Wind speed-power curve abnormal point identification method and device

Abstract

The invention provides a wind speed-power curve abnormal point identification method and device, and belongs to the technical field of wind power generation. The method comprises the following steps: acquiring a plurality of wind speed-power data points of a wind turbine generator in a preset time period; identifying the plurality of wind speed-power data points by using different curve abnormal value identification methods to obtain wind speed-power abnormal data points under the different curve abnormal value identification methods; and determining the same wind speed-power abnormal data points under different curve abnormal value identification methods as wind speed-power target abnormal data points. According to the method, the density clustering algorithm and the wind speed segmentation algorithm are fused, so that accurate identification of the abnormal points of the wind turbine generator is realized, the operation efficiency is improved, the maintenance cost is reduced, and safe and stable operation of the generator is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of wind power generation, and in particular to a method for identifying abnormal points in a wind speed-power curve and a device for identifying abnormal points in a wind speed-power curve. Background Art

[0002] With the advancement of wind power technology and the rapid growth of installed capacity, it is particularly important to quickly and efficiently evaluate the operation and power generation efficiency of wind turbines. The power curve is one of the most important indicators for measuring the operation of wind turbines. Therefore, it is of great significance to develop a universal and accurate method for identifying abnormal points in the power curve.

[0003] At present, the technical methods for identifying abnormal points in power curves mainly include data cleaning and preprocessing, statistical analysis, machine learning algorithms, image recognition technology, and model prediction. These methods identify and judge abnormal points in power curves by cleaning data, calculating statistical indicators, learning historical data to build models, converting curves into images for analysis, and building prediction models.

[0004] However, the existing technical methods for identifying abnormal points on the power curve have the problems of low efficiency and low accuracy, and are not suitable for abnormal judgment of batch wind turbines. Summary of the invention

[0005] The purpose of the embodiments of the present invention is to provide a method and device for identifying abnormal points of a wind speed-power curve to solve the above-mentioned problem.

[0006] In order to achieve the above object, an embodiment of the present invention provides a method for identifying abnormal points of a wind speed-power curve, comprising: Obtain multiple wind speed-power data points of the wind turbine generator set within a preset time period; Using different curve outlier recognition methods, multiple wind speed-power data points are recognized to obtain wind speed-power abnormal data points under different curve outlier recognition methods; The same wind speed-power abnormal data points under different curve outlier recognition methods are determined as wind speed-power target abnormal data points.

[0007] Optionally, different curve outlier recognition methods are used to recognize multiple wind speed-power data points, and wind speed-power abnormal data points under different curve outlier recognition methods are obtained, including: Using the density clustering algorithm, multiple wind speed-power data points are identified to obtain abnormal wind speed-power data points under the density clustering algorithm; The wind speed segmentation algorithm is used to identify multiple wind speed-power data points, and the abnormal wind speed-power data points under the wind speed segmentation algorithm are obtained.

[0008] Optionally, a density clustering algorithm is used to identify multiple wind speed-power data points to obtain abnormal wind speed-power data points under the density clustering algorithm, including: Determine the cluster radius and minimum number of cluster points for the density clustering algorithm; Counting the number of wind speed-power data points other than the wind speed-power data point within the clustering radius of each wind speed-power data point to determine the density of each wind speed-power data point; Determine the wind speed-power data points with a density greater than the minimum number of cluster points as wind speed-power target data points, and determine the wind speed-power data points other than the wind speed-power target data points as wind speed-power abnormal data points; The abnormal wind speed-power data points that are not within the clustering radius of any wind speed-power target data point are determined as abnormal wind speed-power data points under the density clustering algorithm.

[0009] Optionally, the cluster radius of the density clustering algorithm is determined by: Combining multiple wind speed-power data points in pairs to obtain multiple wind speed-power data point groups; Calculate the distance between two wind speed-power data points in each wind speed-power data point group to obtain the Euclidean distance of each wind speed-power data point group; Combining the Euclidean distances of multiple wind speed-power data point groups in pairs to obtain multiple Euclidean distance groups; Calculate the difference of the Euclidean distances between two wind speed-power data point groups in each Euclidean distance group, obtain the difference of each Euclidean distance group, and take the Euclidean distance group with the largest difference as the target Euclidean distance group; The average of the Euclidean distances of the two wind speed-power data point groups in the target Euclidean distance group is calculated, and the average is used as the cluster radius of the density clustering algorithm.

[0010] Optionally, counting the number of wind speed-power data points other than the wind speed-power data point within the clustering radius of each wind speed-power data point to determine the density of each wind speed-power data point includes: For each wind speed-power data point: Determine wind speed-power data points other than the current wind speed-power data point as wind speed-power data points to be identified; Calculate the distance between the current wind speed-power data point and each wind speed-power data point to be identified, and obtain multiple Euclidean distances of the wind speed-power data points; The number of wind speed-power data points to be identified corresponding to the Euclidean distance less than the clustering radius is counted to determine the density of the wind speed-power data points.

[0011] Optionally, a wind speed segmentation algorithm is used to identify multiple wind speed-power data points to obtain abnormal wind speed-power data points under the wind speed segmentation algorithm, including: Determine the wind speed-power data points with a wind speed less than a preset rated wind speed as a first wind speed-power data point group, and determine the wind speed-power data points with a wind speed greater than or equal to the preset rated wind speed as a second wind speed-power data point group; Determine a lower power limit value and an upper power limit value of the first wind speed-power data point group and the second wind speed-power data point group; Determine the first wind speed-power data point in the first wind speed-power data point group whose power is less than the power lower limit value of the first wind speed-power data point group or whose power is greater than the power upper limit value of the first wind speed-power data point group as the first abnormal wind speed-power data point; Determine a second wind speed-power data point in the second wind speed-power data point group whose power is less than a lower power limit value of the second wind speed-power data point group or whose power is greater than an upper power limit value of the second wind speed-power data point group as a second abnormal wind speed-power data point; The first abnormal data point of wind speed-power and the second abnormal data point of wind speed-power are taken as the abnormal data points of wind speed-power under the wind speed segmentation algorithm.

[0012] In a second aspect of an embodiment of the present invention, a device for identifying abnormal points of a wind speed-power curve is provided, comprising: A data acquisition module, used to acquire multiple wind speed-power data points of a wind turbine generator set within a preset time period; An abnormality identification module is used to identify multiple wind speed-power data points using different curve abnormality identification methods to obtain wind speed-power abnormality data points under different curve abnormality identification methods; The cross-validation module is used to determine that the same wind speed-power abnormal data point under different curve abnormal value identification methods is a wind speed-power target abnormal data point.

[0013] Optionally, the anomaly identification module includes: a clustering anomaly identification unit and a segmentation anomaly identification unit; A clustering anomaly identification unit is used to identify multiple wind speed-power data points using a density clustering algorithm to obtain wind speed-power anomaly data points under the density clustering algorithm; The segmented anomaly recognition unit is used to recognize multiple wind speed-power data points using a wind speed segmentation algorithm to obtain wind speed-power anomaly data points under the wind speed segmentation algorithm.

[0014] Optionally, the clustering anomaly identification unit is specifically used for: Determine the cluster radius and minimum number of cluster points for the density clustering algorithm; Counting the number of wind speed-power data points other than the wind speed-power data point within the clustering radius of each wind speed-power data point to determine the density of each wind speed-power data point; Determine the wind speed-power data points with a density greater than the minimum number of cluster points as wind speed-power target data points, and determine the wind speed-power data points other than the wind speed-power target data points as wind speed-power abnormal data points; The abnormal wind speed-power data points that are not within the clustering radius of any wind speed-power target data point are determined as abnormal wind speed-power data points under the density clustering algorithm.

[0015] Beneficial effects of the present invention: (1) Improving the accuracy and robustness of outlier identification: This method combines the density clustering algorithm and the wind speed segmentation algorithm. It can not only identify the outliers in each wind speed segment of the power curve, but also reduce the false alarms in the climbing segment and full-power segment of the power curve. It can also eliminate false identification through cross-validation, thereby achieving accurate and efficient identification of outliers in the power curve.

[0016] (2) Timely discovery and diagnosis of wind turbine operation problems: Abnormal point identification technology can monitor the operating status of wind turbines in real time, provide early warnings, and assist in fault location and diagnosis, effectively avoiding major losses and ensuring safe and stable operation of the units.

[0017] (3) Improve the operating efficiency of wind turbines and reduce maintenance costs: By analyzing abnormal points, optimizing operating strategies, adjusting operating parameters, improving power generation efficiency, and implementing preventive maintenance, the service life of the units can be extended and maintenance costs can be reduced.

[0018] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings: Figure 1 1 is a flow chart of a method for identifying abnormal points on a wind speed-power curve provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of the principle of the density clustering algorithm provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of the structure of a device for identifying abnormal points on a wind speed-power curve provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0022] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0023] Embodiment 1 Please refer to Figure 1 , Figure 1 : is a flow chart of a method for identifying abnormal points on a wind speed-power curve provided by an embodiment of the present invention, the method comprising the following steps: S100, obtaining a plurality of wind speed-power data points of a wind turbine generator set within a preset time period; It is understandable that the known curve is connected by multiple coordinate points, so the wind speed-power curve of the wind turbine generator set in the preset time period can also be understood as multiple wind speed-power data points of the wind turbine generator set in the preset time period.

[0024] It should be noted that the multiple wind speed-power data points are continuous in time, that is, each wind speed-power data point has a timestamp, as shown in Table 1 below.

[0025] Table 1 Wind speed-power data points

[0026] S200, using different curve outlier recognition methods to identify multiple wind speed-power data points, to obtain wind speed-power abnormal data points under different curve outlier recognition methods; In one embodiment, step S200 specifically includes the following steps: S210, using a density clustering algorithm to identify multiple wind speed-power data points, and obtain wind speed-power abnormal data points under the density clustering algorithm; The Density-Based Spatial Clustering of Applications with Noise (DBSCAN) is a density-based data clustering algorithm. This algorithm divides regions with sufficiently high density into clusters and can discover clusters of any shape.

[0027] The basic concepts of the density clustering algorithm include: Core Point: A point p is a core point if the region within its radius ε contains at least MinPts points.

[0028] Border Point: A point p is a border point if it is not a core point, but there exists a core point within its radius ε.

[0029] Noise Point: A point that is neither a core point nor a border point is called a noise point.

[0030] The working process of the density clustering algorithm is as follows: (1) Select an unvisited point p from the database.

[0031] (2) If p is a core point, create a new cluster C.

[0032] (3) For all directly density-reachable points within the radius ε of p, add them to cluster C and mark them as visited.

[0033] (4) For each point newly added to cluster C, repeat steps 2 and 3 until no new points can be added to cluster C.

[0034] (5) If p is not a core point, mark p as a noise point.

[0035] (6) Repeat steps 1 to 5 until all points are visited.

[0036] In the density clustering algorithm, these concepts define the relationships between points and how they form clusters, as Figure 2 shown: Directly Density-Reachable: If point p is a core point and point q is within the ε-neighborhood of p, then point q is said to be directly density-reachable from point p.

[0037] Density-Reachable: If there exists a sequence of points p1, p2,..., pn, where p1 = p, pn = q, and for all 1 ≤ i < n, pi+1 is directly density-reachable from pi, then point q is said to be density-reachable from point p.

[0038] Density-Connected: If there exists a point p such that both points q and r are density-reachable from point p, then points q and r are said to be density-connected.

[0039] Not Density-Connected: If points q and r are not density-connected, they are said to be not density-connected.

[0040] The DBSCAN algorithm uses these definitions to identify and divide clusters. If a point p is a core point and there is a point q in the ε-neighborhood of p, then q and p belong to the same cluster. If a point is not a core point and there is no other point that can directly reach its density, then this point is considered a noise point.

[0041] Specifically, step S210 includes: S211, determining the clustering radius and the minimum number of clustering points of the density clustering algorithm; Specifically, the clustering radius of the density clustering algorithm is determined by: Step 1: Combine multiple wind speed-power data points in pairs to obtain multiple wind speed-power data point groups; Step 2: Calculate the distance between two wind speed-power data points in each wind speed-power data point group to obtain the Euclidean distance of each wind speed-power data point group; Step 3: Combine the Euclidean distances of multiple wind speed-power data point groups in pairs to obtain multiple Euclidean distance groups; Step 4: Calculate the difference in the Euclidean distances between the two wind speed-power data point groups in each Euclidean distance group, obtain the difference between each Euclidean distance group, and take the Euclidean distance group with the largest difference as the target Euclidean distance group; Step 5: Calculate the average of the Euclidean distances of the two wind speed-power data point groups in the target Euclidean distance group, and use the average as the clustering radius of the density clustering algorithm.

[0042] For ease of understanding, the following is an example with actual data: Assume there are 10 wind speed-power data points: [4.5m / s, 90Kw], [4.7m / s, 95Kw], [4.9m / s, 100Kw], [5.1m / s, 120Kw], [5.3m / s, 130Kw], [5.5m, 140Kw], [5.7m / s, 150Kw], [5.9m / s, 160Kw], [6.0m / s, 155Kw], [5.8m / s, 155Kw] and [5.5m / s, 140Kw].

[0043] Step 1: Data point combination.

[0044] There is no need to actually list all the combinations, just know that there are C(12, 2) = 66 combinations.

[0045] Step 2: Calculate the Euclidean distance.

[0046] Taking the first pair of data points [4.5, 90] and [4.7, 95] as an example, calculate the Euclidean distance between them:

[0047] This process is repeated to calculate the Euclidean distances for all combinations.

[0048] Step 3: Euclidean distance group.

[0049] All calculated Euclidean distances are combined pairwise.

[0050] Step 4: Calculate the difference between distance groups.

[0051] For each combination, calculate the difference between the two distances. We only focus on the combination with the largest difference. For example, if there is a combination (5, 10), the difference is 5.

[0052] Step 5: Determine the cluster radius.

[0053] In the combination with the largest difference, the average of the two Euclidean distances is taken as the cluster radius.

[0054] Suppose we find the combination with the largest difference. Suppose this combination is (5, 10) and the mean is 7.5, then the cluster radius will be 7.5.

[0055] In one embodiment, the minimum number of cluster points of the density clustering algorithm is determined by: The dimension of the wind speed-power data point and the sum of the preset number of clustering points are calculated to obtain the minimum number of clustering points for the density clustering algorithm.

[0056] Specifically, the wind speed-power data point is a two-dimensional coordinate, so its dimension is 2. The preset number of clustering points is at least greater than 1 and cannot be zero, otherwise it is meaningless. The embodiment of the present invention does not specifically limit the setting of the preset number of clustering points, and can be reasonably set according to actual needs.

[0057] In one embodiment, the clustering radius and the minimum number of clustering points of the density clustering algorithm may also be determined in the following manner: The cluster radius uses 21 values ​​ranging from 0.5 to 2.5 (data interval is 0.1), the minimum number of cluster points uses 9 values ​​ranging from 2 to 10 (data interval is 1), and the two parameters are combined in pairs, totaling 189 combinations. Different parameter combinations are used for the SCADA data of one wind turbine to identify power curve outliers through density clustering algorithm and observe the outlier identification. It can be found that: (1) When the cluster radius remains unchanged and the minimum number of cluster points increases, more and more points will be marked as outliers; (2) When the number of cluster points remains unchanged and the minimum cluster radius increases, fewer and fewer points are identified as outliers; (3) When the cluster radius is greater than 2.0, there is no significant difference in the identification of outliers; (4) When the minimum number of cluster points exceeds 4, the anomaly recognition results no longer change significantly as the minimum number of cluster points increases.

[0058] Through analysis, a set of clustering radius and minimum number of clustering points that can better identify outliers is selected.

[0059] S212, counting the number of wind speed-power data points other than the wind speed-power data point within the clustering radius of each wind speed-power data point to determine the density of each wind speed-power data point; Specifically, step S212 includes: For each wind speed-power data point: S2121, determining wind speed-power data points other than the current wind speed-power data point as wind speed-power data points to be identified; S2122, calculating the distance between the current wind speed-power data point and each wind speed-power data point to be identified, and obtaining multiple Euclidean distances of the wind speed-power data points; S2123, counting the number of wind speed-power data points to be identified corresponding to the Euclidean distance less than the clustering radius to determine the density of the wind speed-power data points.

[0060] For ease of understanding, the following is an example with actual data: Assume there are 10 wind speed-power data points: [4.5m / s, 90Kw], [4.7m / s, 95Kw], [4.9m / s, 100Kw], [5.1m / s, 120Kw], [5.3m / s, 130Kw], [5.5m, 140Kw], [5.7m / s, 150Kw], [5.9m / s, 160Kw], [6.0m / s, 155Kw], [5.8m / s, 155Kw] and [5.5m / s, 140Kw].

[0061] Step 1: We will select each wind speed-power data point and then treat the remaining 10 data points as data points to be identified.

[0062] Step 2 (Euclidean distance calculation): For each selected data point, we will calculate its Euclidean distance from the remaining data points.

[0063] Taking the data point [4.5m / s, 90Kw] as an example, we need to calculate its distance from the following data points:

[0064]

[0065]

[0066]

[0067]

[0068]

[0069]

[0070]

[0071]

[0072]

[0073] Among them, [5.5m / s, 140Kw] (note that there are two identical points here, but we only calculate them once when calculating the distance).

[0074] The calculation formula of Euclidean distance is: Step 3 (density determination): Suppose we set the clustering radius to ε = 50. We need to count the number of neighbors of each data point within the radius of 50.

[0075] Here is a statistical example: For the data point [4.5m / s, 90Kw], we calculate its distance from all other data points and count the number of data points whose distance is less than the cluster radius (50). For example, for the data point [4.5m / s, 90Kw], we may find: The distance to [4.7m / s, 95Kw] is 5, which is less than the cluster radius (50).

[0076] The distance to [4.9m / s, 100Kw] is 10.008, which is smaller than the cluster radius (50).

[0077] The distance to [5.1m / s, 120Kw] is 34.54, which is smaller than the cluster radius (50).

[0078] The distance to [5.3m / s, 130Kw] is 50.02, which is greater than the cluster radius (50).

[0079] The distance to [5.5m. / s, 140Kw] is 65.42, which is larger than the cluster radius (50).

[0080] The distance to [5.7m / s, 150Kw] is 80.77, which is larger than the cluster radius (50).

[0081] The distance to [5.9m / s, 160Kw] is 95.99, which is greater than the cluster radius (50).

[0082] The distance to [6.0m / s, 155Kw] is 98.68, which is larger than the cluster radius (50).

[0083] The distance to [5.8m / s, 155Kw] is 97.38, which is larger than the cluster radius (50).

[0084] The distance to [5.5m / s, 140Kw] is 65.42, which is greater than the cluster radius (50).

[0085] Therefore, the number of wind speed-power data points of [4.5m / s, 90Kw] in the cluster radius (50) is 3, that is, the density is 3.

[0086] S213, determining wind speed-power data points with a density greater than the minimum number of cluster points as wind speed-power target data points, and determining wind speed-power data points other than the wind speed-power target data points as wind speed-power abnormal data points; S214, determining that the abnormal wind speed-power data points that are not within the clustering radius of any wind speed-power target data point are abnormal wind speed-power data points under the density clustering algorithm.

[0087] Specifically, check whether the wind speed-power anomaly data point is within the cluster radius of any target data point. If not, the point is confirmed as the final wind speed-power anomaly data point. S220, using a wind speed segmentation algorithm, identifying multiple wind speed-power data points to obtain abnormal wind speed-power data points under the wind speed segmentation algorithm.

[0088] It should be noted that the wind speed segmentation algorithm divides the wind speed-power data points into two parts by the rated wind speed, and then sets the minimum and maximum power upper limits of the two parts based on the principle that when the wind speed does not reach the rated wind speed, the power scatter points are often more dispersed, and after the wind speed reaches the rated wind speed, the wind turbine enters the full power stage, and the power should be more concentratedly distributed near the rated power. Then, the wind speed-power data points of each part are judged according to the minimum and maximum power upper limits of each part, and the wind speed-power data points that are not within the minimum and maximum power upper limits are determined as abnormal wind speed-power data points under the wind speed segmentation algorithm.

[0089] In one embodiment, step S220 specifically includes: S221, determining the wind speed-power data points with wind speed less than the preset rated wind speed as the first wind speed-power data point group, and determining the wind speed-power data points with wind speed greater than or equal to the preset rated wind speed as the second wind speed-power data point group; Regarding the rated wind speed, the embodiment of the present invention does not make any specific limitation thereto, and it can be reasonably set according to actual application requirements.

[0090] For ease of understanding, the following example is given: Assume that there are 10 wind speed-power data points: [4.5m / s, 90Kw], [4.7m / s, 95Kw], [4.9m / s, 100Kw], [5.1m / s, 120Kw], [5.3m / s, 130Kw], [5.5m, 140Kw], [5.7m / s, 150Kw], [5.9m / s, 160Kw], [6.0m / s, 155Kw], [5.8m / s, 155Kw] and [5.5m / s, 140Kw], and the rated wind speed is 5.5m / s.

[0091] Wind speed-power data points less than 5.5 m / s include: [4.5 m / s, 90 Kw], [4.7 m / s, 95 Kw], [4.9 m / s, 100 Kw], [5.1 m / s, 120 Kw], and [5.3 m / s, 130 Kw], a total of 5, which are the first wind speed-power data point group.

[0092] The wind speed-power data points greater than or equal to 5.5 m / s include: [5.5 m, 140 Kw], [5.7 m / s, 150 Kw], [5.9 m / s, 160 Kw], [6.0 m / s, 155 Kw], [5.8 m / s, 155 Kw] and [5.5 m / s, 140 Kw], totaling 6, which are the second wind speed-power data point group.

[0093] S222, determining a power lower limit value and a power upper limit value of a first wind speed-power data point group and a second wind speed-power data point group; Specifically, step S222 includes: The preset third power threshold is less than the preset first power threshold and less than the preset second power threshold.

[0094] S2221, calculating the power mean of the first wind speed-power data point group and the second wind speed-power data point; Specifically, the mean power of the first wind speed-power data point group is: (90+95+100+120+130)÷5=107Kw.

[0095] Specifically, the average power of the second wind speed-power data point is: (140+150+160+155+155+140)÷6=150Kw.

[0096] S2222, calculating the difference between the power mean of the first wind speed-power data point group and the preset first power threshold, and obtaining the power lower limit value of the first wind speed-power data point group; Specifically, the power lower limit value of the first wind speed-power data point group=the power average value of the first wind speed-power data point group-the preset first power threshold value=107-10=97w.

[0097] S2223, calculating the sum of the power mean value of the first wind speed-power data point group and the preset second power threshold value to obtain the power upper limit value of the first wind speed-power data point group; Specifically, the power upper limit value of the first wind speed-power data point group=the power average value of the first wind speed-power data point group+the preset second power threshold value=107+15=122Kw.

[0098] S2224, respectively calculating the difference and sum of the power mean value of the second wind speed-power data point group and the preset third power threshold value, to obtain a power lower limit value and a power upper limit value of the second wind speed-power data point group; Specifically, the power lower limit value of the second wind speed-power data point group=the power average value of the second wind speed-power data point group-the preset third power threshold value=150-5=145Kw.

[0099] Specifically, the power upper limit value of the second wind speed-power data point group=the power average value of the second wind speed-power data point group+the preset third power threshold value=150+5=155Kw.

[0100] In this embodiment, the setting of "preset third power threshold < preset first power threshold < preset second power threshold" is taken into consideration to take into account the power variation law and abnormality definition requirements in low and high wind speed intervals. The power in the low wind speed interval is stable and low, with narrow fluctuations. The lower limit is set by a moderate first power threshold, and the upper limit is set by a slightly larger second power threshold to check for abnormalities. The power in the high wind speed interval is high and complex, and the lower limit is set by a smaller third power threshold and the corresponding upper limit is defined to check for abnormalities. In addition, this setting can also clarify the difference in abnormal boundaries according to the different situations of the two intervals, thereby avoiding misjudgment, and then accurately screening abnormal wind speed-power data points.

[0101] S223, determining a first wind speed-power data point in the first wind speed-power data point group whose power is less than a lower power limit value of the first wind speed-power data point group or whose power is greater than an upper power limit value of the first wind speed-power data point group as a first abnormal wind speed-power data point; S224, determining a second wind speed-power data point in the second wind speed-power data point group whose power is less than a lower power limit value of the second wind speed-power data point group or whose power is greater than an upper power limit value of the second wind speed-power data point group as a second abnormal wind speed-power data point; S225, taking the first abnormal wind speed-power data point and the second abnormal wind speed-power data point as abnormal wind speed-power data points under the wind speed segmentation algorithm.

[0102] For ease of understanding, steps S230 to S250 are described below with examples based on actual data: The first wind speed-power data point group: [4.5m / s, 90Kw], [4.7m / s, 95Kw], [4.9m / s, 100Kw], [5.1m / s, 120Kw], [5.3m / s, 130Kw]. The lower power limit of the first wind speed-power data point group is 97Kw, and the upper power limit of the first wind speed-power data point group is 122Kw. By comparison, it can be known that the first abnormal wind speed-power data points include: [4.5m / s, 90Kw], [4.7m / s, 95Kw] and [5.3m / s, 130Kw].

[0103] The second wind speed-power data point group: [5.5m, 140Kw], [5.7m / s, 150Kw], [5.9m / s, 160Kw], [6.0m / s, 155Kw], [5.8m / s, 155Kw] and [5.5m / s, 140Kw]. The power lower limit value of the second wind speed-power data point group is 145Kw, and the power upper limit value of the second wind speed-power data point group is 155Kw. By comparison, it can be known that the second abnormal wind speed-power data points include: [5.5m, 140Kw], [5.9m / s, 160Kw] and [5.5m / s, 140Kw].

[0104] Therefore, the wind speed-power abnormal data points under the wind speed segmentation algorithm are [4.5m / s, 90Kw], [4.7m / s, 95Kw], [5.3m / s, 130Kw], [5.5m, 140Kw], [5.9m / s, 160Kw] and [5.5m / s, 140Kw].

[0105] In one embodiment, in order to make the relationship between the data clearer, Table 2 is given below as an example.

[0106] Table 2 Wind speed-power data point parameter table

[0107] S300: Determine the same wind speed-power abnormal data point under different curve abnormal value identification methods as a wind speed-power target abnormal data point.

[0108] For ease of understanding, the following is an example with actual data: Assume that there are 10 wind speed-power data points: [4.5m / s, 90Kw], [4.7m / s, 95Kw], [4.9m / s, 100Kw], [5.1m / s, 120Kw], [5.3m / s, 130Kw], [5.5m, 140Kw], [5.7m / s, 150Kw], [5.9m / s, 160Kw], [6.0m / s, 155Kw], [5.8m / s, 155Kw] and [5.5m / s, 140Kw], and the rated wind speed is 5.5m / s.

[0109] The wind speed-power anomaly data points under the density clustering algorithm are: [4.5m / s, 90Kw], [4.7m / s, 95Kw], [4.9m / s, 100Kw], [6.0m / s, 155Kw], [5.8m / s, 155Kw] and [5.5m / s, 140Kw].

[0110] The wind speed-power abnormal data points under the wind speed segmentation algorithm are [4.5m / s, 90Kw], [4.7m / s, 95Kw], [5.3m / s, 130Kw], [5.5m, 140Kw], [5.9m / s, 160Kw] and [5.5m / s, 140Kw].

[0111] By comparison, we can know that the intersection of the two is: [4.5m / s, 90Kw], [4.7m / s, 95Kw] and [5.5m / s, 140Kw], that is, the wind speed-power target abnormal data points are [4.5m / s, 90Kw], [4.7m / s, 95Kw] and [5.5m / s, 140Kw].

[0112] In one or more of the above embodiments, different curve outlier identification methods also include: 3-Sigma identification method.

[0113] The 3-Sigma identification method is used to identify multiple wind speed-power data points, and the abnormal wind speed-power data points under the Sigma identification method are obtained.

[0114] The 3-Sigma identification method (3 times the standard deviation identification method) is based on the normal distribution, an important probability distribution. The normal distribution is also called the Gaussian distribution, and the image corresponding to its probability density function presents a bell-shaped curve shape that is high in the middle, gradually decreases on both sides, and is symmetrical on both sides. In real life, the data distribution of a large number of natural and social phenomena is approximately in line with the normal distribution.

[0115] Specifically, the specific application method of 3-Sigma identification method is given below: (1) Calculate the wind speed mean, wind speed standard deviation, power mean and power standard deviation of multiple wind speed-power data points; For example: The wind speed-power data points are [(4,100),(5,120),(6,150),(7,180),(8,220),(9,250), (10,300),(11,350),(12,400),(13,450)].

[0116] Average wind speed: (4+5+6+7+8+9+10+11+12+13) / 10=8.5m / s Wind speed standard deviation: 3.35m / s Average power: (100+120+150+180+220+250+300+350+400+450) / 10=232kW Power standard deviation: 43.3kW (2) Determine the normal range of wind speed data The normal range of wind speed data: [232-3*43.3, 232+3*43.3].

[0117] The normal range of power data: [8.5-3.35*3, 8.5+3.35*3].

[0118] (3) A wind speed-power data point that is not within the normal range of wind speed data or the normal range of power data is determined as an abnormal wind speed-power data point.

[0119] According to the above, the above uses the common wind speed-power abnormal data point under the density clustering algorithm and the wind speed-power abnormal data point under the wind speed segmentation algorithm as the target wind speed-power abnormal data point.

[0120] At present, three types of wind speed-power abnormal data points are known, namely, wind speed-power abnormal data points under the 3-Sigma identification method, wind speed-power abnormal data points under the density clustering algorithm, and wind speed-power abnormal data points under the wind speed segmentation algorithm. In this embodiment, the common wind speed-power abnormal data of any two wind speed-power abnormal data points can be used as the target wind speed-power abnormal data, and the common wind speed-power abnormal data point of the three wind speed-power abnormal data points can also be used as the target wind speed-power abnormal data point.

[0121] In this embodiment, by adding the 3-Sigma identification method, it is possible to further identify abnormal points in each wind speed segment of the power curve, that is, to further improve the accuracy of identifying abnormal values ​​of the power curve.

[0122] Beneficial effects of the present invention: (1) Improving the accuracy and robustness of outlier identification: This method combines the density clustering algorithm and the wind speed segmentation algorithm. It can not only identify the outliers in each wind speed segment of the power curve, but also reduce the false alarms in the climbing segment and full-power segment of the power curve. It can also eliminate false identification through cross-validation, thereby achieving accurate and efficient identification of outliers in the power curve.

[0123] (2) Timely discovery and diagnosis of wind turbine operation problems: Abnormal point identification technology can monitor the operating status of wind turbines in real time, provide early warnings, and assist in fault location and diagnosis, effectively avoiding major losses and ensuring safe and stable operation of the units.

[0124] (3) Improve the operating efficiency of wind turbines and reduce maintenance costs: By analyzing abnormal points, optimizing operating strategies, adjusting operating parameters, improving power generation efficiency, and implementing preventive maintenance, the service life of the units can be extended and maintenance costs can be reduced.

[0125] Embodiment 2 Based on the same inventive concept, Figure 3 As shown, the embodiment of the present invention further provides a wind speed-power curve abnormal point identification device 200, comprising: The data acquisition module 210 is used to acquire a plurality of wind speed-power data points of the wind turbine generator set within a preset time period; The abnormality identification module 220 is used to identify multiple wind speed-power data points using different curve abnormality identification methods to obtain wind speed-power abnormality data points under different curve abnormality identification methods; The cross-validation module 230 is used to determine that the same wind speed-power abnormal data point under different curve abnormal value identification methods is a wind speed-power target abnormal data point.

[0126] Optionally, the anomaly identification module 220 includes: a clustering anomaly identification unit 221 and a segmentation anomaly identification unit 222; The clustering anomaly identification unit 221 is used to identify multiple wind speed-power data points using a density clustering algorithm to obtain wind speed-power anomaly data points under the density clustering algorithm; The segmented abnormality identification unit 222 is used to identify multiple wind speed-power data points using a wind speed segmentation algorithm to obtain wind speed-power abnormality data points under the wind speed segmentation algorithm.

[0127] Optionally, the clustering anomaly identification unit 221 is specifically used for: Determine the cluster radius and minimum number of cluster points for the density clustering algorithm; Counting the number of wind speed-power data points other than the wind speed-power data point within the clustering radius of each wind speed-power data point to determine the density of each wind speed-power data point; Determine the wind speed-power data points with a density greater than the minimum number of cluster points as wind speed-power target data points, and determine the wind speed-power data points other than the wind speed-power target data points as wind speed-power abnormal data points; The abnormal wind speed-power data points that are not within the clustering radius of any wind speed-power target data point are determined as abnormal wind speed-power data points under the density clustering algorithm.

[0128] It should be understood that the device corresponds to the above-mentioned embodiment of the method for identifying abnormal points of wind speed-power curve, and can execute each step involved in the above-mentioned method embodiment. The specific functions of the device can be found in the description above. To avoid repetition, the detailed description is appropriately omitted here. The device includes at least one software function module that can be stored in a memory in the form of software or firmware or solidified in the operating system (OS) of the device.

[0129] Embodiment 3 Based on the same inventive concept, an embodiment of the present invention also provides an electronic device, including: a processor and a memory, the memory storing machine-readable instructions executable by the processor, and the machine-readable instructions executing the above-mentioned wind speed-power curve abnormal point identification method when executed by the processor.

[0130] In a typical configuration, an electronic device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0131] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0132] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0133] Embodiment 4 Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium storing computer instructions. When the computer instructions are executed on a computer, the computer executes the above-mentioned method for identifying abnormal points of a wind speed-power curve.

[0134] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0135] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0136] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0137] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0138] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe various possible combinations.

[0139] In addition, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part.

[0140] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0141] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for identifying abnormal points of a wind speed-power curve, characterized in that: include: Obtain multiple wind speed-power data points of the wind turbine generator set within a preset time period; Using different curve outlier recognition methods, multiple wind speed-power data points are recognized to obtain wind speed-power abnormal data points under different curve outlier recognition methods; The same wind speed-power abnormal data points under different curve outlier recognition methods are determined as wind speed-power target abnormal data points.

2. The method for identifying abnormal points of a wind speed-power curve according to claim 1, characterized in that: Using different curve outlier recognition methods, multiple wind speed-power data points are identified to obtain wind speed-power abnormal data points under different curve outlier recognition methods, including: Using the density clustering algorithm, multiple wind speed-power data points are identified to obtain abnormal wind speed-power data points under the density clustering algorithm; The wind speed segmentation algorithm is used to identify multiple wind speed-power data points, and the abnormal wind speed-power data points under the wind speed segmentation algorithm are obtained.

3. The method for identifying abnormal points of a wind speed-power curve according to claim 2, characterized in that: Using the density clustering algorithm, multiple wind speed-power data points are identified to obtain abnormal wind speed-power data points under the density clustering algorithm, including: Determine the cluster radius and minimum number of cluster points for the density clustering algorithm; Counting the number of wind speed-power data points other than the wind speed-power data point within the clustering radius of each wind speed-power data point to determine the density of each wind speed-power data point; Determine the wind speed-power data points with a density greater than the minimum number of cluster points as wind speed-power target data points, and determine the wind speed-power data points other than the wind speed-power target data points as wind speed-power abnormal data points; The abnormal wind speed-power data points that are not within the clustering radius of any wind speed-power target data point are determined as abnormal wind speed-power data points under the density clustering algorithm.

4. The method for identifying abnormal points of a wind speed-power curve according to claim 3, characterized in that: The cluster radius of the density clustering algorithm is determined as follows: Combining multiple wind speed-power data points in pairs to obtain multiple wind speed-power data point groups; Calculate the distance between two wind speed-power data points in each wind speed-power data point group to obtain the Euclidean distance of each wind speed-power data point group; Combining the Euclidean distances of multiple wind speed-power data point groups in pairs to obtain multiple Euclidean distance groups; Calculate the difference of the Euclidean distances between two wind speed-power data point groups in each Euclidean distance group, obtain the difference of each Euclidean distance group, and take the Euclidean distance group with the largest difference as the target Euclidean distance group; The average of the Euclidean distances of the two wind speed-power data point groups in the target Euclidean distance group is calculated, and the average is used as the cluster radius of the density clustering algorithm.

5. The method for identifying abnormal points of a wind speed-power curve according to claim 3, characterized in that: The number of wind speed-power data points other than the wind speed-power data point within the clustering radius of each wind speed-power data point is counted to determine the density of each wind speed-power data point, including: For each wind speed-power data point: Determine wind speed-power data points other than the current wind speed-power data point as wind speed-power data points to be identified; Calculate the distance between the current wind speed-power data point and each wind speed-power data point to be identified, and obtain multiple Euclidean distances of the wind speed-power data points; The number of wind speed-power data points to be identified corresponding to the Euclidean distance less than the clustering radius is counted to determine the density of the wind speed-power data points.

6. The method for identifying abnormal points of a wind speed-power curve according to claim 2, characterized in that: Using the wind speed segmentation algorithm, multiple wind speed-power data points are identified to obtain abnormal wind speed-power data points under the wind speed segmentation algorithm, including: Determine the wind speed-power data points with a wind speed less than a preset rated wind speed as a first wind speed-power data point group, and determine the wind speed-power data points with a wind speed greater than or equal to the preset rated wind speed as a second wind speed-power data point group; Determine a lower power limit value and an upper power limit value of the first wind speed-power data point group and the second wind speed-power data point group; Determine the first wind speed-power data point in the first wind speed-power data point group whose power is less than the power lower limit value of the first wind speed-power data point group or whose power is greater than the power upper limit value of the first wind speed-power data point group as the first abnormal wind speed-power data point; Determine a second wind speed-power data point in the second wind speed-power data point group whose power is less than a lower power limit value of the second wind speed-power data point group or whose power is greater than an upper power limit value of the second wind speed-power data point group as a second abnormal wind speed-power data point; The first abnormal data point of wind speed-power and the second abnormal data point of wind speed-power are taken as the abnormal data points of wind speed-power under the wind speed segmentation algorithm.

7. The method for identifying abnormal points of a wind speed-power curve according to claim 6, characterized in that: Determining a power lower limit value and a power upper limit value of a first wind speed-power data point group and a second wind speed-power data point group includes: Calculate the power mean of the first wind speed-power data point group and the second wind speed-power data point; Calculate the difference between the power mean of the first wind speed-power data point group and the preset first power threshold value to obtain the power lower limit value of the first wind speed-power data point group; Calculate the sum of the power mean of the first wind speed-power data point group and the preset second power threshold value to obtain the power upper limit value of the first wind speed-power data point group; Calculate the difference and sum of the power mean of the second wind speed-power data point group and the preset third power threshold value respectively, and obtain the power lower limit value and power upper limit value of the second wind speed-power data point group; The preset third power threshold is less than the preset first power threshold and less than the preset second power threshold.

8. A device for identifying abnormal points of a wind speed-power curve, characterized in that: include: A data acquisition module, used to acquire multiple wind speed-power data points of a wind turbine generator set within a preset time period; An abnormality identification module is used to identify multiple wind speed-power data points using different curve abnormality identification methods to obtain wind speed-power abnormality data points under different curve abnormality identification methods; The cross-validation module is used to determine that the same wind speed-power abnormal data point under different curve abnormal value identification methods is a wind speed-power target abnormal data point.

9. The device for identifying abnormal points of wind speed-power curve according to claim 8, characterized in that: The anomaly recognition module includes: a clustering anomaly recognition unit and a segmented anomaly recognition unit; A clustering anomaly identification unit is used to identify multiple wind speed-power data points using a density clustering algorithm to obtain wind speed-power anomaly data points under the density clustering algorithm; The segmented anomaly recognition unit is used to recognize multiple wind speed-power data points using a wind speed segmentation algorithm to obtain wind speed-power anomaly data points under the wind speed segmentation algorithm.

10. The device for identifying abnormal points of a wind speed-power curve according to claim 9, characterized in that: The clustering anomaly identification unit is specifically used for: Determine the cluster radius and minimum number of cluster points for the density clustering algorithm; Counting the number of wind speed-power data points other than the wind speed-power data point within the clustering radius of each wind speed-power data point to determine the density of each wind speed-power data point; Determine the wind speed-power data points with a density greater than the minimum number of cluster points as wind speed-power target data points, and determine the wind speed-power data points other than the wind speed-power target data points as wind speed-power abnormal data points; The abnormal wind speed-power data points that are not within the clustering radius of any wind speed-power target data point are determined as abnormal wind speed-power data points under the density clustering algorithm.

Images