Power prediction method and device for wind turbine generator and computer program product

Abstract

The invention discloses a power prediction method and device for a wind turbine generator and a computer program product, and relates to the technical field of wind power generation, and the method comprises the steps: inputting the weather forecast data of a wind power plant where the wind turbine generator is located into a weather downscaling model, and obtaining the high-resolution weather forecast data outputted by the weather downscaling model; inputting the high-resolution weather forecast data into a first prediction model to obtain first prediction power output by the first prediction model, and inputting the high-resolution weather forecast data into a second prediction model to obtain second prediction power output by the second prediction model; and inputting the first prediction power, the second prediction power and the high-resolution weather forecast data into a third prediction model to obtain target prediction power, output by the third prediction model, of the wind turbine generator. By adopting the technical scheme, the problem of how to accurately predict the generated power of the wind turbine generator is solved.

Description

Technical Field

[0001] The present application relates to the technical field of wind power generation, and in particular to a method, device and computer program product for predicting power of a wind turbine. Background Art

[0002] With the continuous growth of global energy demand and the increasing emphasis on renewable energy, wind power is becoming increasingly important as an environmentally friendly and sustainable energy source. However, the instability and intermittent characteristics of wind power have brought challenges to the operation and planning of power grids. Therefore, accurate wind power forecasting is crucial to optimize power grid dispatch and achieve reliable and economical operation of power grids.

[0003] In recent years, artificial intelligence methods with data-driven as the core have developed rapidly, gradually evolving from shallow traditional machine learning algorithms to a new generation of artificial intelligence methods represented by deep learning. At present, wind power prediction models are mainly artificial intelligence models, which determine the relationship between input and output through the language rules of some fuzzy systems or the training process of learning machines. This type of model has a relatively high accuracy, but requires a large amount of historical data. At the same time, the neural network prediction model is considered a black box, and it is difficult to explain its internal mechanism and decision-making process. It mainly relies on data for training and lacks the guidance of physical laws, which easily leads to a large error between the predicted power and the actual power.

[0004] Therefore, in the related art, there is a problem of how to accurately predict the power generation of a wind turbine.

[0005] With regard to the problem of how to accurately predict the power generation of wind turbines in related technologies, no effective solution has been proposed so far.

[0006] Therefore, it is necessary to improve the related technology to overcome the above-mentioned defects in the related technology. Summary of the invention

[0007] The embodiments of the present application provide a method, device and computer program product for predicting power of a wind turbine generator set, so as to at least solve the problem in the related art of how to accurately predict the power generation of a wind turbine generator set.

[0008] According to one aspect of an embodiment of the present application, a power prediction method for a wind turbine is provided, comprising: inputting meteorological forecast data of a wind farm where the wind turbine is located into a meteorological downscaling model to obtain high-resolution meteorological forecast data output by the meteorological downscaling model; inputting the high-resolution meteorological forecast data into a first prediction model to obtain a first predicted power output by the first prediction model, and inputting the high-resolution meteorological forecast data into a second prediction model to obtain a second predicted power output by the second prediction model, wherein the first prediction model is constructed based on the wake effect of the wind farm, and the second prediction model is constructed based on a power curve of the wind turbine, and the power curve is used to represent the corresponding relationship between wind speed and the power of the wind turbine; inputting the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into a third prediction model to obtain a target predicted power of the wind turbine output by the third prediction model.

[0009] In an exemplary embodiment, the meteorological downscaling model is obtained in the following manner, including: acquiring historical meteorological data of the wind farm, the historical meteorological data including historical low-resolution meteorological data and historical high-resolution meteorological data; using the historical low-resolution meteorological data as input and the historical high-resolution meteorological data as output, training a first initial model to obtain the meteorological downscaling model.

[0010] In an exemplary embodiment, the high-resolution meteorological forecast data is input into a first prediction model to obtain a first predicted power output by the first prediction model, including: dividing the wind farm into grid areas to obtain a plurality of grid areas; setting meteorological parameters of the plurality of grid areas according to the high-resolution meteorological forecast data; calculating the power of the wind turbines in each grid area according to the meteorological parameters of the plurality of grid areas and the wake effect between the plurality of grid areas to obtain the first predicted power.

[0011] In an exemplary embodiment, the second prediction model is obtained in the following manner, including: acquiring historical power data of the wind turbines in the wind farm, the historical power data including the power corresponding to the wind turbines at different wind speeds; determining a power curve of the wind turbines according to the historical power data; and constructing the second prediction model based on the power curve.

[0012] In an exemplary embodiment, the third prediction model includes a third prediction sub-model, and before the first predicted power, the second predicted power, and the high-resolution meteorological forecast data are input into the third prediction model to obtain the target predicted power of the wind turbine output by the third prediction model, the method also includes: acquiring historical high-resolution meteorological data of the wind farm and historical power data of the wind turbines in the wind farm; training the second initial model with the historical high-resolution meteorological data of the wind farm as input and the historical power data of the wind turbines in the wind farm as output to obtain the third sub-prediction model; inputting the historical high-resolution meteorological data into the third prediction sub-model to obtain the first initial predicted power output by the third prediction sub-model; determining a total loss function based on the first initial predicted power, the historical power data, the first predicted power, and the second predicted power, wherein the total loss function is used to calculate the target predicted power.

[0013] In an exemplary embodiment, a total loss function is determined based on the first initial predicted power, the historical power data, the first predicted power, and the second predicted power, including: determining a first loss function of the first initial predicted power and the actual power based on the historical power data; determining a second loss function of the first initial predicted power and the first predicted power; determining a third loss function of the first initial predicted power and the second predicted power; and summing the first loss function, the second loss function, and the third loss function to obtain the total loss function.

[0014] In an exemplary embodiment, the first predicted power, the second predicted power, and the high-resolution meteorological forecast data are input into a third prediction model to obtain the target predicted power of the wind turbine output by the third prediction model, including: inputting the high-resolution meteorological forecast data into the third prediction sub-model to obtain the second initial predicted power output by the third prediction sub-model; and determining the target predicted power based on the second initial predicted power and the total loss function.

[0015] According to another aspect of an embodiment of the present application, a power prediction device for a wind turbine is also provided, comprising: a first input-output module, used to input the meteorological forecast data of the wind farm where the wind turbine is located into a meteorological downscaling model, and obtain high-resolution meteorological forecast data output by the meteorological downscaling model; a second input-output module, used to input the high-resolution meteorological forecast data into a first prediction model, and obtain a first predicted power output by the first prediction model, and to input the high-resolution meteorological forecast data into a second prediction model, and obtain a second predicted power output by the second prediction model, wherein the first prediction model is constructed based on the wake effect of the wind farm, and the second prediction model is constructed based on the power curve of the wind turbine, and the power curve is used to represent the correspondence between wind speed and the power of the wind turbine; a third input-output module, used to input the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into a third prediction model, and obtain the target predicted power of the wind turbine output by the third prediction model.

[0016] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned power prediction method for a wind turbine when running.

[0017] According to another aspect of an embodiment of the present application, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the power prediction method of the wind turbine generator set through the computer program.

[0018] According to another aspect of the embodiments of the present application, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps of the method described in each embodiment of the present application are implemented.

[0019] Through this application, high-resolution meteorological forecast data of a wind farm can be obtained according to a meteorological downscaling model, and the high-resolution meteorological forecast data can be input into a first prediction model based on the wind farm wake effect and a second prediction model based on a power curve to obtain a first predicted power and a second predicted power, and then the first predicted power, the second predicted power, and the high-resolution meteorological forecast data can be input into a third prediction model to obtain a target predicted power. This solves the problem of how to accurately predict the power generation of a wind turbine in the related art, and achieves the effect of improving the power prediction efficiency of a wind turbine. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0022] Figure 1 It is a hardware structure block diagram of a computer terminal of a wind turbine power prediction method according to an embodiment of the present application;

[0023] Figure 2 is a flow chart of a power prediction method for a wind turbine generator set according to an embodiment of the present application;

[0024] Figure 3 is a schematic flow chart of a power prediction method for a wind turbine generator set according to an embodiment of the present application;

[0025] Figure 4 It is a structural block diagram of a power prediction device for a wind turbine generator set according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in 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. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] The method embodiments provided in the embodiments of the present application can be executed in a computer terminal or a similar computing device. Taking running on a computer terminal as an example, Figure 1 FIG. 1 is a hardware structure block diagram of a computer terminal of a wind turbine power prediction method according to an embodiment of the present application. Figure 1 As shown, the computer terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor (Central Processing Unit, MCU) or a programmable logic device (Field Programmable Gate Array, FPGA)) and a memory 104 for storing data, wherein the above-mentioned computer terminal may also include a transmission device 106 and an input and output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above-mentioned computer terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0029] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the power prediction method of the wind turbine in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0030] A wireless network provided by a communication provider of a computer terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (Radio Frequency, referred to as RF) module, which is used to communicate with the Internet wirelessly.

[0031] In this embodiment, a method for predicting power of a wind turbine is provided. Figure 2 is a flow chart of a method for predicting power of a wind turbine according to an embodiment of the present application, such as Figure 2 As shown, the process includes the following steps:

[0032] Step S202, inputting the weather forecast data of the wind farm where the wind turbine is located into the weather downscaling model to obtain high-resolution weather forecast data output by the weather downscaling model;

[0033] It should be noted that in the above step S202, the weather forecast data of the wind farm where the wind turbine is located can be NWP (Numerical Weather Prediction) data, and NWP data refers to data generated by a numerical weather forecast model. These data include various parameters of the atmospheric state, such as temperature, humidity, wind speed, wind direction, air pressure, etc. For example, the weather forecast data of the wind farm is generated by the WRF model (Weather Research and Forecasting, a mesoscale weather forecast model).

[0034] Step S204, inputting the high-resolution meteorological forecast data into a first prediction model to obtain a first predicted power output by the first prediction model, and inputting the high-resolution meteorological forecast data into a second prediction model to obtain a second predicted power output by the second prediction model, wherein the first prediction model is constructed based on the wake effect of the wind farm, and the second prediction model is constructed based on a power curve of the wind turbine generator set, and the power curve is used to represent the corresponding relationship between the wind speed and the power of the wind turbine generator set;

[0035] Step S206: input the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into a third prediction model to obtain a target predicted power of the wind turbine set output by the third prediction model.

[0036] Through the above steps, high-resolution meteorological forecast data of the wind farm can be obtained according to the meteorological downscaling model, and the high-resolution meteorological forecast data is respectively input into the first prediction model constructed based on the wind farm wake effect and the second prediction model constructed based on the power curve to obtain the first predicted power and the second predicted power, and then the first predicted power, the second predicted power, and the high-resolution meteorological forecast data are input into the third prediction model to obtain the target predicted power. Thus, the problem of how to accurately predict the power generation of the wind turbine in the related art is solved, and the effect of improving the power prediction efficiency of the wind turbine is achieved.

[0037] In an exemplary embodiment, the meteorological downscaling model is obtained in the following manner, including: acquiring historical meteorological data of the wind farm, the historical meteorological data including historical low-resolution meteorological data and historical high-resolution meteorological data; using the historical low-resolution meteorological data as input and the historical high-resolution meteorological data as output, training a first initial model to obtain the meteorological downscaling model.

[0038] Optionally, in the above embodiment, the historical low-resolution meteorological data of the above wind farm can be obtained from the ERA5 (European Centre for Medium-Range Weather Forecasts Re-Analysis, version 5) dataset, which is a set of global climate reanalysis datasets provided by the European Centre for Medium-Range Weather Forecasts. It contains meteorological data from 1950 to modern times. Historical high-resolution meteorological data can be generated using the Mesoscale Weather Forecast Model (WRF) and ERA5.

[0039] In an exemplary embodiment, the high-resolution meteorological forecast data is input into a first prediction model to obtain a first predicted power output by the first prediction model, including: dividing the wind farm into grid areas to obtain a plurality of grid areas; setting meteorological parameters of the plurality of grid areas according to the high-resolution meteorological forecast data; calculating the power of the wind turbines in each grid area according to the meteorological parameters of the plurality of grid areas and the wake effect between the plurality of grid areas to obtain the first predicted power.

[0040] Optionally, in the above embodiment, the first prediction model is a physical prediction model constructed based on the wake effect of the wind farm. First, a three-dimensional model of basic information such as terrain height and roughness data of the wind farm area is used. Secondly, the wind farm area is gridded, and boundary and initial conditions are set for conditions such as wind speed, wind direction, temperature and turbulence, and the wind farm flow field is numerically simulated; then, the influence of the wake effect between wind turbines is considered, and the wind speed corresponding to each wind turbine is calculated, and then the power generation power of the wind turbine is calculated.

[0041] In an exemplary embodiment, the second prediction model is obtained in the following manner, including: acquiring historical power data of the wind turbines in the wind farm, the historical power data including the power corresponding to the wind turbines at different wind speeds; determining a power curve of the wind turbines according to the historical power data; and constructing the second prediction model based on the power curve.

[0042] In an exemplary embodiment, the third prediction model includes a third prediction sub-model, and before the first predicted power, the second predicted power, and the high-resolution meteorological forecast data are input into the third prediction model to obtain the target predicted power of the wind turbine output by the third prediction model, the method also includes: acquiring historical high-resolution meteorological data of the wind farm and historical power data of the wind turbines in the wind farm; training the second initial model with the historical high-resolution meteorological data of the wind farm as input and the historical power data of the wind turbines in the wind farm as output to obtain the third sub-prediction model; inputting the historical high-resolution meteorological data into the third prediction sub-model to obtain the first initial predicted power output by the third prediction sub-model; determining a total loss function based on the first initial predicted power, the historical power data, the first predicted power, and the second predicted power, wherein the total loss function is used to calculate the target predicted power.

[0043] Optionally, in the above embodiment, for example, the third prediction sub-model is a deep learning model constructed based on CNN (convolutional neural network) and LSTM (long short-term memory neural network), and is trained based on historical high-resolution meteorological data of the wind farm and historical power data of the wind turbine. The third prediction model is based on the third prediction sub-model with the constraint of the loss function added.

[0044] In an exemplary embodiment, a total loss function is determined based on the first initial predicted power, the historical power data, the first predicted power, and the second predicted power, including: determining a first loss function of the first initial predicted power and the actual power based on the historical power data; determining a second loss function of the first initial predicted power and the first predicted power; determining a third loss function of the first initial predicted power and the second predicted power; and summing the first loss function, the second loss function, and the third loss function to obtain the total loss function.

[0045] Optionally, through the above embodiments, the physical model (i.e., the first prediction model and the second prediction model) can be embedded into the loss function of the deep learning model (i.e., the third prediction model) to ensure that the training process of the deep learning model also conforms to certain physical laws.

[0046] In an exemplary embodiment, the first predicted power, the second predicted power, and the high-resolution meteorological forecast data are input into a third prediction model to obtain the target predicted power of the wind turbine output by the third prediction model, including: inputting the high-resolution meteorological forecast data into the third prediction sub-model to obtain the second initial predicted power output by the third prediction sub-model; and determining the target predicted power based on the second initial predicted power and the total loss function.

[0047] Obviously, the above-described embodiments are only a part of the embodiments of the present application, not all of the embodiments. In order to better understand the above-mentioned method for determining the power generation area, the above-mentioned process is described below in conjunction with the embodiments, but it is not intended to limit the technical solutions of the embodiments of the present application, specifically:

[0048] In an optional embodiment, in combination Figure 3 The power prediction method of the wind turbine generator system of the present application is further described. Figure 3 As shown, the specific process includes the following:

[0049] Data collection and preprocessing: First, the wind speed and power data of each wind turbine in the wind farm are collected; then, NWP data and the 5th generation reanalysis data of the European Center for Medium-Range Weather Forecasts (ERA5) are obtained; finally, data cleaning is performed based on Spark (a big data processing framework) and K-means algorithm (a clustering algorithm) to eliminate unreasonable data and improve data quality.

[0050] A meteorological downscaling model based on the improved U-Net network is constructed: first, the mesoscale weather forecast model (WRF) is used to generate NWP data (equivalent to the meteorological forecast data of the wind farm); secondly, the improved U-Net network is used to construct a meteorological downscaling model, where the improved U-Net network is obtained by simplifying the original network structure, adding an interpolation layer at the beginning of the network, and using the deconvolution algorithm as the upsampling algorithm to improve the traditional U-Net network; then, the ERA5 dataset is used to train the meteorological downscaling model; finally, the NWP data is input into the trained meteorological downscaling model to obtain high-resolution meteorological forecast data.

[0051] A physical prediction model based on the wake effect of wind farms is constructed: first, a three-dimensional model is used based on basic information such as terrain height and roughness data of the wind farm area. Secondly, the wind farm area is gridded, and boundary and initial conditions are set for conditions such as wind speed, wind direction, temperature and turbulence, and the flow field of the wind farm is numerically simulated. Then, the influence of the wake effect between wind turbines is considered, and the wind speed corresponding to each wind turbine is calculated, thereby calculating the power generation capacity of the wind turbine.

[0052] The wind speed and power generation of each wind turbine are calculated by the following formula:

[0053]

[0054] Among them, v' j is the wind speed of wind turbine j after being affected by the wake of the front row M wind turbines; and is the free wind speed at the i-th and j-th wind turbines that is not affected by the wake; v ij is the wind speed of the i-th unit within the influence range of the wake generated by the j-th unit; A ij is the cross-shadow area of ​​the impellers of the i-th and j-th wind turbines; r i is the rotor radius of the i-th unit; r j is the rotor radius of the jth unit; x ij is the vertical distance between the i-th and j-th turbines; λ is the wake propagation coefficient; d is the diameter of the wind turbine rotor; k is the wake drop coefficient; β is the correction coefficient; P j is the predicted output power of wind turbine j; ρ represents the wind field density; S j is the overall swept area of ​​wind turbine j; C p,j is the power conversion efficiency of wind turbine j.

[0055] Construct a prediction model based on the wind speed-power curve: First, use the bin method to establish a wind speed-power curve model for the wind turbine based on the output power data of the wind turbine at different wind speeds, so as to obtain a wind power curve that is closer to the actual situation of the wind farm. The bin method is the wind speed segmentation method, which divides the collected wind speed data into several intervals, each of which is called a "bin". For example, the wind speed interval can be set to [0-1m / s], [1-2m / s], [2-3m / s], etc. Choosing the appropriate interval size depends on the distribution and analysis requirements of the data. For example, the power corresponding to different wind speeds is calculated by the following formula:

[0056]

[0057] Among them, v i is the average value of all wind speed data in the i-th bin; v i,j is the wind speed of the i-th bin data set j; P i is the average value of all power data in the i-th bin; P i,j is the power of the i-th bin data group j; N i is the number of samples in the i-th bin.

[0058] Construct a wind power prediction model based on physical constraints: First, build a deep learning model (i.e., CNN-LSTM model) based on CNN and LSTM; then, train the CNN-LSTM model based on the above-mentioned high-resolution meteorological forecast data and the actual power data of wind turbines in the wind farm. During the training process of the CNN-LSTM model, the physical model is embedded in the loss function of the model to ensure that the training process of the deep learning model also conforms to certain physical laws. The loss function includes a prediction error term (equivalent to the first loss function) and a physical constraint term, wherein the physical constraint term includes a constraint term of the physical prediction model of the wind farm wake effect (equivalent to the second loss function) and a constraint term of the wind speed-power curve prediction model (equivalent to the third loss function). The prediction error term measures the difference between the prediction result of the neural network and the true value, and the physical constraint term measures the difference between the output of the neural network and the output of the physical model. The specific loss function is calculated as follows:

[0059] L=L NN +λL wake +μL curve ;

[0060]

[0061] Among them, L is the total loss function; λ and μ are regularization parameters used to control the weight of the regularization term; L NN

[0062] ^

[0063] is the prediction error term; L wake and L curve are two physical constraints; y i is the real power value of the wind farm; i is the wind farm predicted power obtained by using CNN-LSTM neural network; N is the number of samples; P i-wake is the predicted power of the wind farm obtained by using the physical prediction model based on the wind farm wake effect; P i-curve To predict the power of a wind farm using a prediction model based on the wind speed-power curve.

[0064] Determine the wind power prediction result: Input the above high-resolution meteorological forecast data into the CNN-LSTM model based on the loss function constraint to obtain the wind power prediction result.

[0065] Through the above embodiments, a meteorological downscaling model can be constructed based on the improved U-Net network to obtain high-resolution meteorological forecast data that is more suitable for wind turbine power prediction; then, a physical prediction model based on the wind farm wake effect and a prediction model based on the wind speed-power curve are constructed; finally, a wind power prediction deep learning model CNN-LSTM is constructed, and the wake physical model and the power curve model are used as regularization terms and embedded into the loss function of the CNN-LSTM model, so as to use physical constraints to guide the learning process of the neural network, thereby obtaining a more accurate wind power prediction result.

[0066] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0067] In this embodiment, a power prediction device for a wind turbine is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0068] Figure 4 : is a structural block diagram of a power prediction device for a wind turbine according to an embodiment of the present application, the device comprising:

[0069] The first input-output module 42 is used to input the weather forecast data of the wind farm where the wind turbine is located into the weather downscaling model to obtain high-resolution weather forecast data output by the weather downscaling model;

[0070] A second input-output module 44 is used to input the high-resolution meteorological forecast data into a first prediction model to obtain a first predicted power output by the first prediction model, and to input the high-resolution meteorological forecast data into a second prediction model to obtain a second predicted power output by the second prediction model, wherein the first prediction model is constructed based on the wake effect of the wind farm, and the second prediction model is constructed based on a power curve of the wind turbine generator set, and the power curve is used to represent the corresponding relationship between the wind speed and the power of the wind turbine generator set;

[0071] The third input-output module 46 is used to input the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into a third prediction model to obtain the target predicted power of the wind turbine set output by the third prediction model.

[0072] Through the above device, high-resolution meteorological forecast data of the wind farm can be obtained according to the meteorological downscaling model, and the high-resolution meteorological forecast data can be input into the first prediction model based on the wind farm wake effect and the second prediction model based on the power curve to obtain the first predicted power and the second predicted power, and then the first predicted power, the second predicted power, and the high-resolution meteorological forecast data can be input into the third prediction model to obtain the target predicted power. This solves the problem of how to accurately predict the power generation of the wind turbine in the related art, and achieves the effect of improving the power prediction efficiency of the wind turbine.

[0073] In an exemplary embodiment, the first input-output module 42 is further used to: obtain historical meteorological data of the wind farm, the historical meteorological data including historical low-resolution meteorological data and historical high-resolution meteorological data; use the historical low-resolution meteorological data as input and the historical high-resolution meteorological data as output to train a first initial model to obtain the meteorological downscaling model.

[0074] In an exemplary embodiment, the first input-output module 42 is further used to: divide the wind farm into grid areas to obtain a plurality of grid areas; set meteorological parameters of the plurality of grid areas according to the high-resolution meteorological forecast data; calculate the power of the wind turbines in each grid area according to the meteorological parameters of the plurality of grid areas and the wake effects between the plurality of grid areas to obtain the first predicted power.

[0075] In an exemplary embodiment, the second input-output module 44 is further used to: obtain historical power data of the wind turbines in the wind farm, the historical power data including the power corresponding to the wind turbines at different wind speeds; determine the power curve of the wind turbines according to the historical power data; and construct the second prediction model based on the power curve.

[0076] In an exemplary embodiment, the third prediction model includes a third prediction sub-model, and the third input-output module 46 is also used to: obtain historical high-resolution meteorological data of the wind farm and historical power data of the wind turbines in the wind farm; train the second initial model with the historical high-resolution meteorological data of the wind farm as input and the historical power data of the wind turbines in the wind farm as output to obtain the third sub-prediction model; input the historical high-resolution meteorological data into the third prediction sub-model to obtain the first initial predicted power output by the third prediction sub-model; determine the total loss function based on the first initial predicted power, the historical power data, the first predicted power, and the second predicted power, wherein the total loss function is used to calculate the target predicted power.

[0077] In an exemplary embodiment, the third input-output module 46 is also used to: determine a first loss function between the first initial predicted power and the actual power based on the historical power data; determine a second loss function between the first initial predicted power and the first predicted power; determine a third loss function between the first initial predicted power and the second predicted power; and sum the first loss function, the second loss function, and the third loss function to obtain the total loss function.

[0078] In an exemplary embodiment, the third input-output module 46 is also used to: input the high-resolution meteorological forecast data into the third prediction sub-model to obtain a second initial prediction power output by the third prediction sub-model; and determine the target prediction power based on the second initial prediction power and the total loss function.

[0079] An embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when run.

[0080] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:

[0081] S1, inputting the meteorological forecast data of the wind farm where the wind turbine is located into the meteorological downscaling model to obtain high-resolution meteorological forecast data output by the meteorological downscaling model;

[0082] S2, inputting the high-resolution meteorological forecast data into a first prediction model to obtain a first predicted power output by the first prediction model, and inputting the high-resolution meteorological forecast data into a second prediction model to obtain a second predicted power output by the second prediction model, wherein the first prediction model is constructed based on the wake effect of the wind farm, and the second prediction model is constructed based on the power curve of the wind turbine, and the power curve is used to represent the corresponding relationship between the wind speed and the power of the wind turbine;

[0083] S3, inputting the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into a third prediction model to obtain a target predicted power of the wind turbine set output by the third prediction model.

[0084] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0085] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.

[0086] An embodiment of the present application further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0087] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0088] S1, inputting the meteorological forecast data of the wind farm where the wind turbine is located into the meteorological downscaling model to obtain high-resolution meteorological forecast data output by the meteorological downscaling model;

[0089] S2, inputting the high-resolution meteorological forecast data into a first prediction model to obtain a first predicted power output by the first prediction model, and inputting the high-resolution meteorological forecast data into a second prediction model to obtain a second predicted power output by the second prediction model, wherein the first prediction model is constructed based on the wake effect of the wind farm, and the second prediction model is constructed based on the power curve of the wind turbine, and the power curve is used to represent the corresponding relationship between the wind speed and the power of the wind turbine;

[0090] S3, inputting the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into a third prediction model to obtain a target predicted power of the wind turbine set output by the third prediction model.

[0091] An embodiment of the present application further provides a computer program product, comprising a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program product, and when the computer program is executed by a processor, the steps of the method described in each embodiment of the present application are implemented.

[0092] Optionally, in this embodiment, the above computer program may be configured to implement the following steps when executed by a processor:

[0093] S1, inputting the meteorological forecast data of the wind farm where the wind turbine is located into the meteorological downscaling model to obtain high-resolution meteorological forecast data output by the meteorological downscaling model;

[0094] S2, inputting the high-resolution meteorological forecast data into a first prediction model to obtain a first predicted power output by the first prediction model, and inputting the high-resolution meteorological forecast data into a second prediction model to obtain a second predicted power output by the second prediction model, wherein the first prediction model is constructed based on the wake effect of the wind farm, and the second prediction model is constructed based on the power curve of the wind turbine, and the power curve is used to represent the corresponding relationship between the wind speed and the power of the wind turbine;

[0095] S3, inputting the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into a third prediction model to obtain a target predicted power of the wind turbine set output by the third prediction model.

[0096] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0097] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for predicting power of a wind turbine generator set, characterized in that: include: Inputting the meteorological forecast data of the wind farm where the wind turbine is located into the meteorological downscaling model to obtain high-resolution meteorological forecast data output by the meteorological downscaling model; inputting the high-resolution weather forecast data into a first prediction model to obtain a first prediction power output by the first prediction model, and inputting the high-resolution weather forecast data into a second prediction model, Obtaining a second predicted power output by the second prediction model, wherein the first prediction model is constructed based on a wake effect of a wind farm, and the second prediction model is constructed based on a power curve of the wind turbine generator set, and the power curve is used to represent a corresponding relationship between wind speed and power of the wind turbine generator set; The first predicted power, the second predicted power, and the high-resolution meteorological forecast data are input into a third prediction model to obtain a target predicted power of the wind turbine set output by the third prediction model.

2. The method according to claim 1, characterized in that The meteorological downscaling model is obtained by the following method, including: Acquiring historical meteorological data of the wind farm, wherein the historical meteorological data includes historical low-resolution meteorological data and historical high-resolution meteorological data; The first initial model is trained with the historical low-resolution meteorological data as input and the historical high-resolution meteorological data as output to obtain the meteorological downscaling model.

3. The method according to claim 1, characterized in that Inputting the high-resolution weather forecast data into a first prediction model to obtain a first prediction power output by the first prediction model comprises: Dividing the wind farm into grid areas to obtain a plurality of grid areas; Setting meteorological parameters of the plurality of grid areas according to the high-resolution meteorological forecast data; The power of the wind turbine generator set in each grid area is calculated according to the meteorological parameters of the multiple grid areas and the wake effect between the multiple grid areas to obtain the first predicted power.

4. The method according to claim 1, characterized in that: The second prediction model is obtained by the following method, including: Acquire historical power data of the wind turbines of the wind farm, wherein the historical power data includes power corresponding to the wind turbines at different wind speeds; Determining a power curve of the wind turbine generator system according to the historical power data; The second prediction model is constructed based on the power curve.

5. The method according to claim 1, characterized in that The third prediction model includes a third prediction sub-model. Before inputting the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into the third prediction model and obtaining the target predicted power of the wind turbine set output by the third prediction model, the method further includes: Acquire historical high-resolution meteorological data of the wind farm and historical power data of wind turbines in the wind farm; Taking the historical high-resolution meteorological data of the wind farm as input and the historical power data of the wind turbines in the wind farm as output, training the second initial model to obtain the third sub-prediction model; Inputting the historical high-resolution meteorological data into the third prediction sub-model to obtain a first initial prediction power output by the third prediction sub-model; A total loss function is determined according to the first initial predicted power, the historical power data, the first predicted power, and the second predicted power, wherein the total loss function is used to calculate the target predicted power.

6. The method according to claim 5, characterized in that Determining a total loss function according to the first initial predicted power, the historical power data, the first predicted power, and the second predicted power includes: Determine a first loss function between the first initial predicted power and the actual power according to the historical power data; Determine a second loss function of the first initial predicted power and the first predicted power; Determining a third loss function of the first initial predicted power and the second predicted power; The first loss function, the second loss function, and the third loss function are summed to obtain the total loss function.

7. The method according to claim 5, characterized in that Inputting the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into a third prediction model to obtain a target predicted power of the wind turbine set output by the third prediction model, comprising: Inputting the high-resolution weather forecast data into the third prediction sub-model to obtain a second initial prediction power output by the third prediction sub-model; The target predicted power is determined according to the second initial predicted power and the total loss function.

8. A power prediction device for a wind turbine generator set, characterized in that: include: The first input-output module is used to input the meteorological forecast data of the wind farm where the wind turbine is located into the meteorological downscaling model to obtain high-resolution meteorological forecast data output by the meteorological downscaling model; a second input-output module, configured to input the high-resolution meteorological forecast data into a first prediction model to obtain a first predicted power output by the first prediction model, and to input the high-resolution meteorological forecast data into a second prediction model to obtain a second predicted power output by the second prediction model, wherein the first prediction model is constructed based on a wake effect of a wind farm, and the second prediction model is constructed based on a power curve of the wind turbine generator set, and the power curve is used to represent a corresponding relationship between wind speed and power of the wind turbine generator set; The third input-output module is used to input the first predicted power, the second predicted power, and the high-resolution meteorological forecast data into a third prediction model to obtain the target predicted power of the wind turbine set output by the third prediction model.

9. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 7 through the computer program.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images