WO2023063888A2

WO2023063888A2 - Method and apparatus for predicting wind power, and device and storage medium thereof

Info

Publication number: WO2023063888A2
Application number: PCT/SG2022/050726
Authority: WO
Inventors: Liangguo MENG; Zibo DONG; Hui Yang; Qingsheng ZHAO
Original assignee: Envision Digital International Pte. Ltd.; Shanghai Envision Digital Co., Ltd.
Priority date: 2021-10-14
Filing date: 2022-10-11
Publication date: 2023-04-20
Also published as: CN113937763B; CN113937763A; WO2023063888A3

Abstract

Disclosed are a method and apparatus for predicting wind power, and a device and a storage medium thereof. The method includes: acquiring a plurality of groups of meteorological prediction data corresponding to a target wind power plant based on history meteorological data by different numerical weather prediction systems; acquiring intermediate wind power prediction data corresponding to the target meteorological prediction data based on target meteorological prediction data in the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data; and acquiring wind power prediction data corresponding to the target wind power plant through calculation based on intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data being indicative of a predicted wind power of the target wind power plant in the target prediction time period. According to the present disclosure, the prediction of the wind power of the target wind power plant is achieved by synthesizing the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, which doesn't just depend on meteorological prediction data in a single numerical weather prediction system, thereby improving the prediction accuracy of the wind power.

Description

METHOD AND APPARATUS FOR PREDICTING WIND POWER, AND DEVICE AND STORAGE MEDIUM THEREOF

TECHNICAL FIELD

[0001] Embodiments of the present disclosure relate to the technical field of wind power, and in particular, relate to a method and apparatus for predicting wind power, and a device and a storage medium thereof.

BACKGROUND

[0002] With the increasing proportion of wind power generation in power supply, the power grid has very high requirements on the prediction accuracy of wind power generation in order to ensure the stability of power dispatching.

[0003] Taking ultra-short-term wind power prediction as an example, the related technology firstly performs ultra-short-term wind speed prediction based on a numerical weather prediction system, then performs ultra-short-term wind power prediction based on the predicted wind speed, and finally performs power dispatching based on the predicted ultra-short-term wind power.

[0004] However, the wind speed predicted based on a numerical weather prediction system often has a large deviation, such that the wind power prediction is not accurate enough.

SUMMARY

[0005] Embodiments of the present disclosure provide a method and apparatus for predicting wind power, and a device and a storage medium thereof to improve the prediction accuracy of wind power.

[0006] According to one aspect of the embodiments of the present disclosure, a method for predicting wind power is provided. The method includes:

[0007] acquiring history meteorological data, wind turbine data, history wind power data, and wind condition data of a target wind power plant within a target history time period;

[0008] acquiring, based on the history meteorological data, a plurality of groups of meteorological prediction data corresponding to the target wind power plant by different numerical weather prediction systems, the meteorological prediction data being indicative of predicted meteorological data of the target wind power plant within a target prediction time period, the target prediction time period being a time period after the target history time period; [0009] acquiring, based on target meteorological prediction data in the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, intermediate wind power prediction data corresponding to the target meteorological prediction data; and

[0010] acquiring wind power prediction data corresponding to the target wind power plant by calculation based on intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data being indicative of a predicted wind power of the target wind power plant in the target prediction time period.

[0011] According to another aspect of the embodiments of the present disclosure, an apparatus for predicting wind power is provided. The apparatus includes:

[0012] a history data acquiring module, configured to acquire history meteorological data, wind turbine data, history wind power data, and wind condition data of a target wind power plant within a target history time period;

[0013] a predicted data acquiring module, configured to acquire, based on the history meteorological data a plurality of groups of meteorological prediction data corresponding to the target wind power plant by different numerical weather prediction systems, the meteorological prediction data being indicative of predicted meteorological data of the target wind power plant within a target prediction time period, the target prediction time period being a time period after the target history time period;

[0014] an intermediate power acquiring module, configured to acquire, based on target meteorological prediction data in the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, intermediate wind power prediction data corresponding to the target meteorological prediction data; and

[0015] a predicted power acquiring module, configured to acquire wind power prediction data corresponding to the target wind power plant by calculation based on intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data being indicative of a predicted wind power of the target wind power plant in the target prediction time period.

[0016] According to another aspect of the embodiments of the present disclosure, a computer device is provided. The computer device includes a processor and a memory storing one or more computer programs thereon. The one or more computer programs, when loaded and run by the processor, cause the computer device to perform the method for predicting wind power as described above.

[0017] According to another aspect of the embodiments of the present disclosure, a non-transitory computer-readable storage medium including one or more computer programs stored thereon is provided. The one or more computer programs, when loaded and run by a processor of a computer device, cause the computer device to perform the method for predicting wind power as described above.

[0018] According to another aspect of the embodiments of the present disclosure, a computer program product or a computer program is provided. The computer program product or the computer program includes one or more computer instructions stored in a computer readable storage medium. The one or more computer instructions, when loaded and executed by a processor of a computer device, cause the computer device to perform the method for predicting wind power as described above.

[0019] The technical solutions according to the embodiments of the present disclosure may achieve the following beneficial effects:

[0020] A plurality of groups of meteorological prediction data of a target wind power plant are acquired through different numerical weather prediction systems, then the wind power of the target wind power plant is predicted respectively based on different meteorological prediction data, wind turbine data, history wind power data, and wind condition data, and finally wind power prediction data corresponding to the target wind power plant are acquired by synthesizing a plurality of intermediate wind power prediction data. Therefore, the prediction of the wind power is achieved by synthesizing the plurality of groups of meteorological prediction data, the wind turbine data, the histoiy wind power data, and the wind condition data, which doesn't just depend on meteorological prediction data in a single numerical weather prediction system, thereby improving the prediction accuracy of the wind power.

BRIEF DESCRIPTION OF THE DRAWINGS

[0021] For clearer descriptions of the technical solutions according to the embodiments of the present disclosure, the drawings required to be used in the description of the embodiments are briefly introduced below. It is obvious that the drawings in the description below are only some embodiments of the present disclosure, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.

[0022] FIG. 1 is a schematic diagram of a solution implementation environment according to an embodiment of the present disclosure;

[0023] FIG. 2 is a flowchart of a method for predicting wind power according to an embodiment of the present disclosure;

[0024] FIG. 3 is a table of preprocessed history wind speed data according to an embodiment of the present disclosure;

[0025] FIG. 4 is a schematic diagram of a wind power predicting model according to an embodiment of the present disclosure;

[0026] FIG. 5 is a schematic diagram of a first wind power acquisition network according to an embodiment of the present disclosure;

[0027] FIG. 6 is a block diagram of an apparatus for predicting wind power according to an embodiment of the present disclosure;

[0028] FIG. 7 is a block diagram of an apparatus for predicting wind power according to another embodiment of the present disclosure; and

[0029] FIG. 8 is a block diagram of a computer device according to an embodiment of the present disclosure.

DETAILED DESCRIPTION

[0030] For clearer descriptions of the objects, technical solutions, and advantages of the present disclosure, the embodiments of the present disclosure are further described in detail below with reference to the accompanying drawings.

[0031] FIG. 1 is a schematic diagram of a solution implementation environment according to an embodiment of the present disclosure. The solution implementation environment may include: a wind power plant 10 and a server 20.

[0032] The wind power plant 10 refers to a facility that utilizes wind power to generate electricity, which may include a plurality of wind power generator sets. The wind power generator set refers to a wind turbine generator set, which may include a foundation, a tower, a cabin, hubs, a blade (rotor) assembly, and an anemometer. Optionally, the wind power plant 10 may acquire operation parameters of the wind power generator set through a supervisory control and data acquisition (SCADA) system, wherein the operation parameters may include a wind speed, a wind direction, a rotation speed of a rotor, a wind power, an operation state of the wind power generator set, and the like. The wind power plant 10 may also acquire meteorological data such as a wind speed, a temperature, a pressure, a density, and a humidity through sensors.

[0033] The server 20 may be a server, a server cluster composed of a plurality of servers, or a cloud computing service center. The server 20 may be configured to store data related to the wind power plant 10, such as a wind power, meteorological data, wind turbine data, and wind condition data of the wind power plant 10. Optionally, the server 20 may also predict the wind power of the wind power plant 10 by analyzing the wind power, the meteorological data, the wind turbine data, the wind condition data, and the like of the wind power plant 10 and combining with a numerical weather prediction system, so as to implement reasonable dispatching of the power provided by the wind power plant 10.

[0034] The wind power plant 10 may communicate with the server 20 over a network.

[0035] FIG. 2 illustrates a flowchart of a method for predicting wind power according to an embodiment of the present disclosure. The execution subject of the steps of the method may be the server 20 described above. The method includes the following steps (201 to 204).

[0036] In step 201, history meteorological data, wind turbine data, history wind power data, and wind condition data of a target wind power plant are acquired within a target history time period. [0037] The target wind power plant refers to a wind power plant which needs wind power prediction and may refer to any wind power plant. The target wind power plant includes a plurality of wind power generator sets, and the wind power may be an output power of the wind power generator sets and may also be a wind power of the wind power plant. Optionally, the wind power of the target wind power plant may be calculated based on the wind powers respectively corresponding to the plurality of wind power generator sets in the target wind power plant.

[0038] The meteorological data refer to data describing meteorological parameters such as a wind speed, a pressure, a density, a humidity, and a temperature. The wind turbine data refer to data describing wind power generator sets, and may include data such as a serial number of wind power generator sets and an output power feature (i.e., a wind power). Optionally, the wind power of the wind power generator set is approximately equal to the product of the wind speed multiplied by the air density to the third power. The wind condition data refer to data describing seasonal variation differences, daily variation differences, and the like of wind speeds. For example, daily variation difference data of wind speeds may be acquired by counting the mean wind speed per day.

[0039] The target history time period may refer to any time period of time periods corresponding to the wind power plant having data records. For example, the wind power prediction task is assumed as follows: according to the data of the target wind power plant in the previous 3 hours, the wind power in the future 4 hours is predicted, the previous 3 hours being the target history time period, and the future 4 hours being a target prediction time period in the following text. For example, the current time moment is assumed as 9 o' clock, then the target history time period may be set to 6 o' clock to 9 o' clock, and the target prediction time period may be set to 9 o' clock to 13 o' clock.

[0040] In the embodiments of the present disclosure, the history meteorological data described above are meteorological data (such as history wind speed data and history pressure data) of a target wind power plant within a target history time period, the history wind power data described above are wind power data of the target wind power plant in the target history time period, the wind turbine data described above are wind turbine data of the target wind power plant in the target history time period, and the wind condition data described above are wind condition data of the target wind power plant in the target history time period.

[0041] In some embodiments, the process of acquiring the history meteorological data, the wind turbine data, the history wind power data, and the wind condition data may be as follows:

[0042] 1. A meteorological timing data set, a wind turbine timing data set, a power timing data set, and a wind condition timing data set corresponding to the target wind power plant are acquired.

[0043] The meteorological timing data set includes meteorological data of the target wind power plant within a threshold time period. The wind turbine timing data set includes wind turbine data of the target wind power plant within a threshold time period. The power timing data set includes wind power data of the target wind power plant within a threshold time period. The wind condition timing data set includes wind condition data of the target wind power plant within a threshold time period. Optionally, the threshold time period may be greater than or equal to one month (e.g., 2021.7.1 to 2021.9.1), and the data time interval corresponding to the power timing data set may be 15 minutes. The meteorological data, the wind turbine data, and the wind condition data may be acquired in real time every day, with a data time interval not limited, and the data time interval respectively corresponding to the meteorological data, the wind turbine data, and the wind condition data may be adjusted to 15 minutes as needed.

[0044] 2. The meteorological timing data set, the wind turbine timing data set, the power timing data set, and the wind condition timing data set are preprocessed respectively.

[0045] The preprocessing includes at least one of abnormal data processing, default data interpolation, and data normalization.

[0046] The abnormal data processing is designed to clear abnormal data in the above four data sets. The default data interpolation is designed to interpolate to complement the default data in the above four data sets. The data normalization is designed to adjust the data in the above four data sets to a range of [0-1] to facilitate data comparison and analysis.

[0047] In some embodiments, taking a wind speed timing data set in a meteorological timing data set as an example, referring to FIG. 3, a wind speed timing data set 301 is a wind speed timing data set with a data time interval of 15 minutes within 3 hours, and the wind speed timing data set 301 is acquired through abnormal data processing and default data interpolation. Optionally, the wind speed timing data set 301 may be further subjected to data normalization.

[0048] 3. The history meteorological data, the wind turbine data, the history wind power data, and the wind condition data of the target wind power plant are acquired in the target history time period by respectively performing data extraction on the preprocessed meteorological timing data set, the preprocessed wind turbine timing data set, the preprocessed power timing data set, and the preprocessed wind condition timing data set by adopting a sliding window method.

[0049] In some embodiments, the length of the sliding window is assumed as 3 hours and the step size is assumed as 1 hour, then history meteorological data, wind turbine data, history wind power data, and wind condition data in time periods of 0-3 hours, 1-4 hours, 2-5 hours, and the like may be acquired through the sliding window method, respectively. Optionally, in the case of training the model in the following text, a plurality of groups of sample data may be acquired by adopting a sliding window method. In the case of wind power prediction of a target wind power plant, the data in the target history time period may be acquired by adopting a sliding window method.

[0050] In step 202, a plurality of groups of meteorological prediction data corresponding to the target wind power plant are acquired based on the history meteorological data through different numerical weather prediction systems, the meteorological prediction data being indicative of predicted meteorological data of the target wind power plant within a target prediction time period, and the target prediction time period referring to a time period after the target history time period. [0051] The numerical weather prediction system refers to a system which performs numerical calculation through a large computer under certain initial value and boundary value conditions according to the actual atmospheric conditions, solves the fluid mechanics and thermodynamics equation set describing the weather evolution process, and predicts the atmospheric motion state and weather phenomena in a certain time period. Optionally, the source of the numerical weather prediction system is not limited in the embodiments of the present disclosure, and may be from different organizations, different countries, and the like. For example, the source of the numerical weather prediction system may be China, the Netherlands, the United States, the United Kingdom, etc. Due to diversity and difference existing among different numerical weather prediction systems, the diversity and difference of meteorological prediction data may be guaranteed, and therefore the applicability of wind power prediction is improved.

[0052] In some embodiments, the meteorological prediction data may include wind speed prediction data, pressure prediction data, density prediction data, humidity prediction data, temperature prediction data, and the like. Meanwhile, due to the diversity and the difference existing among different numerical weather prediction systems, the diversity and the difference also exist among a plurality of groups of meteorological prediction data.

[0053] In some embodiments, in case of ultra-short-term wind power prediction, the target prediction time period may refer to a time period with a data time interval of 15 minutes in the future 4 hours, i.e., 16 predicted data points may be present. In the case of short-term wind power prediction, the target prediction time period may be in units of days. In the case of mid-term wind power prediction, the target prediction time period may be in units of weeks. In the case of longterm wind power prediction, the target prediction time period may be in units of months.

[0054] In step 203, intermediate wind power prediction data corresponding to the target meteorological prediction data are acquired based on target meteorological prediction data in the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data.

[0055] According to the embodiments of the present disclosure, different meteorological prediction data are respectively combined with wind turbine data, history wind power data, and wind condition data to form different input data, and then a plurality of intermediate wind power prediction data are acquired through the different input data. The intermediate wind power prediction data refer to wind power prediction data of the target wind power plant predicted based on single input data.

[0056] The target meteorological prediction data may refer to any meteorological prediction data in the plurality of groups of meteorological prediction data. In one example, the process of acquiring the intermediate wind power prediction data may be as follows:

[0057] 1. A corrected wind speed corresponding to the target meteorological prediction data is acquired based on the target meteorological prediction data, the wind turbine data, and the wind condition data.

[0058] The corrected wind speed is a wind speed acquired by correcting a deviation of the predicted wind speed in the meteorological prediction data. Optionally, the corrected wind speed may be represented by the deviation-corrected wind speed timing, or may be represented by an offset timing between the corrected wind speed and the predicted wind speed, which is not limited herein. For example, in the case of ultra-short-term wind power prediction, the corrected wind speed may include 16 corrected wind speed values in the future 4 hours, i.e., one corrected wind speed value every 15 minutes.

[0059] In some embodiments, the acquisition process of the corrected wind speed may be as follows: acquiring a first feature set corresponding to the meteorological prediction data by performing extraction of timing statistical features on the meteorological prediction data, the timing statistical features including at least one of a variance, a multi-order difference, a mean value, and a quantile value; constructing a second feature set based on the meteorological prediction data and the history meteorological data, the second feature set being indicative of the relationship between the meteorological prediction data and the history meteorological data; constructing an artificial high-order feature based on the wind condition data, the history meteorological data, and the wind turbine data, the artificial high-order feature including at least one of a date, a wind frequency, a wind energy, and turbulence; and acquiring the corrected wind speed corresponding to the target meteorological prediction data based on the first feature set, the second feature set, and the artificial high-order feature.

[0060] The first feature set may include a variance, a multi-order difference, a mean value and a quantile values (or one or more of a variance, a multi-order difference, a mean value and a quantile value) respectively corresponding to the wind speed prediction data, the pressure prediction data, the density prediction data, the humidity prediction data, the temperature prediction data, and the like.

[0061] In some embodiments, the second feature set may include cross timing statistical features respectively corresponding to a plurality of meteorological parameters, the cross timing statistical features being indicative of a relationship between the predicted data and the history data corresponding to the meteorological parameters. In one example, the specific acquisition process of the second feature set may be as follows: acquiring a deviation value set between data of a target meteorological parameter in the meteorological prediction data and data of the target meteorological parameter in the history meteorological data; determining a mean value corresponding to the deviation value set as a cross timing statistical feature corresponding to the target meteorological parameter; and constructing the second feature set based on cross timing statistical features corresponding to all meteorological parameters in the meteorological prediction data. The target meteorological parameter may refer to any meteorological parameter in meteorological prediction data or history meteorological data, such as a wind speed, a pressure, a density, a humidity, and a temperature. According to the embodiments of the present disclosure, the corrected wind speed is acquired based on the cross timing statistical features, and the relationship between the meteorological prediction data and the history meteorological data is comprehensively considered, thereby improving the accuracy of acquiring the corrected wind speed.

[0062] For example, the wind speed of the target meteorological parameter is taken as an example. Assuming that the history wind speed data are 1, 2, and 3 and the wind speed prediction data are 2, 3, and 4, then the mean deviation between them is ((2-1) + (3-2) + (4-3))/3 = 1, and thus 1 is determined as the cross timing statistical feature.

[0063] The artificial high-order feature refers to a feature constructed by using relevant knowledge (e.g., artificial experience) in the field of data, which may improve the utilization effect of the data. Optionally, the wind frequency is a percentage of the number of times of a certain wind direction to the total number of observed statistics. The wind energy acquisition formula may be as follows: W = 0.5 x p x v³ x s, where p is the air density, v is the wind speed, and S is the swept area. The turbulence acquisition formula may be as follows: I = o/v, where o is the standard deviation of wind speed over 10 minutes, and v is the mean value of wind speed over 10 minutes. According to the embodiments of the present disclosure, the corrected wind speed is acquired based on the artificial high-order feature, and the artificial experience is comprehensively considered, so that the accuracy of acquiring the corrected wind speed is further improved.

[0064] 2. The intermediate wind power prediction data are acquired based on the corrected wind speed and the history wind power data.

[0065] In some embodiments, a third feature set corresponding to the history wind power data is acquired by performing extraction of timing statistical features on the history wind power data; and the intermediate wind power prediction data are acquired based on the corrected wind speed and the third feature set.

[0066] The third feature set may include a variance, a multi-order difference, a mean value, and a quantile value (or one or more of a variance, a multi-order difference, a mean value, and a quantile value) respectively corresponding to the history wind power data.

[0067] In step 204, wind power prediction data corresponding to the target wind power plant are acquired through calculation based on intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data being indicative of a predicted wind power of the target wind power plant in the target prediction time period.

[0068] In some embodiments, fused wind power prediction data are acquired by fusing the intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data; and wind power prediction data corresponding to the target wind power plant are acquired based on the fused wind power prediction data.

[0069] The plurality of intermediate wind power prediction powers may be fused by methods such as a linear regression model, a support vector machine (SVM), and segmented weighting, so as to acquire the fused wind power prediction data.

[0070] In some embodiments, in case of ultra-short-term wind power prediction, the wind power prediction data include 16 wind power prediction values within the future 4 hours, with a time interval of 15 minutes.

[0071] In some embodiments, the corrected wind speed, the intermediate wind power data, and the wind power data described above may be acquired through a model, and the detailed acquisition process and a network structure of the model are described below, and are not repeated herein. [0072] In summary, according to the technical solutions according to the embodiments of the present disclosure, a plurality of groups of meteorological prediction data of a target wind power plant are acquired through different numerical weather prediction systems, then the wind power of the target wind power plant is predicted respectively based on different meteorological prediction data, wind turbine data, history wind power data, and wind condition data, and finally wind power prediction data corresponding to the target wind power plant are acquired by synthesizing a plurality of intermediate wind power prediction data. Therefore, the prediction of the wind power is achieved by synthesizing the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, which doesn't just depend on meteorological prediction data in a single numerical weather prediction system, thereby improving the prediction accuracy of the wind power.

[0073] In addition, the corrected wind speed is acquired based on the predicted wind speed data, and the intermediate wind power prediction data are acquired based on the corrected wind speed and the history wind power data, so that the combination of direct prediction (prediction is directly performed based on the history wind power data) and rolling prediction (prediction is performed based on the predicted wind speed data) is achieved, and the problems of lack of timing information in prediction stage under single direct prediction and error accumulation under single rolling prediction are solved, thereby improving the prediction accuracy of the wind power. Meanwhile, the predicted wind speed data are corrected, so that the prediction accuracy of the wind speed can be improved, thereby further improving the prediction accuracy of the wind power. [0074] In addition, the corrected wind speed is acquired based on the cross timing statistical features, and the relationship between the meteorological prediction data and the history meteorological data is comprehensively considered, thereby improving the accuracy of acquiring the corrected wind speed. Meanwhile, the corrected wind speed is acquired based on the artificial high-order feature, and the artificial experience is comprehensively considered, so that the accuracy of acquiring the corrected wind speed is further improved.

[0075] In some embodiments, the wind power prediction data are acquired by a wind power predicting model. FIG. 4 illustrates a schematic diagram of a wind power predicting model according to an embodiment of the present disclosure. The following description will take the process of acquiring the wind power prediction data through the wind power predicting model 400 as an example, and the specific content may be as follows:

[0076] The wind power predicting model 400 includes a plurality of first wind power acquisition networks 401 and a second wind power acquisition network 402, each of the first wind power acquisition networks 401 is configured to acquire intermediate wind power prediction data corresponding to different meteorological prediction data.

[0077] In some embodiments, referring to FIG. 4, different meteorological prediction data correspond to different first wind power acquisition networks, that is, different first wind power acquisition networks are trained for different numerical weather prediction systems, and the different first wind power acquisition networks have the same network structure, but different network parameters. The intermediate wind power prediction data 1 are acquired based on the meteorological prediction data 1, the wind condition data, the wind turbine data, and the history wind power data through the first wind power acquisition network corresponding to the meteorological prediction data 1. The intermediate wind power prediction data 2 are acquired based on the meteorological prediction data 2, the wind condition data, the wind turbine data, and the history wind power data through the first wind power acquisition network corresponding to the meteorological prediction data 2. The intermediate wind power prediction data corresponding to n (n is an integer greater than 1) groups of meteorological prediction data are acquired by adopting the same method. Fused wind power prediction data are acquired by fusing the intermediate wind power prediction data respectively corresponding to the n groups of meteorological prediction data, and finally, the fused wind power prediction data are input into the second wind power acquisition network 402, such that wind power prediction data corresponding to the target wind power plant can be acquired.

[0078] In some embodiments, the first wind power acquisition network 401 may be a network structure of two-layer Stacking (a hierarchical model integration framework). For example, referring to FIG. 5, the first wind power acquisition network 401 includes a corrected wind speed acquisition sub-network 401a and an intermediate power acquisition sub-network 402b. The corrected wind speed acquisition sub-network 401a is a first layer network, and the network structure thereof may be a network structure of a radial basis function-support vector machine (RBF-SVM), a random forest, an extreme gradient boosting (XGBOOST), a neural network, e.g., a gate recurrent unit (GRU), and the like, and the intermediate power acquisition sub-network 402b is a second layer network, and the network structure thereof may also be a network structure of an RBF-SVM, a random forest, an XGBOOST, a neural network, and the like, that is, the network structure of the first wind power acquisition network 401 may be a combination of the above network structures. For example, the network structure of the corrected wind speed acquisition sub-network 401a may be any one of the network structures of an RBF-SVM, a random forest, an XGBOOST, a neural network, and the like, and the network structure of the intermediate power acquisition sub-network 402b may be any one of the network structures of an RBF-SVM, a random forest, an XGBOOST, a neural network, and the like.

[0079] The second wind power acquisition network 402 may be a network constructed by method such as linear regression, SVM, and segment weighting.

[0080] The corrected wind speed acquisition sub-network 401a is configured to acquire, based on the target meteorological prediction data, the wind turbine data, and the wind condition data, a corrected wind speed corresponding to the target meteorological prediction data. The intermediate power acquisition sub-network 401b is configured to acquire, based on the corrected wind speed and the history wind power data, intermediate wind power prediction data corresponding to the target meteorological prediction data. The second wind power acquisition network 402 is configured to acquire, based on intermediate wind power prediction data respectively corresponding to a plurality of groups of meteorological prediction data, wind power prediction data corresponding to the target wind power plant.

[0081] In some embodiments, the process of acquiring the intermediate wind power prediction data corresponding to the target meteorological prediction data is taken as an example. Firstly, a first feature set corresponding to the meteorological prediction data is acquired by performing extraction of timing statistical features on the meteorological prediction data. A second feature set is constructed based on the meteorological prediction data and the history meteorological data. An artificial high-order feature is constructed based on the wind condition data, the history meteorological data, and the wind turbine data. A corrected wind speed corresponding to the target meteorological prediction data is acquired through the corrected wind speed acquisition subnetwork 401a based on the first feature set, the second feature set, and the artificial high-order feature, wherein the corrected wind speed includes m corrected wind speed values, and m is 16 in the case of ultra-short-term wind power prediction. [0082] In some embodiments, before the first feature set, the second feature set, and the artificial high-order feature are input into the corrected wind speed acquisition sub-network 401a, category features in the first feature set, the second feature set, and the artificial high-order feature may be subjected to one-hot-coding, wherein the category features may include a serial number, a date, a time, a wind frequency, and the like of the wind power generator set.

[0083] A third feature set corresponding to the history wind power data is acquired by performing extraction of timing statistical features on the history wind power data, and intermediate wind power prediction data are acquired by the intermediate power acquisition sub-network 401b based on the corrected wind speed and the third feature set. The intermediate wind power prediction data include m intermediate wind power prediction values, and m is 16 in the case of ultra-short-term wind power prediction.

[0084] Finally, fused wind power prediction data are acquired by fusing the intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, and wind power prediction data corresponding to the target wind power plant are acquired through the second wind power acquisition network 402 based on the fused wind power prediction data.

[0085] It should be noted that, the network structures of the wind power predicting model in FIGS. 4 and 5 and the above are only exemplary and explanatory, and the structure of the wind power predicting model may be adjusted according to actual situations. For example, the number of the first wind power acquisition networks is increased or decreased appropriately, and the network structure of the first wind power acquisition network or the second wind power acquisition network is adjusted adaptively. The specific network structure of the wind power predicting model is not limited in the present disclosure, and any network structure with the wind power prediction function should fall within the protection scope of the present disclosure.

[0086] In summary, according to the technical solutions according to the embodiments of the present disclosure, a plurality of groups of meteorological prediction data of a target wind power plant are acquired through different numerical weather prediction systems, then the wind power of the target wind power plant is predicted respectively based on different meteorological prediction data, wind turbine data, history wind power data, and wind condition data, and finally wind power prediction data corresponding to the target wind power plant are acquired by synthesizing a plurality of intermediate wind power prediction data. Therefore, the prediction of the wind power is achieved by synthesizing the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, which doesn't just depend on meteorological prediction data in a single numerical weather prediction system, thereby improving the prediction accuracy of the wind power.

[0087] In addition, the first wind power acquisition network is set to be a two-layer Stacking network structure, such that data leakage can be prevented, and meanwhile, the diversity of input features in the model can be ensured. In addition, the wind speed and the wind power are respectively predicted through the corrected wind speed acquisition sub-network and the intermediate power acquisition sub-network, and compared with the prediction of the wind speed and the wind power through a single network, the prediction accuracy of the two-layer Stacking network structure provided herein is higher, and therefore the prediction accuracy of the wind power is further improved.

[0088] FIG. 6 illustrates a block diagram of an apparatus for predicting wind power according to an embodiment of the present disclosure. The apparatus has the function of implementing the above method for predicting wind power, and the function may be achieved by hardware or by hardware executing corresponding software. The apparatus may be a computer device, or may be arranged in a computer device. The apparatus 600 may include: a history data acquiring module 601, a predicted data acquiring module 602, an intermediate power acquiring module 603, and a predicted power acquiring module 604.

[0089] The history data acquiring module 601 is configured to acquire history meteorological data, wind turbine data, history wind power data, and wind condition data of a target wind power plant within a target history time period.

[0090] The predicted data acquiring module 602 is configured to acquire, based on the history meteorological data, a plurality of groups of meteorological prediction data corresponding to the target wind power plant by different numerical weather prediction systems, the meteorological prediction data being indicative of predicted meteorological data of the target wind power plant within a target prediction time period, the target prediction time period being a time period after the target history time period.

[0091] The intermediate power acquiring module 603 is configured to acquire, based on target meteorological prediction data in the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, intermediate wind power prediction data corresponding to the target meteorological prediction data; and

[0092] The predicted power acquiring module 604 is configured to acquire wind power prediction data corresponding to the target wind power plant by calculation based on intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data being indicative of a predicted wind power of the target wind power plant in the target prediction time period.

[0093] In an exemplary embodiment, as shown in FIG. 7, the intermediate power acquiring module 603 includes: a corrected wind speed acquiring sub-module 603a and an intermediate power acquiring sub-module 603b.

[0094] The corrected wind speed acquiring sub-module 603a is configured to acquire, based on the target meteorological prediction data, the wind turbine data, and the wind condition data, a corrected wind speed corresponding to the target meteorological prediction data.

[0095] The intermediate power acquiring sub-module 603b is configured to acquire the intermediate wind power prediction data based on the corrected wind speed and the history wind power data.

[0096] In some embodiments, the corrected wind speed acquiring sub-module 603a is configured to:

[0097] acquire a first feature set corresponding to the meteorological prediction data by performing extraction of timing statistical features on the meteorological prediction data, the timing statistical features comprising at least one of a variance, a multi-order difference, a mean value, and a quantile value;

[0098] construct a second feature set based on the meteorological prediction data and the history meteorological data, the second feature set being indicative of a relationship between the meteorological prediction data and the history meteorological data;

[0099] construct an artificial high-order feature based on the wind condition data, the history meteorological data, and the wind turbine data, the artificial high-order feature comprising at least one of a date, a wind frequency, a wind energy, and turbulence; and

[0100] The corrected wind speed corresponding to the target meteorological prediction data is acquired based on the first feature set, the cross timing statistical feature, and the artificial high- order feature. [0101] In some embodiments, the corrected wind speed acquiring sub-module 603a is also configured to:

[0102] acquire a deviation value set between data of a target meteorological parameter in the meteorological prediction data and data of the target meteorological parameter in the history meteorological data;

[0103] determine a mean value corresponding to the deviation value set as a cross timing statistical feature corresponding to the target meteorological parameter; and

[0104] construct the second feature set based on cross timing statistical features corresponding to all meteorological parameters in the meteorological prediction data.

[0105] In some embodiments, the intermediate power acquiring sub-module 603b is configured to:

[0106] acquire a third feature set corresponding to the history wind power data by performing extraction of timing statistical features on the history wind power data; and

[0107] acquire the intermediate wind power prediction data based on the corrected wind speed and the third feature set.

[0108] In some embodiments, the predicted power acquiring module 604 is configured to:

[0109] acquire fused wind power prediction data by fusing the intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data; and [0110] acquire, based on the fused wind power prediction data, the wind power prediction data corresponding to the target wind power plant.

[0111] In some embodiments, the history data acquiring module 601 is configured to:

[0112] acquire a meteorological timing data set, a wind turbine timing data set, a power timing data set, and a wind condition timing data set corresponding to the target wind power plant;

[0113] respectively preprocess the meteorological timing data set, the wind turbine timing data set, the power timing data set, and the wind condition timing data set; and

[0114] acquire the history meteorological data, the wind turbine data, the history wind power data, and the wind condition data of the target wind power plant in the target history time period by respectively performing data extraction on the preprocessed meteorological timing data set, the preprocessed wind turbine timing data set, the preprocessed power timing data set, and the preprocessed wind condition timing data set by adopting a sliding window method, [0115] the preprocessing including at least one of abnormal data processing, default data interpolation, and data normalization.

[0116] In some embodiments, the wind power prediction data are acquired from a wind power predicting model; wherein the wind power predicting model comprises a plurality of first wind power acquisition networks and a second wind power acquisition network, each of the first wind power acquisition networks is configured to acquire intermediate wind power prediction data corresponding to different meteorological prediction data, and the first wind power acquisition networks comprise a corrected wind speed acquisition sub-network and an intermediate power acquisition sub-network;

[0117] the corrected wind speed acquisition sub-network is configured to acquire, based on the target meteorological prediction data, the wind turbine data, and the wind condition data, a corrected wind speed corresponding to the target meteorological prediction data;

[0118] the intermediate power acquisition sub-network is configured to acquire, based on the corrected wind speed and the history wind power data, the intermediate wind power prediction data corresponding to the target meteorological prediction data; and

[0119] the second wind power acquisition network is configured to acquire, based on the intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data corresponding to the target wind power plant.

[0120] In summary, according to the technical solutions according to the embodiments of the present disclosure, a plurality of groups of meteorological prediction data of a target wind power plant are acquired through different numerical weather prediction systems, then the wind power of the target wind power plant is predicted respectively based on different meteorological prediction data, wind turbine data, history wind power data, and wind condition data, and finally wind power prediction data corresponding to the target wind power plant are acquired by synthesizing a plurality of intermediate wind power prediction data. Therefore, the prediction of the wind power is achieved by synthesizing the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, which doesn't just depend on meteorological prediction data in a single numerical weather prediction system, thereby improving the prediction accuracy of the wind power. [0121] It should be noted that, in a case that the apparatus according to the above embodiment implements the functions thereof, the division of the functional modules is merely exemplary. In practical application, the above functions can be assigned to different functional modules according to needs, that is, the internal structure of the device may be divided into different functional modules, so as to implement all or a part of the above functions. In addition, the apparatus and the method according to the above embodiments belong to the same concept, and specific implementation processes thereof are described in the method embodiments in detail, and are not repeated herein.

[0122] FIG. 8 is a block diagram of a computer device according to an embodiment of the present disclosure. The computer device may be configured to perform the method for predicting wind power according to the above embodiments.

[0123] The computer device 800 includes a central processing unit a CPU 801, for example, a graphics processing unit (GPU) or a field programmable gate array (FPGA), a system memory 804 including a random-access memory (RAM) 802 and a read-only memory (ROM) 803, and a system bus 805 configured to connect the system memory 804 to the central processing unit 801. The computer device 800 also includes a basic input/output system (I/O system) 806 configured to facilitate information transfer between devices within the server, and a mass storage device 807 configured to store an operating system 813, application programs 814, and other program modules 815.

[0124] The basic input/output system 806 includes a display 808 configured to display information and an input device 810 configured to input information by a user, such as a mouse and a keyboard. The display 808 and the input device 809 are connected to the central processing unit 801 through the input/output controller 810 connected to the system bus 805. The basic input/output system 806 may further include an input/output controller 810 configured to receive and process inputs from a plurality of other devices, such as a keyboard, a mouse, or electronic stylus. Similarly, the input/output controller 810 further provides output devices outputting onto a display screen and a printer, or other type of output devices.

[0125] The mass storage device 807 is connected to the CPU 801 through a mass storage controller (not shown) connected to the system bus 805. The mass storage device 807 and computer readable media associated therewith provide non-volatile storage for the computer device 800. That is, the mass storage device 807 may include a computer readable medium (not shown) such as a hard disk or compact disc read-only memory (CD-ROM) drive.

[0126] Generally, the computer-readable medium may include a computer storage medium and a communication medium. The computer storage media include a volatile and non-volatile, removable and non-removable medium implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. The computer storage medium includes a RAM, a ROM, an erasable programmable read only memory (EPROM), an electrically erasable programmable read only memory (EEPROM), a flash memory or other solid state storage techniques, a CD-ROM, a digital versatile disc (DVD), or other optical storage, magnetic cassette, magnetic tape, magnetic disc storage or magnetic storage devices. It will be appreciated by those skilled in the art that the computer storage medium is not limited to the foregoing. The system memory 804 and the mass storage device 807 described above may be collectively referred to as memory.

[0127] According to the embodiments of the present disclosure, the computer device 800 may be further connected to a remote computer on a network through the network, such as the Internet, for running. That is, the computer device 800 may be connected to the network 812 through a network interface unit 811 connected to the system bus 805, or may be connected to other types of networks or remote computer systems (not shown) using the network interface unit 811.

[0128] The memory also includes one or more computer programs stored therein. The one or more computer programs, when loaded and run by one or more processors, cause the computer device to perform the method for predicting wind power as described above.

[0129] An exemplary embodiment of the present disclosure provides a non-transitory computer- readable storage medium including one or more computer programs stored therein. The one or more computer programs, when loaded and run by a processor of a server, cause the server to perform the method for predicting wind power as described above.

[0130] Optionally, the computer readable storage medium may include: a read-only memory (ROM), a random-access memory (RAM), a solid state drive (SSD), an optical disc, or the like. The RAM may include a resistance random-access memory (ReRAM) and a dynamic randomaccess memory (DRAM).

[0131] An exemplary embodiment of the present disclosure provides a computer program product or a computer program including one or more computer instructions. The one or more computer instructions are stored in a computer readable storage medium. The one or more computer instructions, when loaded and run by a processor of a computer device, cause the computer device to perform the method for predicting wind power as described above.

[0132] The term "a plurality of means two or more. The term "and/or" describes the association relationship of the associated objects, and indicates that three relationships may be present. For example, A and/or B may indicate that: only A is present, both A and B are present, and only B is present. The symbol "/" usually indicates an "or" relationship between the associated objects. In addition, serial numbers of the steps described herein only show an exemplary possible execution sequence among the steps, and in some other embodiments, the steps may also be executed out of the numbering sequence, for example, two steps with different serial numbers are executed simultaneously, or two steps with different serial numbers are executed in a reverse order to the illustrated sequence, which is not limited in the present disclosure.

[0133] Described above are merely exemplary embodiments of the present disclosure and are not intended to limit the present disclosure. Any modifications, equivalents, improvements, and the like, made within the spirit and principle of the present disclosure should fall within the protection scope of the present disclosure.

Claims

CLAIMS What is claimed is:

1. A method for predicting wind power, comprising: acquiring history meteorological data, wind turbine data, history wind power data, and wind condition data of a target wind power plant within a target history time period; acquiring, based on the history meteorological data, a plurality of groups of meteorological prediction data corresponding to the target wind power plant by different numerical weather prediction systems, the meteorological prediction data being indicative of predicted meteorological data of the target wind power plant within a target prediction time period, the target prediction time period being a time period after the target history time period; acquiring, based on target meteorological prediction data in the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, intermediate wind power prediction data corresponding to the target meteorological prediction data; and acquiring wind power prediction data corresponding to the target wind power plant by calculation based on intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data being indicative a predicted wind power of the target wind power plant in the target prediction time period.

2. The method according to claim 1, wherein acquiring, based on the target meteorological prediction data in the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, the intermediate wind power prediction data corresponding to the target meteorological prediction data comprises: acquiring, based on the target meteorological prediction data, the wind turbine data, and the wind condition data, a corrected wind speed corresponding to the target meteorological prediction data; and acquiring the intermediate wind power prediction data based on the corrected wind speed and the history wind power data.

3. The method according to claim 2, wherein acquiring, based on the target meteorological

23 prediction data, the wind turbine data, and the wind condition data, the corrected wind speed corresponding to the target meteorological prediction data comprises: acquiring a first feature set corresponding to the meteorological prediction data by performing extraction of timing statistical features on the meteorological prediction data, the timing statistical features comprising at least one of a variance, a multi-order difference, a mean value, and a quantile value; constructing a second feature set based on the meteorological prediction data and the history meteorological data, the second feature set being indicative of a relationship between the meteorological prediction data and the history meteorological data; constructing an artificial high-order feature based on the wind condition data, the history meteorological data, and the wind turbine data, the artificial high-order feature comprising at least one of a date, a wind frequency, a wind energy, and turbulence; and acquiring, based on the first feature set, the second feature set, and the artificial high-order feature, the corrected wind speed corresponding to the target meteorological prediction data.

4. The method according to claim 3, wherein constructing the second feature set based on the meteorological prediction data and the history meteorological data comprises: acquiring a deviation value set between data of a target meteorological parameter in the meteorological prediction data and data of the target meteorological parameter in the history meteorological data; determining a mean value corresponding to the deviation value set as a cross timing statistical feature corresponding to the target meteorological parameter; and constructing the second feature set based on cross timing statistical features corresponding to all meteorological parameters in the meteorological prediction data.

5. The method according to claim 2, wherein acquiring the intermediate wind power prediction data based on the corrected wind speed and the history wind power data comprises: acquiring a third feature set corresponding to the history wind power data by performing extraction of timing statistical features on the history wind power data; and acquiring the intermediate wind power prediction data based on the corrected wind speed and the third feature set.

6. The method according to claim 1, wherein acquiring, based on the intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data corresponding to the target wind power plant comprises: acquiring fused wind power prediction data by fusing the intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data; and acquiring, based on the fused wind power prediction data, the wind power prediction data corresponding to the target wind power plant.

7. The method according to claim 1, wherein acquiring the history meteorological data, the wind turbine data, the history wind power data, and the wind condition data of the target wind power plant in the target history time period comprises: acquiring a meteorological timing data set, a wind turbine timing data set, a power timing data set, and a wind condition timing data set corresponding to the target wind power plant; respectively preprocessing the meteorological timing data set, the wind turbine timing data set, the power timing data set, and the wind condition timing data set; and acquiring the history meteorological data, the wind turbine data, the history wind power data, and the wind condition data of the target wind power plant in the target history time period by respectively performing data extraction on the preprocessed meteorological timing data set, the preprocessed wind turbine timing data set, the preprocessed power timing data set, and the preprocessed wind condition timing data set by adopting a sliding window technique, wherein the preprocessing comprises at least one of abnormal data processing, default data interpolation, and data normalization.

8. The method according to claim 1, wherein the wind power prediction data are acquired from a wind power predicting model; wherein the wind power predicting model comprises a plurality of first wind power acquisition networks and a second wind power acquisition network, each of the first wind power acquisition networks is configured to acquire intermediate wind power prediction data corresponding to different meteorological prediction data, and the first wind power acquisition networks comprise a corrected wind speed acquisition sub-network and an intermediate power acquisition sub-network; the corrected wind speed acquisition sub-network is configured to acquire, based on the target meteorological prediction data, the wind turbine data, and the wind condition data, a corrected wind speed corresponding to the target meteorological prediction data; the intermediate power acquisition sub-network is configured to acquire, based on the corrected wind speed and the history wind power data, the intermediate wind power prediction data corresponding to the target meteorological prediction data; and the second wind power acquisition network is configured to acquire, based on the intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data corresponding to the target wind power plant.

9. An apparatus for predicting wind power, comprising: a history data acquiring module, configured to acquire history meteorological data, wind turbine data, history wind power data, and wind condition data of a target wind power plant within a target history time period; a predicted data acquiring module, configured to acquire, based on the history meteorological data, a plurality of groups of meteorological prediction data corresponding to the target wind power plant by different numerical weather prediction systems, the meteorological prediction data being indicative of indicate predicted meteorological data of the target wind power plant within a target prediction time period, and the target prediction time period being a time period after the target history time period; an intermediate power acquiring module, configured to acquire, based on target meteorological prediction data in the plurality of groups of meteorological prediction data, the wind turbine data, the history wind power data, and the wind condition data, intermediate wind power prediction data corresponding to the target meteorological prediction data; and a predicted power acquiring module, configured to acquire wind power prediction data corresponding to the target wind power plant by calculation based on intermediate wind power prediction data respectively corresponding to the plurality of groups of meteorological prediction data, the wind power prediction data being indicative of indicate a predicted wind power of the target wind power plant in the target prediction time period.

26

10. A computer device, comprising a processor and a memory storing one or more computer programs therein, wherein the one or more computer programs, when loaded and run by the processor, cause the processor to computer device to perform the method for predicting wind power as defined in any one of claims 1 to 8.

11. A non-transitory computer-readable storage medium storing one or more computer programs therein, wherein the one or more computer programs, when loaded and run by a processor of a computer device, cause the computer device to perform the method for predicting wind power as defined in any one of claims 1 to 8.

27