Power prediction method and system for wind farm, and storage medium
Patent Information
- Authority / Receiving Office
- AU · AU
- Patent Type
- Applications
- Current Assignee / Owner
- YUANJIAN WIND POWER JIANGYINENVISION ENERGY CO LTD
- Filing Date
- 2024-12-19
- Publication Date
- 2026-07-30
AI Technical Summary
The existing wind farm power prediction methods have large calculation volume, high application cost, and rely on a large amount of historical data to train, making it difficult to achieve minute-level power prediction.
By setting the wind farm station in the wind farm, collecting wind speed and other meteorological data, reconstructing the airflow distribution within the wind farm range, using the video repair model to predict the future airflow distribution, and converting it into power predicted value based on the predicted wind speed value.
It realizes high-precision minute-level wind farm power prediction, reduces calculation volume and application cost, improves the real-time and accuracy of prediction, and brings huge safety and economic benefits.
Smart Images
Figure 00000000_0000_ABST
Abstract
Description
A wind farm power prediction method, system and storage medium CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is based on the Chinese patent application with application number "202311766432.9" and application date of December 20, 2023, and claims the priority of the above-mentioned Chinese patent application. The entire content of the above-mentioned Chinese patent application is hereby incorporated into this application by introduction. Technical Field
[0002] The present application relates to the technical field of wind power generation, and in particular to a method, system and storage medium for predicting power of a wind farm. Background Art
[0003] Drastic changes in wind speed can cause huge fluctuations in wind farm output, thereby affecting the balance of the power grid. In countries and regions where the power grid is not strong enough and in scenarios where wind farms directly supply the load side (such as zero-carbon industrial parks), effectively predicting wind farm power changes can bring huge safety and economic benefits.
[0004] To achieve advance prediction of wind farm power, model-based prediction methods have been proposed in the existing technology. However, most wind farm power prediction methods are only suitable for achieving power predictions of more than 15 minutes. However, power predictions of more than 15 minutes cannot make timely adjustments to the power grid, which will cause power fluctuations and affect the safety and economic benefits of the power grid.
[0005] Some methods for predicting wind farm power within 15 minutes rely on physical simulation models based on CFD (Computational Fluid Dynamics). These methods generate a database of wind flow and turbulence patterns in advance, then match the wind flow and turbulence patterns to the aircraft position and real-time observation data for wind power forecasting. These solutions are computationally intensive, costly, and rely on extensive historical data training. Furthermore, they require model matching, which is expected to be challenging.
[0006] Another method is to train two models and use lidar to build a model to obtain real-time wind speed information for multiple positions and perform real-time feedforward control of the units. This type of solution is based on second-level signal data for control and is used for power prediction within 1 minute. It relies on the actual measured data of the lidar and has high equipment costs.
[0007] In summary, existing wind farm power forecasting solutions are mostly based on prediction models that are computationally intensive and rely on large amounts of historical data for training, or rely on high-precision and high-cost wind speed detection equipment; and most of them find it difficult to achieve high-precision minute-level power forecasting. Summary of the Invention
[0008] The purpose of this application is to provide a wind farm power prediction method, system and storage medium in order to overcome the defects of the above-mentioned existing technologies, such as large computational complexity, high application cost, reliance on a large amount of historical data training, and difficulty in achieving minute-level power prediction.
[0009] The purpose of this application can be achieved through the following technical solutions:
[0010] A method for predicting power in a wind farm comprises the following steps: setting up wind farm positions in the wind farm to collect wind speed in the area where wind turbines are located and collecting other meteorological data in the wind farm; reconstructing the airflow distribution within the wind farm based on the collected data; predicting the future airflow distribution in the wind farm based on the reconstructed airflow distribution within the wind farm; selecting a wind farm position and selecting a predicted wind speed value corresponding to the wind farm position from the predicted future airflow distribution; and converting the predicted wind speed value corresponding to the wind farm position into a power prediction value.
[0011] In some embodiments, the reconstruction process of the airflow distribution within the wind field includes the following steps: dividing the wind field into multiple grid points, and the grid spacing of the grid points is set according to the wind field machine positions and other meteorological data collection points; forming continuous images of each collection point in the wind field based on historical continuous data of wind speed and other meteorological data; using a pre-trained video restoration model to reconstruct the continuous images of each collection point in the wind field to obtain the airflow distribution of the entire wind field.
[0012] In some embodiments, the reconstruction method of the video restoration model is one or more of interpolation, image restoration, and video restoration.
[0013] In some embodiments, the continuous time of the continuous images is between 0 and 1 hour; and the predicted time of the future airflow distribution of the wind farm is between 0 and 1 hour.
[0014] In some embodiments, the predicted time of the future airflow distribution of the wind farm is within 1-15 minutes.
[0015] In some embodiments, the prediction method for the future airflow distribution of the wind farm is one of an optical flow method and a neural network training method.
[0016] In some embodiments, the method for obtaining the power prediction value includes: according to the predicted wind speed value corresponding to the machine site, looking up the power curve table to obtain the corresponding power prediction value.
[0017] In some embodiments, the wind speed collected at the wind farm machine position is a measurement value of an anemometer, or a current wind speed estimated by a wind turbine of a wind farm generator based on its own speed, torque and blade angle;
[0018] The other meteorological data may be derived from one or more of a wind tower and a weather station.
[0019] The present application also provides a power prediction system for a wind farm, comprising: a wind farm machine site observation and supplementary meteorological observation module, which is used to set the wind farm machine site to collect the wind speed in the area where the wind turbines in the wind farm are located, and to collect other meteorological data; a wind farm range airflow distribution reconstruction module that integrates observation data, which is used to reconstruct the airflow distribution within the wind farm range based on the collected data; a wind farm future airflow distribution prediction module, which is used to predict the future airflow distribution of the wind farm based on the reconstructed airflow distribution within the wind farm range; a wind speed prediction value module that selects the machine site based on the predicted airflow distribution, which is used to select the machine site and the corresponding predicted wind speed value from the predicted future airflow distribution; a wind speed prediction value to power prediction value conversion module, which is used to convert the predicted wind speed value corresponding to the machine site into a power prediction value.
[0020] The present application also provides a computer-readable storage medium, on which a computer program is stored, and the computer program is used by a processor to execute the above-mentioned method.
[0021] Compared with the prior art, this application has at least one of the following advantages:
[0022] (1) The present application sets up a wind farm station in a wind farm, and the wind farm station is used to collect the wind speed in the area where the wind turbines in the wind farm are located, collect other meteorological data in the wind farm, reconstruct the airflow distribution within the wind farm according to the wind speed and other meteorological data, predict the future airflow distribution of the wind farm according to the reconstructed airflow distribution within the wind farm, select the station location of the wind farm station, select the predicted wind speed value corresponding to the station location from the predicted future airflow distribution, and convert the predicted wind speed value corresponding to the station location into the power prediction value of the wind turbines in the wind farm, that is, the present application obtains the future airflow distribution of the wind farm by airflow prediction, thereby predicting the power prediction value of the wind turbines in the wind farm, and the prediction accuracy is relatively high; at the same time, the present application does not need to use a prediction model with large computational complexity, high application cost, and reliance on a large amount of historical data training, thereby reducing the investment cost.
[0023] (2) This application takes into account that drastic changes in wind speed will cause huge fluctuations in wind farm output, thereby affecting the balance of the power grid. By collecting wind speed and other meteorological data at multiple points in the wind farm to form historical continuous data, the airflow distribution within the wind farm can be reconstructed through a video repair model, and the future airflow distribution of the wind farm can be predicted. According to the corresponding predicted wind speed value, the power prediction value of the wind turbine can be obtained. This solution adopts the airflow prediction method, which can effectively predict the wind farm power change of 1-15 minutes and realize minute-level power prediction. The prediction is more real-time and accurate, which brings huge safety and economic benefits to countries and regions with weak power grids and scenarios where wind farms directly supply the load side (such as zero-carbon industrial parks).
[0024] (3) Considering that the evolution mechanism of airflow at the microscopic scale is very complex and cannot be observed on a large scale, this application is based on wind turbine observations and supplementary meteorological observations to effectively locate and track the trajectory of wind propagation.
[0025] (4) Considering that wind farm power forecasting requires a large amount of historical data accumulation, it cannot be applied to newly built wind farms. In the case that a new wind farm does not have a large amount of historical data accumulation, this application can still achieve effective forecasting based on a small amount of data. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] FIG1 is a schematic flow chart of a method for predicting power of a wind farm provided in an embodiment of the present application;
[0027] FIG2 is a schematic structural diagram of a wind farm power prediction system provided in an embodiment of the present application. DETAILED DESCRIPTION
[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Generally, the components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.
[0029] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for protection, but merely represents selected embodiments of the present application. All other embodiments obtained by persons of ordinary skill in the art based on the embodiments in the present application without creative work are within the scope of protection of the present application.
[0030] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.
[0031] Example 1
[0032] As shown in FIG1 , this embodiment provides a method for predicting wind farm power, including the following steps:
[0033] S1: Wind farm position observation and supplementary meteorological observation; that is, setting up wind farm positions in the wind farm to collect wind speed in the area where the wind turbines are located and collecting other meteorological data in the wind farm.
[0034] The collected wind speed includes but is not limited to the measurement value of the anemometer, and the current wind speed estimated by the wind turbine based on its own operating conditions such as speed, torque, and blade angle; other sources of meteorological data include but are not limited to wind towers, nearby weather stations, etc.
[0035] S2: Fuse the observation data to reconstruct the airflow distribution within the wind field; that is, reconstruct the airflow distribution within the wind field based on the collected data.
[0036] The reconstruction process of the airflow distribution within the wind farm includes the following steps:
[0037] The wind farm is divided into multiple grid points, and the grid spacing of the grid points is set according to the wind farm machine locations and other meteorological data collection points;
[0038] Based on the historical continuous data of wind speed and other meteorological data, a continuous image of each collection point in the wind farm is formed;
[0039] A pre-trained video restoration model is used to reconstruct continuous images of each collection point in the wind farm to obtain the airflow distribution of the entire wind farm.
[0040] It can be understood that the historical continuous data is composed of wind speed and other meteorological data; the collection points are grid points; and the wind turbines are wind turbines in the wind farm.
[0041] The reconstruction method of the video restoration model includes but is not limited to one or more of interpolation, image restoration, and video restoration. It can be understood that the reconstruction method of the video restoration model is a method for the video restoration model to reconstruct continuous images.
[0042] The continuous time of the continuous images is between 0 and 1 hour, for example, 30 minutes;
[0043] For example, before processing the collected data, the collected data is also cleaned;
[0044] In this embodiment, the specific implementation process of the above step S2 includes:
[0045] After the collected data is cleaned, the airflow distribution within the wind farm is reconstructed. The reconstruction method here mainly refers to first dividing the wind farm into grid points. The grid spacing depends on the aircraft position and the meteorological observation method, and can be selected as 200m or 500m. Then the wind farm aircraft position observation data and the supplementary meteorological observation data are filled into the grid, and a continuous image is formed based on a historical continuous time observation. The continuous time is between 0-1 hours, and optionally 30 minutes. The pre-trained video repair model is used for reconstruction. Before reconstruction, only the wind speed values of the aircraft position and the observation point are available. After reconstruction, the airflow distribution of the entire wind farm is obtained. The reconstruction methods include but are not limited to interpolation, image repair, video repair, etc. The video repair model does not require actual operation data of the wind farm. It can be trained using public domain video data or wind farm simulated airflow video.
[0046] It can be understood that the above-mentioned collected data are the wind speed in the area where the wind turbine is located, and other meteorological data; the above-mentioned wind farm machine position observation data are the wind speed in the area where the wind turbine is located collected by the wind farm machine position, the above-mentioned supplementary meteorological observation data are other meteorological data, and the above-mentioned grids are grid points.
[0047] S3: predicting the future airflow distribution of the wind farm; that is, predicting the future airflow distribution of the wind farm based on the reconstructed airflow distribution within the wind farm.
[0048] The prediction time of the predicted future airflow distribution of the wind farm is between 0 and 1 hour, for example, within 1 to 15 minutes.
[0049] Prediction methods include but are not limited to optical flow method (without data accumulation) and neural network training method (with data accumulation).
[0050] It is understandable that the prediction method includes a neural network training method in which a neural network model is trained to obtain a trained neural network model, and prediction is performed using the trained neural network model.
[0051] That is, this step is based on the reconstructed airflow distribution within the wind field, and further predicts the airflow distribution in the wind field within the future time t, where t can be between 0 and 1 hour, preferably within 1 to 15 minutes.
[0052] S4: extracting the predicted wind speed value of the station site according to the predicted airflow distribution; that is, selecting a station site and selecting the predicted wind speed value corresponding to the station site from the predicted future airflow distribution.
[0053] That is, based on the predicted future airflow distribution of the wind field, the grid corresponding to the machine site is selected from the grid as the predicted wind speed value of the machine site.
[0054] It can be understood that the machine site is the location where the wind farm machine is located.
[0055] S5: Converting the wind speed prediction value to the power prediction value; that is, converting the predicted wind speed value corresponding to the generator site into the power prediction value. The conversion method includes but is not limited to power curve table lookup.
[0056] Example 2
[0057] As shown in FIG2 , this embodiment provides a wind farm power prediction system, including:
[0058] Wind farm position observation and supplementary meteorological observation module 01 is used to set the wind farm position to collect wind speed in the area where the wind turbines are located, as well as to collect other meteorological data;
[0059] The module 02 for reconstructing the airflow distribution within the wind field range by integrating the observed data is used to reconstruct the airflow distribution within the wind field range based on the collected data;
[0060] The wind farm future airflow distribution prediction module 03 is used to predict the future airflow distribution of the wind farm based on the reconstructed airflow distribution within the wind farm range;
[0061] The wind speed prediction value module 04 is used to select the aircraft site and the corresponding predicted wind speed value from the predicted future airflow distribution;
[0062] The wind speed prediction value to power prediction value conversion module 05 is used to convert the predicted wind speed value corresponding to the machine site into a power prediction value.
[0063] The functions of the above modules are relatively independent and performed serially, and each module depends on the output results of the previous module.
[0064] It should be noted that the specific content and beneficial effects of the implementation process of each module in the device of the present application can be found in the above method embodiments and will not be repeated here.
[0065] It is worth noting that all modules involved in this embodiment are logical modules. In actual applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, to highlight the innovations of the present invention, this embodiment does not include units that are not closely related to solving the technical problems proposed by the present invention. However, this does not mean that other units do not exist in this embodiment.
[0066] The following describes an optimal operating state of the wind farm power prediction system: During a strong wind, the wind farm turbine observation and supplementary meteorological observation module 01 collects data from each turbine and meteorological observation normally, transmits data back normally, integrates the observation data, and reconstructs the wind farm-wide airflow distribution module 02 based on historical data. The wind farm-wide airflow distribution is reconstructed normally, and the propagation trajectory of the strong wind is clear. The wind farm future airflow distribution prediction module 03 captures and calculates this propagation trajectory, correctly calculates its direction and speed, and accurately predicts the wind farm airflow distribution for a period of time in the future. Based on the predicted airflow distribution, the wind speed prediction module 04 selects the turbine site as the wind speed prediction value, and the wind speed prediction value to power prediction value conversion module 05 converts the predicted value into power. This wind speed prediction model accurately predicts that the wind speed will increase rapidly in the future, and appropriate control measures are taken to respond, ensuring stable operation on the load side or grid end.
[0067] Example 2
[0068] This embodiment provides a computer-readable storage medium, on which a computer program is stored. The computer program is executed by a processor to implement the wind farm power prediction method described in Embodiment 1.
[0069] The program code for implementing the methods of the present application can be written in any combination of one or more programming languages. Such program code can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the program code is executed by the processor or controller, the functions / operations specified in the flow charts and / or block diagrams are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.
[0070] In the context of the present application, a machine-readable medium can be a tangible medium that can contain or store a program for use by an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared or semiconductor system, device or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM) or a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0071] The above describes in detail the preferred embodiments of the present application. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present application without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art based on the concepts of the present application through logical analysis, reasoning, or limited experimentation on the basis of the prior art should be within the scope of protection defined by the claims.
Claims
1. A method for predicting power of a wind farm, comprising the following steps: Collecting wind speed in the area where the wind turbines are located in the wind farm, and collecting other meteorological data in the wind farm; the wind speed is collected by the wind farm positions set in the wind farm; Reconstructing the airflow distribution within the wind field according to the wind speed and the other meteorological data; Predicting the future airflow distribution of the wind farm according to the reconstructed airflow distribution within the wind farm; Selecting a machine site of the wind farm, and selecting a predicted wind speed value corresponding to the machine site from the predicted future airflow distribution; The predicted wind speed value corresponding to the machine site is converted into a power prediction value of the wind turbine generator in the wind farm.
2. A method for predicting wind farm power according to claim 1, wherein: The reconstruction process of the airflow distribution within the wind field includes the following steps: Dividing the wind farm into a plurality of grid points, wherein the grid spacing of the grid points is set according to the wind farm machine positions and the collection points of the other meteorological data; Filling the wind field and the other meteorological data into the grid points; forming a continuous image of each of the grid points in the wind field according to the historical continuous data composed of the wind speed and the other meteorological data; The continuous images of each grid point in the wind field are reconstructed using a pre-trained video restoration model to obtain the airflow distribution of the entire wind field.
3. A method for predicting wind farm power according to claim 2, wherein: The method by which the video restoration model reconstructs the continuous images is one or more of interpolation, image restoration and video restoration.
4. A wind farm power prediction method according to claim 2 or 3, wherein: The continuous time for forming the continuous images is between 0 and 1 hour; The predicted time of the future airflow distribution of the predicted wind farm is between 0 and 1 hour.
5. A method for predicting wind farm power according to any one of claims 1 to 4, wherein: The predicted time of the future airflow distribution of the wind farm is predicted within 1-15 minutes.
6. A method for predicting power of a wind farm according to any one of claims 1 to 5, wherein: The prediction method of the future airflow distribution of the wind farm is one of an optical flow method and a neural network training method.
7. A method for predicting wind farm power according to any one of claims 1 to 6, wherein: The method for converting the power prediction value includes: according to the predicted wind speed value corresponding to the machine site, obtaining the corresponding power prediction value by looking up a power curve table.
8. A method for predicting power of a wind farm according to any one of claims 1 to 7, wherein: The wind speed collected at the wind farm position is a measurement value of an anemometer, or the wind speed collected at the wind farm position is a current wind speed estimated by the wind turbine of the wind farm wind turbine according to its own rotation speed, torque and blade angle; The data sources of the other meteorological data are one or more of a wind tower and a meteorological station.
9. A power prediction system for a wind farm, comprising: Wind farm position observation and supplementary meteorological observation module, used to collect wind speed in the area where wind turbines are located, as well as other meteorological data; The wind speed is collected by a wind farm machine station arranged in the wind farm; A module for reconstructing airflow distribution within the wind field range by fusing observation data, which is used to reconstruct the airflow distribution within the wind field range according to the wind speed and other meteorological data; A module for predicting future airflow distribution in a wind farm, used for predicting future airflow distribution in the wind farm according to the reconstructed airflow distribution within the wind farm; A module for predicting wind speed at a station site is selected according to the predicted airflow distribution, and is used to select a station site and a corresponding predicted wind speed value from the predicted future airflow distribution; The wind speed prediction value to power prediction value conversion module is used to convert the predicted wind speed value corresponding to the machine site into the power prediction value of the wind turbine generator in the wind farm.
10. A computer-readable storage medium having a computer program stored thereon, wherein the method according to any one of claims 1 to 8 is executed by a processor through the computer program.