A wind farm outlier screening method and system

Abstract

The application discloses a wind farm abnormal value screening method and system, and the method comprises the following steps: acquiring a wind speed-power first data set of a wind farm, screening to obtain a second data set; calculating a first set inner limit of the second data set by using a horizontal four quantile method, and deleting points in the second data set whose wind speed is outside the first set inner limit; calculating a second set inner limit of the second data set by using a vertical four quantile method, and deleting points in the second data set whose power is outside the second set inner limit to obtain a third data set; performing a first operation on all points in the third data set by using a clustering algorithm, and outputting a wind speed-power data set which is identified as normal. The horizontal and vertical four quantile methods are used to accurately define the normal range of wind speed and power, remove abnormal values, and obtain an accurate and reliable third data set; the clustering algorithm is further used for screening, so that only normal wind speed-power data points are contained in the data set, and the data quality is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wind farm data processing, and particularly relates to a wind farm abnormal value screening method and system. BACKGROUND

[0002] In the process of data acquisition and transmission in wind farms, some abnormal data often appear, which may be caused by sensor failure, environmental interference or data processing error, etc. If these abnormal data are not effectively screened and processed, it may cause deviation in wind farm power prediction, operation optimization and dispatching strategy execution, thereby affecting the overall operation efficiency of the power system. Therefore, accurately and effectively screening the abnormal data existing in the wind farm is a key step to ensure the reliable operation of the wind farm and improve the stability of the system.

[0003] Currently, the screening methods for abnormal data in wind farms mainly include quartile method based on statistical characteristics, isolated forest algorithm, K-means clustering algorithm and robust regression method (such as RANSAC) etc. The quartile method identifies abnormal points beyond the range by calculating the interquartile range of data, has the advantages of simple calculation and easy implementation, and is suitable for processing isolated discrete abnormal data. The isolated forest algorithm constructs an anomaly detection model through random sampling and tree structure, and can efficiently identify isolated abnormal points in high-dimensional data. The K-means clustering algorithm divides the data into multiple clusters, and identifies outliers according to the distance of cluster centers, which is suitable for processing abnormal samples with uniform distribution. The RANSAC robust regression algorithm can identify and fit the model under the condition that most of the data contain noise, and is suitable for processing data sets with many abnormal values.

[0004] The existing technology has certain limitations in screening abnormal values of wind farm power data. The quartile method often has difficulty in effectively processing horizontally stacked abnormal data. The K-means clustering algorithm has poor effect when processing clusters with complex shapes or uneven sizes, especially for bias cluster type abnormal points. The RANSAC robust regression algorithm may not be able to effectively fit when processing data with complex distribution, and may easily ignore nonlinear abnormal points deviating from the mainstream model. Therefore, the existing methods have deficiencies in processing various types of abnormal data, and cannot comprehensively and accurately screen all abnormal values. SUMMARY

[0005] This section is intended to summarize some aspects of the embodiments of the present application and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of the specification of the present application in order to avoid obscuring the purpose of this section, the abstract and the title, and such simplifications or omissions cannot be used to limit the scope of the present application.

[0006] In view of the above existing problems, the present application is proposed. Therefore, the present application provides a wind farm outlier screening method to solve the problem of low accuracy of abnormal value identification in the prior art in the background art.

[0007] To solve the above technical problems, the present application provides the following technical solutions.

[0008] In a first aspect, the present application provides a wind farm outlier screening method, comprising:

[0009] Obtaining a wind speed-power first data set of a wind farm, and screening the wind speed-power first data set to obtain a second data set;

[0010] Calculating a first set inner limit of the second data set by using a horizontal quartile method, and deleting points in the second data set whose wind speed is outside the first set inner limit;

[0011] Calculating a second set inner limit of the second data set by using a vertical quartile method, and deleting points in the second data set whose power is outside the second set inner limit to obtain a third data set;

[0012] Performing a first operation on all points in the third data set by using a clustering algorithm, and outputting a wind speed-power data set identified as normal.

[0013] As a preferred scheme of the wind farm outlier screening method of the present application, wherein: calculating the first set inner limit of the second data set by using the horizontal quartile method comprises,

[0014] Calculating the first set inner limit of the wind speed in the second data set by using the horizontal quartile method, which is expressed as:

[0015]

[0016] Wherein, Q s1 , Q s3 are the first and third quantiles of the wind speed data set, IQR s , IQR p are the quartile ranges of the wind speed data set and the power data set, [W s，l , W s，h ] are the set inner limits of the wind speed data set.

[0017] As a preferred scheme of the wind farm outlier screening method of the present application, wherein: calculating the second set inner limit of the second data by using the vertical quartile method comprises,

[0018] Calculating the second set inner limit of the power in the second data set by using the vertical quartile method, which is expressed as:

[0019]

[0020] wherein Q p1 , Q p3 are the first and third quantiles of the power dataset, IQR, IQRp are the interquartile ranges of the wind speed dataset and the power dataset, [W p，l , W p，h ] is the inner set limit of the power dataset.

[0021] As a preferred scheme of the wind farm outlier screening method, the clustering algorithm comprises,

[0022] The DBSCAN clustering algorithm is used for secondary outlier screening on the third data set, and the radius eps and the minimum inner point MinPts parameters are selected.

[0023] As a preferred scheme of the wind farm outlier screening method, the first operation comprises,

[0024] All points in the third data set are scanned and marked as core points, boundary points or noise points respectively, and the noise points are removed.

[0025] As a preferred scheme of the wind farm outlier screening method, the core points, boundary points or noise points comprise,

[0026] The points with more than MinPts number of points within the radius eps are represented as core points;

[0027] The points with less than MinPts number of points within the radius eps but located in the field of core points are represented as boundary points;

[0028] The points with less than MinPts number of points within the radius eps but located outside the field of any core points are represented as noise points.

[0029] As a preferred scheme of the wind farm outlier screening method, the normal wind speed-power data set is represented as:

[0030] A Q-D = {(S i , P i )|i = 1, 2,..., l}

[0031] wherein S i is the wind speed data, P i is the power data corresponding to the wind speed, l is the number of wind speed-power points screened by DBSCAN clustering, and A Q-D is the normal wind speed-power data set.

[0032] In a second aspect, the present application provides a system for screening abnormal values of a wind farm, comprising,

[0033] obtaining a first wind speed-power data set of the wind farm, screening the first wind speed-power data set of the wind farm to obtain a second data set;

[0034] calculating a first inner limit of the second data set by using a horizontal quartile method, and deleting points in the second data set whose wind speed is outside the first inner limit;

[0035] calculating a second inner limit of the second data set by using a vertical quartile method, and deleting points in the second data set whose power is outside the second inner limit to obtain a third data set;

[0036] performing a first operation on all points in the third data set by using a clustering algorithm, and outputting a wind speed-power data set identified as normal.

[0037] In a third aspect, the present application provides a computing device, comprising:

[0038] a memory and a processor;

[0039] the memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions, so as to implement the steps of the method for screening abnormal values of a wind farm.

[0040] In a fourth aspect, the present application provides a computer readable storage medium storing computer executable instructions, which, when executed by a processor, implement the steps of the method for screening abnormal values of a wind farm.

[0041] Compared with the prior art, the present application has the following beneficial effects: the present application systematically processes wind speed-power data of a wind farm by using a method for screening abnormal values of a wind farm, screens a reasonable second data set from an original data set, and then accurately defines normal ranges of wind speed and power by using horizontal and vertical quartile methods, so as to eliminate abnormal values and obtain a more accurate and reliable third data set; the DBSCAN clustering algorithm is further used for screening, so as to ensure that only normal wind speed-power data points are contained in the data set, improve data quality, and provide a solid foundation for subsequent performance analysis, operation and maintenance management of the wind farm, and improvement of energy output efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor. Among them:

[0043] Figure 1 The overall flowchart of the wind farm outlier screening method according to one embodiment of the present application is shown in the figure.

[0044] Figure 2 The wind speed-power scatter diagram after screening outliers by the quartile method in the wind farm outlier screening method according to one embodiment of the present application is shown in the figure.

[0045] Figure 3 The wind speed-power scatter diagram after screening outliers by the DBSCAN clustering method in the wind farm outlier screening method according to one embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0046] In order to make the above objectives, features and advantages of the present application more apparent, specific embodiments of the present application will be described in detail below with the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work should fall within the scope of protection of the present application.

[0047] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application. However, the present application can be practiced without the specific details, other than those described herein, and it is understood that the present application is not limited to the embodiments described herein and can be practiced with or without other apparatuses, systems, structures, methodologies, etc. The present application is therefore understood to extend to all alternatives, modifications, and equivalents falling within the scope of the application.

[0048] Secondly, the term "one embodiment" or "an embodiment" as used herein means that a particular implementation can include a particular feature, structure, or characteristic. However, such appearances of the phrase "in one embodiment" or "in an embodiment" are not necessarily referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in one embodiment, it is understood that such features, structures, or characteristics can be combined with one or more other features, structures, or characteristics in the same or other embodiments.

[0049] The present application is described in detail with reference to the accompanying drawings. In the detailed description of the embodiments of the present application, the sectional view of the device structure is locally enlarged without the general proportion for the convenience of description, and the schematic diagram is only an example, which should not limit the scope of protection of the present application. In addition, the three-dimensional spatial dimensions of length, width and depth should be included in the actual manufacture.

[0050] Meanwhile, in the description of the present application, it should be noted that the terms "upper, lower, inner and outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the systems or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first, second or third" are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0051] Unless otherwise clearly specified and limited, the terms "mounting, connecting, connecting" in the present application should be understood broadly, for example: it can be fixed connection, detachable connection or integral connection; it can also be mechanical connection, electrical connection or direct connection, it can also be indirectly connected through an intermediate medium, or it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0052] Embodiment 1

[0053] Reference Figure 1 For an embodiment of the present application, a wind farm outlier screening method is provided, comprising:

[0054] S100: Obtain a wind farm wind speed-power first data set, and screen the wind farm wind speed-power first data set to obtain a second data set;

[0055] Specifically, input the wind farm wind speed data S and the corresponding power data P;

[0056] Assuming that the number of wind farm wind speed-power points is n, the wind farm wind speed-power first data set is obtained, denoted as:

[0057] A0={(S i ,P i )|i=1,2,...,n}

[0058] When the wind speed s∈[s ci ,s co ], wherein s ci ,s co represent the cut-in and cut-out wind speeds of the wind turbine respectively;

[0059] In the embodiments of the present application, the screening operation is: modifying the wind power to 0 when the wind power is less than 0, and eliminating the data greater than the rated power of the wind farm, to obtain m wind speed-power points, i.e. the second data set, denoted as A1={(S i ,P i )|i=1,2,...,m};

[0060] It should be noted that the present invention effectively cleans up unreasonable data points in the wind speed-power data set of wind farms, such as correcting negative power to zero and eliminating outliers that exceed the rated power of wind farms. This processing step not only ensures the accuracy and reliability of the data set.

[0061] S102: Calculate the first set inner limit of the second data set using the lateral quartile method, and delete the points in the second data set whose wind speed is outside the first set inner limit;

[0062] Preferably, the first set inner limit of wind speed in the second data set is calculated using the lateral quartile method, and is expressed as:

[0063]

[0064] Among them, Q s1 Q s3 The first and third quantiles of the wind speed dataset, IQR s IQR p For the interquartile range of the wind speed dataset and the power dataset, [W s，l W s，h [ ] represents the inner bound of the wind speed dataset;

[0065] In this embodiment, the power is first divided into p intervals to obtain p sets of wind speed data samples, which are represented as follows:

[0066]

[0067] Where q j This represents the number of wind speed samples contained in the j-th sample set;

[0068] In this embodiment of the application, the first and third quantiles Q of the j-th wind speed data sample set are calculated. s1，j Q s3，j , is represented as:

[0069]

[0070]

[0071] In this embodiment of the application, based on the first and third quantiles Q s1，j Q s3，j The interquartile range (IQR) of the j-th wind speed data sample set was calculated. s，j , is represented as:

[0072] IQR s，j =Q s3 -Q s1

[0073] In this embodiment of the application, based on the interquartile range (IQR)s，j The upper limit W of the jth wind speed data sample set Whisker is calculated s，j，h The lower limit W s，j，l is expressed as:

[0074]

[0075] The ith wind speed data set S is defined j The samples between W s，j，l and W s，j，h are normal samples, and the remaining points in the set are considered abnormal samples and need to be screened and removed.

[0076] According to the above process, the abnormal points in the p wind speed data sample sets are removed.

[0077] It should be noted that through the above process of calculating the first set limit of wind speed in the second data set by using the horizontal quartile method and removing the abnormal points outside the limit, the wind speed-power data set of the wind farm is further purified. This method can accurately identify and exclude abnormal wind speed data caused by factors such as sensor failure, data transmission error or extreme weather conditions, ensuring that the remaining data points can better reflect the actual operating conditions of the wind farm.

[0078] S104: The second set limit of the second data set is calculated by using the longitudinal quartile method, and the points in the second data set whose power is outside the second set limit are deleted to obtain a third data set;

[0079] Preferably, the second set limit of the power in the second data set is calculated by using the longitudinal quartile method, and is expressed as:

[0080]

[0081] Wherein, Q p1 , Q p3 are the first and third quantiles of the power data set, IQR s , IQR p are the interquartile ranges of the wind speed data set and the power data set, [W p，l , W p，h ] is the set limit of the power data set.

[0082] In the embodiments of the present application, the wind speed is divided into a intervals, that is, a power data sample sets can be divided, which is expressed as:

[0083]

[0084] Wherein b j represents the number of power samples contained in the jth sample set.

[0085] In this embodiment of the application, the first and third quantiles Q of the j-th power data sample set are calculated. p1，j Q p3，j , is represented as:

[0086]

[0087]

[0088] In this embodiment of the application, based on the first and third quantiles Q p1，j Q p3，j The interquartile range (IQR) of the j-th power data sample set was calculated. p，j , is represented as:

[0089] IQR p，j =Q p3 -Q p1

[0090] In this embodiment of the application, based on the interquartile range (IQR) p，j The Whisker upper bound W for the j-th power data sample set is calculated. p，j，h Lower limit W p，j，l , is represented as:

[0091]

[0092] Define the i-th power dataset P j In the middle of W p，j，l and W p，j，h The samples between are considered normal samples, while the remaining points in the set are considered abnormal samples and need to be filtered and removed.

[0093] The outliers in the set of power data samples a are removed following the above process.

[0094] By employing the lateral and vertical quartile methods, points in the second dataset whose wind speed and power values ​​fall outside the set limits are removed, resulting in the wind speed-power set filtered by the quartile method, which is the third dataset, denoted as A. Q ={(S i P i |i = 1, 2, ..., k};

[0095] like Figure 2 As shown, this is a scatter plot of wind speed-power after outlier screening using the interquartile range method. Green sample points indicate data identified as normal, while red sample points indicate data identified as outlier. The interquartile range method screens out most isolated discrete and biased clustered outlier data. Outlier residual fragment 1 and residual fragment 2 in the figure are some unscreened horizontally stacked outlier data.

[0096] It should be noted that by employing both horizontal and vertical quartile methods to perform dual filtering on the second dataset, this invention achieves deep purification of wind farm wind speed-power data. This method not only eliminates outlier data from the wind speed dimension but also further filters data from the power dimension, ensuring the rationality and matching of wind speed and power values ​​in the dataset. Through this step, the resulting third dataset more accurately reflects the actual operating characteristics of the wind farm.

[0097] S106: A clustering algorithm is used to perform the first operation on all points in the third data set, and the output is identified as a normal wind speed-power data set;

[0098] Preferably, the DBSCAN clustering algorithm is used to perform secondary outlier screening on the third dataset, selecting the radius eps and the minimum interior point MinPts parameters.

[0099] Preferably, all points in the third data set are scanned and marked as core points, boundary points, or noise points respectively, and noise points are removed.

[0100] Preferably, points containing more than MinPts within the radius eps are represented as core points;

[0101] Preferably, the number of points within the radius eps is less than MinPts, but points located within the neighborhood of the core point are represented as boundary points;

[0102] Preferably, the number of points within the radius eps is less than MinPts, but points located outside the neighborhood of any core point are represented as noise points.

[0103] The preferred, normal wind speed-power data set is represented as follows:

[0104] A Q-D ={(S i P i |i = l, 2, ..., l}

[0105] Among them, S i For wind speed data, P i Here, l represents the power data corresponding to wind speed, and l represents the number of wind speed-power points selected by DBSCAN clustering. Q-D This is a normal set of wind speed-power data;

[0106] like Figure 3 The image shows a wind speed-power scatter plot after secondary screening of outliers using DBSCAN clustering. Figure 3 It can be seen that the remaining horizontally stacked value segments after the quartile method screening have been completely screened out after DBSCAN clustering;

[0107] It should be noted that the process of performing secondary outlier screening on the third data set by using the DBSCAN clustering algorithm further improves the accuracy and reliability of the wind speed-power data set. The DBSCAN algorithm effectively distinguishes core points, boundary points and noise points by scanning data points and based on the preset radius and minimum inner point parameters, thereby realizing accurate removal of abnormal data. This process not only ensures that the remaining data points can better reflect the actual operating conditions of the wind farm.

[0108] The above is a schematic scheme of a wind farm outlier screening method of the embodiment. It should be noted that the technical scheme of the wind farm outlier screening system belongs to the same concept as the technical scheme of the wind farm outlier screening method described above. The technical scheme of the wind farm outlier screening system in the embodiment is not described in detail, and the description of the technical scheme of the wind farm outlier screening method described above can be referred to.

[0109] The wind farm outlier screening system in the embodiment comprises:

[0110] Obtain a wind speed-power first data set of a wind farm, and screen the wind speed-power first data set of the wind farm to obtain a second data set;

[0111] A first set inner limit of the second data set is calculated by using a horizontal quartile method, and points in the second data set whose wind speed is outside the first set inner limit are deleted;

[0112] A second set inner limit of the second data set is calculated by using a vertical quartile method, and points in the second data set whose power is outside the second set inner limit are deleted to obtain a third data set;

[0113] A first operation is performed on all points in the third data set by using a clustering algorithm, and a wind speed-power data set recognized as normal is output.

[0114] The embodiment also provides a computing device suitable for wind farm outlier screening, comprising:

[0115] A memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to realize the wind farm outlier screening method proposed in the above embodiment.

[0116] The embodiment also provides a storage medium having a computer program stored thereon, and the program is executed by a processor to realize the wind farm outlier screening method proposed in the above embodiment.

[0117] The storage medium proposed in the embodiment belongs to the same inventive concept as the method for screening abnormal values of a wind farm proposed in the above embodiment, and the technical details not described in detail in the embodiment can be seen from the above embodiment, and the embodiment has the same beneficial effects as the above embodiment.

[0118] Through the above description of the embodiments, those skilled in the art can clearly understand that the present application can be realized by software and necessary general hardware, and of course can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a FLASH, a hard disk or an optical disk, etc., including a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the method of each embodiment of the present application.

[0119] Embodiment 2

[0120] Referring to Table 1, for an embodiment of the present application, a method for screening abnormal values of a wind farm is provided, and comparative results of several schemes are provided to verify the beneficial effects.

[0121] In order to verify the superiority of the method of the present application, four models are used to select one-year actual operation data of a wind farm in South China for research, to identify and screen abnormal values, and to compare their effects.

[0122] Table 1 lists the comparative results of the correct identification rate of abnormal values by different methods;

[0123] Table 1 Comparative of correct identification rate of abnormal values by different methods

[0124] Outlier screening method Outlier identification accuracy Quartile method and DBSCAN clustering 93.75％ k-means clustering 87.29％ Isolation forest algorithm 82.72％ RANSAC robust regression algorithm 89.83％

[0125] As can be seen from Table 1, the correct identification rate is increased by about 6% compared with the K-means clustering, is increased by about 11% compared with the Isolation Forest algorithm, and is increased by about 4% compared with the RANSAC robust regression algorithm.

[0126] Through the above data analysis, it can be seen that the performance and accuracy of the wind farm abnormal value identification and screening by combining the quartile method and DBSCAN clustering are obviously better than those of other models.

[0127] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced, without departing from the spirit and scope of the technical solutions of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. A wind farm outlier screening method, characterized in that, The method comprises the following steps: obtaining a first wind speed-power data set of a wind farm, and screening the first wind speed-power data set to obtain a second data set; calculating a first set inner limit of the second data set by using a horizontal quartile method, and deleting points in the second data set whose wind speed is outside the first set inner limit; calculating the first set inner limit of the wind speed in the second data set by using the horizontal quartile method, and representing the first set inner limit as: where Q s1 , Q s3 are the first and third quantiles of the wind speed dataset, IQR s is the interquartile range of the wind speed dataset, [W s,l , W s,h ] is the inner set limit of the wind speed dataset; calculating a second set inner limit of the second data set by using a vertical quartile method, and deleting points in the second data set whose power is outside the second set inner limit to obtain a third data set; calculating the second set inner limit of the power in the second data set by using the vertical quartile method, and representing the second set inner limit as: where Q p1 , Q p3 are the first and third quartiles of the power dataset, IQR p is the interquartile range of the power dataset, [W p,l , W p,h ] is the inner set limit of the power dataset; performing a first operation on all points in the third data set by using a clustering algorithm, and outputting a normal wind speed-power data set; performing secondary outlier screening on the third data set by using a DBSCAN clustering algorithm, and selecting a radius eps and a minimum inner point MinPts parameter; scanning all points in the third data set and marking the points as core points, boundary points or noise points respectively, and deleting the noise points.

2. The wind farm outlier screening method of claim 1, wherein, The core points, boundary points or noise points comprise: a point having more than MinPts points within a radius eps is a core point; a point having less than MinPts points within the radius eps but located in a domain of a core point is a boundary point; a point having less than MinPts points within the radius eps but located outside the domain of any core point is a noise point.

3. The wind farm outlier screening method of claim 2, wherein, The normal wind speed-power data set is represented as: A Q-D = {(S i , P i )|i = 1, 2,..., I} Wherein, S i is wind speed data, P i is wind speed corresponding to the power data, l is the number of wind speed-power point filtered by DBSCAN clustering, A Q-D is the normal wind speed-power data set.

4. A system for wind farm outlier screening, characterized in that, The method comprises the following steps: obtaining a first wind speed-power data set of a wind farm, and screening the first wind speed-power data set to obtain a second data set; calculating a first set inner limit of the second data set by using a horizontal quartile method, and deleting points in the second data set whose wind speed is outside the first set inner limit; calculating the first set inner limit of the wind speed in the second data set by using the horizontal quartile method, and representing the first set inner limit as: where Q s1 , Q s3 are the first and third quartiles of the wind speed dataset, IQR s is the interquartile range of the wind speed dataset, [W s,l , W s,h ] is the inner set limit of the wind speed dataset; calculating a second set inner limit of the second data set by using a vertical quartile method, and deleting points in the second data set whose power is outside the second set inner limit to obtain a third data set; calculating the second set inner limit of the power in the second data set by using the vertical quartile method, and representing the second set inner limit as: where Q p1 , Q p3 are the first and third quartiles of the power dataset, IQR p is the interquartile range of the power dataset, [W p,l , W p,h ] is the inner set limit of the power dataset; performing a first operation on all points in the third data set by using a clustering algorithm, and outputting a normal wind speed-power data set; performing secondary outlier screening on the third data set by using a DBSCAN clustering algorithm, and selecting a radius eps and a minimum inner point MinPts parameter; scanning all points in the third data set and marking the points as core points, boundary points or noise points respectively, and deleting the noise points. 5.An electronic device comprising: a memory and a processor; the memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions, and the computer executable instructions, when executed by the processor, implement the steps of the wind farm outlier screening method according to any one of claims 1 to 3.

6. A computer readable storage medium storing computer executable instructions which, when executed by a processor, implement the steps of the wind farm outlier screening method of any one of claims 1 to 3.

Images