Wind power prediction method of wind power plant, related equipment and computer program product
By clustering wind turbines within a wind farm into regional wind turbine units and independently modeling each regional wind turbine unit, the problem of low wind power prediction accuracy in existing technologies is solved, achieving higher accuracy wind power prediction.
Patent Information
- Application Number
- CN202511671334.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-14
- Publication Date
- 2026-02-03
AI Technical Summary
Existing wind power prediction technologies model the entire wind farm, resulting in coarse spatial granularity and low accuracy of prediction results, which is difficult to meet the needs of practical applications.
The wind turbines in the wind farm are clustered according to the similarity of their spatial coordinates to form regional wind turbine groups. Each regional wind turbine group is independently modeled, and a deep neural network model is trained using historical average actual power and weather forecast data to predict the average power of the regional wind turbine groups, and finally determine the total power of the wind farm.
It improves the accuracy and applicability of wind power forecasting, making it suitable for practical application scenarios, especially short-term wind power forecasting.
Smart Images
Figure CN121457735A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of wind power prediction, and more particularly to a wind power prediction method for a wind farm, related equipment and a computer program product. BACKGROUND
[0002] Due to the time-varying characteristics of wind resources, wind power has strong randomness, intermittency and uncertainty. With the increasing proportion of wind power in installed capacity year by year, the uncertainty of power system operation increases, and the contradiction between safe operation of power grid and efficient consumption of new energy becomes increasingly prominent. Improving the accuracy of wind power prediction technology has become an effective means to address the above problems.
[0003] The current wind power prediction technology generally takes the actual power of the entire wind farm or a region composed of multiple wind farms as the modeling object. By obtaining the latitude and longitude information of the entire wind farm (or a region composed of multiple wind farms), high-precision weather forecasts are matched, weather forecast data is aligned with historical actual power of the wind farm (or a region composed of multiple wind farms), and then a deep neural network model is used to train and model the meteorological-power data pairs. The input of the model is the meteorological forecast data of the next day, and the output is the power of the entire wind farm (or a region composed of multiple wind farms) on the next day.
[0004] The spatial granularity of the above modeling method is relatively rough, and the prediction result accuracy is not high. SUMMARY
[0005] In view of the above problems, the present application is proposed to provide a wind power prediction method for a wind farm, related equipment and a computer program product, and to improve the wind power prediction accuracy of the wind farm. The specific scheme is as follows:
[0006] In a first aspect, a wind power prediction method for a wind farm is provided, comprising:
[0007] Obtaining meteorological forecast data corresponding to the spatial coordinate positions of each regional wind turbine group in a to-be-predicted period, each regional wind turbine group being obtained by clustering wind turbines in the wind farm according to the similarity of their spatial coordinate positions, and the spatial coordinate position of each regional wind turbine group being the spatial coordinate position of the cluster center of the corresponding cluster;
[0008] Predicting the average power of each regional wind turbine group in the to-be-predicted period by using the power prediction model corresponding to each regional wind turbine group and based on the meteorological forecast data of the to-be-predicted period corresponding to each regional wind turbine group, wherein the power prediction model corresponding to each regional wind turbine group is obtained by separately modeling the historical average actual power and meteorological forecast data of each regional wind turbine group;
[0009] Determine the total power of the wind farm in the to-be-predicted period according to the average power of each regional wind turbine group in the to-be-predicted period, and the number of wind turbines in each group.
[0010] In a possible design, in a further implementation manner of the first aspect of the embodiment of the present application, before the meteorological prediction data corresponding to the spatial coordinate position of each regional wind turbine group in the to-be-predicted period is acquired, the method further includes:
[0011] Acquire the spatial coordinate position of each wind turbine in the wind farm.
[0012] Cluster the wind turbines in the wind farm according to the similarity of the spatial coordinate positions of the wind turbines, to obtain a plurality of clusters, each cluster corresponding to a regional wind turbine group, and the spatial coordinate position corresponding to the cluster center is taken as the spatial coordinate position of the regional wind turbine group.
[0013] In a possible design, in a further implementation manner of the first aspect of the embodiment of the present application, the meteorological prediction data corresponding to each regional wind turbine group in the to-be-predicted period is meteorological factor prediction data given by a plurality of meteorological sources, and the meteorological factors include wind speed and wind direction.
[0014] In a possible design, in a further implementation manner of the first aspect of the embodiment of the present application, the power prediction model corresponding to each regional wind turbine group includes two or more power prediction models corresponding to each meteorological group, and the power prediction model corresponding to each meteorological group is obtained by modeling the historical average actual power of the regional wind turbine group and the meteorological prediction data of each meteorological source in the corresponding meteorological group, and each meteorological source in the same meteorological group has the same wind speed prediction characteristic.
[0015] The prediction process of the average power of each regional wind turbine group in the to-be-predicted period includes:
[0016] For each regional wind turbine group, acquire the meteorological prediction data of each meteorological group corresponding to the regional wind turbine group in the to-be-predicted period, and the power prediction model corresponding to the regional wind turbine group.
[0017] Send the meteorological prediction data of each meteorological group in the to-be-predicted period into the power prediction model corresponding to the meteorological group, to obtain the average power of the regional wind turbine group in the to-be-predicted period output by the power prediction model corresponding to each meteorological group.
[0018] Determine the wind speed interval of each sub-period in the to-be-predicted period according to the meteorological prediction data of each meteorological group in the to-be-predicted period.
[0019] For each sub-period, the prediction correlations of different meteorological groups on the wind speed interval of the sub-period are compared, a target meteorological group with the highest prediction correlation is determined, and the average power of the sub-period corresponding to the output of the power prediction model corresponding to the target meteorological group is selected as the final average power of the sub-period. The final average power of each sub-period is used to form the final average power of the regional wind turbine group in the to-be-predicted period.
[0020] In a possible design, in another implementation manner of the first aspect of the embodiment of the present application, the training process of the power prediction model corresponding to each regional wind turbine group comprises:
[0021] The historical actual power of each wind turbine in each regional wind turbine group and the meteorological forecast data corresponding to the spatial coordinate position of the regional wind turbine group are obtained.
[0022] The historical actual power of each wind turbine in the regional wind turbine group is subjected to outlier processing to obtain corrected historical actual power, and the historical average actual power of the regional wind turbine group is calculated.
[0023] The historical average actual power of the regional wind turbine group and the meteorological forecast data are used to train the power prediction model corresponding to the regional wind turbine group.
[0024] In a possible design, in another implementation manner of the first aspect of the embodiment of the present application, the process of subjecting the historical actual power of each wind turbine in the regional wind turbine group to outlier processing to obtain corrected historical actual power comprises:
[0025] The historical actual data pair of each wind turbine is obtained, and the historical actual data pair comprises historical actual wind speed and historical actual power.
[0026] The historical actual data pairs of each wind turbine in the wind farm are divided into several intervals according to wind speed intervals, and each point in the interval represents a wind speed-power data pair.
[0027] Outlier points with power values exceeding a set range in each interval are determined, and the power values of the outlier points are replaced with the median value of all historical actual powers in the current interval to obtain corrected historical actual power.
[0028] In a possible design, in another implementation manner of the first aspect of the embodiment of the present application, the spatial coordinate position of the wind turbine comprises longitude and latitude position information and machine head altitude information of the wind turbine.
[0029] In a possible design, in a further implementation manner of the first aspect of the embodiment of the present application, the process of determining the total power of the wind farm in the to-be-predicted period based on the average power of each regional fan group in the to-be-predicted period and the number of fans in the group comprises:
[0030] the product of the average power of the regional fan group in the to-be-predicted period and the number of fans in the group as the total power of the regional fan group;
[0031] integrating the total power of each regional fan group in the to-be-predicted period to obtain the total power of the wind farm in the to-be-predicted period.
[0032] In a second aspect, an electronic device is provided, comprising a memory and a processor.
[0033] The memory is configured to store a program.
[0034] The processor is configured to execute the program to implement each step of the wind power method of the wind farm described in any one of the preceding first aspects of the present application.
[0035] In a third aspect, a readable storage medium is provided, having a computer program stored thereon, which, when executed by a processor, implements each step of the wind power method of the wind farm described in any one of the preceding first aspects of the present application.
[0036] In a fourth aspect, a computer program product is provided, comprising a computer program which, when executed by a processor, implements each step of the wind power method of the wind farm described in any one of the preceding first aspects of the present application.
[0037] According to the above technical solution, the wind turbines in the wind farm are clustered according to the similarity of spatial coordinate positions, each cluster corresponds to a regional fan group, and the spatial coordinate positions of the wind turbines in the same regional fan group are similar. Meanwhile, a corresponding power prediction model is pre-modeled for each regional fan group. Compared with modeling the entire wind farm as a unit, the present application models the historical average actual power and meteorological prediction data of each regional fan group separately, with a smaller spatial granularity and higher modeling accuracy. On this basis, the meteorological prediction data corresponding to the spatial coordinate positions of each regional fan group in the to-be-predicted period is obtained, the average power of each regional fan group in the to-be-predicted period is predicted through the corresponding power prediction model of each regional fan group based on the meteorological prediction data of the to-be-predicted period corresponding to each regional fan group, and finally the total power of the wind farm in the to-be-predicted period is determined based on the average power of each regional fan group in the to-be-predicted period and the number of fans in the group. The present application models the wind farm specifically in different regions, improves the power prediction accuracy, and is more suitable for practical application scenarios. Attached Figure Description
[0038] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:
[0039] Figure 1 A schematic diagram of an implementation system architecture for the wind power prediction method for wind farms provided in this application embodiment;
[0040] Figure 2 A schematic flowchart of a wind power prediction method for a wind farm provided in an embodiment of this application;
[0041] Figure 3 A schematic flowchart of another wind farm power prediction method provided in this application embodiment;
[0042] Figure 4 A schematic diagram of a regional wind turbine grouping modeling process provided in an embodiment of this application;
[0043] Figure 5 This is a schematic diagram of a power prediction model structure provided in an embodiment of this application;
[0044] Figure 6 A schematic diagram illustrating the division of historical measured data according to wind speed intervals, as provided in this application embodiment;
[0045] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0046] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0047] Wind power forecasting is mainly divided into four categories: long-term, medium-term, short-term, and ultra-short-term, corresponding to forecast lengths of years, months, days, and hours, respectively. Taking short-term power forecasting technology as an example, existing solutions mainly obtain a specific meteorological-power mapping relationship by jointly modeling daily numerical weather forecasts and historical power data of wind farms, and then predict the daily power generation of wind farms by inputting the same source weather forecast for the next day. Existing solutions have the following drawbacks:
[0048] The model modeling takes the actual power of the entire wind farm as a benchmark. This modeling method has a relatively coarse spatial granularity and requires that the wind turbines in the wind farm are uniformly distributed in a meteorological grid and the altitudes of all wind turbines remain consistent, which is difficult to meet in actual application scenarios, thereby resulting in low prediction accuracy.
[0049] In view of the above defects of the existing system, the present application provides a high-precision power prediction scheme based on regional wind turbine groups (which can be applied to a short-term wind power prediction scenario). The present application clusters the wind turbines in the entire wind farm into multiple regional wind turbine groups by clustering, and then models each regional wind turbine group separately. In the reasoning process, the power of each regional wind turbine group is predicted one by one, and finally the predicted power of the entire wind farm is integrated, which can improve the wind farm power prediction accuracy and be more suitable for actual application scenarios.
[0050] The present application provides a wind farm wind power prediction method, which can be applied to the system architecture as shown in Figure 1 The system can include a terminal 100 and a server 200. The server 200 can include one or more servers Figure 1 In the present application, one server is taken as an example for illustration.
[0051] The terminal 100 or the server 200 can be used alone to execute the wind farm wind power prediction method provided in the present application. In addition, the terminal 100 and the server 200 can also be used cooperatively to execute the wind farm wind power prediction method provided in the present application.
[0052] Next, the product form of the terminal 100 in Figure 1 will be described.
[0053] The terminal 100 in the present application can be a mobile phone, a tablet computer, a notebook computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), etc., and the present application does not make any limitation in this regard.
[0054] The present application provides a wind farm wind power prediction method. This method is taken as an example for illustration when applied to a computer device, which can be specifically the terminal 100 in Figure 1 or a system composed of the terminal 100 and the server 200. Referring to Figure 2 , the wind farm wind power prediction method specifically includes the following steps:
[0055] Step S100, obtaining meteorological forecast data corresponding to the spatial coordinate positions of each regional wind turbine group in a to-be-predicted period.
[0056] wherein each regional wind turbine group is obtained by clustering the wind turbines in the wind farm according to the similarity of the spatial coordinate positions of the wind turbines, and the spatial coordinate position of each regional wind turbine group is the spatial coordinate position of the cluster center of the corresponding cluster.
[0057] The present application can first cluster the wind turbines in the wind farm before predicting the power of the wind farm. Referring to Figure 3 As shown in FIG. 1, in some possible embodiments, a clustering step of the wind turbines in the wind farm can be added before step S100, which specifically includes:
[0058] Step S200, obtaining the spatial coordinate position of each wind turbine in the wind farm.
[0059] Step S210, clustering the wind turbines in the wind farm according to the similarity of the spatial coordinate positions of the wind turbines to obtain a plurality of clusters, each cluster corresponding to a regional wind turbine group, and the spatial coordinate position of the cluster center is taken as the spatial coordinate position of the regional wind turbine group.
[0060] Wherein the spatial coordinate position of the wind turbine can include the latitude and longitude position information and the head altitude information of the wind turbine. Since the wind turbine power generation is time-varying, the power generation power is seriously dependent on the wind speed, wind direction and other meteorological data of the wind turbine head, and obtaining accurate meteorological forecast data becomes an important factor to improve the prediction accuracy of the wind turbine power generation. Therefore, in this step, the head altitude information of the wind turbine is used to form the spatial coordinate position of the wind turbine. Compared with the altitude of the bottom of the wind turbine, the altitude of the head is more accurate, and the meteorological forecast data obtained accordingly can guide the power prediction with higher accuracy.
[0061] When clustering the wind turbines in the wind farm, a variety of clustering algorithms can be used. For example, the K-means clustering algorithm can be used to cluster the wind turbines in the wind farm according to the similarity of the spatial coordinate positions, and each cluster is taken as a regional wind turbine group. The spatial coordinate position of the cluster center can be taken as the spatial coordinate position of the regional wind turbine group. Wherein the spatial coordinate position of the cluster center can be calculated based on the spatial coordinate positions of the wind turbines in the cluster, for example, the average of the spatial coordinate positions of the wind turbines in the cluster. In addition, the spatial coordinate position of the wind turbine at the center position of the cluster can also be taken as the spatial coordinate position of the cluster center.
[0062] By clustering the wind turbines, wind turbines with similar spatial coordinate positions can be grouped into the same regional wind turbine group, and then the regional wind turbine group can be modeled to obtain a power prediction model corresponding to each regional wind turbine group. Compared with modeling the entire wind farm, the granularity of modeling the regional wind turbine group is finer, and the meteorological forecast data corresponding to the same regional wind turbine group is more consistent with the actual meteorological data of each wind turbine in the group.
[0063] On this basis, the meteorological forecast data corresponding to the spatial coordinate position of each regional wind turbine group in the to-be-predicted period is obtained.
[0064] The to-be-predicted period is a future time period for which power prediction is to be performed. For example, in a short-term power prediction scenario, the to-be-predicted period can be one day (or other period) in the future.
[0065] The meteorological forecast data can be meteorological factor prediction data that affects the wind power of the wind turbine, including but not limited to wind speed, wind direction, etc.
[0066] In step S110, the average power of the regional wind turbine group in the to-be-predicted period is predicted based on the meteorological forecast data corresponding to each regional wind turbine group in the to-be-predicted period by using the power prediction model corresponding to each regional wind turbine group.
[0067] The power prediction model corresponding to each regional wind turbine group is obtained by separately modeling the historical average actual power of each regional wind turbine group and the meteorological forecast data.
[0068] Referring to Figure 4 As shown in the figure, the present application separately models different regional wind turbine groups. For each regional wind turbine group, the historical average actual power of the regional wind turbine group and the meteorological forecast data are used for modeling.
[0069] The weather forecast data is aligned with the historical average actual power of the regional wind turbine group, and then a deep neural network model is used to train and model the meteorological-power data pair. For example, in a short-term power prediction scenario, the input of the model is the meteorological forecast data of the next day, and the output is the historical average actual power of the regional wind turbine group on the next day.
[0070] For each regional wind turbine group, the meteorological forecast data corresponding to the to-be-predicted period of the regional wind turbine group is input into the power prediction model corresponding to the regional wind turbine group to obtain the average power of the regional wind turbine group in the to-be-predicted period output by the model.
[0071] In step S120, the total power of the wind farm in the to-be-predicted period is determined based on the average power of each regional wind turbine group in the to-be-predicted period and the number of wind turbines in the group.
[0072] Specifically, the total power of the regional wind turbine group in the to-be-predicted period can be obtained by multiplying the average power of the regional wind turbine group in the to-be-predicted period by the number of wind turbines in the group. The total power of the wind farm in the to-be-predicted period can be obtained by integrating the total power of each regional wind turbine group in the to-be-predicted period.
[0073] The method provided in the embodiments of the present application clusters the wind turbines in the wind farm according to the similarity of the spatial coordinate positions of the wind turbines, and each cluster corresponds to a regional wind turbine group. The spatial coordinate positions of the wind turbines in the same regional wind turbine group are similar. Meanwhile, a corresponding power prediction model is pre-modeled for each regional wind turbine group. Compared with modeling the entire wind farm, the historical average actual power and the meteorological prediction data of each regional wind turbine group are used to model the power prediction model of the regional wind turbine group, so that the modeling is performed at a smaller spatial granularity and the modeling accuracy is higher. On this basis, the meteorological prediction data corresponding to the spatial coordinate positions of each regional wind turbine group in the to-be-predicted period is obtained, the average power of each regional wind turbine group in the to-be-predicted period is predicted by using the corresponding power prediction model of the regional wind turbine group and based on the meteorological prediction data of the regional wind turbine group in the to-be-predicted period, and finally the total power of the wind farm in the to-be-predicted period is determined based on the average power of each regional wind turbine group in the to-be-predicted period and the number of wind turbines in the group. The power prediction model is specifically divided into different regions for the wind farm, the power prediction accuracy is improved, and the model is more suitable for practical application scenarios.
[0074] In some optional embodiments, considering that a single meteorological source is difficult to take into account the prediction accuracy of all wind speeds and wind directions, meteorological prediction errors often occur, which interfere with the modeling effect of the power prediction model. Therefore, multiple meteorological sources can be used in the embodiments. That is, the meteorological prediction data of each regional wind turbine group in the to-be-predicted period is meteorological factor prediction data given by multiple meteorological sources, and the multiple meteorological sources constitute a meteorological group. By integrating the meteorological factor prediction data given by multiple meteorological sources, the power prediction accuracy can be improved.
[0075] In some embodiments of the present application, the power prediction model is described.
[0076] The power prediction model can use various neural network structures. In the embodiments, a neural network model with a CNN_LSTM structure is taken as an example, and reference is made to FIG. 1. Figure 5
[0077] The input of the model is the second-day 15-minute-level multi-source weather factor, the input scale is (96, N), 96 is the total time sequence point number of a day, and N is the number of multi-source weather factors. The multi-source weather factors are sent into the power prediction model in a feature stacking manner, then high-dimensional features containing the multi-source weather factors are obtained through 2d CNN convolution and feature dimension mean calculation, then the high-dimensional features are sent into a bidirectional long short-term memory network BiLSTM to capture time sequence information, and finally, time sequence point prediction is performed through a linear layer FC layer to obtain the second-day 15-minute-level power prediction result.
[0078] In some embodiments of the present application, the training process of the power prediction model corresponding to each regional wind turbine group is further described, combined with Figure 4 as shown in the following figure:
[0079] First, the historical actual output power of each wind turbine in each regional wind turbine group and the weather forecast data corresponding to the spatial coordinate position of the regional wind turbine group are obtained.
[0080] It should be noted that when the multi-weather source scheme is adopted, the weather forecast data given by the weather group composed of multiple weather sources for the spatial coordinate position of the regional wind turbine group can be obtained.
[0081] Further, the historical actual output power of each wind turbine in the regional wind turbine group is subjected to outlier processing to obtain the repaired historical actual output power, and the historical average actual output power of the regional wind turbine group is calculated.
[0082] Finally, the historical average actual output power of the regional wind turbine group and the weather forecast data are used to train the power prediction model corresponding to the regional wind turbine group.
[0083] In the embodiments of the present application, an optional implementation manner of the above outlier processing on the historical actual output power of the wind turbine is introduced.
[0084] For the obtained historical actual output power of each wind turbine in the wind farm, it is disturbed by multiple factors, and the obtained historical actual output power of the wind turbine may have outliers. In order to ensure the accuracy of the training data, the historical actual output power of the wind turbine can be subjected to outlier correction processing in the present embodiment.
[0085] In the present embodiment, the historical actual measurement data pair of each wind turbine can be obtained, the historical actual measurement data pair including historical actual measurement wind speed and historical actual output power. The historical actual measurement wind speed can be obtained through a wind speed measuring device and is accurate.
[0086] Further, the historical actual measurement data pairs of each wind turbine in the wind farm are divided into several intervals according to the wind speed interval, and each point in the interval represents a wind speed-power data pair. Combined with Figure 6As shown, it illustrates a diagram of dividing historical measured data according to wind speed intervals. The abscissa represents wind speed, and the ordinate represents power.
[0087] It should be noted that the fan generally has a wind speed starting threshold, and when the wind speed is lower than the starting threshold, the fan does not rotate and does not generate power. When the wind speed exceeds the starting threshold, the power of the fan gradually increases with the increase of the wind speed.
[0088] Figure 6 In the diagram, there are several outliers in each wind speed interval, which represent power abnormal points.
[0089] In this embodiment, the outliers with power values exceeding the set range in each interval can be determined, and the power values of the outliers are replaced by the median of all historical actual power in the current interval to obtain the corrected historical actual power.
[0090] For example, the IQR outlier rejection method can be used according to the wind speed interval, the points with power values exceeding the range of [Q1-1.5IQR, Q3+1.5IQR] in each wind speed interval are marked as outliers, and the outliers are replaced by the median of all historical actual power in the interval to complete the correction of abnormal points.
[0091] The fan historical actual power abnormal point correction strategy provided in this embodiment can improve the accuracy of the historical actual power of the fan by introducing the historical measured wind speed of the fan, dividing the historical measured data (wind speed-power data pairs) according to the wind speed interval, and then detecting and correcting the historical actual power in each wind speed interval. This can effectively improve the accuracy of the historical actual power of the fan, and provide high-quality training data for subsequent model training.
[0092] In some embodiments of the present application, further considering that the meteorological forecast data (predicted wind speed) of different meteorological sources has different prediction characteristics for the measured wind speed of the fan of the wind farm, for example: the meteorological source a is more accurate in predicting the low wind speed interval (1 m / s~3 m / s) of the measured wind speed, and is less accurate in predicting the medium-high wind speed interval (3 m / s~12 m / s), while the meteorological source c is more accurate in predicting the medium-high wind speed interval, and is less accurate in predicting the low wind speed interval. If two kinds of meteorological sources are fused for modeling, although the modeling short board in the interval with poor prediction of wind speed can be improved to some extent, the upper limit of modeling in the interval with good prediction of wind speed may also be lowered.
[0093] Therefore, in this embodiment, the wind speed prediction characteristics of different meteorological sources for the wind farm are retained, the meteorological sources with the same wind speed prediction characteristics are divided into the same group, and then different meteorological groups are used for modeling of the same area fan group.
[0094] The power prediction model corresponding to each regional wind turbine group includes two or more power prediction models corresponding to each meteorological group. The power prediction model corresponding to each meteorological group is obtained by modeling the historical average actual power of the regional wind turbine group and the meteorological prediction data of each meteorological source in the corresponding meteorological group. Each meteorological source in the same meteorological group has the same wind speed prediction characteristic. The risk prediction characteristic represents the prediction characteristic of the predicted wind speed of the meteorological source to the measured wind speed of the wind turbine of the wind farm.
[0095] The application can statistically analyze the historical predicted wind speed and the measured wind speed of each meteorological source, calculate the prediction correlation of each meteorological source with different wind speed intervals, and the higher the prediction correlation, the higher the prediction accuracy of the meteorological source to the corresponding wind speed interval. On this basis, meteorological sources with the same wind speed prediction characteristic can be screened out to form a meteorological group.
[0096] Table 1 below shows the corresponding relationship of the power prediction model modeled for different meteorological groups and wind turbine groups.
[0097] Table 1
[0098]
[0099] Table 1 shows n meteorological groups, and the meteorological sources in each meteorological group have the same wind speed prediction characteristic, for example, the prediction of the measured wind speed in the low wind speed interval is accurate, or the prediction of the measured wind speed in the high wind speed interval is accurate.
[0100] Taking meteorological group 1 as an example, Table 1 shows that it includes two meteorological sources: meteorological source a and meteorological source b.
[0101] For the jth regional wind turbine group:
[0102] Obtain the meteorological prediction data of the ith meteorological group, and use the meteorological prediction data of the ith meteorological group and the historical average actual power of the jth regional wind turbine group to train the power prediction model j_i.
[0103] Each regional wind turbine group includes n power prediction models corresponding to each meteorological group, that is, each regional wind turbine group corresponds to n power prediction models.
[0104] On this basis, the prediction process of the average power of each regional wind turbine group in the to-be-predicted period includes the following steps:
[0105] S1, for each regional wind turbine group, obtain the meteorological prediction data of each meteorological group corresponding to the to-be-predicted period of the regional wind turbine group, and the power prediction model corresponding to the regional wind turbine group.
[0106] S2, input the meteorological forecast data of each meteorological group of the to-be-predicted period into the power prediction model corresponding to the meteorological group, to obtain the average power of the regional fan group in the to-be-predicted period output by the power prediction model corresponding to each meteorological group.
[0107] Taking the meteorological group i and the regional fan group j as an example:
[0108] The meteorological forecast data of the meteorological group i for the regional fan group j in the to-be-predicted period is input into the power prediction model j_i, to obtain the average power of the regional fan group j in the to-be-predicted period output by the power prediction model j_i.
[0109] S3, determine the wind speed interval of each sub-period in the to-be-predicted period according to the meteorological forecast data of each meteorological group in the to-be-predicted period.
[0110] The to-be-predicted period can be divided into a plurality of sub-periods according to a set unit time length. For example, the to-be-predicted period is one day in the future, which can be divided into 96 sub-periods according to 15 minutes (or other granularity) as a unit.
[0111] Since the wind speed interval is an interval value, although the prediction characteristics of different meteorological groups for different wind speed intervals are different, there is a slight deviation in the prediction accuracy of the wind speed, but the deviation generally does not exceed the wind speed interval. Therefore, the meteorological forecast data of a meteorological group can be randomly selected, and the wind speed interval of each sub-period in the to-be-predicted period can be determined according to the meteorological forecast data.
[0112] In another optional implementation, for each sub-period, the wind speed interval to which the wind speed of the current sub-period belongs in the meteorological forecast data of different meteorological groups is obtained, and the prediction correlation of each meteorological group and the corresponding wind speed interval is determined, and the wind speed interval to which the wind speed of the current sub-period belongs in the meteorological forecast data of the meteorological group with the highest prediction correlation is selected as the final result.
[0113] For example, for the t1-t2 sub-period, the wind speed of the sub-period predicted by the meteorological group 1 belongs to the low wind speed interval, and the wind speed of the sub-period predicted by the meteorological group 2 belongs to the medium-high wind speed interval. It is known that the prediction correlation of each meteorological source in the meteorological group 1 and the low wind speed interval is 90%, and the prediction correlation of each meteorological source in the meteorological group 2 and the medium-high wind speed interval is 60%, so the prediction result of the meteorological group 1 with higher prediction correlation can be selected, that is, the wind speed interval of the t1-t2 sub-period is finally determined as: low wind speed interval.
[0114] S4, for each sub-period, compare the prediction relevance of different meteorological groups on the wind speed interval of the sub-period, determine the target meteorological group with the highest prediction relevance, and select the average power of the sub-period corresponding to the output of the power prediction model corresponding to the target meteorological group as the final average power of the sub-period. The final average power of each sub-period is used to form the final average power of the regional wind turbine group in the to-be-predicted period.
[0115] Taking the regional wind turbine group j and the sub-period t1-t2 as an example:
[0116] The wind speed interval of the period t1-t2 is a low wind speed interval. If the highest prediction relevance of the meteorological group 1-n on the low wind speed interval is the meteorological group n, then the average power P(t1-t2) of the sub-period t1-t2 in the output of the power prediction model j_n corresponding to the meteorological group n can be selected as the final average power of the sub-period t1-t2.
[0117] Similarly, the final average power of each sub-period is obtained, and the final average power of each sub-period is used to form the final average power of the regional wind turbine group in the to-be-predicted period.
[0118] In this embodiment, the output results of the power prediction models corresponding to different meteorological groups are post-processed and fused according to the wind speed relevance. For each sub-period, the output result of the power prediction model corresponding to the meteorological group with the highest prediction relevance is selectively used by comparing the prediction relevance of different meteorological groups on the wind speed interval of the sub-period, thereby improving the accuracy of the finally obtained power prediction result.
[0119] The embodiments of the present application also provide an electronic device. Referring to Figure 7 , a structural schematic diagram suitable for implementing the electronic device in the embodiments of the present application is shown. The electronic device in the embodiments of the present application can include but is not limited to fixed terminals such as mobile phones, tablet computers, desktop computers, notebooks, etc. Figure 7 The electronic device shown is only an example, and should not impose any limitation on the functions and use range of the embodiments of the present application.
[0120] As shown in Figure 7 , the electronic device can include a processing device (such as a central processing unit, a graphics processing unit, etc.) 1, which can perform various appropriate actions and processes according to programs stored in a read-only memory (ROM) 2 or loaded from a storage device 8 to a random access memory (RAM) 3, to implement the wind power method of the wind farm in the aforementioned embodiments of the present application. In the state that the electronic device is powered on, various programs and data required for the operation of the electronic device are also stored in the RAM 3. The processing device 1, the ROM 2, and the RAM 3 are connected to each other through a bus 4. An input / output (I / O) interface 5 is also connected to the bus 4.
[0121] Generally, the following devices can be connected to the I / O interface 5: input devices 6 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, and the like; output devices 7 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, and the like; storage devices 8 including, for example, a memory card, a hard disk, and the like; and communication devices 9. The communication devices 9 can allow the electronic device to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 7 The electronic device is shown with various devices, but it is understood that all of the shown devices need not be implemented or present. More or fewer devices can alternatively be implemented or present.
[0122] The embodiment of the present application further provides a computer program product comprising computer readable instructions, which, when executed on an electronic device, cause the electronic device to implement any of the wind power methods of a wind farm provided by the embodiment of the present application.
[0123] The embodiment of the present application further provides a computer readable storage medium, which carries one or more computer programs, and when the one or more computer programs are executed by an electronic device, the electronic device can implement any of the wind power methods of a wind farm provided by the embodiment of the present application.
[0124] In addition, it should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e. they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. In addition, the connection relationship between the modules in the device embodiment provided by the present application indicates that there is a communication connection between them, which can be implemented as one or more communication buses or signal lines.
[0125] Those skilled in the art can clearly understand, through the description of the foregoing embodiments, that the present application can be implemented by means of software and the necessary universal hardware, and of course can also be implemented by means of special hardware including special integrated circuits, special CPUs, special memories, special components, etc. Generally, functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can be various, such as analog circuits, digital circuits, or special circuits. However, for the present application, software program implementation is a better embodiment. Based on this understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer's floppy disk, U disk, mobile hard disk, ROM, RAM, magnetic disk or optical disk, etc., including a plurality of instructions to make a computer device (which can be a personal computer, training device, or network device, etc.) execute the methods described in various embodiments of the present application.
[0126] In the above embodiments, all or part can be implemented by software, hardware, firmware, or any combination thereof. When implemented by software, it can be implemented in the form of a computer program product in whole or in part.
[0127] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, training device or data center to another through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. integrated with one or more available media. The available media can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid state disk (SSD)), etc.
[0128] The embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The embodiments can be combined as needed, and the same or similar parts refer to each other.
Claims
1. A method for predicting wind power in a wind farm, characterized in that, include: The meteorological forecast data corresponding to the spatial coordinates of wind turbine units in each region during the forecast period is obtained. Each wind turbine unit in the region is obtained by clustering wind turbines in the wind farm according to the similarity of their spatial coordinates. The spatial coordinates of each wind turbine unit in the region are the spatial coordinates of the cluster center of the corresponding cluster. By configuring the power prediction model corresponding to each of the regional wind turbine units, and based on the meteorological forecast data of the forecast period corresponding to each of the regional wind turbine units, the average power of the regional wind turbine units in the forecast period is predicted. The power prediction model corresponding to each of the regional wind turbine units is obtained by modeling the historical average actual power and meteorological forecast data of each of the regional wind turbine units separately. The total power of the wind farm during the predicted period is determined based on the average power of each wind turbine group in the region and the number of wind turbines in the group.
2. The method according to claim 1, characterized in that, Before obtaining the meteorological forecast data corresponding to the spatial coordinates of wind turbine units in each region for the forecast period, the following steps are also included: Obtain the spatial coordinates of each wind turbine within the wind farm; The wind turbines in the wind farm are clustered based on the similarity of their spatial coordinate positions, resulting in several clusters. Each cluster corresponds to a regional wind turbine group, and the spatial coordinate position of the cluster center is used as the spatial coordinate position of the regional wind turbine group.
3. The method according to claim 1, characterized in that, The meteorological forecast data for the forecast period corresponding to each wind turbine unit in the region is meteorological factor forecast data given by multiple meteorological sources, and the meteorological factors include: wind speed and wind direction.
4. The method according to claim 3, characterized in that, The power prediction model corresponding to each regional wind turbine unit includes power prediction models corresponding to two or more meteorological groups. The power prediction model corresponding to each meteorological group is modeled using the historical average actual power generated by the regional wind turbine unit and the meteorological forecast data of each meteorological source within the corresponding meteorological group. Each meteorological source within the same meteorological group has the same wind speed prediction characteristics. The prediction process for the average power of each regional wind turbine unit during the predicted period includes: For each of the aforementioned regional wind turbine units, obtain the meteorological forecast data of each meteorological group corresponding to the forecast period for the regional wind turbine unit, and the respective power prediction models corresponding to the regional wind turbine unit. The meteorological forecast data of each meteorological group in the period to be predicted is sent into the power prediction model corresponding to the meteorological group to obtain the average power of the regional wind turbines in the period to be predicted from the power prediction model corresponding to each meteorological group. Based on the meteorological forecast data of each meteorological group during the period to be predicted, the wind speed range for each sub-period within the period to be predicted is determined. For each sub-period, the predictive correlation of different meteorological groups for the wind speed range of the sub-period is compared, the target meteorological group with the highest predictive correlation is determined, and the average power corresponding to the sub-period in the output of the power prediction model corresponding to the target meteorological group is selected as the final average power of the sub-period. The final average power of the regional wind turbine group in the period to be predicted is composed of the final average power of each sub-period.
5. The method according to claim 1, characterized in that, The training process for the power prediction model corresponding to each of the aforementioned regional wind turbine units includes: Obtain the historical actual power generation of each wind turbine in each of the aforementioned regional wind turbine units, and the meteorological forecast data corresponding to the spatial coordinates of the regional wind turbine units; Outlier processing is performed on the historical actual power of each wind turbine in the area to obtain the corrected historical actual power, and the historical average actual power of the area wind turbine group is calculated. Using the historical average actual power generation of the wind turbines in the region and meteorological forecast data, a power prediction model corresponding to the wind turbines in the region is trained.
6. The method according to claim 5, characterized in that, The process of processing outliers in the historical actual power generation of each wind turbine in the aforementioned area to obtain the corrected historical actual power generation includes: Acquire historical measured data pairs for each wind turbine, the historical measured data pairs including historical measured wind speed and historical measured power; According to the wind speed range, the historical measured data of each wind turbine in the wind farm is divided into several intervals, and each point in the interval represents a wind speed-power data pair. Identify outliers in each interval whose power values exceed a set range, and replace the power values of these outliers with the median of all historical actual power values within the current interval to obtain the corrected historical actual power values.
7. The method according to any one of claims 1-6, characterized in that, The spatial coordinates of the wind turbine include its latitude and longitude and its altitude at the turbine head.
8. The method according to any one of claims 1-6, characterized in that, The process of determining the total power of the wind farm during the forecast period based on the average power of each wind turbine group in the region and the number of wind turbines in the group includes: The total power of the regional wind turbine group is the product of the average power of the regional wind turbine group during the predicted period and the number of wind turbines in the group. The total power of the wind turbines in each region during the predicted period is integrated to obtain the total power of the wind farm during the predicted period.
9. An electronic device, characterized in that, include: Memory and processor; The memory is used to store programs; The processor is configured to execute the program to implement the various steps of the wind power generation method for a wind farm as described in any one of claims 1 to 8.
10. A readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the various steps of the wind power generation method for a wind farm as described in any one of claims 1 to 8.
11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the various steps of the wind power generation method for a wind farm as described in any one of claims 1 to 8.