Method for determining wind power forecast error, processor and storage medium

Abstract

This method for determining wind power generation forecast errors includes obtaining data necessary for wind power generation forecast error decoupling based on historical data of the wind power plant; dividing the wind power generation forecast into three operations, namely numerical weather forecast, wind-power conversion model, and prediction result correction, in accordance with the actual wind power plant forecast flow; calculating the error due to the correction of the prediction result; fitting the wind-power conversion model, fitting the output of the wind-power conversion model under actual weather conditions, comparing it with the output of the wind-power conversion model under numerical weather forecast conditions, and calculating the change in error of the wind-power conversion model under different weather conditions; constructing an error decoupling equation for wind power generation by combining the multiple operations; and solving the equation to calculate the error percentage.

Description

[Technical field]

[0001] This application claims priority to a Chinese patent application bearing application number 202211336429.9, filed with the China Patent Office on October 28, 2022, the entire contents of which are incorporated herein by reference.

[0002] The present application relates to the field of wind power forecasting, such as to a method, a processor and a storage medium for determining wind power forecast errors. [Background technology]

[0003] Wind energy is an important renewable clean energy, and because of its abundant storage and wide distribution, wind power generation has become an important research direction for the development and utilization of renewable energy. However, natural wind resources have characteristics such as randomness and intermittency, and the output power of wind power plants will exhibit obvious fluctuations due to random wind speed and direction, which poses a great threat to the stability of the voltage and frequency of the power system. When the wind power accessed by the power system exceeds a certain ratio, the safety and stability of the entire system will face a huge challenge. In order to effectively absorb new energy such as wind power, which is characterized by intermittency and randomness, the power system needs to add extra reserve capacity, which will increase the total operation cost of the power system. In order to reduce the allocation cost of the reserve capacity of the power system and improve the safety and reliability of the operation of the entire power system, how to accurately predict the wind power generation has become an important challenge that needs to be broken through in the field of power system operation and control.

[0004] The wind power forecasting technology in the related art has difficulty in accurately describing and expressing the inherent characteristics of wind power, and errors inevitably exist in the forecasting results. The development of an objective and comprehensive analysis and evaluation of the errors in wind power forecasting can help identify key influencing factors and weak steps in the wind power forecasting process, guide the further improvement and optimization of wind power forecasting technology, and is of great significance to improving the accuracy of wind power forecasting. However, the wind power forecasting error analysis method in the related art is less reliable due to the influence of environmental, weather and other conditions. Summary of the Invention

[0005] SUMMARY OF THE DISCLOSURE Embodiments of the present application provide a method, a processor and a storage medium for determining a wind power forecast error.

[0006] The present application provides a method for determining a wind power forecast error, and the wind power forecast process includes a numerical weather forecast operation, a wind-electricity conversion operation and a forecast result correction operation. The method for determining the wind power forecast error includes: Obtaining data required for wind power generation forecast error analysis from the historical operation data of the wind power plant, including wind power plant actual measured weather data, numerical weather forecast data, installed capacity, real-time start capacity, planned start capacity, wind power generation forecast data, and wind power plant actual output power; determining a first target error by a prediction result correction operation; Performing fitting to a wind-power conversion model of the wind power plant according to necessary data to construct a wind-power conversion fitting model; Inputting the wind power plant actual weather data into a wind-electricity conversion fitting model to obtain model output under actual weather conditions; determining a wind power forecast power under real weather conditions according to the installed capacity, the planned start-up capacity, and the model output under real weather conditions; Determining a correction error of a prediction result under actual weather conditions according to the real-time starting capacity, the planned starting capacity, and the predicted wind power generation power under actual weather conditions; Determining a wind-power conversion model error under real weather conditions according to the wind power forecast power under real weather conditions, the wind power station actual output power, and the forecast result correction error under real weather conditions; Obtaining a model output in a numerical weather prediction that corresponds to the model output under real weather conditions; and determining a second target error by wind-electricity conversion operation according to a model output under real weather conditions, a model output in a numerical weather forecast, and a wind-electricity conversion model error under real weather conditions; determining a third target error by a numerical weather forecast operation according to the wind power forecast data, the actual output power of the wind power station, the first target error, and the second target error; determining a proportion of each target error according to the first target error, the second target error, and the third target error.

[0007] The present application further provides a processor configured to perform the wind power forecast error determination method described above.

[0008] The present application further provides a machine-readable storage medium having stored thereon instructions which, when executed by a processor, cause the processor to implement the wind power forecast error determination method described above.

[0009] The drawings are used to interpret the examples of the present application together with the following specific embodiments, but are not intended to limit the examples of the present application. [Brief description of the drawings]

[0010] [Figure 1] 1 shows a schematic flow diagram of a method for determining a wind power forecast error according to an embodiment of the present application; [Diagram 2] 1 is a schematic diagram showing the internal structure of a computer device according to an embodiment of the present application; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, specific embodiments of the present application will be described with reference to the drawings. It should be understood that the specific embodiments described herein are merely for the purpose of explaining and interpreting the present application, and are not intended to limit the present application.

[0012] The embodiments of the present application put forward a wind power forecast error determination method taking into account the error change characteristics, which takes into account the error changes of multiple operations in wind power forecasting after conditions such as weather data change, and provides a more reliable basis for improving the wind power forecasting method and identifying the key influencing factors of the forecast error.

[0013] This application decomposes the wind power prediction process into three key operations: numerical weather forecast, wind-power conversion model, and prediction result correction, constructs a fitting model for the wind-power conversion operation, fits the output of the wind-power conversion model under actual weather conditions, takes into account the prediction results under actual weather conditions and numerical weather forecast conditions, makes predictions taking into account the error changes of multiple operations, solves the prediction errors of the same operations under different conditions, constructs an error decoupling equation, and solves the errors of each operation based on the error decoupling equation, thereby realizing accurate positioning of weak operations in wind power prediction.

[0014] A schematic flow diagram of a wind power forecast error determination method according to an embodiment of the present application is illustrated in Figure 1. In an embodiment of the present application, the wind power forecast error determination method may include the following steps:

[0015] In step S101, the wind power plant actual weather data, numerical weather forecast data, installed capacity C, real-time startup capacity C0, planned startup capacity C0', and wind power generation forecast data P are calculated from the historical operation data of the wind power plant. p , wind power plant actual output power P m Obtain data necessary for wind power forecast error analysis, including:

[0016] The wind farm actual weather data may include actual data of the wind farm collected by a collecting device (e.g., a wind gauge tower). In one example, the wind farm actual weather data may include actual wind speed (e.g., wind speed at different heights, e.g., wind speed at 10 meters height, 70 meters height, etc.) of the wind farm at multiple times within a certain time period (e.g., one month, one quarter, one year, or multiple years, etc.). The wind farm actual weather data may further include actual wind direction, actual temperature, actual humidity, actual pressure, etc. of the wind farm.

[0017] The numerical weather prediction data may include numerical weather prediction data for the location of the wind farm and may include various meteorological factors, such as wind farm forecast wind speed, wind farm forecast wind direction, wind farm forecast temperature, wind farm forecast humidity, wind farm forecast air pressure, etc.

[0018] In one example, the numerical weather prediction data may include wind farm forecast wind speed (e.g., forecast wind speed at different heights, e.g., at 10 meter height, at 70 meter height, etc.). The numerical weather prediction data may further include wind farm forecast wind direction, wind farm forecast temperature, wind farm forecast humidity, wind farm forecast air pressure, etc.

[0019] Numerical Weather Prediction (NWP) is a method of predicting atmospheric movement and weather phenomena for a certain period of time in the future by solving simultaneous equations of fluid mechanics and thermodynamics that describe the weather transition process using a large-scale computer to perform numerical calculations under certain initial and boundary value conditions according to the actual atmospheric conditions.

[0020] Wind power generation forecast data P p is obtained by prediction using a forecast model of the wind power plant or by a road prediction service provider, and the wind power plant submits the stored wind power forecast data (forecast power) to the dispatching side server.

[0021] Wind power plant actual output power P mis obtained by measuring on-site using a power measuring device.

[0022] In step S102, a first target error E r Determine.

[0023] According to the wind power forecast flow, the wind power forecast process can be divided into three key operations: numerical weather forecast, wind-power conversion model, and forecast result correction. The errors of the above three key operations are E n , E m , E r and the wind power generation forecast data P p can be expressed as follows: JPEG0007634115000001.jpg849Formula (1)

[0024] As shown in formula (2), the wind power generation forecast data P p , the equivalent wind power forecast value P under accurate (real-time) starting capacity conditions according to the real-time starting capacity C0 and the planned starting capacity C0' of the wind power plant cap can be calculated. JPEG0007634115000002.jpg1026Formula (2)

[0025] As shown in formula (3), the wind power generation forecast data P p and the equivalent wind power forecast value P cap By subtracting and, the error in the prediction result correction (i.e., the first target error) E r is obtained. JPEG0007634115000003.jpg950Formula (3)

[0026] In step S103, fitting is performed on the wind-power conversion model of the wind power plant according to the necessary data, and a wind-power conversion fitting model is constructed.

[0027] E n , E mIn order to decouple the wind power generation, a wind power generation model of the wind power plant is fitted based on the data in step S101. Examples of the wind power generation model may include one of a wind power generation model based on a back propagation (BP) neural network, a wind power generation model based on a long short-term memory (LSTM) neural network, and a wind power generation model based on a convolutional neural network-long short-term memory (CNN-LSTM) hybrid neural network.

[0028] The wind-power conversion model is a mathematical model that describes the relationship between wind resource meteorological factors and the effective power of wind power equipment. In actual production, due to the influence of factors such as weather conditions and plant power generation performance, the wind resource meteorological factors and power often have a complex mapping relationship. In order to ensure the accuracy of power forecasting, a wind-power conversion fitting model can be established for a specific wind power plant and a specific time period, so as to better adapt to the requirements of power forecasting.

[0029] Step S103 may include the following sub-steps.

[0030] In the first sub-step of step S103, as shown in formula (4), the installed capacity C, the planned startup capacity C0′, and the wind power generation forecast data P p Using the wind-to-electricity conversion model output f m Calculate. JPEG0007634115000004.jpg1025Formula (4)

[0031] In a second sub-step of step S103, the numerical weather forecast data is selected as an input of the wind-electricity conversion model (hereinafter, abbreviated as model input); mis the output of the wind-electricity conversion model (model output), and the model input and model output data of the wind-electricity conversion model are normalized. Numerical weather forecast data (e.g., wind power plant forecast wind speed, and preferably, wind power plant forecast wind direction, wind power plant forecast temperature, wind power plant forecast humidity, wind power plant forecast air pressure, etc.) at multiple times during a scheduled time period (e.g., one month, one quarter, one year, or multiple years, etc.) are selected from the historical operation data of the wind power plant, and the calculated output f m The dataset can be divided into a training set and a test set in a certain ratio.

[0032] In the third sub-step of step S103, training is performed on the training set to fit the wind power conversion model. XGBoost can be used to train the training set.

[0033] In the fourth sub-step of step S103, after the training is completed (for example, a set number of training times is reached), it is determined whether the fitted wind-electricity conversion model satisfies an accuracy requirement using the test set.

[0034] In the fifth sub-step of step S103, if the accuracy requirement is not met, the initial parameters of XGBoost can be modified and the fourth sub-step of step S103 can be repeated until the fitted wind-electricity conversion model meets the accuracy requirement. The fitted wind-electricity conversion model that meets the accuracy requirement can be used as the wind-electricity conversion fitting model to be constructed.

[0035] The construction (or fitting) aspect of the wind-electricity conversion fitting model according to the embodiment of the present application can automatically obtain an accurate wind-electricity conversion model when it is impossible or difficult to obtain an existing wind-electricity conversion model.

[0036] In step S104, the wind power plant actual weather data is input to the wind power conversion fitting model to obtain the model output fwind,m get.

[0037] In step S105, the installed capacity C, the planned startup capacity C0', and the model output f wind,m Predicted wind power P under actual weather conditions according to wind Determine.

[0038] The wind power plant actual weather data (e.g. data collected by a wind gauge tower, e.g. actual wind speed, etc.) are input into the wind-electricity conversion fitting model to obtain the model output f under actual weather conditions. wind,m The wind power forecast power P under actual weather conditions can be obtained according to formula (5). wind can be calculated. JPEG0007634115000005.jpg832Formula(5)

[0039] In step S106, a prediction result correction error E under actual weather conditions is calculated according to the real-time starting capacity C0, the planned starting capacity C0′, and the predicted wind power generation power under actual weather conditions. wind,r Determine.

[0040] In actual weather conditions, there is no error due to numerical weather forecasting, so at this time, the wind power generation forecast power P wind The prediction error of is the error E in the wind-electricity conversion model. wind,m , the error in correcting the predicted result E wind,r Therefore, we have the following formula: JPEG0007634115000006.jpg856Formula(6)

[0041] In addition, there is the following formula that is similar to formula (3). JPEG0007634115000007.jpg946Formula(7)

[0042] In step S107, the predicted wind power generation power P wind , wind power plant actual output power P m and the correction error E of the prediction results under actual weather conditions wind,rThe wind-electricity conversion model error E under actual weather conditions according to wind,m Determine.

[0043] The error E calculated according to formula (7) wind,r Substituting into equation (6), the wind-electricity conversion model error E wind,m can be obtained.

[0044] In step S108, the model output f wind,m The model output f in numerical weather prediction corresponds to m Get the.

[0045] Here, the "correspondence" of the model output in the numerical weather forecast that corresponds to the model output under actual weather conditions means the value under two conditions for the same reference object. For example, for one wind power plant, the actual wind speed at one reference time is x, and the model output obtained after inputting x into the wind-power conversion fitting model is f wind,m (x), the corresponding model output in numerical weather prediction can be calculated according to equation (4), where the wind power forecast data P p is the wind power generation power forecast data obtained at the reference time.

[0046] In step S109, the model output f wind,m , model output in numerical weather prediction f m and wind-electricity conversion model error E under actual weather conditions wind,m The second target error E by wind-electricity conversion operation according to m Determine.

[0047] When only the error in the wind-power conversion model is considered, the second target error E due to the wind-power conversion operation is m , wind-electricity conversion model error E under actual weather conditions wind,m , the output of the wind-electricity conversion model when the input data is numerical weather forecast data and actual weather data, respectively. m , f wind,m The following relationship exists: JPEG0007634115000008.jpg542Formula(8)

[0048] In step S110, the wind power generation forecast data P p , wind power plant actual output power P m , first target error E r and the second target error E m Depending on the numerical weather forecast operation, the third target error E n Determine.

[0049] According to formula (1), the third target error E n can be calculated.

[0050] According to the formulas in the above steps, an error decoupling equation can be constructed, JPEG0007634115000009.jpg4674 Using this error decoupling equation, the first target error E r , second target error E m , third target error E n can be calculated.

[0051] In step S111, the first target error E r , second target error E m and the third target error E n The ratio of each target error is determined according to the

[0052] The percentage of each target error can be calculated according to formula (9), JPEG0007634115000010.jpg4562Formula(9) Among them, R n , R m , R r are the proportions of the third target error, the second target error, and the first target error, respectively.

[0053] After determining the proportions of the multiple target errors, the method may further include: analyzing the operations that cause errors among the multiple operations in the wind power forecasting process according to the proportions of the multiple target errors, i.e., analyzing according to the proportions of each target error, and determining the target error corresponding to any of the prediction result correction operation, the wind-to-electricity conversion operation, and the numerical weather forecast operation in the wind power forecasting process as the main error. For example, R n , R m , R r Compare and R n If is the maximum value, the numerical weather prediction is the cause of the error, and R m If is the maximum value, the wind-electricity conversion model is a source of error, and R r If is a maximum value, the prediction result correction is the cause of the error.

[0054] The input meteorological data affects the training of the actual forecast model of the wind power plant, which in turn affects the output results of the "wind-electricity conversion model", which in turn affects the correction of the forecast results, so the error (second target error) E due to the wind-electricity conversion operation in the numerical weather forecast is simply m and wind-electricity conversion model error E under actual weather conditions wind,m and equivalent (i.e., E m =E wind,m ) and the error due to the correction operation of the forecast results in the numerical weather forecast (first target error) E r and the error E of the correction operation of the prediction result under actual weather conditions wind,r and equivalent (i.e., E r =E wind,r ) simply, the results of analyzing the error in such cases are not sufficiently reasonable.

[0055] In one specific example of the present application, the data from August 2021 to December 2021 of the Baiyunxian Wind Power Station in Hunan Chinzhou (113.5576°E, 25.3931°N) is based on the sampling interval time of 15 minutes, and a total of 96 samplings are performed per day. Table 1 shows the error decoupling results and error rates in this specific example.

[0056] [Table 1]

[0057] An embodiment of the present application provides a processor configured to perform the method for determining a wind power forecast error according to any of the embodiments described above.

[0058] The processor obtains data required for wind power prediction error analysis from the historical operation data of the wind power station, including wind power station actual measured weather data, numerical weather forecast data, installed capacity, real-time start-up capacity, planned start-up capacity, wind power prediction data, and wind power station actual output power, determines a first target error through a prediction result correction operation, performs fitting on the wind-power conversion model of the wind power station according to the required data, builds a wind-power conversion fitting model, inputs the wind power station actual measured weather data into the wind-power conversion fitting model to obtain model output under actual weather conditions, determines wind power prediction power under actual weather conditions according to the installed capacity, planned start-up capacity, and model output under actual weather conditions, and calculates the real-time start-up capacity, planned start-up capacity, and wind power prediction under actual weather conditions. The method may be configured to execute the following methods: determine a correction error of the prediction result under actual weather conditions according to the measured power; determine a wind-power conversion model error under actual weather conditions according to the wind power prediction power under actual weather conditions, the actual output power of the wind farm, and the correction error of the prediction result under actual weather conditions; obtain a model output in the numerical weather forecast corresponding to the model output under real weather conditions; determine a second target error due to the wind-power conversion operation according to the model output under real weather conditions, the model output in the numerical weather forecast, and the wind-power conversion model error under real weather conditions; determine a third target error due to the numerical weather forecast operation according to the wind power prediction data, the actual output power of the wind farm, the first target error, and the second target error; and determine the proportion of each target error according to the first target error, the second target error, and the third target error.

[0059] In an embodiment of the present application, the processor may be further configured to execute the following method: analyzing an operation that causes an error in the wind power forecasting process according to the proportions of the multiple target errors.

[0060] In an embodiment of the present application, determining the first target error by the prediction result correction operation includes: determining an equivalent wind power prediction value according to the wind power prediction data, the real-time startup capacity and the planned startup capacity; and subtracting the equivalent wind power prediction value from the wind power prediction data to obtain the first target error.

[0061] In an embodiment of the present application, fitting the wind-power conversion model of a wind power plant according to required data to construct a wind-power conversion fitting model includes: determining the output of the wind-power conversion model according to the installed capacity, planned start-up capacity, and wind power generation forecast data; using numerical weather forecast data as the input of the wind-power conversion model, and forming a training set and a test set together with the output of the wind-power conversion model; using XGBoost to train on the training set to fit the wind-power conversion model; after training is completed, using the test set to determine whether the fitted wind-power conversion model meets the accuracy requirement; modifying the initial parameters of XGBoost and repeating training on the training set until the fitted wind-power conversion model meets the accuracy requirement; and taking the fitted wind-power conversion model that meets the accuracy requirement as the wind-power conversion fitting model.

[0062] In an embodiment of the present application, the wind-power conversion model includes one of a wind-power conversion model based on a BP neural network, a wind-power conversion model based on a LSTM neural network, and a wind-power conversion model based on a CNN-LSTM hybrid neural network.

[0063] In an embodiment of the present application, determining a predicted wind power power under real weather conditions according to an installed capacity, a planned start-up capacity, and a model output under real weather conditions includes calculating a predicted wind power power under real weather conditions according to the following formula: JPEG0007634115000012.jpg1031 Among them, P wind is the predicted wind power under actual weather conditions, C0' is the planned start-up capacity, C is the installed capacity, and f wind,m is the model output under real weather conditions.

[0064] In an embodiment of the present application, determining a correction error of the prediction result under the actual weather conditions according to the real-time starting capacity, the planned starting capacity, and the predicted wind power generation power under the actual weather conditions includes calculating a correction error of the prediction result under the actual weather conditions according to the following formula: JPEG0007634115000013.jpg1249 Among them, E wind,r is the prediction result correction error under actual weather conditions, C0 is the real-time start capacity, C0' is the planned start capacity, and P wind is the predicted wind power generation under actual weather conditions.

[0065] In an embodiment of the present application, determining a wind-power conversion model error under real weather conditions according to a wind power forecast power, an actual output power of a wind power station, and a forecast result correction error under real weather conditions includes calculating a wind-power conversion model error under real weather conditions according to the following formula: JPEG0007634115000014.jpg865 Among them, P wind is the predicted wind power under actual weather conditions, and P m is the actual output power of the wind power plant, and E wind,m is the wind-electricity conversion model error under actual weather conditions, and E wind,r is the correction error of the forecast result under actual weather conditions.

[0066] In an embodiment of the present application, determining a second target error for wind-power conversion according to a model output under real weather conditions, a model output in numerical weather forecasting, and a wind-power conversion model error under real weather conditions includes calculating the second target error according to the following formula: JPEG0007634115000015.jpg753 Among them, Em is the second target error, and f m is the model output in numerical weather prediction, and E wind,m is the wind-electricity conversion model error under actual weather conditions, and f wind,m is the model output under real weather conditions.

[0067] In an embodiment of the present application, determining a third target error by a numerical weather forecast operation according to the wind power forecast data, the wind farm actual output power, the first target error, and the second target error includes calculating the third target error according to the following formula: JPEG0007634115000016.jpg641 Among them, P p is the wind power forecast data, and E r is the first target error, E m is the second target error, E n is the third target error.

[0068] The present application considers the change of each operation error in the wind power forecasting process after the conditions such as weather data change, and improves the reliability of the wind power forecast error analysis method.

[0069] An embodiment of the present application provides a machine-readable storage medium having stored thereon instructions which, when executed by a processor, cause the processor to implement the method for determining wind power forecast error according to any of the embodiments described above.

[0070] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in FIG. 2. The computer device includes a processor A01, a network interface A02, a memory (not shown), and a database (not shown) connected by a system bus. The processor A01 of the computer device is configured to provide calculation and control capabilities. The memory of the computer device includes an internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 in the non-volatile storage medium A04. The database of the computer device is used to store data, such as historical meteorological measurement data, numerical weather forecast data, wind power plant installed capacity, real-time startup capacity, planned startup capacity, wind power generation power forecast data, actual output power, etc. The network interface A02 of the computer device is configured to communicate with an external terminal through a network connection. When the computer program B02 is executed by the processor A01, it realizes a method for determining a wind power forecast error.

[0071] An embodiment of the present application provides an apparatus including a processor, a memory, and a program stored in the memory and operable by the processor, and when the processor executes the program, the method for determining a wind power forecast error of any of the above-mentioned embodiments is realized.

[0072] The above-mentioned technical aspects of the embodiments of the present application are based on the prediction results under actual weather conditions and numerical weather forecast conditions, and comprehensively consider the impact on each operation when weather conditions change, thereby accurately decoupling multiple operations that cause errors, and calculating the error proportion of each operation based on this, so as to provide theoretical guidance for improving the prediction method of wind power plant.

[0073] According to the wind power forecast flow, the wind farm power forecast process can be divided into three key operations: numerical weather forecast, wind-power conversion model, and forecast result correction. Changes in meteorological conditions will not only affect numerical weather forecast, but also affect the wind-power conversion model and forecast result correction.

[0074] The wind power generation forecasting method of the embodiment of the present application includes:

[0075] (1) Regarding the error of the numerical weather forecast operation, the embodiment of the present application takes into account the influence of the wind-power conversion operation and the prediction result correction operation on the error of the numerical weather forecast operation in the process of solving for the predicted power.

[0076] (2) For the error of the wind-power conversion operation, the embodiment of the present application solves based on the equation (Equation 8) related to the wind-power conversion model error.

[0077] (3) The embodiment of the present application additionally supplements the calculation formula of the error of the prediction result correction operation under actual weather conditions in order to improve the error decoupling method.

[0078] Therefore, the technical aspects of the embodiments of the present application comprehensively consider the error change characteristics of multiple operations, supplement the estimation of the error change of the wind-electricity conversion operation and the prediction result correction operation under actual weather conditions, and perform a more comprehensive and reasonable decoupling analysis of the errors of multiple operations in wind power forecasting, and the obtained error analysis result can provide more reliable guidance for the improvement and optimization of wind power forecasting technology. Those skilled in the art should understand that the embodiments of the present application may be a method, a system, or a computer program product. Therefore, the present application may adopt the form of an all-hardware embodiment, an all-software embodiment, or an embodiment in which software and hardware aspects are combined. Furthermore, the present application may adopt the form of a computer program product embodied in a computer usable storage medium (including, but not limited to, a magnetic disk memory, a portable compact magnetic disk read-only memory (CD-ROM), an optical memory, etc.) having computer usable program code thereon.

[0079] The present application will be described with reference to flow diagrams and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flow diagrams and / or block diagrams, and combinations of flows and / or blocks in the flow diagrams and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, a special purpose computer, an embedded processor, or other programmable data processing device to generate a machine, whereby the instructions executed by the processor of the computer or other programmable data processing device generate an apparatus configured to implement the function specified in one or more flows in the flow diagrams and / or one or more blocks in the block diagrams.

[0080] These computer program instructions may be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner such that the instructions stored in the computer readable memory produce an article of manufacture that includes an instruction apparatus that implements the functions specified in one or more flows of the flow diagrams and / or one or more blocks of the block diagrams.

[0081] These computer program instructions may be loaded onto a computer or other programmable data processing device, causing the computer or other programmable device to execute a series of operational steps to generate a computer-implemented process, such that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows of the flow diagrams and / or one or more blocks of the block diagrams.

[0082] In one arrangement, a computing device includes one or more processors, such as a Central Processing Unit (CPU), input / output interfaces, network interfaces, and internal memory.

[0083] Memory may include forms of computer readable media, such as non-persistent memory, such as Random Access Memory (RAM), and / or non-volatile internal memory, such as Read-Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer readable medium.

[0084] Computer-readable media includes persistent and non-persistent, portable and non-portable media and may implement storage of information in any manner or technology. Information may be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, Phase-change Random Access Memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read-Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Flash storage or other internal memory technology, portable Compact Magnetic Disk Read-Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, magnetic cased magnetic tape, magnetic tape magnetic disk storage or other magnetic storage device or any other non-transmission medium that can be used to store information accessible by a computing device. As defined herein, computer readable media does not include transitory computer readable media, such as modulated data signals and carrier waves.

[0085] The terms "comprising," "including," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or device comprising a set of elements includes not only those elements, but also other elements not expressly listed or inherent in such process, method, article, or device. In the absence of further limitations, an element qualified by the phrase "comprising a..." does not exclude the presence of further identical elements in the process, method, article, or device that includes the element.

Claims

1. The dispatching server acquires data including wind power plant actual measured weather data, numerical weather forecast data, installed capacity, real-time start capacity, planned start capacity, wind power generation power forecast data, and wind power plant actual output power from the historical operation data stored in the wind power plant; A dispatching side server determines a first target error by a prediction result correction operation in a wind power prediction process; A dispatching side server performs fitting to a wind-power conversion model of the wind power plant according to the acquired data to construct a wind-power conversion fitting model; A dispatching side server inputs the wind power plant actual measured weather data into the wind-power conversion fitting model to obtain a model output under actual weather conditions; A dispatching side server determines a wind power generation forecast power under actual weather conditions according to the installed capacity, the planned start-up capacity, and a model output under the actual weather conditions; A dispatching side server determines a correction error of a prediction result under actual weather conditions according to the real-time activation capacity, the planned activation capacity, and the wind power generation prediction power under the actual weather conditions; A dispatching side server determines a wind-power conversion model error under the actual weather conditions according to the wind power generation forecast power under the actual weather conditions, the wind power plant actual output power, and the forecast result correction error under the actual weather conditions; A dispatching server obtains a model output in a numerical weather forecast corresponding to the model output under the actual weather conditions; A dispatching side server determines a second target error due to wind-power conversion operation in the wind power forecast process according to a model output under the actual weather conditions, a model output in the numerical weather forecast, and a wind-power conversion model error under the actual weather conditions; A dispatching side server determines a third target error by a numerical weather forecast operation in the wind power forecast process according to the wind power forecast data, the wind power actual output power, the first target error, and the second target error; A dispatching side server determines a ratio value of each target error in the sum of the first target error, the second target error, and the third target error according to the first target error, the second target error, and the third target error. A method for determining wind power forecast errors.

2. The method further includes determining a target error corresponding to any one of the prediction result correction operation, the wind-electricity conversion operation, and the numerical weather forecast operation in the wind power generation power prediction process as a main error according to a ratio value of each target error. The method of claim 1.

3. The above-mentioned determining a first target error by a prediction result correction operation in the wind power prediction process includes: determining an equivalent wind power forecast value according to the wind power forecast data, the real-time startup capacity, and the planned startup capacity; and subtracting the equivalent wind power forecast value from the wind power forecast data to obtain the first target error. The method of claim 1.

4. The above-mentioned fitting to the wind-power conversion model of the wind power plant according to the acquired data to construct a wind-power conversion fitting model is Determining an output of a wind-power conversion model according to the installed capacity, the planned startup capacity, and the wind power generation forecast data; forming a training set and a test set using the numerical weather forecast data as an input to a wind-electricity conversion model and an output of the wind-electricity conversion model; training the training set using XGBoost to fit a wind-power conversion model; After completing the training, using the test set to determine whether the fitted wind-power conversion model meets an accuracy requirement; and Modifying the initial parameters of XGBoost and repeating training on the training set until the fitted wind-electricity conversion model meets the accuracy requirement; and setting the fitted wind-electricity conversion model that satisfies the accuracy requirement as the wind-electricity conversion fitting model. The method of claim 1.

5. The wind-electricity conversion model is The wind-electricity conversion model includes one of a wind-electricity conversion model based on a feedforward BP neural network, a wind-electricity conversion model based on a long short-term memory LSTM neural network, and a wind-electricity conversion model based on a convolution-long short-term memory CNN-LSTM hybrid neural network. The method according to claim 4.

6. Determining a predicted wind power generation power under actual weather conditions according to the installed capacity, the planned startup capacity, and the model output under the actual weather conditions, Calculating a forecasted wind power generation power under actual weather conditions according to the following formula: P ｗｉｎｄ is the predicted wind power generation under actual weather conditions, and C ０ ' is the planned startup capacity, C is the installed capacity, and f ｗｉｎｄ，ｍ is the model output under the actual weather conditions, The method of claim 1.

7. Determining a correction error for a prediction result under actual weather conditions according to the real-time startup capacity, the planned startup capacity, and the predicted wind power generation power under the actual weather conditions, Calculating a correction error of the forecast result under actual weather conditions according to the following formula: E ｗｉｎｄ，ｒ is the correction error of the prediction result under actual weather conditions, and C ０ is the real-time starting capacity, C ０ ' is the planned startup capacity, P ｗｉｎｄ is the predicted wind power generation power under the actual weather conditions, The method according to claim 6.

8. Determining the wind-power conversion model error under the actual weather conditions according to the wind power generation forecast power under the actual weather conditions, the wind power plant actual output power, and the forecast result correction error under the actual weather conditions, Calculating the wind-electricity conversion model error under real weather conditions according to the following formula: P ｗｉｎｄ is the predicted wind power generation power under the actual weather conditions, and P ｍ is the actual output power of the wind power plant, and E ｗｉｎｄ，ｍ is the wind-electricity conversion model error under actual weather conditions, and E ｗｉｎｄ，ｒ is the correction error of the forecast result under actual weather conditions, The method according to claim 7.

9. Determining a second target error due to wind-power conversion in the wind power forecast process according to the model output under the actual weather conditions, the model output in the numerical weather forecast, and the wind-power conversion model error under the actual weather conditions, calculating the second target error according to the following formula: E ｍ is the second target error, and f ｍ is the model output in the numerical weather forecast, and E ｗｉｎｄ，ｍ is the wind-electricity conversion model error under the actual weather conditions, and f ｗｉｎｄ，ｍ is the model output under the actual weather conditions, The method according to claim 8.

10. Determining a third target error by a numerical weather forecast operation in the wind power forecast process according to the wind power forecast data, the wind power plant actual output power, the first target error, and the second target error, calculating the third target error according to the following formula: P ｐ is the wind power generation forecast data, and E ｒ is the first target error, E ｍ is the second target error, E ｎ is the third target error, 10. The method of claim 9.

11. 11. A method for determining a wind power forecast error, comprising: Processor.

12. A method for determining a wind power forecast error comprising: storing instructions, the instructions being stored on a processor, the instructions being stored on a processor when executed by the processor causing the processor to implement a method for determining a wind power forecast error according to any one of claims 1 to 10; A machine-readable storage medium.

Images