Wind power generation power prediction method, device, equipment, medium and product

Through the wind power generation power prediction method combined with residual channel attention network model and convolutional neural network, the wind power generation power prediction accuracy problem under the influence of wind speed data randomness is solved, and higher prediction accuracy is achieved.

CN119093328APending Publication Date: 2024-12-06GUANGDONG POWER GRID CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411069024.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-06
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

In the prior art, due to the randomness and volatility of wind speed data in the prediction of wind power, the prediction accuracy is not high, making it difficult to accurately predict wind power power.

Method used

The residual channel attention network model is used to predict wind speed, and the wind speed data is decomposed through the empirical modal decomposition algorithm, and the prediction value is corrected to improve accuracy. Then the corrected wind speed data is input into the convolutional neural network model for wind power prediction.

Benefits of technology

Through data preprocessing and error correction, the accuracy of wind power generation power prediction is effectively improved, the gap between the predicted value and the real value is reduced, and the accuracy of wind power prediction is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119093328A_ABST
    Figure CN119093328A_ABST
Patent Text Reader

Abstract

The invention discloses a wind power generation power prediction method, device and equipment, a medium and a product. The method comprises the following steps: acquiring predicted wind speed data of a to-be-predicted time period; the predicted wind speed data of the to-be-predicted time period are input into a target wind speed prediction model, actual wind speed data of the to-be-predicted time period are obtained, the target wind speed prediction model is a trained initial wind speed prediction model, and the initial wind speed prediction model is a residual channel attention network model; the actual wind speed data of the to-be-predicted time period is input into a target wind power generation power prediction model, wind power generation power corresponding to the to-be-predicted time period is obtained, and the target wind power generation power prediction model is a trained convolutional neural network model. The accuracy of wind power generation power prediction can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of new energy technology, and in particular to a method, device, equipment, medium and product for predicting wind power generation. Background Art

[0002] The urgent need for a low-carbon economy and the advancement of wind power technology are driving the rapid and sustainable transformation of the energy industry and the development of wind power worldwide. Due to the difficulty of connecting distributed power sources to the grid due to intermittency and randomness, the use of active distribution network technology is considered to be an effective technical approach to large-scale utilization of distributed energy, and one of the key technologies is the ability to actively analyze and predict energy. Accurate wind power forecasting can improve wind power utilization and system reliability, reduce operating costs, and allow for flexible scheduling strategies.

[0003] The existing technology has conducted a lot of research on the prediction of wind power generation, and adopted a variety of prediction models for power generation prediction. For example, the use of a multi-task learning framework to build a prediction network based on LSTM can have higher prediction accuracy than a single LSTM. However, since the size of wind power is greatly affected by weather factors, its output power data has strong randomness and volatility. Directly inputting relevant data into the prediction model will affect the accuracy of wind power generation prediction. Summary of the invention

[0004] The embodiments of the present invention provide a method, device, equipment, medium and product for predicting wind power generation to solve at least one technical problem existing in the prior art and improve the accuracy of wind power generation prediction.

[0005] According to one aspect of the present invention, a method for predicting wind power generation is provided, comprising:

[0006] Obtain the predicted wind speed data for the time period to be predicted;

[0007] Inputting the predicted wind speed data of the time period to be predicted into the target wind speed prediction model to obtain the actual wind speed data of the time period to be predicted, wherein the target wind speed prediction model is a trained initial wind speed prediction model, and the initial wind speed prediction model is a residual channel attention network model;

[0008] The actual wind speed data of the time period to be predicted is input into the target wind power prediction model to obtain the wind power corresponding to the time period to be predicted, wherein the target wind power prediction model is a trained convolutional neural network model.

[0009] According to another aspect of the present invention, a wind power generation power prediction device is provided, the wind power generation power prediction device comprising:

[0010] A predicted wind speed data acquisition module is used to obtain the predicted wind speed data for the time period to be predicted;

[0011] An actual wind speed data determination module is used to input the predicted wind speed data of the time period to be predicted into a target wind speed prediction model to obtain the actual wind speed data of the time period to be predicted, wherein the target wind speed prediction model is a trained initial wind speed prediction model, and the initial wind speed prediction model is a residual channel attention network model;

[0012] The wind power generation determination module is used to input the actual wind speed data of the time period to be predicted into the target wind power generation prediction model to obtain the wind power generation corresponding to the time period to be predicted, wherein the target wind power generation prediction model is a trained convolutional neural network model.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the wind power generation power prediction method described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the wind power prediction method described in any embodiment of the present invention when executed.

[0018] According to another aspect of the present invention, a computer program product is provided. When the computer program is executed by a processor, the computer program implements the wind power prediction method as described in any one of the embodiments of the present invention.

[0019] The embodiment of the present invention obtains predicted wind speed data for a time period to be predicted; inputs the predicted wind speed data for the time period to be predicted into a target wind speed prediction model to obtain actual wind speed data for the time period to be predicted, wherein the target wind speed prediction model is a trained initial wind speed prediction model, and the initial wind speed prediction model is a residual channel attention network model; inputs the actual wind speed data for the time period to be predicted into a target wind power prediction model to obtain the wind power corresponding to the time period to be predicted, wherein the target wind power prediction model is a trained convolutional neural network model, which can improve the accuracy of wind power prediction.

[0020] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 is a flow chart of a method for predicting wind power generation in an embodiment of the present invention;

[0023] Figure 2 is a structural schematic diagram of a wind power generation power prediction device in an embodiment of the present invention;

[0024] Figure 3 It is a structural schematic diagram of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0028] Embodiment 1

[0029] Figure 1 A flow chart of a wind power generation prediction method provided in an embodiment of the present invention. This embodiment is applicable to the case of wind power generation prediction. The method can be executed by a wind power generation prediction device in an embodiment of the present invention. The device can be implemented in software and / or hardware. Figure 1 As shown, the method specifically comprises the following steps:

[0030] S110, obtaining predicted wind speed data for a time period to be predicted.

[0031] The predicted wind speed data may be wind speed data in a weather forecast. The time period to be predicted may be a pre-selected time period or a time period set by a user. The unit of the time period to be predicted may be minutes, hours, days, weeks, months, seasons, years, etc., which is not limited in the embodiment of the present invention.

[0032] S120, inputting the predicted wind speed data of the time period to be predicted into a target wind speed prediction model to obtain actual wind speed data of the time period to be predicted.

[0033] Among them, the target wind speed prediction model is a trained initial wind speed prediction model, and the initial wind speed prediction model is a residual channel attention network model.

[0034] Among them, the training process of the target wind speed prediction model can be: obtaining historical predicted wind speed data and historical actual wind speed data corresponding to the historical predicted wind speed data; generating a first training sample set based on the historical predicted wind speed data and the historical actual wind speed data corresponding to the historical predicted wind speed data, wherein the first training sample set includes: historical predicted wind speed data and historical actual wind speed data corresponding to the historical predicted wind speed data; inputting the historical predicted wind speed data in the first training sample set into the residual channel attention network model to obtain the first predicted wind speed data; according to the difference between the first predicted wind speed data and the historical actual wind speed data corresponding to the historical predicted wind speed data, training the parameters of the residual channel attention network model to obtain the target wind speed prediction model.

[0035] Specifically, the predicted wind speed data of the time period to be predicted is input into the target wind speed prediction model to obtain the actual wind speed data of the time period to be predicted, which can be: input the predicted wind speed data of the time period to be predicted into the target wind speed prediction model to obtain the predicted wind speed data of the time period to be predicted, and correct the predicted wind speed data of the time period to be predicted based on the error sequence to obtain the actual wind speed data of the time period to be predicted. The predicted wind speed data of the time period to be predicted is input into the target wind speed prediction model to obtain the actual wind speed data of the time period to be predicted, which can also be: decompose the predicted wind speed data of the time period to be predicted using the empirical mode decomposition algorithm to obtain multiple predicted wind speed subsequences of the time period to be predicted; input the multiple predicted wind speed subsequences of the time period to be predicted into the target wind speed prediction model in sequence to obtain the first wind speed data corresponding to each predicted wind speed subsequence of the time period to be predicted, and determine the actual wind speed data of the time period to be predicted according to the first wind speed data corresponding to each predicted wind speed subsequence of the time period to be predicted. The predicted wind speed data of the time period to be predicted is input into the target wind speed prediction model to obtain the actual wind speed data of the time period to be predicted. The method can also be as follows: the predicted wind speed data of the time period to be predicted is decomposed by using the empirical mode decomposition algorithm to obtain multiple predicted wind speed subsequences of the time period to be predicted; the multiple predicted wind speed subsequences of the time period to be predicted are input into the target wind speed prediction model in sequence to obtain the first wind speed data corresponding to each predicted wind speed subsequence of the time period to be predicted; based on the error sequence, the first wind speed data corresponding to each predicted wind speed subsequence of the time period to be predicted is corrected to obtain the second wind speed data corresponding to each predicted wind speed subsequence of the time period to be predicted; the sum of the second wind speed data corresponding to the multiple predicted wind speed subsequences of the time period to be predicted is determined as the actual wind speed data of the time period to be predicted.

[0036] Optionally, the target wind speed prediction model is trained in the following manner:

[0037] Obtain historical predicted wind speed data and historical actual wind speed data corresponding to the historical predicted wind speed data;

[0038] Using the empirical mode decomposition algorithm, the historical forecast wind speed data is decomposed to obtain the historical forecast wind speed subsequence, and the historical actual wind speed data is decomposed to obtain the historical actual wind speed subsequence corresponding to the historical forecast wind speed subsequence;

[0039] Generate a first training sample set according to the historical predicted wind speed subsequence and the historical actual wind speed subsequence corresponding to the historical predicted wind speed subsequence;

[0040] Inputting the historical predicted wind speed subsequence in the first training sample set into the residual channel attention network model to obtain a first predicted wind speed subsequence;

[0041] According to the difference between the first predicted wind speed subsequence and the historical actual wind speed subsequence corresponding to the historical predicted wind speed subsequence, the parameters of the residual channel attention network model are trained to obtain a target wind speed prediction model.

[0042] The historical predicted wind speed data is the weather forecast wind speed data corresponding to the historical time period, and the historical actual wind speed data is the actual wind speed data corresponding to the historical time period. The length of the historical time period is the same as the length of the time period to be predicted. Alternatively, the length of the time period to be predicted is an integer multiple of the length of the historical time period.

[0043] In a specific example, the historical predicted wind speed data may be the weather forecast wind speed data for each day of the previous month. The historical actual wind speed data corresponding to the historical predicted wind speed data may be the actual wind speed data for each day of the previous month.

[0044] In another specific example, the predicted wind speed data for the time period to be predicted is input into the target wind speed prediction model to obtain the actual wind speed data for the time period to be predicted as follows:

[0045] Construct a weather forecast wind speed sequence and a historical actual wind speed sequence, decompose the weather forecast wind speed sequence and the historical actual wind speed sequence using the empirical mode decomposition algorithm to obtain subsequences with different frequencies, and divide the decomposed subsequences into a training set and a test set;

[0046] The subsequences in the training set are divided into a weather forecast wind speed sequence training set and a historical actual wind speed sequence training set. The weather forecast wind speed sequence training set is used to train the residual channel attention network model, and the trained residual channel attention network model is used to predict the data in the historical actual wind speed sequence training set.

[0047] Subtract the predicted result from the previous step from the actual value in the historical actual wind speed sequence training set to obtain the corresponding error sequence, and bring the obtained error sequence into the residual channel attention network model for training;

[0048] The trained residual channel attention network model is used to predict the wind speed data in the weather forecast wind speed sequence test set, and the residual channel attention network model trained by the error sequence is used to predict the error generated by the preliminary prediction;

[0049] The obtained error prediction sequence is used to correct the initial prediction results of the weather forecast wind speed sequence test set to obtain the final prediction results of each IMF component and residual sequence. Then the final prediction values ​​of all IMF components and residual sequences are added together to obtain the total wind speed prediction value, which is the corrected wind speed data for the time period to be predicted.

[0050] Specifically, the method of using the empirical mode decomposition algorithm to decompose the historical forecast wind speed data to obtain the historical forecast wind speed subsequences can be: using the empirical mode decomposition algorithm to decompose the forecast wind speed data of each historical time period to obtain multiple historical forecast wind speed subsequences. For example, the weather forecast wind speed data of each day of the previous month can be decomposed using the empirical mode decomposition algorithm to obtain multiple historical forecast wind speed subsequences corresponding to each day of the previous month.

[0051] Specifically, the method of using the empirical mode decomposition algorithm to decompose the historical actual wind speed data to obtain the historical actual wind speed subsequence corresponding to the historical predicted wind speed subsequence can be: using the empirical mode decomposition algorithm to decompose the actual wind speed data of each historical time period to obtain multiple historical actual wind speed subsequences. For example, the empirical mode decomposition algorithm can be used to decompose the actual wind speed data of each day of the previous month to obtain multiple historical actual wind speed subsequences corresponding to each day of the previous month.

[0052] Specifically, the historical predicted wind speed subsequence in the first training sample set is input into the residual channel attention network model to obtain the first predicted wind speed subsequence: multiple historical predicted wind speed subsequences for each historical time period in the first training sample set are input into the residual channel attention network model in sequence to obtain the first predicted wind speed subsequence corresponding to each historical predicted wind speed subsequence.

[0053] Optionally, the predicted wind speed data for the time period to be predicted is input into a target wind speed prediction model to obtain actual wind speed data for the time period to be predicted, including:

[0054] Decomposing the predicted wind speed data for the time period to be predicted by using an empirical mode decomposition algorithm to obtain a plurality of predicted wind speed subsequences for the time period to be predicted;

[0055] Inputting the plurality of predicted wind speed subsequences of the time period to be predicted into the target wind speed prediction model in sequence, and obtaining first wind speed data corresponding to each predicted wind speed subsequence of the time period to be predicted;

[0056] Based on the error sequence, correcting the first wind speed data corresponding to each predicted wind speed subsequence in the time period to be predicted, to obtain the second wind speed data corresponding to each predicted wind speed subsequence in the time period to be predicted;

[0057] The sum of the second wind speed data corresponding to the multiple predicted wind speed subsequences in the time period to be predicted is determined as the actual wind speed data in the time period to be predicted.

[0058] Optionally, before correcting the first wind speed data corresponding to each predicted wind speed subsequence of the time period to be predicted based on the error sequence to obtain the second wind speed data corresponding to each predicted wind speed subsequence of the time period to be predicted, the method further includes:

[0059] Acquire a first test sample set, wherein the first test sample set includes: a historical predicted wind speed subsequence and a historical actual wind speed subsequence corresponding to the historical predicted wind speed subsequence;

[0060] Input the historical predicted wind speed subsequence in the test sample set into the target wind speed prediction model to obtain the predicted wind speed subsequence corresponding to the historical predicted wind speed subsequence;

[0061] The error sequence is determined according to the difference between the predicted wind speed subsequence corresponding to the historical predicted wind speed subsequence in the test sample set and the historical actual wind speed subsequence corresponding to the historical predicted wind speed subsequence.

[0062] S130, inputting the actual wind speed data of the time period to be predicted into a target wind power generation prediction model to obtain the wind power generation corresponding to the time period to be predicted.

[0063] Among them, the target wind power generation prediction model is a trained convolutional neural network model.

[0064] In an embodiment of the present invention, the training process of the target wind power generation prediction model may be: obtaining the historical wind power generation and historical actual wind speed data of the wind farm; generating a target sample set according to the historical wind power generation and historical actual wind speed data of the wind farm, wherein the target sample set includes: the historical actual wind speed data and historical wind power generation corresponding to the sample time period; inputting the historical actual wind speed data corresponding to the sample time period into the initial wind power generation prediction model to obtain the predicted wind power generation; training the initial wind power generation prediction model according to the difference between the predicted wind power generation and the historical wind power generation corresponding to the sample time period to obtain the target wind power generation prediction model. The sample time period may include: the first time period and a related time period of the first time period; or a related time period of the first time period.

[0065] It should be noted that if the sample time period includes: a first time period and a time period related to the first time period, the training process of the target wind power generation prediction model can be: inputting the historical actual wind speed data corresponding to the first time period into the initial wind power generation prediction model to obtain a first predicted wind power generation, and inputting the historical actual wind speed data corresponding to the related time period of the first time period into the initial wind power generation prediction model to obtain a second predicted wind power generation; based on the difference between the first predicted wind power generation and the historical wind power generation corresponding to the first time period, and the difference between the second predicted wind power generation and the historical wind power generation corresponding to the related time period of the first time period, train the initial wind power generation prediction model to obtain the target wind power generation prediction model.

[0066] In an embodiment of the present invention, before inputting the historical actual wind speed data corresponding to the first time period or the historical actual wind speed data corresponding to the time period related to the first time period into the initial wind power prediction model, the historical actual wind speed data corresponding to the first time period and the historical actual wind speed data corresponding to the time period related to the first time period can be decomposed in advance using an empirical mode decomposition algorithm to obtain a historical actual wind speed subsequence.

[0067] It should be noted that the embodiment of the present invention decomposes the wind speed time series through the empirical mode decomposition algorithm to generate a series of inherent mode components of different scales, which can reduce the non-stationarity of the wind speed series, and then uses the back propagation algorithm to adjust the model parameters. Finally, the predicted values ​​of each component are superimposed to obtain the ultra-short-term wind speed prediction value, which can better predict the change law of wind speed and improve the accuracy of wind power generation prediction.

[0068] It should be noted that wind speed is the most important factor affecting wind power output. Currently, most wind power prediction models use weather forecast data and wind farm original operation data for modeling. Therefore, the accuracy of weather forecast wind speed will greatly affect the accuracy of wind power prediction.

[0069] However, the forecast error of weather forecast and the difference in spatial distance between the weather forecast point and the wind farm can easily lead to errors between the wind speed data in the weather forecast and the wind speed data actually measured in the wind farm. Therefore, after data preprocessing, it is also necessary to analyze the error between the wind speed data in the weather forecast and the actual wind speed, so that the wind speed data can be corrected when the weather forecast data is used to predict wind power generation.

[0070] In the embodiment of the present invention, by collecting the historical power generation data, historical actual wind speed data and predicted wind speed data of the time period to be predicted of the wind farm; correcting the predicted wind speed data of the time period to be predicted to obtain the corrected wind speed data of the time period to be predicted; determining the target similar time period of the first time period and the adjacent similar time period of the first time period based on the historical wind power generation and historical actual wind speed data of the wind farm; constructing a training data set and a test sample set, using the training data set to train the wind power generation prediction model; inputting the corrected wind speed data of the time period to be predicted into the target wind power generation prediction model to obtain the predicted value of the wind power generation of the time period to be predicted. The gap between the predicted value and the true value can be effectively narrowed, more accurate wind speed prediction data can be obtained, and the prediction accuracy of the subsequent wind power generation function can be improved.

[0071] The embodiment of the present invention applies the residual channel attention network model to perform error sequence prediction on the wind speed prediction results, which can effectively correct the error deviation of the primary wind speed prediction value and improve the prediction accuracy of the model.

[0072] The wind speed prediction model based on data preprocessing and error correction proposed in the embodiment of the present invention can better handle the random characteristics of the wind speed sequence. Combined with the prediction of the error sequence and the correction of the initial prediction, it can effectively narrow the gap between the predicted value and the true value, obtain more accurate wind speed prediction data, and thus improve the prediction accuracy of subsequent wind power generation functions.

[0073] Optionally, the target wind power generation prediction model is trained in the following manner:

[0074] Obtain historical wind power generation and historical actual wind speed data of wind farms;

[0075] Generate a target sample set according to the historical wind power generation and historical actual wind speed data of the wind farm, wherein the target sample set includes: the historical actual wind speed data and the historical wind power generation corresponding to the sample time period;

[0076] Inputting the historical actual wind speed data corresponding to the sample time period into the initial wind power prediction model to obtain the predicted wind power;

[0077] According to the difference between the predicted wind power generation and the historical wind power generation corresponding to the sample time period, the initial wind power generation prediction model is trained to obtain a target wind power generation prediction model.

[0078] Specifically, according to the historical wind power generation and historical actual wind speed data of the wind farm, the method of generating the target sample set can be: cleaning the historical wind power generation and historical actual wind speed data of the wind farm, determining the historical actual wind speed data and historical wind power generation corresponding to the sample time period according to the cleaned historical wind power generation and historical actual wind speed data, and generating the target sample set according to the historical actual wind speed data and historical wind power generation corresponding to the sample time period.

[0079] Optionally, the sample time period includes: a time period related to the first time period and the first time period.

[0080] The historical actual wind speed data corresponding to the sample time period is input into the initial wind power prediction model to obtain the predicted wind power, including:

[0081] Acquire historical actual wind speed data and historical wind power generation corresponding to the first time period, historical actual wind speed data and historical wind power generation corresponding to a target similar time period of the first time period, and historical actual wind speed data and historical wind power generation corresponding to an adjacent similar time period of the first time period;

[0082] Decomposing the historical actual wind speed data corresponding to the first time period using an empirical mode decomposition algorithm to obtain a plurality of historical actual wind speed subsequences corresponding to the first time period;

[0083] Decomposing the historical actual wind speed data corresponding to the target similar time period of the first time period by using an empirical mode decomposition algorithm to obtain a plurality of historical actual wind speed subsequences corresponding to the target similar time period of the first time period;

[0084] Decomposing the historical actual wind speed data corresponding to the adjacent similar time periods of the first time period by using an empirical mode decomposition algorithm to obtain a plurality of historical actual wind speed subsequences corresponding to the adjacent similar time periods of the first time period;

[0085] A second training sample set is generated according to the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the first time period, the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the target similar time period of the first time period, and the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the adjacent similar time period of the first time period, wherein the second training sample set includes: a historical actual wind speed subsequence training set corresponding to the first time period, a historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period, and a historical actual wind speed subsequence training set corresponding to the adjacent similar time period of the first time period, the historical actual wind speed subsequence training set corresponding to the first time period includes: the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the first time period; the historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period includes: the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the target similar time period of the first time period; the historical actual wind speed subsequence training set corresponding to the adjacent similar time period of the first time period includes: the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the adjacent similar time period of the first time period;

[0086] Inputting the historical actual wind speed subsequence corresponding to the first time period in the historical actual wind speed subsequence training set corresponding to the first time period into the convolutional neural network model to obtain the predicted wind power generation corresponding to the first time period;

[0087] Inputting the historical actual wind speed subsequence corresponding to the target similar time period of the first time period in the historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period into the convolutional neural network model to obtain the predicted wind power corresponding to the target similar time period of the first time period;

[0088] Inputting the historical actual wind speed subsequence corresponding to the adjacent similar time period of the first time period in the historical actual wind speed subsequence training set corresponding to the adjacent similar time period of the first time period into the convolutional neural network model to obtain the predicted wind power corresponding to the adjacent similar time period of the first time period;

[0089] Accordingly, according to the difference between the predicted wind power generation and the historical wind power generation corresponding to the sample time period, the initial wind power generation prediction model is trained to obtain a target wind power generation prediction model, including:

[0090] According to the difference between the predicted wind power generation and the historical wind power generation corresponding to the first time period, the difference between the predicted wind power generation and the historical wind power generation corresponding to the target similar time period of the first time period, and the difference between the predicted wind power generation and the historical wind power generation corresponding to the adjacent similar time period of the first time period, the convolutional neural network model is trained to obtain a target wind power generation prediction model.

[0091] Among them, the difference between the predicted wind power generation power and the historical wind power generation power corresponding to the first time period is the difference between the predicted wind power generation power corresponding to the first time period and the historical wind power generation power corresponding to the first time period. The difference between the predicted wind power generation power and the historical wind power generation power corresponding to the adjacent similar time periods of the first time period is the difference between the predicted wind power generation power corresponding to the adjacent similar time periods of the first time period and the historical wind power generation power corresponding to the adjacent similar time periods of the first time period. The difference between the predicted wind power generation power and the historical wind power generation power corresponding to the adjacent similar time periods of the first time period is the difference between the predicted wind power generation power corresponding to the adjacent similar time periods of the first time period and the historical wind power generation power corresponding to the adjacent similar time periods of the first time period.

[0092] In some embodiments, a specific method for determining a target similar time period close to the first time period and searching for an adjacent similar time period in data of several days before the first time period is as follows:

[0093] Based on the statistical characteristics of historical wind power generation, a clustering algorithm is used to cluster historical wind power generation, and combined with historical weather types, multiple clusters are obtained, each cluster includes: multiple samples, each sample includes: historical actual wind speed samples and wind power generation corresponding to historical actual wind speed samples. The clustering results are divided into different clustering types according to weather types;

[0094] The statistical features may include at least one of: the maximum power generation in the first time period, the total power generation in the first time period, the average power generation in the first time period, the variance of the power generation in the first time period, and the standard deviation of the power generation in the first time period. For example, the statistical features may include: the maximum power generation in a day, the total power generation in a day, the average power generation in a day, the variance of the power generation in a day, and the standard deviation of the power generation in a day;

[0095] According to the meteorological characteristic value of each sample in each cluster, the center point position of each cluster is determined, and the distance between the first time period and each center point is calculated by using the Euclidean distance, so as to determine the weather type to which the first time period belongs; the meteorological characteristic value may be at least one of temperature, humidity and atmospheric pressure;

[0096] By calculating the Euclidean distance between the first time period and the meteorological characteristic values ​​of each center point, the set corresponding to the minimum Euclidean distance is the weather type to which the first time period belongs;

[0097] Find a target similar time period close to the first time period from the same weather type, and find an adjacent similar time period in the data of several days before the first time period.

[0098] It should be noted that the above-mentioned target similar time period is determined in the following manner, namely, taking the meteorological characteristic values ​​of the first time period (sample day, for example, July 2, 2024) as a reference sequence, taking the meteorological characteristic values ​​of historical days (for example, January 2, 2024-July 1, 2024) as a subsequence, and normalizing the meteorological characteristic values ​​of the first time period with the meteorological characteristic values ​​of the historical days; then, calculating the characteristic correlation coefficient of each meteorological parameter of the first time period and the historical days; then, calculating the mean of the characteristic correlation coefficients of each historical day with the first time period, wherein the mean of the characteristic correlation coefficients is the average value after summing the characteristic correlation coefficients of each meteorological parameter; finally, sorting these means from large to small, and selecting the historical day with the smallest difference in wind force level with the first time period from the historical days corresponding to these means as the target similar time period, that is, the best similar day.

[0099] The above-mentioned search for the adjacent similar time period (that is, the adjacent similar day) in the data of the first several days of the first time period can be seven days, ten days or fifteen days, and can be specifically set in combination with the local weather change law.

[0100] It should be noted that the adjacent similar days mentioned above are determined in the following manner, namely, taking the meteorological characteristic values ​​of the first time period as the reference sequence, taking the meteorological characteristic values ​​of the previous several days as the subsequence, and normalizing the meteorological characteristic values ​​of the first time period with the meteorological characteristic values ​​of the previous several days; then, calculating the characteristic correlation coefficient of each meteorological parameter of the first time period and the previous several days; then, calculating the mean of the characteristic correlation coefficients of the previous several days with the first time period, wherein the mean of the characteristic correlation coefficients is the average value after summing the characteristic correlation coefficients of each meteorological parameter; finally, sorting these means from large to small, and selecting the historical day with the smallest difference in wind force level with the first time period from the previous several days corresponding to these means as the adjacent similar day.

[0101] Optionally, the sample time period includes: a time period related to the first time period.

[0102] Optionally, the related time periods of the first time period include: a target similar time period of the first time period and an adjacent similar time period of the first time period;

[0103] The historical actual wind speed data corresponding to the sample time period is input into the initial wind power prediction model to obtain the predicted wind power, including:

[0104] Acquire historical actual wind speed data and historical wind power generation corresponding to a target similar time period of the first time period, and historical actual wind speed data and historical wind power generation corresponding to a similar time period adjacent to the first time period;

[0105] Decomposing the historical actual wind speed data corresponding to the target similar time period of the first time period by using an empirical mode decomposition algorithm to obtain a plurality of historical actual wind speed subsequences corresponding to the target similar time period of the first time period;

[0106] Decomposing the historical actual wind speed data corresponding to the adjacent similar time periods of the first time period by using an empirical mode decomposition algorithm to obtain a plurality of historical actual wind speed subsequences corresponding to the adjacent similar time periods of the first time period;

[0107] Generate a second training sample set according to the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the target similar time period of the first time period, and the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the adjacent similar time period of the first time period, wherein the second training sample set includes: a historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period and a historical actual wind speed subsequence training set corresponding to the adjacent similar time period of the first time period;

[0108] Inputting the historical actual wind speed subsequence corresponding to the target similar time period of the first time period in the historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period into the convolutional neural network model to obtain the predicted wind power corresponding to the target similar time period of the first time period;

[0109] Inputting the historical actual wind speed subsequence corresponding to the adjacent similar time period of the first time period in the historical actual wind speed subsequence training set corresponding to the adjacent similar time period of the first time period into the convolutional neural network model to obtain the predicted wind power corresponding to the adjacent similar time period of the first time period;

[0110] Accordingly, according to the difference between the predicted wind power generation and the historical wind power generation corresponding to the sample time period, the initial wind power generation prediction model is trained to obtain a target wind power generation prediction model, including:

[0111] The convolutional neural network model is trained according to the difference between the predicted wind power generation and the historical wind power generation corresponding to the target similar time period of the first time period, and the difference between the predicted wind power generation and the historical wind power generation corresponding to the adjacent similar time period of the first time period to obtain a target wind power generation prediction model.

[0112] Among them, the historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period includes: the historical wind power generation power and multiple historical actual wind speed subsequences corresponding to the target similar time period of the first time period; the historical actual wind speed subsequence training set corresponding to the adjacent similar time period of the first time period includes: the historical wind power generation power and multiple historical actual wind speed subsequences corresponding to the adjacent similar time period of the first time period.

[0113] Specifically, the historical actual wind speed subsequence corresponding to the target similar time period of the first time period in the historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period is input into the convolutional neural network model to obtain the predicted wind power generation corresponding to the target similar time period of the first time period. The method can be: multiple historical actual wind speed subsequences corresponding to the target similar time period of the first time period in the historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period are input into the convolutional neural network model in sequence to obtain the predicted wind power generation corresponding to each historical actual wind speed subsequence, and the predicted wind power generation corresponding to the target similar time period of the first time period is determined according to the sum of the predicted wind power generation corresponding to each historical actual wind speed subsequence.

[0114] In the embodiment of the present invention, the method for determining the predicted wind power generation corresponding to the adjacent similar time period of the first time period is similar to the method for determining the predicted wind power generation corresponding to the target similar time period of the first time period, which will not be repeated here.

[0115] It is understandable that when predicting wind power generation, the quality of the data used will affect the effect and accuracy of subsequent predictions. Under the same prediction model, real and reliable data will greatly improve the prediction accuracy. However, most of these historical power generation data are obtained through system collection. Abnormal data generated during the data collection process, such as data collection errors of the data collection system, communication system transmission errors, communication interruptions, and wind turbine system failures, lead to abnormal data points in the monitoring data of the monitoring system; and the volatility of wind farm output will lead to absorption problems. When the power grid cannot effectively absorb the electricity generated by wind power, it needs to abandon wind and limit power, which will cause abnormal operation data of the wind farm.

[0116] Therefore, if the wind power generation prediction model is expected to achieve higher prediction accuracy, it is necessary to first clean the historical data (historical wind power generation and historical actual wind speed data) extracted from the wind power system, and then use the cleaned data to carry out the learning, training and prediction process of the wind power generation prediction model.

[0117] In order to clean the historical data, the embodiment of the present invention sets the following abnormal data processing method to pre-process the historical power generation data:

[0118] First, the missing values ​​in the wind farm operation data are supplemented, and the supplemented operation data are normalized; then, the abnormal data points in the wind farm operation data are identified, the abnormal data are marked, and the identified abnormal values ​​are removed.

[0119] It should be noted that the missing values ​​in the operation data of the wind farm may be supplemented by an interpolation method or a polynomial fitting method.

[0120] In some embodiments, the method for identifying abnormal data points in wind farm operation data (that is, historical wind power generation and historical actual wind speed data) specifically includes the following steps:

[0121] Filter the power-wind speed data set from the wind farm operation data, and set the identification radius and the number of identification data points;

[0122] Select any data point (power-wind speed data pair) from the power-wind speed data set. If the number of data points contained in the range with the data point as the center and the radius as the identification radius is greater than the number of identification data points, then determine the data point as the core point; if the data point is the core point, create a new data cluster and add the data point to the new data cluster;

[0123] The data set with the core point as the center and the radius equal to the recognition radius is traversed, and all data points whose distance from the core point is less than the recognition radius are added to the data cluster of the core point;

[0124] All data points in the power-wind speed data set are judged in turn, and the data points that are not added to any data cluster are noise points.

[0125] In some embodiments, the target wind power generation prediction model is trained in the following manner:

[0126] Construct the wind speed sequence of the best similar day and the wind speed sequence of the adjacent similar day, use the empirical mode decomposition algorithm to decompose the wind speed sequence of the best similar day and the wind speed sequence of the adjacent similar day to obtain subsequences with different frequencies, and divide the decomposed subsequences into a training set and a test set;

[0127] The subsequences in the training set are divided into the best similar day wind speed sequence training set and the adjacent similar day wind speed sequence training set. The convolutional neural network model is trained using the best similar day wind speed sequence training set and the adjacent similar day wind speed sequence training set respectively, and the model parameters are adjusted using the back propagation algorithm to obtain the first wind power prediction model.

[0128] The subsequences of the test set are divided into a test set of the best similar day wind speed sequence and a test set of the adjacent similar day wind speed sequence, and the test set of the best similar day wind speed sequence and the test set of the adjacent similar day wind speed sequence are respectively input into the first wind power prediction model to obtain the prediction result of each IMF component, and then all IMF components are added together to obtain the predicted wind power of the best similar day and the predicted wind power of the adjacent similar day;

[0129] The prediction error is calculated based on the predicted wind power generation power on the best similar day, the predicted wind power generation power on the adjacent similar days, the actual wind power generation power on the best similar day and the actual wind power generation power on the adjacent similar days. When the prediction error is less than a threshold, the target wind power generation model is obtained.

[0130] It should be noted that the embodiment of the present invention decomposes the wind speed time series through the empirical mode decomposition algorithm to generate a series of inherent mode components of different scales, which can reduce the non-stationarity of the wind speed series, and then uses the back propagation algorithm to adjust the model parameters. Finally, the predicted values ​​of each component are superimposed to obtain the ultra-short-term wind speed prediction value, which can better predict the change law of wind speed and improve the accuracy of wind power generation prediction.

[0131] The technical solution of this embodiment is to obtain the predicted wind speed data for the time period to be predicted; input the predicted wind speed data for the time period to be predicted into the target wind speed prediction model to obtain the actual wind speed data for the time period to be predicted, wherein the target wind speed prediction model is a trained initial wind speed prediction model, and the initial wind speed prediction model is a residual channel attention network model; input the actual wind speed data for the time period to be predicted into the target wind power prediction model to obtain the wind power corresponding to the time period to be predicted, wherein the target wind power prediction model is a trained convolutional neural network model, which can improve the accuracy of wind power prediction.

[0132] Embodiment 2

[0133] Figure 2 The present invention provides a schematic diagram of the structure of a wind power generation prediction device. This embodiment is applicable to the case of wind power generation prediction. The device can be implemented in software and / or hardware. The device can be integrated in any device that provides wind power generation prediction function, such as Figure 2 As shown, the wind power generation power prediction device specifically includes: a predicted wind speed data acquisition module 210, an actual wind speed data determination module 220 and a wind power generation power determination module 230.

[0134] Among them, the predicted wind speed data acquisition module is used to obtain the predicted wind speed data for the time period to be predicted;

[0135] An actual wind speed data determination module is used to input the predicted wind speed data of the time period to be predicted into a target wind speed prediction model to obtain the actual wind speed data of the time period to be predicted, wherein the target wind speed prediction model is a trained initial wind speed prediction model, and the initial wind speed prediction model is a residual channel attention network model;

[0136] The wind power generation determination module is used to input the actual wind speed data of the time period to be predicted into the target wind power generation prediction model to obtain the wind power generation corresponding to the time period to be predicted, wherein the target wind power generation prediction model is a trained convolutional neural network model.

[0137] The above-mentioned product can execute the method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0138] Embodiment 3

[0139] Figure 3 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0140] like Figure 3 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0141] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0142] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a wind power generation power prediction method.

[0143] In some embodiments, the wind power generation power prediction method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the wind power generation power prediction method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the wind power generation power prediction method in any other appropriate manner (e.g., by means of firmware).

[0144] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0145] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0146] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0147] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0148] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0149] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0150] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0151] An embodiment of the present invention further provides a computer program product, including a computer program, wherein when the computer program is executed by a processor, the method for predicting wind power generation according to any embodiment of the present invention is implemented.

[0152] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0153] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting wind power generation, characterized in that: include: Obtain the predicted wind speed data for the time period to be predicted; Inputting the predicted wind speed data of the time period to be predicted into the target wind speed prediction model to obtain the actual wind speed data of the time period to be predicted, wherein the target wind speed prediction model is a trained initial wind speed prediction model, and the initial wind speed prediction model is a residual channel attention network model; The actual wind speed data of the time period to be predicted is input into the target wind power prediction model to obtain the wind power corresponding to the time period to be predicted, wherein the target wind power prediction model is a trained convolutional neural network model.

2. The method according to claim 1, characterized in that The training method of the target wind speed prediction model is: Obtain historical predicted wind speed data and historical actual wind speed data corresponding to the historical predicted wind speed data; Using the empirical mode decomposition algorithm, the historical forecast wind speed data is decomposed to obtain the historical forecast wind speed subsequence, and the historical actual wind speed data is decomposed to obtain the historical actual wind speed subsequence corresponding to the historical forecast wind speed subsequence; Generate a first training sample set according to the historical predicted wind speed subsequence and the historical actual wind speed subsequence corresponding to the historical predicted wind speed subsequence; Inputting the historical predicted wind speed subsequence in the first training sample set into the residual channel attention network model to obtain a first predicted wind speed subsequence; According to the difference between the first predicted wind speed subsequence and the historical actual wind speed subsequence corresponding to the historical predicted wind speed subsequence, the parameters of the residual channel attention network model are trained to obtain a target wind speed prediction model.

3. The method according to claim 2, characterized in that Inputting the predicted wind speed data of the time period to be predicted into the target wind speed prediction model to obtain the actual wind speed data of the time period to be predicted, including: Decomposing the predicted wind speed data for the time period to be predicted by using an empirical mode decomposition algorithm to obtain a plurality of predicted wind speed subsequences for the time period to be predicted; Inputting the plurality of predicted wind speed subsequences of the time period to be predicted into the target wind speed prediction model in sequence, and obtaining first wind speed data corresponding to each predicted wind speed subsequence of the time period to be predicted; Based on the error sequence, correcting the first wind speed data corresponding to each predicted wind speed subsequence in the time period to be predicted, to obtain the second wind speed data corresponding to each predicted wind speed subsequence in the time period to be predicted; The sum of the second wind speed data corresponding to the multiple predicted wind speed subsequences in the time period to be predicted is determined as the actual wind speed data in the time period to be predicted.

4. The method according to claim 1, characterized in that The training method of the target wind power generation prediction model is: Obtain historical wind power generation and historical actual wind speed data of wind farms; Generate a target sample set according to the historical wind power generation and historical actual wind speed data of the wind farm, wherein the target sample set includes: the historical actual wind speed data and the historical wind power generation corresponding to the sample time period; Inputting the historical actual wind speed data corresponding to the sample time period into the initial wind power prediction model to obtain the predicted wind power; According to the difference between the predicted wind power generation and the historical wind power generation corresponding to the sample time period, the initial wind power generation prediction model is trained to obtain a target wind power generation prediction model.

5. The method according to claim 4, characterized in that The sample time period includes: a time period related to the first time period.

6. The method according to claim 5, characterized in that The relevant time periods of the first time period include: a target similar time period of the first time period and an adjacent similar time period of the first time period; The historical actual wind speed data corresponding to the sample time period is input into the initial wind power prediction model to obtain the predicted wind power, including: Acquire historical actual wind speed data and historical wind power generation corresponding to a target similar time period of the first time period, and historical actual wind speed data and historical wind power generation corresponding to a similar time period adjacent to the first time period; Decomposing the historical actual wind speed data corresponding to the target similar time period of the first time period by using an empirical mode decomposition algorithm to obtain a plurality of historical actual wind speed subsequences corresponding to the target similar time period of the first time period; Decomposing the historical actual wind speed data corresponding to the adjacent similar time periods of the first time period by using an empirical mode decomposition algorithm to obtain a plurality of historical actual wind speed subsequences corresponding to the adjacent similar time periods of the first time period; Generate a second training sample set according to the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the target similar time period of the first time period, and the historical wind power generation and multiple historical actual wind speed subsequences corresponding to the adjacent similar time period of the first time period, wherein the second training sample set includes: a historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period and a historical actual wind speed subsequence training set corresponding to the adjacent similar time period of the first time period; Inputting the historical actual wind speed subsequence corresponding to the target similar time period of the first time period in the historical actual wind speed subsequence training set corresponding to the target similar time period of the first time period into the convolutional neural network model to obtain the predicted wind power corresponding to the target similar time period of the first time period; Inputting the historical actual wind speed subsequence corresponding to the adjacent similar time period of the first time period in the historical actual wind speed subsequence training set corresponding to the adjacent similar time period of the first time period into the convolutional neural network model to obtain the predicted wind power corresponding to the adjacent similar time period of the first time period; Accordingly, according to the difference between the predicted wind power generation and the historical wind power generation corresponding to the sample time period, the initial wind power generation prediction model is trained to obtain a target wind power generation prediction model, including: The convolutional neural network model is trained according to the difference between the predicted wind power generation and the historical wind power generation corresponding to the target similar time period of the first time period, and the difference between the predicted wind power generation and the historical wind power generation corresponding to the adjacent similar time period of the first time period to obtain a target wind power generation prediction model.

7. A wind power generation power prediction device, characterized in that: include: A predicted wind speed data acquisition module is used to obtain the predicted wind speed data for the time period to be predicted; An actual wind speed data determination module is used to input the predicted wind speed data of the time period to be predicted into a target wind speed prediction model to obtain the actual wind speed data of the time period to be predicted, wherein the target wind speed prediction model is a trained initial wind speed prediction model, and the initial wind speed prediction model is a residual channel attention network model; The wind power generation determination module is used to input the actual wind speed data of the time period to be predicted into the target wind power generation prediction model to obtain the wind power generation corresponding to the time period to be predicted, wherein the target wind power generation prediction model is a trained convolutional neural network model.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the wind power generation power prediction method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the wind power generation power prediction method according to any one of claims 1 to 6 when executed.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the computer program implements the wind power prediction method according to any one of claims 1 to 6.