, The cleaning dataset is divided into wind speed intervals to obtain N wind speed intervals to be processed, including: Obtain the preset interval length, cut-in wind speed, and rated wind speed; The cut-in wind speed is used as the starting wind speed, the rated wind speed is used as the cut-off wind speed, and the cleaning dataset is divided into wind speed intervals based on the preset interval length to obtain N wind speed intervals to be processed.

4. The method for calculating the yaw error of a wind turbine based on goodness of fit according to claim 3, characterized in that, The DBSCAN clustering algorithm, a density-based spatial clustering method with noise, is used to remove outlier data points in each wind speed interval to be processed, resulting in N wind speed intervals.

5. The method for calculating the yaw error of a wind turbine based on goodness of fit according to claim 1, characterized in that, The wind speed range is divided into yaw angle ranges, and each wind speed range is further divided into M sub-ranges, including: Obtain the preset yaw angle interval length, preset starting yaw angle, and preset ending yaw angle for each wind speed interval; The datasets in each wind speed interval are divided into yaw angle intervals based on the preset yaw angle interval length, preset starting yaw angle, and preset ending yaw angle, resulting in M ​​sub-intervals for each wind speed interval.

6. The method for calculating the yaw error of a wind turbine based on goodness of fit according to claim 1, characterized in that, The coefficient of determination for the fitted result is calculated using the following formula: Among them, R 2 The coefficient of determination for the fitting result, y i This represents the average power in the i-th wind direction interval. This represents the fitted value of the average power in the i-th wind direction interval. This represents the average power across n wind direction intervals.

7. A system for calculating the yaw error of a wind turbine based on goodness-of-fit, characterized in that, The system includes: SCADA data acquisition module: Acquires SCADA data from wind turbines; Data preprocessing module: preprocesses the SCADA data to obtain N wind speed ranges; Distribution detection module: Performs KS normal distribution detection and Jarque-Bera skewed distribution hypothesis detection on the power data in each wind speed range to obtain wind speed ranges that follow KS normal distribution, wind speed ranges that follow Jarque-Bera skewed distribution, and / or wind speed ranges that do not conform to the standard distribution. Yaw angle interval division module: Divide each wind speed interval into yaw angle intervals, and each wind speed interval will have M sub-intervals; Average power calculation module: Calculates the average power of M sub-intervals in each wind speed interval, and obtains M average power datasets for each wind speed interval; Power-wind-direction fitting module: For wind speed intervals that follow a KS normal distribution, the number of valid data points in the wind speed intervals that follow a KS normal distribution is used as the weight to perform binomial fitting on the average power and wind direction of each sub-interval of the wind speed intervals that follow a KS normal distribution, and a quadratic function curve describing the change of power with wind direction is obtained. If the vertex of the quadratic function curve is within the preset wind direction range and the determination coefficient of the fitting result is greater than 0.9, then the yaw error angle of the wind speed range that follows the KS normal distribution is the x-coordinate of the vertex of the quadratic function curve, which is the yaw error angle of the wind speed range that follows the KS normal distribution. For wind speed intervals that follow a Jarque-Bera skewed distribution, the number of valid data points in the wind speed intervals that follow a Jarque-Bera skewed distribution is used as the weight. A trinomial fit is performed on the average power and wind direction of each sub-interval of the wind speed intervals that follow a Jarque-Bera skewed distribution to obtain a cubic function curve describing the change of power with wind direction. If the vertex of the cubic function curve is within the preset wind direction range and the determination coefficient of the fitting result is greater than 0.9, then the yaw error angle of the wind speed range that follows the Jarque-Bera skewed distribution is the x-coordinate of the vertex of the cubic function curve, which is the yaw error angle of the wind speed range that follows the Jarque-Bera skewed distribution. Yaw error angle calculation module: If there are N wind speed ranges with yaw error angles in the unit * The number of wind speed ranges N across all units exceeds 70%, and N * Number of wind speed intervals with yaw error angles within the range of N°±2° More than N * If the yaw rate is 90% and N° is greater than 5°, then the yaw error angle of the wind turbine is N°, where N is... ° For N * The average yaw error angle for each wind speed range.

8. A terminal, characterized in that, The system includes a processor, an input device, an output device, and a memory, which are interconnected. The memory is used to store a computer program, which includes program instructions. The processor is configured to invoke the program instructions to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the method as described in any one of claims 1-6.

Images