Training method and device of wind speed prediction model, equipment and medium

Abstract

The invention relates to a training method and device of a wind speed prediction model, equipment and a medium. The method comprises the following steps: acquiring first historical wind speed data, first future wind speed data and historical meteorological data of a wind power plant at a historical moment; performing clustering processing on the first historical wind speed data to obtain a plurality of historical clusters of the first historical wind speed data; target meteorological data corresponding to each historical cluster is determined from the historical meteorological data, and the target meteorological data is the historical meteorological data with the maximum target association degree with the first future wind speed data; and performing iterative training on a preset network corresponding to each historical cluster by using the plurality of historical clusters, the target meteorological data corresponding to each cluster and the first future wind speed data to obtain a wind speed prediction model corresponding to each historical cluster. Therefore, wind speed prediction is carried out by using the wind speed prediction models corresponding to different clusters, wind speed fluctuation conditions of different degrees can be accurately identified, and the wind speed prediction precision is improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of wind speed prediction, and in particular to a training method, device, equipment and medium for a wind speed prediction model. Background Art

[0002] In the field of wind power development, wind farm wake control technology is one of the key technologies to improve the power generation efficiency of wind farms and reduce wake interference between units. This technology is highly dependent on future wind speed data with high temporal resolution as its core input.

[0003] However, current wind speed prediction technologies mostly focus on historical wind speeds over longer time periods to predict future wind speeds, which cannot meet the needs of real-time control of wind farms. At the same time, historical wind speeds may change rapidly and significantly in a short period of time, increasing the difficulty of predicting future wind speeds.

[0004] Obviously, this wind speed prediction method has low accuracy, which in turn affects the wake control effect of the wind farm. Summary of the invention

[0005] In order to solve the above technical problems, the present disclosure provides a training method, device, equipment and medium for a wind speed prediction model.

[0006] In a first aspect, the present disclosure provides a method for training a wind speed prediction model, comprising:

[0007] Acquire first historical wind speed data, first future wind speed data, and historical meteorological data of the wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are measured wind speed data of the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold;

[0008] Performing clustering processing on the first historical wind speed data to obtain a plurality of historical clusters of the first historical wind speed data;

[0009] Determining target meteorological data corresponding to each historical cluster from the historical meteorological data, wherein the target meteorological data is the historical meteorological data having the greatest target correlation with the first future wind speed data;

[0010] By using the multiple historical clusters, the target meteorological data corresponding to each cluster, and the first future wind speed data, the preset network corresponding to each historical cluster is iteratively trained to obtain a wind speed prediction model corresponding to each historical cluster.

[0011] In a second aspect, the present disclosure provides a training device for a wind speed prediction model, comprising:

[0012] A first acquisition module is used to acquire first historical wind speed data, first future wind speed data and historical meteorological data of the wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are measured wind speed data of the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold;

[0013] A second acquisition module, configured to perform clustering processing on the first historical wind speed data to obtain a plurality of historical clusters of the first historical wind speed data;

[0014] A first determination module is used to determine target meteorological data corresponding to each historical cluster from the historical meteorological data, wherein the target meteorological data is the historical meteorological data with the greatest target correlation with the first future wind speed data;

[0015] The second determination module is used to iteratively train the preset network corresponding to each historical cluster using the multiple historical clusters, the target meteorological data corresponding to each cluster and the first future wind speed data to obtain a wind speed prediction model corresponding to each historical cluster.

[0016] In a third aspect, an embodiment of the present disclosure further provides an electronic device, the device comprising:

[0017] one or more processors;

[0018] a storage device for storing one or more programs,

[0019] When one or more programs are executed by one or more processors, the one or more processors implement the method provided in the first aspect.

[0020] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method provided in the first aspect is implemented.

[0021] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:

[0022] A training method, device, equipment and medium for a wind speed prediction model of an embodiment of the present disclosure obtains first historical wind speed data, first future wind speed data and historical meteorological data of a wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are actually measured wind speed data of the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold; clusters the first historical wind speed data to obtain multiple historical clustering clusters of the first historical wind speed data; determines target meteorological data corresponding to each historical clustering cluster from the historical meteorological data, wherein the target meteorological data is the historical meteorological data with the greatest target correlation with the first future wind speed data; and iteratively trains a preset network corresponding to each historical clustering cluster using multiple historical clustering clusters, the target meteorological data corresponding to each clustering cluster and the first future wind speed data to obtain a wind speed prediction model corresponding to each historical clustering cluster. In this way, historical wind speed data with high temporal resolution and similar fluctuation characteristics can be clustered into the same cluster, and for clusters with different fluctuation characteristics, the meteorological features most relevant to future wind speed can be selected, and then the wind speed prediction models corresponding to different clusters can be obtained through machine learning methods. Therefore, wind speed prediction using wind speed prediction models corresponding to different clusters can accurately identify wind speed fluctuations of different degrees, thus achieving accurate prediction of high temporal resolution wind speed data, which is beneficial to improving the control effect of wind farm wake. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0025] Figure 1 A schematic diagram of a flow chart of a training method for a wind speed prediction model provided in an embodiment of the present disclosure;

[0026] Figure 2 A schematic diagram of the structure of a wind speed prediction model training device provided in an embodiment of the present disclosure;

[0027] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0028] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0029] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0030] In order to solve the above problems, the following Figure 1 The training method of the wind speed prediction model provided in the embodiment of the present disclosure is described. In the embodiment of the present disclosure, the training method of the wind speed prediction model can be executed by an electronic device or a server. Among them, the electronic device may include a device with a communication function such as a tablet computer, a desktop computer, a laptop computer, and may also include a device simulated by a virtual machine or a simulator. The server may include a single server or a server cluster. This embodiment is specifically explained with an electronic device as the execution subject.

[0031] Figure 1 A flow chart of a training method for a wind speed prediction model provided by an embodiment of the present disclosure is shown.

[0032] like Figure 1 As shown, the training method of the wind speed prediction model may include the following steps.

[0033] S110. Acquire first historical wind speed data, first future wind speed data, and historical meteorological data of the wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are measured wind speed data of the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold.

[0034] In this embodiment, when performing model training, the electronic device obtains the wind farm's measured wind speed data and related meteorological data over a long period of time in the past to meet the comprehensiveness requirements of the subsequent model training.

[0035] The first historical wind speed data is wind speed data actually measured in the wind farm before the historical moment. Optionally, the first historical wind speed data can be measured by any one of a wind turbine, a laser radar, and a wind tower.

[0036] The first future wind speed data is the measured wind speed data after the historical moment and before the current moment.

[0037] The historical meteorological data is the relevant meteorological data before the historical moment. Optionally, the historical meteorological data includes but is not limited to one or more of temperature, humidity, air pressure, etc.

[0038] In order to improve the time resolution of the measured wind speed data, the electronic device can divide the wind farm measured wind speed data into a plurality of short-duration first historical wind speed data and first future wind speed data according to a preset sliding window. Since the time length of the preset time window is less than the preset length threshold, the first historical wind speed data and the first future wind speed data are wind farm measured wind speed data with high time resolution, so as to facilitate the subsequent use of the wind farm measured wind speed data with high time resolution to train the wind speed prediction model.

[0039] The preset length threshold is a minimum time length determined based on experience, for example, the preset length threshold is 30 seconds, 1 minute, etc.

[0040] In order to improve the accuracy of model training, the electronic device may also preprocess the first historical wind speed data and the historical meteorological data to improve the quality and consistency of the historical measured wind speed and the historical predicted wind speed data. Optionally, the preprocessing method includes but is not limited to one or more methods of missing value completion, outlier processing, and time correction processing of the data used.

[0041] S120 , clustering the first historical wind speed data to obtain a plurality of historical clusters of the first historical wind speed data.

[0042] It can be understood that, considering that the first historical wind speed data may change rapidly and significantly in a short period of time, this embodiment clusters the first historical wind speed data with high time resolution and similar fluctuation characteristics into the same cluster to obtain multiple historical clustering clusters.

[0043] Among them, the specific implementation method of S120 includes but is not limited to the following method: obtaining the target cluster number corresponding to the first historical wind speed data; clustering the first historical wind speed data according to the target cluster number to obtain multiple historical clusters of the first historical wind speed data.

[0044] Among them, the target cluster number corresponding to the first historical wind speed data is obtained, and the specific implementation method includes but is not limited to the following methods: obtaining multiple different initial cluster numbers; based on the initial cluster numbers, clustering the first historical wind speed data to obtain the initial cluster clusters of the first historical wind speed data; calculating the evaluation index for the initial cluster clusters corresponding to each initial cluster number to obtain the evaluation index value corresponding to each initial cluster cluster; from multiple different initial cluster numbers, obtaining the cluster number with the highest evaluation index value as the target cluster number.

[0045] Specifically, the electronic device obtains multiple different initial clustering data, clusters the first historical wind speed data based on multiple different initial clustering numbers and using a preset clustering algorithm to obtain multiple different initial clustering clusters, and then uses a preset clustering evaluation index algorithm to calculate the evaluation index value corresponding to each initial clustering cluster, and then obtains the clustering data with the highest evaluation index value from the multiple different initial clustering clusters as the target clustering number, thereby obtaining the optimal clustering number, and then based on the target clustering number and using a preset clustering algorithm, clusters the first historical wind speed data to obtain multiple historical clustering clusters.

[0046] Optionally, the preset clustering algorithm includes but is not limited to any one or more of a deep embedding clustering algorithm, a clustering method based on time series data (such as a K-shape clustering algorithm), and the like.

[0047] Optionally, the evaluation index value includes but is not limited to one or more of a silhouette coefficient obtained based on a silhouette coefficient calculation algorithm, a ratio of inter-cluster distance to intra-cluster distance obtained based on an index algorithm for evaluating clustering quality (such as the Davies-Bouldin index), and the like.

[0048] Therefore, by performing evaluation index analysis on multiple initial clustering clusters determined by multiple initial clustering clusters, the optimal number of clusters for clustering the first historical wind speed data is determined, and multiple historical clustering clusters are determined based on the optimal number of clusters, thereby achieving clustering of high-time-resolution historical wind speed data with similar fluctuation characteristics into the same cluster, which is beneficial to overcoming the influence of wind speed fluctuations when predicting wind speed.

[0049] S130. Determine target meteorological data corresponding to each historical cluster from the historical meteorological data, wherein the target meteorological data is the historical meteorological data having the greatest target correlation with the first future wind speed data.

[0050] In order to further improve the prediction accuracy of wind speed data with high temporal resolution, electronic equipment also needs to accurately identify the meteorological data that has the greatest impact on future wind speed to train the wind speed prediction model. Specifically, for each historical cluster, the corresponding target meteorological data is determined from the historical meteorological data.

[0051] Among them, the specific implementation method of S130 includes but is not limited to the following methods: for each historical clustering cluster, determine the target correlation between the historical meteorological data and the first future wind speed data; obtain the meteorological data with the largest target correlation from the historical meteorological data as the target meteorological data corresponding to each historical clustering cluster.

[0052] Among them, for each historical clustering cluster, the target correlation between the historical meteorological data and the first future wind speed data is determined, and the specific implementation method includes but is not limited to the following method: determining the first correlation between each historical clustering cluster and each historical meteorological data; for each historical clustering cluster, determining the second correlation between each historical meteorological data and the first future wind speed data; performing weighted averaging calculation on the first correlation and the second correlation to obtain the target correlation.

[0053] Specifically, first, for each historical clustering cluster, the electronic device calculates the correlation between each first historical wind speed data and each historical meteorological data, and averages all the correlations corresponding to each historical clustering cluster to obtain the first correlation between each historical clustering cluster and each historical meteorological data; then, since each historical clustering cluster corresponds to a series of historical meteorological data, for each historical clustering cluster, a baseline model is established for the historical meteorological data and the first future wind speed data, and the second correlation between each historical meteorological data and the first future wind speed data is calculated by adding historical meteorological data to each historical clustering cluster one by one or removing historical meteorological data from each historical clustering cluster one by one; then, the electronic device performs weighted averaging calculation on the first correlation and the second correlation based on the weights corresponding to the first correlation and the second correlation, to obtain the target correlation.

[0054] Optionally, the correlation between each first historical wind speed data and each historical meteorological data can be determined by any one of a Pearson correlation coefficient calculation method and a Spearman rank correlation coefficient calculation method.

[0055] Optionally, the benchmark model includes but is not limited to a recurrent neural network such as a gated recurrent unit (GRU), a long short-term memory network (LSTM), etc. The second correlation between each historical meteorological data and the first future wind speed data includes but is not limited to a root mean square error (RMSE), a coefficient of determination (R 2 )

[0056] Optionally, the weights corresponding to the first degree of association and the second degree of association respectively may be preset based on experience.

[0057] Therefore, for clusters with different fluctuation characteristics, the meteorological features most relevant to future wind speed are selected to utilize the meteorological features most relevant to future wind speed and improve the prediction accuracy of high temporal resolution wind speed.

[0058] S140, using multiple historical clusters, target meteorological data corresponding to each cluster, and first future wind speed data, iteratively train a preset network corresponding to each historical cluster to obtain a wind speed prediction model corresponding to each historical cluster.

[0059] In this embodiment, for each historical clustering cluster, the electronic device uses the first historical wind speed data contained in the historical clustering cluster and the target meteorological data corresponding to the clustering cluster as input data of a preset network corresponding to the historical clustering cluster, and uses the first future wind speed data as output data of the preset network corresponding to the historical clustering cluster, and then uses the input data and the output data to iteratively train the preset network corresponding to the clustering cluster until the network parameters of the preset network reach a stable state, thereby obtaining a wind speed prediction model corresponding to each historical clustering cluster.

[0060] In order to improve the generalization ability and reliability of the model, before model training, the target meteorological data and the first future wind speed data corresponding to each historical cluster can be normalized to use the normalized data for model training. After model training, the trained wind speed prediction model can be tested using the test samples corresponding to each historical cluster to verify the accuracy of the trained prediction model.

[0061] Optionally, the wind speed prediction model includes but is not limited to any one of recurrent neural networks such as a gated recurrent unit (GRU for short), a long short-term memory network (LSTM for short), etc.

[0062] A training method for a wind speed prediction model in an embodiment of the present disclosure obtains first historical wind speed data, first future wind speed data and historical meteorological data of a wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are actually measured wind speed data of the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold; clustering is performed on the first historical wind speed data to obtain multiple historical clustering clusters of the first historical wind speed data; target meteorological data corresponding to each historical clustering cluster is determined from the historical meteorological data, wherein the target meteorological data is the historical meteorological data with the greatest target correlation with the first future wind speed data; and iterative training is performed on a preset network corresponding to each historical clustering cluster using multiple historical clustering clusters, the target meteorological data corresponding to each clustering cluster and the first future wind speed data to obtain a wind speed prediction model corresponding to each historical clustering cluster. In this way, historical wind speed data with high temporal resolution and similar fluctuation characteristics can be clustered into the same cluster, and for clusters with different fluctuation characteristics, the meteorological features most relevant to future wind speed can be selected, and then the wind speed prediction models corresponding to different clusters can be obtained through machine learning methods. Therefore, wind speed prediction using wind speed prediction models corresponding to different clusters can accurately identify wind speed fluctuations of different degrees, thus achieving accurate prediction of high temporal resolution wind speed data, which is beneficial to improving the control effect of wind farm wake.

[0063] In some embodiments of the present disclosure, after the electronic device has trained the wind speed prediction model corresponding to each historical clustering cluster, the method also includes the following steps: obtaining the second historical wind speed data of the wind farm at the current moment; determining the target clustering cluster corresponding to the second historical wind speed data from multiple historical clustering clusters; and performing wind speed prediction processing on the second historical wind speed data using the wind speed prediction model corresponding to the target clustering cluster to obtain the second future wind speed data at the current moment.

[0064] Among them, from multiple historical clustering clusters, a target clustering cluster corresponding to the second historical wind speed data is determined, and the specific implementation method includes but is not limited to the following method: calculating the average distance between the second historical wind speed data and each historical clustering cluster; obtaining the clustering cluster with the smallest average distance from multiple historical clustering clusters as the target clustering cluster.

[0065] Specifically, after the wind speed prediction model corresponding to each historical clustering cluster is trained, the wind speed prediction model corresponding to each historical clustering cluster can be put into use. First, the average distance between the second historical wind speed data and each historical clustering cluster is calculated, and the clustering cluster with the smallest average distance is obtained from multiple historical clustering clusters as the target clustering cluster with the highest closeness to the second historical wind speed data. The second historical wind speed data is then input into the wind speed prediction model corresponding to the target clustering cluster, and the information wind speed is predicted to obtain the second future wind speed data at the current moment.

[0066] The average distance is used to characterize the closeness between the second historical wind speed data and the historical clustering clusters. Specifically, the smaller the average distance, the higher the closeness between the second historical wind speed data and the historical clustering clusters. On the contrary, the larger the average distance, the lower the closeness between the second historical wind speed data and the historical clustering clusters.

[0067] Specifically, the average distance may be calculated based on the distance between the second historical wind speed data and each first historical data in the historical cluster and then the average value is obtained. Optionally, the average distance includes but is not limited to the Euclidean distance.

[0068] Therefore, after the wind speed prediction model corresponding to each historical cluster is put into use, for the second historical wind speed data at the current moment, the target cluster with the highest degree of closeness is selected, and the wind speed prediction model corresponding to the target cluster is used to predict the second future wind speed data. Obviously, this method improves the prediction accuracy of high temporal resolution wind speed.

[0069] The present disclosure also provides a wind speed prediction model training device for implementing the wind speed prediction model training method described above, and the device is configured in an electronic device or a server. Figure 2The electronic device may include a tablet computer, a desktop computer, a laptop computer, or other devices with communication functions, and may also include a virtual machine or a device simulated by a simulator. The server may include a single server or a server cluster.

[0070] Figure 2 A schematic structural diagram of a wind speed prediction model training device provided by an embodiment of the present disclosure is shown.

[0071] like Figure 2 As shown, the training device 200 of the wind speed prediction model may include:

[0072] A first acquisition module 210 is used to acquire first historical wind speed data, first future wind speed data and historical meteorological data of the wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are wind speed data actually measured in the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold;

[0073] A second acquisition module 220 is used to perform clustering processing on the first historical wind speed data to obtain a plurality of historical clusters of the first historical wind speed data;

[0074] A first determination module 230 is used to determine target meteorological data corresponding to each historical cluster from the historical meteorological data, wherein the target meteorological data is the historical meteorological data with the greatest target correlation with the first future wind speed data;

[0075] The second determination module 240 is used to iteratively train the preset network corresponding to each historical cluster using the multiple historical clusters, the target meteorological data corresponding to each cluster and the first future wind speed data to obtain a wind speed prediction model corresponding to each historical cluster.

[0076] A training device for a wind speed prediction model of an embodiment of the present disclosure obtains first historical wind speed data, first future wind speed data and historical meteorological data of a wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are actually measured wind speed data of the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold; clustering is performed on the first historical wind speed data to obtain multiple historical clustering clusters of the first historical wind speed data; target meteorological data corresponding to each historical clustering cluster is determined from the historical meteorological data, wherein the target meteorological data is the historical meteorological data with the greatest target correlation with the first future wind speed data; and a preset network corresponding to each historical clustering cluster is iteratively trained using multiple historical clustering clusters, the target meteorological data corresponding to each clustering cluster and the first future wind speed data to obtain a wind speed prediction model corresponding to each historical clustering cluster. In this way, historical wind speed data with high temporal resolution and similar fluctuation characteristics can be clustered into the same cluster, and for clusters with different fluctuation characteristics, the meteorological features most relevant to future wind speed can be selected, and then the wind speed prediction models corresponding to different clusters can be obtained through machine learning methods. Therefore, wind speed prediction using wind speed prediction models corresponding to different clusters can accurately identify wind speed fluctuations of different degrees, thus achieving accurate prediction of high temporal resolution wind speed data, which is beneficial to improving the control effect of wind farm wake.

[0077] In some embodiments of the present disclosure, the second acquisition module 220 includes:

[0078] An acquisition unit, used to acquire the target cluster number corresponding to the first historical wind speed data;

[0079] A clustering unit is used to perform clustering processing on the first historical wind speed data according to the target clustering number to obtain multiple historical clusters of the first historical wind speed data.

[0080] In some embodiments of the present disclosure, the acquisition unit is specifically configured to:

[0081] Get multiple different initial cluster numbers;

[0082] Based on the initial cluster number, clustering processing is performed on the first historical wind speed data to obtain initial clusters of the first historical wind speed data;

[0083] Calculate the evaluation index for each initial cluster corresponding to the initial cluster number to obtain the evaluation index value corresponding to each initial cluster;

[0084] From the multiple different initial cluster numbers, the cluster number with the highest evaluation index value is obtained as the target cluster number.

[0085] In some embodiments of the present disclosure, the first determination module 230 includes:

[0086] a first determining unit, configured to determine, for each of the historical clusters, a target correlation between the historical meteorological data and the first future wind speed data;

[0087] The second determining unit is used to obtain meteorological data with the largest target correlation from the historical meteorological data as the target meteorological data corresponding to each historical cluster.

[0088] In some embodiments of the present disclosure, the first determining unit is specifically configured to: determine a first degree of association between each historical cluster and each historical meteorological data;

[0089] For each of the historical clusters, determining a second degree of association between each of the historical meteorological data and the first future wind speed data;

[0090] A weighted average calculation is performed on the first degree of association and the second degree of association to obtain the target degree of association.

[0091] In some embodiments of the present disclosure, the device further includes:

[0092] The third acquisition module is used to acquire the second historical wind speed data of the wind farm at the current moment;

[0093] A third determination module, configured to determine a target cluster corresponding to the second historical wind speed data from the plurality of historical clusters;

[0094] The fourth determination module is used to use the wind speed prediction model corresponding to the target cluster to perform wind speed prediction processing on the second historical wind speed data to obtain the second future wind speed data at the current moment.

[0095] In some embodiments of the present disclosure, the third determination module is specifically configured to:

[0096] Calculating the average distance between the second historical wind speed data and each historical cluster;

[0097] A cluster with the smallest average distance is obtained from the multiple historical clusters as the target cluster.

[0098] It should be noted that Figure 2 The wind speed prediction model training device 200 shown can be executed Figure 1 The various steps in the method embodiment shown in the figure are implemented Figure 1 The various processes and effects in the method embodiment shown are not described in detail here.

[0099] Figure 3A schematic structural diagram of an electronic device provided by an embodiment of the present disclosure is shown.

[0100] like Figure 3 As shown, the electronic device may include a processor 301 and a memory 302 storing computer program instructions.

[0101] Specifically, the processor 301 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0102] The memory 302 may include a large capacity memory for information or instructions. By way of example and not limitation, the memory 302 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a tape, or a universal serial bus (USB) drive or a combination of two or more of these. Where appropriate, the memory 302 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 302 may be inside or outside the integrated gateway device. In a particular embodiment, the memory 302 is a non-volatile solid-state memory. In a particular embodiment, the memory 302 includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (Electrically Erasable Programmable ROM, EEPROM), an electrically rewritable ROM (EAROM) or a flash memory, or a combination of two or more of these.

[0103] The processor 301 reads and executes the computer program instructions stored in the memory 302 to perform the steps of the training method of the wind speed prediction model provided in the embodiment of the present disclosure.

[0104] In one example, the electronic device may further include a transceiver 303 and a bus 304. Figure 3 As shown, the processor 301, the memory 302 and the transceiver 303 are connected via a bus 304 and communicate with each other.

[0105] The bus 304 includes hardware, software, or both. For example, but not limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industrial Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a Memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, the bus 304 may include one or more buses. Although embodiments of the present application describe and illustrate a particular bus, the present application contemplates any suitable bus or interconnect.

[0106] The following is an embodiment of a computer-readable storage medium provided in an embodiment of the present disclosure. The computer-readable storage medium and the training method of the wind speed prediction model in the above-mentioned embodiments belong to the same inventive concept. For details not described in detail in the embodiment of the computer-readable storage medium, reference can be made to the embodiment of the training method of the wind speed prediction model in the above-mentioned embodiments.

[0107] This embodiment provides a storage medium containing computer executable instructions. When the computer executable instructions are executed by a computer processor, they are used to perform a training method for a wind speed prediction model, including:

[0108] Acquire first historical wind speed data, first future wind speed data, and historical meteorological data of the wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are measured wind speed data of the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold;

[0109] Performing clustering processing on the first historical wind speed data to obtain a plurality of historical clusters of the first historical wind speed data;

[0110] Determining target meteorological data corresponding to each historical cluster from the historical meteorological data, wherein the target meteorological data is the historical meteorological data having the greatest target correlation with the first future wind speed data;

[0111] By using the multiple historical clusters, the target meteorological data corresponding to each cluster, and the first future wind speed data, the preset network corresponding to each historical cluster is iteratively trained to obtain a wind speed prediction model corresponding to each historical cluster.

[0112] Of course, the storage medium containing computer executable instructions provided in the embodiment of the present disclosure is not limited to the above method operations, and the computer executable instructions can also execute related operations in the training method of the wind speed prediction model provided in any embodiment of the present disclosure.

[0113] Through the above description of the implementation methods, technicians in the relevant field can clearly understand that the present disclosure can be implemented with the help of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present disclosure can be essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer cloud platform (which can be a personal computer, a server, or a network cloud platform, etc.) to execute the training method of the wind speed prediction model provided in each embodiment of the present disclosure.

[0114] Note that the above are only preferred embodiments of the present disclosure and the technical principles used. Those skilled in the art will understand that the present disclosure is not limited to the specific embodiments herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present disclosure. Therefore, although the present disclosure is described in more detail through the above embodiments, the present disclosure is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present disclosure, and the scope of the present disclosure is determined by the scope of the appended claims.

Claims

1. A training method for a wind speed prediction model, characterized in that: include: Acquire first historical wind speed data, first future wind speed data, and historical meteorological data of the wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are measured wind speed data of the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold; Performing clustering processing on the first historical wind speed data to obtain a plurality of historical clusters of the first historical wind speed data; Determining target meteorological data corresponding to each historical cluster from the historical meteorological data, wherein the target meteorological data is the historical meteorological data having the greatest target correlation with the first future wind speed data; By using the multiple historical clusters, the target meteorological data corresponding to each cluster, and the first future wind speed data, the preset network corresponding to each historical cluster is iteratively trained to obtain a wind speed prediction model corresponding to each historical cluster.

2. The method according to claim 1, characterized in that The clustering process is performed on the first historical wind speed data to obtain a plurality of historical clusters of the first historical wind speed data, including: Obtaining the target cluster number corresponding to the first historical wind speed data; The first historical wind speed data is clustered according to the target cluster number to obtain a plurality of historical clusters of the first historical wind speed data.

3. The method according to claim 2, characterized in that The obtaining the target cluster number corresponding to the first historical wind speed data includes: Get multiple different initial cluster numbers; Based on the initial cluster number, clustering processing is performed on the first historical wind speed data to obtain initial clusters of the first historical wind speed data; Calculate the evaluation index for each initial cluster corresponding to the initial cluster number to obtain the evaluation index value corresponding to each initial cluster; From the multiple different initial cluster numbers, the cluster number with the highest evaluation index value is obtained as the target cluster number.

4. The method according to claim 1, characterized in that: The step of determining target meteorological data corresponding to each historical cluster from the historical meteorological data includes: For each of the historical clusters, determining a target correlation between the historical meteorological data and the first future wind speed data; The meteorological data with the greatest target correlation is obtained from the historical meteorological data as the target meteorological data corresponding to each historical cluster.

5. The method according to claim 4, characterized in that The step of determining, for each of the historical clusters, a target correlation between the historical meteorological data and the first future wind speed data comprises: Determining a first degree of association between each of the historical clusters and each of the historical meteorological data; For each of the historical clusters, determining a second degree of association between each of the historical meteorological data and the first future wind speed data; A weighted average calculation is performed on the first degree of association and the second degree of association to obtain the target degree of association.

6. The method according to claim 1, characterized in that Also includes: Obtain the second historical wind speed data of the wind farm at the current moment; Determining a target cluster corresponding to the second historical wind speed data from the multiple historical clusters; The wind speed prediction model corresponding to the target cluster is used to perform wind speed prediction processing on the second historical wind speed data to obtain the second future wind speed data at the current moment.

7. The method according to claim 6, characterized in that The step of determining a target cluster corresponding to the second historical wind speed data from the plurality of historical clusters includes: Calculating the average distance between the second historical wind speed data and each historical cluster; A cluster with the smallest average distance is obtained from the multiple historical clusters as the target cluster.

8. A training device for a wind speed prediction model, characterized in that: include: A first acquisition module is used to acquire first historical wind speed data, first future wind speed data and historical meteorological data of the wind farm at a historical moment, wherein the first historical wind speed data and the first future wind speed data are measured wind speed data of the wind farm within a preset sliding window, and the time length of the preset time window is less than a preset length threshold; A second acquisition module, configured to perform clustering processing on the first historical wind speed data to obtain a plurality of historical clusters of the first historical wind speed data; A first determination module is used to determine target meteorological data corresponding to each historical cluster from the historical meteorological data, wherein the target meteorological data is the historical meteorological data with the greatest target correlation with the first future wind speed data; The second determination module is used to iteratively train the preset network corresponding to each historical cluster using the multiple historical clusters, the target meteorological data corresponding to each cluster and the first future wind speed data to obtain a wind speed prediction model corresponding to each historical cluster.

9. An electronic device, characterized in that: include: processor; A memory for storing executable instructions; The processor is used to read the executable instructions from the memory and execute the executable instructions to implement the method described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements the method according to any one of claims 1 to 7.

Images