Wind power prediction method and device, electronic equipment and storage medium

Abstract

The invention is suitable for the technical field of wind power prediction, and provides a wind power prediction method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a first target data set, a second target data set and a third target data set, the data comprise environment data and wind power data of the target wind power plant in a first preset time period, a second preset time period and a third preset time period before the current moment; respectively inputting the first target data set, the second target data set and the third target data set into a first prediction model, a second prediction model and a third prediction model in a trained wind power prediction model to obtain a first prediction result, a second prediction result and a third prediction result; and fusing the first prediction result, the second prediction result and the third prediction result based on the target prediction model in the wind power prediction models to obtain the target wind power prediction result, thereby improving the accuracy of wind power prediction.

Description

Technical Field

[0001] This application belongs to the technical field of wind power prediction, and particularly relates to a wind power prediction method, a prediction device, an electronic device, and a storage medium. Background Art

[0002] In the rapidly developing field of clean energy, wind energy, as a clean and renewable energy source, is increasing in proportion. However, due to the volatility of wind speed, it is difficult to accurately estimate the wind power generation, which not only affects the stable operation of the power grid, but also limits the large-scale grid connection of wind energy, bringing huge challenges to the dispatching of the power system. The traditional power system relies on a relatively stable energy supply, and the instability of wind power requires more accurate prediction technologies to balance supply and demand.

[0003] Currently, the commonly used wind power prediction methods are mainly machine learning model methods, such as neural networks, support vector machines, etc., which have strong non-linear fitting ability and adaptive learning ability and are widely used in wind power prediction. However, when training these machine learning models, the existing solutions fail to fully explore and utilize information such as rich historical meteorological data and historical power generation data, ignoring the potential regular features and mutation features in various historical data, resulting in low accuracy of model training and thus low accuracy of wind power prediction.

[0004] Therefore, how to improve the accuracy of wind power prediction has become an urgent problem to be solved. Summary of the Invention

[0005] The embodiments of this application provide a wind power prediction method, a prediction device, an electronic device, and a storage medium, aiming to improve the accuracy of wind power prediction.

[0006] In a first aspect, an embodiment of the present application provides a wind power prediction method, the method comprising: obtaining a first target data set, a second target data set, and a third target data set, respectively including environmental data and wind power data of a target wind farm within a first preset period, a second preset period, and a third preset period before the current moment, the first preset period being in days, the second preset period being in hours, the third preset period being in minutes, the first preset period being greater than the second preset period, and the second preset period being greater than the third preset period; inputting the first target data set, the second target data set, and the third target data set into a first prediction model, a second prediction model, and a third prediction model in a trained wind power prediction model respectively to obtain a first prediction result, a second prediction result, and a third prediction result, the first prediction result being a prediction result representing the long-term wind power characteristics of the target wind farm, the second prediction result being a prediction result representing the short-term wind power characteristics of the target wind farm, and the third prediction result being a prediction result representing the ultra-short-term wind power characteristics of the target wind farm; fusing the first prediction result, the second prediction result, and the third prediction result based on a target prediction model in the wind power prediction model to obtain a target wind power prediction result, the target wind power prediction result including one or more of a long-term wind power prediction result, a short-term wind power prediction result, and an ultra-short-term wind power prediction result.

[0007] In a possible implementation manner, the fusing the first prediction result, the second prediction result, and the third prediction result based on a target prediction model in the wind power prediction model to obtain a target wind power prediction result includes: obtaining a wind power prediction instruction carrying a target prediction result type, the target prediction result type being used to indicate the type of prediction result included in the target wind power prediction result; in response to the wind power prediction instruction, fusing the first prediction result, the second prediction result, and the third prediction result by using the target prediction model based on the target prediction result type to obtain the target wind power prediction result.

[0008] In a possible implementation, the training process of the wind power prediction model includes: obtaining a first historical data set, a second historical data set, and a third historical data set. The first historical data set includes the environmental data and wind power data corresponding to each month in any past year of the target wind farm. The second historical data set includes the environmental data and wind power data corresponding to each hour in any past day of the target wind farm. The third historical data set includes the environmental data and wind power data corresponding to each minute within any past 30 minutes of the target wind farm. Based on the first historical data set, a seasonal autoregressive integrated moving average model is constructed and trained to obtain the first prediction model. Based on the second historical data set, a vector autoregressive model is constructed and trained to obtain the second prediction model. Based on the third historical data set, a Prophet model is constructed and trained to obtain the third prediction model. Based on the first historical data set, the second historical data set, and the third historical data set, a feedforward neural network model is trained to obtain the target prediction model. Based on the first prediction model, the second prediction model, the third prediction model, and the target prediction model, the wind power prediction model is constructed.

[0009] In a possible implementation, the wind power prediction model includes a first model input module, a second model input module, a third model input model, the first prediction model, the second prediction model, the third prediction model, the target prediction model, and a model output module. The first prediction model is a seasonal autoregressive integrated moving average model, and its input is connected to the first model input module. The second prediction model is a vector autoregressive model, and its input is connected to the second model input module. The third prediction model is a Prophet model, and its input is connected to the third model input module. The target prediction model is a feedforward neural network, and its input is connected to the outputs of the first prediction model, the second prediction model, and the third prediction model. The model output module is connected to the output of the target prediction model.

[0010] In a possible implementation manner, the target prediction model includes any one or more of a first target prediction sub-model, a second target prediction sub-model, and a third target prediction sub-model; training the feedforward neural network model based on the first historical data set, the second historical data set, and the third historical data set to obtain the target prediction model includes: constructing and training a feedforward neural network model based on the first historical data set to obtain the first target prediction sub-model; constructing and training a feedforward neural network model based on the second historical data set to obtain the second target prediction sub-model; constructing and training a feedforward neural network model based on the third historical data set to obtain the third target prediction sub-model; constructing the target prediction model based on the first target prediction sub-model, the second target prediction sub-model, and the third target prediction sub-model.

[0011] In a possible implementation manner, the first target data set includes the environmental data and wind power data corresponding to each day within 30 days before the current moment of the target wind farm, the second target data set includes the environmental data and wind power data corresponding to each hour within 24 hours before the current moment of the target wind farm, and the third target data set includes the environmental data and wind power data corresponding to each minute within 30 minutes before the current moment of the target wind farm; the long-term wind power prediction result includes the wind power prediction values corresponding to each day within the next 30 days of the target wind farm, the short-term wind power prediction result includes the wind power prediction values corresponding to each hour within the next 24 hours of the target wind farm, and the ultra-short-term wind power prediction result includes the wind power prediction values corresponding to each minute within the next 30 minutes of the target wind farm.

[0012] In a possible implementation manner, the environmental data includes any one or more of wind speed, wind direction, air pressure, temperature, and humidity.

[0013] Second aspect, an embodiment of the present application provides a wind power prediction device, which includes: a data acquisition module, configured to acquire a first target data set, a second target data set, and a third target data set, respectively including environmental data and wind power data of a target wind farm within a first preset period, a second preset period, and a third preset period before the current moment. The first preset period is based on days as a node, the second preset period is based on hours as a node, the third preset period is based on minutes as a node, the first preset period is greater than the second preset period, and the second preset period is greater than the third preset period; a first prediction module, configured to input the first target data set, the second target data set, and the third target data set into a first prediction model, a second prediction model, and a third prediction model in a trained wind power prediction model respectively to obtain a first prediction result, a second prediction result, and a third prediction result. The first prediction result is a prediction result representing the long-term wind power characteristics of the target wind farm, the second prediction result is a prediction result representing the short-term wind power characteristics of the target wind farm, and the third prediction result is a prediction result representing the ultra-short-term wind power characteristics of the target wind farm; a second prediction module, configured to fuse the first prediction result, the second prediction result, and the third prediction result based on a target prediction model in the wind power prediction model to obtain a target wind power prediction result, where the target wind power prediction result includes one or more of a long-term wind power prediction result, a short-term wind power prediction result, and an ultra-short-term wind power prediction result.

[0014] Third aspect, an embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method described in the first aspect or any one of its implementation manners is implemented.

[0015] Fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the method described in the first aspect or any one of its implementation manners is implemented.

[0016] Fifth aspect, an embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the method described in the first aspect or any one of its implementation manners are implemented.

[0017] The beneficial effects of the embodiments of the present application compared with the prior art are as follows: The first target data set including the environmental data and wind power data of the target wind farm within the first preset period before the current moment, the second target data set including the environmental data and wind power data of the target wind farm within the second preset period before the current moment, and the third target data set including the environmental data and wind power data of the target wind farm within the third preset period before the current moment are respectively input into the first prediction model, the second prediction model, and the third prediction model in the wind power prediction model to obtain the first prediction result representing the long-term wind power characteristics of the target wind farm, the second prediction result representing the short-term wind power characteristics of the target wind farm, and the third prediction result representing the ultra-short-term wind power characteristics of the target wind farm; then, based on the target prediction model in the wind power prediction model, the first prediction result, the second prediction result, and the third prediction result are fused to obtain the target wind power prediction result, which includes one or more of the long-term wind power prediction result, the short-term wind power prediction result, and the ultra-short-term wind power prediction result, so that the long-term wind power prediction result, the short-term wind power prediction result, and the ultra-short-term wind power prediction result of the obtained target wind farm all incorporate the long-term wind power characteristics, short-term wind power characteristics, and ultra-short-term wind power characteristics of the target wind farm, fully considering various regular characteristics and mutation situations of the target wind farm during the wind power prediction process, improving the accuracy of the wind power prediction result, and thus enhancing the accuracy of the wind power prediction.

[0018] It can be understood that a wind power prediction device, an electronic device, a computer-readable storage medium, and a computer program product provided by the embodiments of the present application have the same beneficial effects as the above-mentioned wind power prediction method, and will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 It is a flowchart of a wind power prediction method provided by an embodiment of the present application;

[0021] Figure 2 It is a structural diagram of a wind power prediction model provided by an embodiment of the present application;

[0022] Figure 3 It is a flowchart of a training method of a wind power prediction model provided by an embodiment of the present application;

[0023] Figure 4 The structural block diagram of a wind power prediction device provided by an embodiment of the present application;

[0024] Figure 5 The structural schematic diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0025] In the following description, for the purpose of illustration rather than limitation, specific details such as specific system architectures, technologies, etc. are presented to thoroughly understand the embodiments of the present application. However, those skilled in the art should clearly understand that the present application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of the present application.

[0026] It should be understood that when used in the specification and claims of the present application, the term "comprising" indicates the presence of the described features, wholes, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.

[0027] It should also be understood that the term "and / or" as used in the specification and claims of the present application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0028] As used in the specification and claims of the present application, the term "if" can be interpreted as "when", "once", "in response to determining", or "in response to detecting" depending on the context. Similarly, the phrase "if determined" or "if [the described condition or event] is detected" can be interpreted as meaning "once determined", "in response to determining", "once [the described condition or event] is detected", or "in response to detecting [the described condition or event]" depending on the context.

[0029] In addition, in the description of the specification and claims of the present application, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0030] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that a specific feature, structure, or characteristic described in connection with that embodiment is included in one or more embodiments of this application. Thus, statements such as "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments", etc. that appear in different places in this specification do not necessarily all refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized. The terms "comprising", "including", "having" and their variants all mean "including but not limited to", unless otherwise specifically emphasized.

[0031] In the rapidly developing field of clean energy, wind energy, as a clean and renewable energy source, is increasing in proportion. However, due to the volatility of wind speed, it is difficult to accurately predict wind power generation, which not only affects the stable operation of the power grid but also limits the large-scale grid connection of wind energy. Therefore, wind power prediction systems have emerged, aiming to accurately predict wind power generation over a period of time in the future through complex meteorological models, big data analysis, and machine learning algorithms. Wind power prediction systems are an advanced management tool designed for the wind power generation industry, and their development background mainly stems from the urgent need of the wind power industry to improve grid stability and economic benefits. With the increase in the proportion of renewable energy, especially the uncertainty of wind power generation (greatly affected by weather conditions), it has brought huge challenges to the dispatching of the power system. Traditional power systems rely on relatively stable energy supplies, while the instability of wind requires more accurate prediction technologies to balance supply and demand.

[0032] Traditional wind power prediction systems have shown various limitations in practical applications, mainly in the following aspects:

[0033] 1) Simplicity of the prediction model: Traditional wind power prediction systems usually adopt linear prediction methods based on physical models, such as time series analysis, regression analysis, etc. The disadvantage of these methods is that they are too simplistic and difficult to capture the complexity and non-linear characteristics of wind speed changes. Especially under extreme weather conditions, the prediction accuracy is greatly reduced.

[0034] 2) Limitations in data processing: Traditional wind power prediction systems are unable to handle big data effectively, failing to fully explore and utilize the potential value of rich meteorological, geographical, and historical power generation data. They often ignore the subtle differences in big data, limiting the fineness of the prediction.

[0035] 3) Weak dynamic adaptability: Facing external factors such as climate change, newly built wind power projects, equipment aging, or maintenance activities, traditional prediction models are difficult to quickly adapt to these changes, resulting in an increase in prediction deviation and requiring frequent manual adjustment and calibration.

[0036] 4) Lack of real-time feedback mechanism: Traditional wind power prediction systems often do not establish an effective real-time data feedback loop, and are unable to update the prediction model in real time to reflect the latest on-site conditions, such as sudden changes in wind speed, equipment failures, etc. This weakens their performance in dealing with emergencies.

[0037] 5) Integration and scalability issues: Traditional wind power prediction systems may lack sufficient openness and modularity in design, making it difficult to deeply integrate with other systems of modern power grids, which limits their application and function expansion in a wider network.

[0038] To solve the above technical problems, this application provides a wind power prediction method, which obtains a first target data set, a second target data set, and a third target data set, respectively including environmental data and wind power data of a target wind farm within a first preset period, a second preset period, and a third preset period before the current moment. The first preset period is based on days as nodes, the second preset period is based on hours as nodes, and the third preset period is based on minutes as nodes. The first preset period is greater than the second preset period, and the second preset period is greater than the third preset period; the first target data set, the second target data set, and the third target data set are respectively input into the first prediction model, the second prediction model, and the third prediction model in the trained wind power prediction model to obtain a first prediction result, a second prediction result, and a third prediction result. The first prediction result is a prediction result representing the long-term wind power characteristics of the target wind farm, the second prediction result is a prediction result representing the short-term wind power characteristics of the target wind farm, and the third prediction result is a prediction result representing the ultra-short-term wind power characteristics of the target wind farm; based on the target prediction model in the wind power prediction model, the first prediction result, the second prediction result, and the third prediction result are fused to obtain a target wind power prediction result, which includes one or more of the long-term wind power prediction result, the short-term wind power prediction result, and the ultra-short-term wind power prediction result, improving the accuracy of wind power prediction.

[0039] For ease of understanding, the technical solution of this application will be introduced in detail below with reference to the accompanying drawings.

[0040] Figure 1 The following is a schematic flowchart of a wind power prediction method provided by an embodiment of this application. For ease of explanation, only parts related to this embodiment are shown. The method provided by this embodiment includes the following steps:

[0041] S110. Obtain a first target data set, a second target data set, and a third target data set, which respectively include environmental data and wind power data of a target wind farm within a first preset period, a second preset period, and a third preset period before the current moment. The first preset period is based on days, the second preset period is based on hours, and the third preset period is based on minutes. The first preset period is greater than the second preset period, and the second preset period is greater than the third preset period.

[0042] Specifically, the target wind farm is any wind farm, and the obtained environmental data of the target wind farm includes environmental data within the target wind farm and environmental data within a preset range outside the target wind farm.

[0043] In a possible implementation manner, the first target data set includes the environmental data and wind power data corresponding to each day within 30 days before the current moment of the target wind farm, the second target data set includes the environmental data and wind power data corresponding to each hour within 24 hours before the current moment of the target wind farm, and the third target data set includes the environmental data and wind power data corresponding to each minute within 30 minutes before the current moment of the target wind farm.

[0044] As an example, obtain the environmental data and wind power data corresponding to multiple time points in each day within 30 days before the current moment of the target wind farm, and use the average value of each environmental data and the average value of the wind power data corresponding to these multiple time points in each day as the first target data set; obtain the environmental data and wind power data corresponding to multiple time points in each hour within 24 hours before the current moment of the target wind farm, and use the average value of each environmental data and the average value of the wind power data corresponding to these multiple time points in each hour as the second target data set; obtain the environmental data and wind power data corresponding to multiple time points in each minute within 30 minutes before the current moment of the target wind farm, and use the average value of each environmental data and the average value of the wind power data corresponding to these multiple time points in each minute as the third target data set.

[0045] As another example, obtain the environmental data and wind power data corresponding to the same time point in each day within 30 days before the current moment of the target wind farm, and use all the obtained environmental data and wind power data as the first target data set; obtain the environmental data and wind power data corresponding to the same time point in each hour within 24 hours before the current moment of the target wind farm, and use all the obtained environmental data and wind power data as the second target data set; obtain the environmental data and wind power data corresponding to the same time point in each minute within 30 minutes before the current moment of the target wind farm, and use all the obtained environmental data and wind power data as the third target data set.

[0046] Exemplarily, the environmental data includes any one or more of wind speed, wind direction, air pressure, temperature, and humidity.

[0047] S120, input the first target data set, the second target data set, and the third target data set into the first prediction model, the second prediction model, and the third prediction model in the trained wind power prediction model respectively, to obtain a first prediction result, a second prediction result, and a third prediction result. The first prediction result is a prediction result representing the long-term wind power characteristics of the target wind farm, the second prediction result is a prediction result representing the short-term wind power characteristics of the target wind farm, and the third prediction result is a prediction result representing the ultra-short-term wind power characteristics of the target wind farm.

[0048] Specifically, the first prediction model is a prediction model for predicting based on the long-term wind power characteristics of the target wind farm, the second prediction model is a prediction model for predicting based on the short-term wind power characteristics of the target wind farm, and the third prediction model is a prediction model for predicting based on the ultra-short-term wind power characteristics of the target wind farm.

[0049] As an example, the long-term wind power characteristics represent the change characteristics shown by the target wind farm in each month of a year. Usually, there are obvious differences in the change characteristics corresponding to each quarter. The long-term wind power characteristics can also be understood as the characteristics shown in different seasons within a year; the short-term wind power characteristics represent the change characteristics shown by the target wind farm in each hour of a day. Usually, there are relatively large differences in the morning and evening; the ultra-short-term wind power characteristics represent the change characteristics shown by the target wind farm every minute within 30 minutes. Usually, the wind power differences shown every minute are very small, but there will be sudden changes in wind power caused by certain unexpected situations, such as equipment failures, etc. Therefore, the ultra-short-term wind power characteristics can also be understood as mutation characteristics.

[0050] In a specific implementation, the first target data set, the second target data set, and the third target data set are cleaned, verified, and preprocessed. After removing noise and filling in missing values, they are then input into the first prediction model, the second prediction model, and the third prediction model in the trained wind power prediction model respectively, to obtain a first prediction result, a second prediction result, and a third prediction result, so as to ensure the quality and consistency of the input data.

[0051] In a possible implementation manner, such as Figure 2As shown in the figure, the wind power prediction model includes a first model input module, a second model input module, a third model input model, a first prediction model, a second prediction model, a third prediction model, a target prediction model, and a model output module. Among them, the first prediction model is a seasonal autoregressive integrated moving average model, and its input is connected to the first model input module; the second prediction model is a vector autoregressive model, and its input is connected to the second model input module; the third prediction model is a Prophet model, and its input is connected to the third model input module; the target prediction model is a feedforward neural network, and its input is connected to the outputs of the first prediction model, the second prediction model, and the third prediction model; the model output module is connected to the output of the target prediction module.

[0052] As an example, based on Figure 2 the wind power prediction model shown in the figure, the first target data set is input into the first prediction model through the first model input module to obtain a first prediction result, the second target data set is input into the second prediction model through the second model input module to obtain a second prediction result; the third target data set is input into the third prediction model through the third model input module to obtain a third prediction result.

[0053] S130, based on the target prediction model in the wind power prediction model, fuse the first prediction result, the second prediction result, and the third prediction result to obtain a target wind power prediction result, which includes one or more of a long-term wind power prediction result, a short-term wind power prediction result, and an ultra-short-term wind power prediction result.

[0054] Specifically, the long-term wind power prediction result, the short-term wind power prediction result, and the ultra-short-term wind power prediction result are all obtained by fusing the first prediction result, the second prediction result, and the third prediction result. The difference is that for different prediction result types, the weights used for fusing the first prediction result, the second prediction result, and the third prediction result are different, and the weight value corresponding to each prediction result type is obtained by training the target prediction model.

[0055] As an example, when calculating the long-term wind power prediction result, the weight corresponding to the first prediction result is larger; when calculating the short-term wind power prediction result, the weight corresponding to the second prediction result is larger; when calculating the ultra-short-term wind power prediction result, the weight corresponding to the third prediction result is larger.

[0056] In a possible implementation, a wind power prediction instruction is obtained. The wind power prediction instruction carries a target prediction result type, and the target prediction result type is used to indicate the prediction result type included in the target wind power prediction result; in response to the wind power prediction instruction, based on the target prediction result type, a target prediction model is used to fuse the first prediction result, the second prediction result, and the third prediction result to obtain the target wind power prediction result.

[0057] As an example, a user can select the prediction result type to be output through an interactive visualization interface, including one or more of a long-term wind power prediction result, a short-term wind power prediction result, and an ultra-short-term wind power prediction result. In response to the user's operation, a wind power prediction instruction is obtained. The wind power prediction instruction carries identification information of the prediction result type (i.e., the target prediction result type) selected by the user. Based on the target prediction result type, a target prediction model is used to fuse the first prediction result, the second prediction result, and the third prediction result to obtain the target wind power prediction result.

[0058] Exemplarily, if the wind power prediction instruction carries identification information of the long-term wind power prediction result and the short-term wind power prediction result, then a target prediction model is used to fuse the first prediction result, the second prediction result, and the third prediction result to obtain the long-term wind power prediction result and the short-term wind power prediction result.

[0059] As an example, the long-term wind power prediction result includes the wind power prediction values corresponding to each day within the next 30 days of the target wind farm, the short-term wind power prediction result includes the wind power prediction values corresponding to each hour within the next 24 hours of the target wind farm, and the ultra-short-term wind power prediction result includes the wind power prediction values corresponding to each minute within the next 30 minutes of the target wind farm.

[0060] As an example, based on Figure 2 the shown wind power prediction model, the first prediction result output by the first prediction model, the second prediction result output by the second prediction model, and the third prediction result output by the third prediction model are all input into the target prediction model. The target prediction model fuses the first prediction result, the second prediction result, and the third prediction result to obtain the target wind power prediction result, and outputs the target wind power prediction result through the model output module.

[0061] In addition, after obtaining the target wind power prediction result, an intuitive data analysis and monitoring interface is provided for the user through the provided interactive visualization interface, the prediction result is intuitively displayed, a detailed prediction report is automatically generated at the same time, and necessary intervention and adjustment are carried out, which enhances the friendliness and practicality of human-computer interaction.

[0062] The technical solution provided by this application inputs the first target data set including environmental data and wind power data of the target wind farm within the first preset period before the current moment, the second target data set including environmental data and wind power data of the target wind farm within the second preset period before the current moment, and the third target data set including environmental data and wind power data of the target wind farm within the third preset period before the current moment into the first prediction model, the second prediction model, and the third prediction model in the wind power prediction model respectively, to obtain the first prediction result representing the long-term wind power characteristics of the target wind farm, the second prediction result representing the short-term wind power characteristics of the target wind farm, and the third prediction result representing the ultra-short-term wind power characteristics of the target wind farm; then, based on the target prediction model in the wind power prediction model, the first prediction result, the second prediction result, and the third prediction result are fused to obtain the target wind power prediction result. The target wind power prediction result includes one or more of the long-term wind power prediction result, the short-term wind power prediction result, and the ultra-short-term wind power prediction result, so that the long-term wind power prediction result, the short-term wind power prediction result, and the ultra-short-term wind power prediction result of the obtained target wind farm all incorporate the long-term wind power characteristics, short-term wind power characteristics, and ultra-short-term wind power characteristics of the target wind farm. During the process of wind power prediction, various regular characteristics and mutation situations of the target wind farm are fully considered, improving the accuracy of the wind power prediction result and further enhancing the accuracy of wind power prediction.

[0063] Figure 3 It is a schematic flowchart of a method for training a wind power prediction model provided by an embodiment of this application. In combination with Figure 3 As shown, on the basis of the above embodiment, this embodiment further explains and optimizes the technical solution. Specifically, in this embodiment, the training process of the wind power prediction model includes:

[0064] S310, obtain the first historical data set, the second historical data set, and the third historical data set. The first historical data set includes environmental data and wind power data corresponding to each month of the target wind farm in any past year. The second historical data set includes environmental data and wind power data corresponding to each hour of the target wind farm in any past day. The third historical data includes environmental data and wind power data corresponding to each minute within any past 30 minutes of the target wind farm.

[0065] As an example, obtain the environmental data and wind power data corresponding to the same time point of each day in any past year of the target wind farm. Based on the obtained environmental data and wind power data, calculate the average values of various environmental data corresponding to the number of days included in each month of this year and the average value of wind power data as the environmental data and wind power corresponding to each month, and use them as the first historical data set; obtain the environmental data and wind power data corresponding to the same time point of each hour within any past day of the target wind farm as the second historical data set; obtain the environmental data and wind power data corresponding to the same time point of each minute within any past 30 minutes of the target wind farm as the third historical data set.

[0066] S320, based on the first historical data set, construct and train a Seasonal Autoregressive Integrated Moving Average (SARIMA) model to obtain the first prediction model.

[0067] Specifically, the Seasonal Autoregressive Integrated Moving Average (SARIMA) model is a seasonal extension of the Autoregressive Integrated Moving Average (ARIMA) model. The ARIMA model is a combination of Autoregressive (AR) and Moving Average (MA), while the SARIMA model also takes into account the seasonal component. Its parameters include the seasonal autoregressive order, seasonal moving average order, non-seasonal autoregressive order, non-seasonal moving average order, and differencing order. In this representation, the SARIMA model can simultaneously handle non-seasonal and seasonal components in the time series, making the model more suitable for time series data with obvious seasonal changes.

[0068] As an example, perform preprocessing operations such as handling missing values and outliers on the first historical data set to ensure the integrity of the time series data. Perform seasonal differencing according to the seasonal period of the preprocessed first historical data set to eliminate the non-stationarity caused by seasonal changes. Identify a suitable SARIMA model by observing the time series plot, Autocorrelation Function (ACF), and Partial Autocorrelation Function (PACF) plot. Use maximum likelihood estimation or other methods to estimate the model parameters. Check whether the residual sequence is white noise to ensure the goodness of fit of the model. After completion of the fitting, determine the first prediction model.

[0069] S330. Based on the second historical data set, construct and train a vector autoregression model to obtain a second prediction model.

[0070] Specifically, the vector autoregression model (VAR) is a statistical model used to analyze and predict the dynamic relationships between multiple time series variables. It is established based on the statistical properties of the data. In this model, each endogenous variable in the system is constructed as a function of the lagged values of all endogenous variables in the system, thus generalizing the univariate autoregressive model to a "vector" autoregressive model composed of multiple time series variables. The general expression of the VAR(p) model is:

[0071] Y t =A 1 Y t-1 +A 2 Y t-2 +...+A p Y t-p +μ t ,

[0072] where Y t is a vector containing multiple time series variables, which is the second historical data set in this application. A 1 ,A 2 ,...,A p are coefficient matrices, μ t is the error term vector, and p is the lag order.

[0073] In specific implementation, first preprocess the second historical data set, such as missing value processing, outlier processing, etc. Then conduct a stationarity test on the second historical data set to ensure that the data meets the requirements of the VAR model. If the data is not stationary, it can be transformed into a stationary sequence through methods such as differencing. Determine the lag order of the VAR model according to the information criterion or model diagnosis results, estimate the coefficient matrix of the VAR model using the least squares method or other optimization methods, and diagnose the fitted VAR model to check the stability of the model and the normality of the residuals, etc., to obtain the second prediction model.

[0074] S340. Based on the third historical data set, construct and train a Prophet model to obtain a third prediction model.

[0075] Specifically, the Prophet model is a time series prediction algorithm open-sourced by Facebook. Based on the ideas of time series decomposition and machine learning, it realizes accurate prediction of time series by fitting multiple components such as trend terms, seasonal terms, and holiday terms. The basic form of the Prophet model can be expressed as:

[0076] y(t) = g(t) + s(t) + h(t) + ε(t),

[0077] where y(t) is the value of the time series at time t, g(t) is the trend component, representing the non-periodic changes of the time series, which can be non-linear and can automatically identify changes in the growth rate, s(t) is the seasonal factor, representing the periodic changes of the time series, which can be fixed (such as weekly or annual repeating patterns) or varying (such as seasonal changes that increase or decrease year by year), h(t) is the impact of holidays and other known events, which can be custom-added as needed, and ε(t) is the noise term, representing random and unpredictable fluctuations, and it is usually assumed to follow a normal distribution.

[0078] In a specific implementation, collect and preprocess the third historical data set, including handling missing values, outliers, etc., initialize the Prophet object using the preprocessed third historical data set, and set model parameters (such as holiday information), fit the model, that is, use the third historical data set to train the Prophet model, specify the future time period for prediction, call the predict method of the model to obtain the prediction result, use the built-in plotting function of Prophet to visually display the predicted value, confidence interval, and model components, and further analyze the prediction effect to obtain the third prediction model.

[0079] S350, train a feedforward neural network model based on the first historical data set, the second historical data set, and the third historical data set to obtain the target prediction model.

[0080] In a possible implementation, the target prediction model includes any one or more of the first target prediction sub-model, the second target prediction sub-model, and the third target prediction sub-model; based on the first historical data set, construct and train a feedforward neural network model to obtain the first target prediction sub-model; based on the second historical data set, construct and train a feedforward neural network model to obtain the second target prediction sub-model; based on the third historical data set, construct and train a feedforward neural network model to obtain the third target prediction sub-model; based on the first target prediction sub-model, the second target prediction sub-model, and the third target prediction sub-model, construct the target prediction model.

[0081] In a specific implementation, before the training starts, the weights and biases of the feedforward neural network are randomly initialized. The first historical data set, the second historical data set, and the third historical data set are propagated forward through the feedforward neural network to obtain a prediction result. An appropriate loss function (such as mean squared error, cross-entropy loss, etc.) is used to calculate the difference between the prediction result and the true result. According to the gradient information of the loss function, the weights and biases of the network are adjusted through the backpropagation algorithm to reduce the loss value. The processes of forward propagation, loss calculation, and backpropagation are repeated until the stopping condition is met (such as reaching the maximum number of iterations, the loss value converges, etc.), obtaining the first target prediction sub-model, the second target prediction sub-model, and the third target prediction sub-model, and constructing the target prediction model.

[0082] S360. Based on the first prediction model, the second prediction model, the third prediction model, and the target prediction model, construct a wind power prediction model.

[0083] In a specific implementation, compare the prediction results obtained from the trained first prediction model, second prediction model, third prediction model, and target prediction model with the actual observed values to evaluate the prediction accuracy, and accordingly adjust and optimize each prediction model. Based on the adjusted first prediction model, second prediction model, third prediction model, and target prediction model, construct a wind power prediction model.

[0084] The technical solution provided in this embodiment constructs and trains an autoregressive integrated moving average (ARIMA) model based on the first historical data set to obtain the first prediction model; constructs and trains a vector autoregressive (VAR) model based on the second historical data set to obtain the second prediction model; constructs and trains a Prophet model based on the third historical data set to obtain the third prediction model; trains a feedforward neural network model based on the first historical data set, the second historical data set, and the third historical data set to obtain the target prediction model; adjusts and optimizes each prediction model through actual observed values, and constructs a wind power prediction model based on the adjusted first prediction model, second prediction model, third prediction model, and target prediction model, improving the accuracy of the wind power prediction model, and thus improving the accuracy of wind power prediction.

[0085] In summary, the wind power prediction method provided in this application has the following beneficial effects:

[0086] 1. Significantly improve the prediction accuracy: By means of various advanced machine learning algorithms, complex patterns are mined from a large amount of historical data, including microclimate effects and long-term climate change trends, thus greatly improving the prediction accuracy of long-term, short-term, and ultra-short-term wind power. This high accuracy provides a reliable basis for power grid dispatching and reduces economic losses caused by prediction errors.

[0087] 2. Enhance real-time response ability: It can be adjusted according to the latest meteorological data and operating status within an extremely short time, achieving minute-level or even second-level prediction updates. This enables the power grid to respond quickly, effectively cope with emergencies such as sudden wind speed changes, and ensure the stability of the power system.

[0088] 3. Optimize resource allocation: By making refined predictions of wind power output and market demand, it supports more scientific resource planning, including reasonably allocating energy storage facilities, adjusting the output of other power generation sources, and optimizing the operation and maintenance strategies of wind farms, ultimately achieving the goal of reducing costs and improving efficiency.

[0089] 4. Promote cross-regional collaboration: It helps to integrate wind energy resources in different regions. By analyzing wind farm data within a wide area, it assists in formulating more reasonable power transmission plans, enhancing cross-regional power complementarity, and improving the overall energy utilization efficiency.

[0090] 5. Boost new energy investment confidence: High-precision wind power prediction enhances the confidence of investors and operators in wind power projects. Stable power output and more reliable predictions are conducive to accelerating the deployment and expansion of new energy projects.

[0091] 6. Environmental protection and social responsibility: By maximizing the utilization of wind energy, it indirectly promotes the reduction of carbon emissions, in line with the global green and low-carbon development goal. In addition, improving the economic benefits of the wind power industry also demonstrates the social responsibility of enterprises on the path of sustainable development.

[0092] Generally speaking, the technical solution provided by this application not only greatly promotes the technological innovation of the wind power industry, but also provides strong technical support for the transformation and upgrading of the energy structure. It is an indispensable part in promoting the wind power industry towards intelligence and high efficiency.

[0093] Figure 4 The following is a structural block diagram of a wind power prediction device provided by an embodiment of this application. For the sake of convenience of description, only the parts related to the embodiment of this application are shown. Referring to Figure 4 , the wind power prediction device 400 may include a data acquisition module 401, a first prediction module 402, and a second prediction module 403.

[0094] The data acquisition module 401 is used to acquire a first target data set, a second target data set, and a third target data set, which respectively include environmental data and wind power data of the target wind farm within a first preset period, a second preset period, and a third preset period before the current moment. The first preset period is based on days as a node, the second preset period is based on hours as a node, and the third preset period is based on minutes as a node. The first preset period is greater than the second preset period, and the second preset period is greater than the third preset period.

[0095] The first prediction module 402 is configured to input the first target data set, the second target data set, and the third target data set into the first prediction model, the second prediction model, and the third prediction model in the trained wind power prediction model respectively, to obtain a first prediction result, a second prediction result, and a third prediction result. The first prediction result is a prediction result representing the long-term wind power characteristics of the target wind farm, the second prediction result is a prediction result representing the short-term wind power characteristics of the target wind farm, and the third prediction result is a prediction result representing the ultra-short-term wind power characteristics of the target wind farm.

[0096] The second prediction module 403 is configured to fuse the first prediction result, the second prediction result, and the third prediction result based on the target prediction model in the wind power prediction model to obtain a target wind power prediction result, where the target wind power prediction result includes one or more of a long-term wind power prediction result, a short-term wind power prediction result, and an ultra-short-term wind power prediction result.

[0097] A wind power prediction device provided by an embodiment of the present application has the same beneficial effects as the above-mentioned wind power prediction method.

[0098] It should be noted that for the information interaction, execution process, etc. between the above-mentioned device / units, since they are based on the same concept as the method embodiment of the present application, their specific functions and the technical effects brought can be specifically referred to the method embodiment part, and will not be elaborated here.

[0099] Those skilled in the art can clearly understand that for the convenience and conciseness of description, only the above-mentioned division of each functional unit and module is used as an example. In actual application, the above-mentioned functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of the present application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the foregoing method embodiment, and will not be elaborated here.

[0100] Figure 5 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As Figure 5 shown, the electronic device 5 in this embodiment includes: at least one processor 50 ( Figure 5only one is shown), a memory 51, and a computer program 52 stored in the memory 51 and executable on at least one processor 50. When the processor 50 executes the computer program 52, it implements the above Figure 1 or Figure 3 steps in the method embodiments, or implements the functions of the various modules / units in the above Figure 4 device embodiments.

[0101] The electronic device 5 may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The electronic device 5 may include, but is not limited to, a processor 50 and a memory 51. Those skilled in the art can understand that Figure 5 merely examples of the electronic device 5, which do not constitute a limitation on the electronic device 5, and may include more or fewer components than shown in the figure, or combine some components, or different components. For example, it may also include input / output devices, network access devices, etc.

[0102] The processor 50 may be a central processing unit (CPU), and the processor 50 may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0103] In some embodiments, the memory 51 may be an internal storage unit of the electronic device 5, such as the hard disk or memory of the electronic device 5. In other embodiments, the memory 51 may also be an external storage device of the electronic device 5, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 5. Further, the memory 51 may also include both the internal storage unit and the external storage device of the electronic device 5. The memory 51 is used to store an operating system, application programs, a boot loader, data, and other programs, such as the program code of the computer program. The memory 51 may also be used to temporarily store data that has been output or will be output.

[0104] The embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0105] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-mentioned method embodiments of the present application, a computer program can be used to instruct the relevant hardware to complete. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium can at least include: any entity or device that can carry the computer program code to an electronic device, a recording medium, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), an electrical carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disc, etc.

[0106] The computer-readable storage medium provided by the embodiments of the present application has the same beneficial effects as the above-mentioned wind power prediction method.

[0107] The embodiments of the present application provide a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0108] The computer program product provided by the embodiments of the present application has the same beneficial effects as the above-mentioned wind power prediction method.

[0109] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0110] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. A professional person can use different methods to implement the described functions for each specific application, but such an implementation should not be considered to exceed the scope of the present application.

[0111] In the embodiments provided in the present application, it should be understood that the disclosed device / equipment and method can be implemented in other ways. For example, the device / equipment embodiments described above are merely illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of the device or unit can be in electrical, mechanical or other forms.

[0112] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0113] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A wind power prediction method, characterized in that: The method comprises: Acquire a first target data set, a second target data set, and a third target data set, respectively including environmental data and wind power data of the target wind farm within a first preset period, a second preset period, and a third preset period before the current moment, wherein the first preset period is based on days, the second preset period is based on hours, and the third preset period is based on minutes, the first preset period is greater than the second preset period, and the second preset period is greater than the third preset period; Inputting the first target data set, the second target data set and the third target data set into the first prediction model, the second prediction model and the third prediction model in the trained wind power prediction model respectively, to obtain a first prediction result, a second prediction result and a third prediction result, wherein the first prediction result is a prediction result representing the long-term wind power characteristics of the target wind farm, the second prediction result is a prediction result representing the short-term wind power characteristics of the target wind farm, and the third prediction result is a prediction result representing the ultra-short-term wind power characteristics of the target wind farm; Based on the target prediction model in the wind power prediction model, the first prediction result, the second prediction result and the third prediction result are integrated to obtain a target wind power prediction result, and the target wind power prediction result includes one or more of a long-term wind power prediction result, a short-term wind power prediction result and an ultra-short-term wind power prediction result.

2. The method according to claim 1, characterized in that: The step of fusing the first prediction result, the second prediction result, and the third prediction result based on the target prediction model in the wind power prediction model to obtain a target wind power prediction result includes: Acquire a wind power prediction instruction, wherein the wind power prediction instruction carries a target prediction result type, and the target prediction result type is used to indicate a prediction result type included in the target wind power prediction result; In response to the wind power prediction instruction, based on the target prediction result type, the target prediction model is used to fuse the first prediction result, the second prediction result and the third prediction result to obtain the target wind power prediction result.

3. The method according to claim 1, characterized in that The training process of the wind power prediction model includes: Acquire a first historical data set, a second historical data set, and a third historical data set, wherein the first historical data set includes the environmental data and wind power data corresponding to each month of the target wind farm in any past year, the second historical data set includes the environmental data and wind power data corresponding to each hour of the target wind farm in any past day, and the third historical data set includes the environmental data and wind power data corresponding to each minute of the target wind farm in any past 30 minutes; Based on the first historical data set, construct and train a seasonal difference autoregressive moving average model to obtain the first prediction model; Based on the second historical data set, construct and train a vector autoregression model to obtain the second prediction model; Based on the third historical data set, construct and train a Prophet model to obtain the third prediction model; Training a feedforward neural network model based on the first historical data set, the second historical data set, and the third historical data set to obtain the target prediction model; The wind power prediction model is constructed based on the first prediction model, the second prediction model, the third prediction model and the target prediction model.

4. The method according to claim 3, characterized in that The wind power prediction model comprises a first model input module, a second model input module, a third model input module, the first prediction model, the second prediction model, the third prediction model, the target prediction model and a model output module; The first prediction model is a seasonal difference autoregressive moving average model, and its input is connected to the first model input module; The second prediction model is a vector autoregression model, and its input is connected to the second model input module; The third prediction model is a Prophet model, and its input is connected to the third model input module; The target prediction model is a feedforward neural network, whose input is connected to the outputs of the first prediction model, the second prediction model and the third prediction model; The model output module is connected to the output of the target prediction model.

5. The method according to claim 4, characterized in that The target prediction model includes any one or more of a first target prediction sub-model, a second target prediction sub-model and a third target prediction sub-model; The training of a feedforward neural network model based on the first historical data set, the second historical data set, and the third historical data set to obtain the target prediction model includes: Based on the first historical data set, construct and train a feedforward neural network model to obtain the first target prediction sub-model; Based on the second historical data set, construct and train a feedforward neural network model to obtain the second target prediction sub-model; Based on the third historical data set, construct and train a feedforward neural network model to obtain the third target prediction sub-model; The target prediction model is constructed based on the first target prediction sub-model, the second target prediction sub-model and the third target prediction sub-model.

6. The method according to any one of claims 1 to 5, characterized in that: The first target data set includes the environmental data and wind power data corresponding to each day of the target wind farm within 30 days before the current moment, the second target data set includes the environmental data and wind power data corresponding to each hour of the target wind farm within 24 hours before the current moment, and the third target data set includes the environmental data and wind power data corresponding to each minute of the target wind farm within 30 minutes before the current moment; The long-term wind power forecast result includes the wind power forecast value corresponding to each day of the target wind farm in the next 30 days, the short-term wind power forecast result includes the wind power forecast value corresponding to each hour of the target wind farm in the next 24 hours, and the ultra-short-term wind power forecast result includes the wind power forecast value corresponding to each minute of the target wind farm in the next 30 minutes.

7. The method according to any one of claims 1 to 5, characterized in that: The environmental data includes any one or more of wind speed, wind direction, air pressure, temperature and humidity.

8. A wind power prediction device, characterized in that: The device comprises: A data acquisition module, used to acquire a first target data set, a second target data set and a third target data set, respectively including environmental data and wind power data of a target wind farm within a first preset period, a second preset period and a third preset period before a current moment, wherein the first preset period is based on days, the second preset period is based on hours, the third preset period is based on minutes, the first preset period is greater than the second preset period, and the second preset period is greater than the third preset period; A first prediction module is used to input the first target data set, the second target data set and the third data target data set into the first prediction model, the second prediction model and the third prediction model in the trained wind power prediction model, respectively, to obtain a first prediction result, a second prediction result and a third prediction result, wherein the first prediction result is a prediction result representing the long-term wind power characteristics of the target wind farm, the second prediction result is a prediction result representing the short-term wind power characteristics of the target wind farm, and the third prediction result is a prediction result representing the ultra-short-term wind power characteristics of the target wind farm; The second prediction module is used to fuse the first prediction result, the second prediction result and the third prediction result based on the target prediction model in the wind power prediction model to obtain a target wind power prediction result, wherein the target wind power prediction result includes one or more of a long-term wind power prediction result, a short-term wind power prediction result and an ultra-short-term wind power prediction result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images