Wind power plant low-temperature condition power prediction method based on data-experience fusion

Abstract

The invention discloses a data-experience fusion-based power prediction method for a wind power plant under a low-temperature condition. The method comprises the steps of establishing an experience-based wind power plant engineering wake flow field model; establishing a wind power plant low-temperature condition power prediction model based on the wind power plant engineering wake flow field model; and utilizing historical data to optimize coefficients in the power prediction model of the wind power plant under the low-temperature condition, establishing a data-experience fusion-based power prediction model of the wind power plant under the low-temperature condition, and utilizing the data-experience fusion-based power prediction model of the wind power plant under the low-temperature condition to predict the power of the wind power plant. According to the method, the influence of the temperature on the wind power plant engineering wake flow field model is considered, and the prediction deviation of the wind power plant engineering wake flow field model under the temperature influence is corrected according to the actual data, so that the power prediction deviation of the wind power plant under the low-temperature condition is effectively reduced, and the power prediction precision of the wind power plant is improved.

Description

Technical Field

[0001] The present invention belongs to the field of wind farm power prediction, and in particular relates to a wind farm power prediction method, and more specifically to a wind farm power prediction method under low temperature conditions based on data-experience fusion. Background Art

[0002] Wind power forecasting is of great significance in the wind energy industry and meteorology. Accurate wind power forecasting helps the operation of wind power plants, and also plays a key role in weather forecasting, storm warning, etc. Wind power forecasting can usually be performed using two main methods: data-driven method and empirical formula method.

[0003] The advantage of the data-driven approach is that it is based on machine learning and deep learning technology, and can process a large amount of meteorological data, including multi-dimensional information such as wind speed, temperature, and humidity, to capture complex wind speed patterns. It can automatically learn and adapt to wind speed prediction patterns in different environments, including seasonality, geographical differences, etc. It provides higher prediction accuracy, especially in long-term and short-term predictions. However, its disadvantages are also obvious. It requires a large amount of historical meteorological data for training, and there may be data scarcity in certain regions and time periods. Complex data-driven models require a lot of computing resources and time for training. For special cases, such as rapidly changing storms, data-driven methods may not provide sufficient response speed.

[0004] The advantage of the empirical formula method is that it is based on physical principles or empirical laws, and can usually provide fast and simple wind speed forecasts without requiring a large amount of training data. In some specific cases, such as stable meteorological conditions, the empirical formula method may perform well. It is usually more suitable for short-term wind speed forecasts. Its disadvantage is that the empirical formula method may not be very accurate for complex meteorological and geographical conditions because it ignores the interaction between multidimensional data. It is also not adaptable and cannot automatically learn or adapt to changes in different environments. The accuracy is low in the long term and under special conditions. Summary of the invention

[0005] Purpose of the invention: In view of the problems and shortcomings of the above-mentioned two existing technologies, the purpose of the present invention is to provide a power prediction method for wind farms under low temperature conditions based on data-experience fusion, so as to effectively reduce the power prediction deviation under low temperature conditions of wind farms and improve the power prediction accuracy of wind farms.

[0006] Technical solution: To achieve the above-mentioned invention object, the technical solution adopted by the present invention is:

[0007] In a first aspect, the present application provides a method for predicting power under low temperature conditions of a wind farm based on data-experience fusion, comprising:

[0008] Acquire meteorological data within a period of history, wherein the meteorological data includes temperature, wind speed, and wind direction at a corresponding time point;

[0009] Perform eigenmode decomposition on temperature, wind speed and wind direction respectively to obtain a series of corresponding eigenmode components and residual components;

[0010] According to a series of intrinsic mode components and residual components of temperature, wind speed and wind direction in the previous time period, they are input into the LSTM recurrent neural network respectively to obtain a series of intrinsic mode components and residual components of the predicted values ​​of temperature, wind speed and wind direction in the next time period;

[0011] According to a series of intrinsic mode components and residual components of the temperature, wind speed and wind direction prediction values, linear superposition and reconstruction are performed to obtain the temperature, wind speed and wind direction prediction values ​​for the next time period;

[0012] According to the predicted values ​​of temperature, wind speed and wind direction, the power prediction model of the wind farm under low temperature conditions based on data-experience fusion is used to obtain the predicted power value of the wind farm at the corresponding time point; wherein the method for constructing the power prediction model of the wind farm under low temperature conditions based on data-experience fusion includes:

[0013] 1) Establish an experience-based wind farm engineering wake field model;

[0014] 2) Based on the wind farm engineering tail flow field model, a power prediction model for the wind farm under low temperature conditions is established;

[0015] 3) Utilizing historical meteorological data and corresponding historical power data, the coefficients in the power prediction model of the wind farm under low temperature conditions are optimized to obtain a power prediction model of the wind farm under low temperature conditions based on data-experience fusion.

[0016] In a second aspect, the present application provides a device for predicting power under low temperature conditions of a wind farm based on data-experience fusion, including a processor and a storage medium;

[0017] The storage medium is used to store instructions;

[0018] The processor is configured to operate according to the instructions to execute the method according to the first aspect.

[0019] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the method described in the first aspect when executed by a processor.

[0020] In a fourth aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the method of the first aspect when executing the computer program.

[0021] Beneficial effects: The power prediction method and device for wind farms under low temperature conditions based on data-experience fusion provided in this application have the following advantages: The current research status of wind power prediction shows the advantages and limitations of two main methods. However, research combining these two methods, that is, the fusion of data-driven method and empirical formula method, has potential advantages: (1) The data-driven method can capture complex patterns and long-term trends, providing more accurate input data for the empirical formula method. (2) The empirical formula method can provide fast short-term predictions to make up for the shortcomings of the data-driven method in some cases. (3) Combining the two can improve the comprehensive wind speed prediction performance, especially in complex environments.

[0022] Therefore, the research combining data-driven method and empirical formula method has potential value in improving the accuracy and robustness of wind power forecasting, but in-depth research and experiments are needed to determine the best fusion strategy and coefficients. Research in this field is of great significance for improving the efficiency and reliability of wind farms and the accuracy of meteorological forecasts. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 The present invention is a flow chart of a method for predicting power under low temperature conditions in a wind farm based on data-experience fusion according to an embodiment of the present invention.

[0024] Figure 2 This is a thrust coefficient curve diagram of a wind turbine generator system according to an embodiment of the present invention.

[0025] Figure 3 This is a power curve diagram of a wind turbine generator system according to an embodiment of the present invention.

[0026] Figure 4 This is a temperature prediction data diagram under low temperature conditions of a wind farm based on data in an embodiment of the present invention.

[0027] Figure 5 This is a data diagram of wind speed prediction under low temperature conditions in a wind farm based on data in an embodiment of the present invention.

[0028] Figure 6 This is a data diagram of wind direction prediction under low temperature conditions in a wind farm based on data in an embodiment of the present invention.

[0029] Figure 7 Root mean square error diagram of wind farm power prediction methods based on experience in various schemes in the embodiments of the present invention.

[0030] Figure 8 This is a comparison chart between predicted power data and actual power data of a wind farm under low temperature conditions in the optimal solution in an embodiment of the present invention. DETAILED DESCRIPTION

[0031] The present invention is further explained below in conjunction with the accompanying drawings and specific examples. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0032] In the description of this application, "several" means more than one, "more" means more than two, "greater than", "less than", "exceed", etc. are understood to exclude the number itself, and "above", "below", "within", etc. are understood to include the number itself. If there is a description of "first" or "second", it is only used for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0033] In the description of the present application, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0034] The term "and / or" is only a description of the association relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " generally indicates that the related objects are in an "or" relationship.

[0035] Embodiment 1: This embodiment provides a method for predicting power under low temperature conditions of a wind farm based on data-experience fusion, such as Figure 1 As shown, including:

[0036] Acquire meteorological data within a period of history, wherein the meteorological data includes temperature, wind speed, and wind direction at a corresponding time point;

[0037] Perform eigenmode decomposition on temperature, wind speed and wind direction respectively to obtain a series of corresponding eigenmode components and residual components;

[0038] According to a series of intrinsic mode components and residual components of temperature, wind speed and wind direction in the previous time period, they are input into the LSTM recurrent neural network respectively to obtain a series of intrinsic mode components and residual components of the predicted values ​​of temperature, wind speed and wind direction in the next time period;

[0039] According to a series of intrinsic mode components and residual components of the temperature, wind speed and wind direction prediction values, linear superposition and reconstruction are performed to obtain the temperature, wind speed and wind direction prediction values ​​for the next time period;

[0040] According to the predicted values ​​of temperature, wind speed and wind direction, the power prediction model of the wind farm under low temperature conditions based on data-experience fusion is used to obtain the predicted power value of the wind farm at the corresponding time point; wherein the method for constructing the power prediction model of the wind farm under low temperature conditions based on data-experience fusion includes:

[0041] 1) Establish an experience-based wind farm engineering wake field model;

[0042] 2) Based on the wind farm engineering tail flow field model, a power prediction model for the wind farm under low temperature conditions is established;

[0043] 3) Utilizing historical meteorological data and corresponding historical power data, the coefficients in the power prediction model of the wind farm under low temperature conditions are optimized to obtain a power prediction model of the wind farm under low temperature conditions based on data-experience fusion.

[0044] In some embodiments, the method for constructing a power prediction model for a wind farm under low temperature conditions based on data-experience fusion includes:

[0045] 1) Establishing an experience-based wind farm engineering wake field model:

[0046] ① Establish a single-machine wake model:

[0047]

[0048] Where u0 is the wind speed at the center of the wind turbine hub, Δu(x,y,z) is the speed reduction at the position (x,y,z) behind the wind turbine, and x, y, z are the coordinates of the X, Y, and Z axes in the relative coordinate system, respectively; C T is the thrust coefficient of the wind turbine, ε is the initial wake diameter, is the wake growth rate, is the standard deviation of the Gaussian distribution velocity reduction σ that changes with the downstream distance w; D is the wind turbine rotor diameter, z h is the hub height of the wind turbine, and β is the ratio of the initial wake area to the swept area of ​​the wind rotor. ② Establish a wake superposition model:

[0049] In a wind farm, the downwind wind turbines are affected by the wakes of multiple wind turbines upstream. By evaluating the wake effects of multiple wind turbines, the wind farm flow model can be obtained. There are currently four main wake superposition models, namely the square sum superposition model, the geometric sum superposition model, the linear superposition model, and the energy conservation model. Using any wake superposition model, take the linear superposition model as an example:

[0050]

[0051] Where u e+1 represents the wind speed at the target wind turbine, u0 represents the inflow wind speed of the wind farm, e represents the number of upstream wind turbines, Δu i Represents the speed reduction at the i-th wind turbine.

[0052] It should be understood that the linear superposition model is only used to illustrate the present invention and is not used to limit the scope of the present invention. The use of other forms of wake superposition models or modifications to various equivalent forms of the wake superposition model are all within the scope defined by the claims.

[0053] 2) Establishing a power prediction method for a wind farm under low temperature conditions based on the model described in step 1):

[0054] ① The tail flow field model of the wind farm project is modified according to the ambient temperature:

[0055] Under low temperature conditions, the wake of the wind turbine will change, and the wake growth rate correction is used:

[0056] k * =a0+a1T+a2T 2 +…+a n-1 T n-1 +a n T n (4)

[0057] Where k * is the wake growth rate, T is the ambient temperature in Celsius, a0, a1, …, a n is the unknown coefficient, which will be determined in step 3), and n is the highest order term of the ambient temperature in Celsius in equation (4).

[0058] ② Wind speed-thrust coefficient curve:

[0059] In formula (1), the thrust coefficient C T It can be determined by the wind speed at the center of the wind turbine hub.

[0060] The wind turbine power curve is closely related to the selection of wind turbines. For a specific model, the wind turbine manufacturer will provide the thrust coefficient of the model at each wind speed between the cut-in wind speed and the cut-out wind speed:

[0061] C T =f(u) (5)

[0062] Where C T is the thrust coefficient, and f(u) is the function of the thrust coefficient with respect to wind speed u, which is usually provided by the wind turbine manufacturer. Figure 2 shown.

[0063] ③ Wind speed-power curve:

[0064] The output power of each wind turbine in the wind farm can be determined by the wind speed obtained by equation (1).

[0065] For a specific model, the wind turbine manufacturer will provide the power of the model at various wind speeds between the cut-in wind speed and the cut-out wind speed:

[0066] P=g(u) (6)

[0067] Where P is the output power of the wind turbine, g(u) is the function of the output power of the wind turbine with respect to the wind speed u, which is usually provided by the wind turbine manufacturer. Figure 3 shown.

[0068] 3) Using historical data, optimize the coefficients in the method described in step 2) and establish a power prediction method for wind farms under low temperature conditions based on data-experience fusion:

[0069] ① Preprocessing of temperature, wind speed, wind direction and power data in wind farms based on actual data:

[0070] Historical power data, including the power output of the wind farm at different time points, is usually actual operating data. Meteorological data, including temperature, wind speed, wind direction and other information, is obtained from meteorological stations, satellite data or other sources. The initial actual data is preprocessed and decomposed step by step according to different characteristic scales to form a series of intrinsic mode functions (IMFs). After analyzing each component, it can be linearly superimposed and reconstructed to form the original signal. The decomposed sequence has a stronger regularity than the original sequence of meteorological data, which is conducive to improving the prediction accuracy. The data preprocessing relationship in this system is as follows:

[0071]

[0072] In the formula, x(t) is an original data sequence, q j (t) is the jth eigenmode component, r d (t) is a residual component.

[0073] Then, the actual data after preprocessing is processed by low-temperature feature engineering. Feature engineering is to extract useful features from the original data for use in the prediction model. Design and create features related to power prediction under low-temperature conditions. Specifically, it includes:

[0074] (1) Create temperature-related features, such as average temperature, minimum temperature, maximum temperature, etc.

[0075] (2) Consider temporal features, such as time of day, day of the week, etc.

[0076] (3) Consider time-lag characteristics, such as power data at past time points, to account for the lagged effect of temperature changes on power.

[0077] ② Establish a data-based method for predicting temperature, wind speed and wind direction under low temperature conditions in wind farms

[0078] According to ① the collection and processing of initial data and feature engineering, the LSTM recurrent neural network is used to establish a data-driven power prediction model under low temperature conditions of wind farms. The prediction model relationship in this system is as follows:

[0079] i t =σ(W xi x t +W hi h t-1 +W ci c t-1 +b i ) (8)

[0080] ft=σ(Wxfxt+Whfht-1+Wcfct-1+bf) (9)

[0081] c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c ) (10)

[0082] o t =σ(W xo x t +W ho h t-1 +W co c t +b o ) (11)

[0083] h t =o t tanh(c t ) (12)

[0084] Where: i t is the input gate, f t is the forget gate, c t is the cell state, o t is the output gate, W xi , W hi , W ci , W xf , W hf , W cf , Wxc , W hc , W xo , W ho , W co is the weight coefficient matrix, x t For a given sequence, h t is the hidden layer sequence, t is the time, b i 、b f 、b c 、b o is the corresponding bias term, σ is the sigmoid activation function, and tanh is the hyperbolic tangent activation function.

[0085] The prediction results of temperature, wind speed and wind direction are as follows: Figure 4 , Figure 5 , Figure 6 shown.

[0086] ③ Initialize the undetermined coefficients:

[0087] In formula (4), the range of values ​​of the ambient temperature T is determined by the meteorological conditions of the area where the wind farm is located. According to the reality that the power of wind turbines increases as the ambient temperature decreases, the wake growth rate should increase when the ambient temperature decreases. According to the above change rules, the initial value scheme of the unknown coefficient is determined:

[0088] Taking n=1 as an example, 30 plans are established.

[0089]

[0090] Where p01, p02,…, p30 are the scheme numbers.

[0091] It should be understood that the aforementioned initial value scheme for the undetermined coefficients is only used to illustrate the present invention and is not used to limit the scope of the present invention. The use of other initial value schemes for the undetermined coefficients or modifications to various equivalent forms of the initial value schemes for the undetermined coefficients are all within the scope defined by the claims.

[0092] ④Using different unknown coefficients, establish an experience-based wind farm power prediction method:

[0093] Input different undetermined coefficient schemes into the method described in step 2) to obtain the predicted wind speed time series:

[0094]

[0095] Where u0(t) is the wind speed time series at the center of the wind turbine hub, Δu(x,y,z,t) is the speed reduction time series at the position (x,y,z) behind the wind turbine, and C T is the thrust coefficient of the wind turbine, ε is the initial wake diameter, a0, a1, …, an is the unknown coefficient, D is the wind turbine rotor diameter, z h is the hub height of the wind turbine, and β is the ratio of the initial wake area to the rotor swept area.

[0096] In a wind farm, the downwind wind turbines are affected by the wakes of multiple wind turbines upstream. By evaluating the wake effects of multiple wind turbines, the wind farm flow model can be obtained. There are currently four main wake superposition models, namely the square sum superposition model, the geometric sum superposition model, the linear superposition model, and the energy conservation model. Using any wake superposition model, take the linear superposition model as an example:

[0097]

[0098] In the formula, +1 (t) represents the wind speed at the target wind turbine, u0(t) represents the inflow wind speed of the wind farm, e represents the number of upstream wind turbines, and Δu(x, y, z, t) is the speed reduction time series at the position (x, y, z) behind the wind turbine.

[0099] It should be understood that the linear superposition model is only used to illustrate the present invention and is not used to limit the scope of the present invention. The use of other forms of wake superposition models or modifications to various equivalent forms of the wake superposition model are all within the scope defined by the claims.

[0100] ⑤ According to historical data, select the optimal coefficient combination and establish a power prediction method for wind farms under low temperature conditions based on data-experience fusion

[0101] The predicted wind speed time series obtained by the data-based model trained by the training set wind speed is used as the input of the experience-based model, and the wind farm power time series obtained by different undetermined coefficient schemes with the same length as the prediction set is output. The result is compared with the wind farm power in the prediction set and the error is calculated. The undetermined coefficient scheme with the smallest error is selected as the determination coefficient, and a power prediction method for wind farms under low temperature conditions based on data-experience fusion is established.

[0102] To evaluate the approximation of the model, the root mean square error (RMSE) can be used to measure it.

[0103]

[0104] Where x t for y t The approximation coefficient of .

[0105] The RMSE of each solution is as follows Figure 7As shown in Figure 2, Scheme 21 has the lowest RMSE, so Scheme 21 is selected as the optimal coefficient combination. According to Scheme 21, a wind farm power prediction method under low temperature conditions based on data-experience fusion is established. The comparison between the predicted power data of the wind farm and the actual power data is as follows: Figure 8 .

[0106] Embodiment 2: Based on Embodiment 1, this embodiment provides a power prediction device for a wind farm under low temperature conditions based on data-experience fusion, including a processor and a storage medium;

[0107] The storage medium is used to store instructions;

[0108] The processor is used to operate according to the instruction to execute the method according to embodiment 1.

[0109] Embodiment 3: Based on Embodiment 1, this embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method described in Embodiment 1 is implemented.

[0110] Embodiment 4: Based on Embodiment 1, this embodiment provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the method described in Embodiment 1 when executing the computer program.

[0111] Embodiment 5: Based on Embodiment 1, this embodiment provides a computer program product, including a computer program, which implements the method described in Embodiment 1 when executed by a processor.

[0112] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0113] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram and the combination of the processes and / or blocks in the flowchart and / or block diagram 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, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0116] 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 wind farm power under low temperature conditions based on data-experience fusion, characterized in that: include: Acquire meteorological data within a period of history, wherein the meteorological data includes temperature, wind speed, and wind direction at a corresponding time point; Perform eigenmode decomposition on temperature, wind speed and wind direction respectively to obtain a series of corresponding eigenmode components and residual components; According to a series of intrinsic mode components and residual components of temperature, wind speed and wind direction in the previous time period, they are input into the LSTM recurrent neural network respectively to obtain a series of intrinsic mode components and residual components of the predicted values ​​of temperature, wind speed and wind direction in the next time period; According to a series of intrinsic mode components and residual components of the temperature, wind speed and wind direction prediction values, linear superposition and reconstruction are performed to obtain the temperature, wind speed and wind direction prediction values ​​for the next time period; According to the predicted values ​​of temperature, wind speed and wind direction, the power prediction model of the wind farm under low temperature conditions based on data-experience fusion is used to obtain the predicted power value of the wind farm at the corresponding time point; The method for constructing a power prediction model for a wind farm under low temperature conditions based on data-experience fusion includes: 1) Establish an experience-based wind farm engineering wake field model; 2) Based on the wind farm engineering tail flow field model, a power prediction model for the wind farm under low temperature conditions is established; 3) Utilizing historical meteorological data and corresponding historical power data, the coefficients in the power prediction model of the wind farm under low temperature conditions are optimized to obtain a power prediction model of the wind farm under low temperature conditions based on data-experience fusion.

2. The method for predicting power of a wind farm under low temperature conditions according to claim 1, characterized in that: In step 1), the wind farm engineering wake field model based on experience specifically includes: ① Establish a single-machine wake model: Where u0 is the wind speed at the center of the wind turbine hub, Δu(x,y,z) is the speed reduction at the position (x,y,z) behind the wind turbine, and x, y, z are the coordinates of the X, Y, and Z axes in the relative coordinate system, respectively; C T is the thrust coefficient of the wind turbine, ε is the initial wake diameter, k * is the wake growth rate; D is the wind turbine rotor diameter, z h is the hub height of the wind turbine, β is the ratio of the initial wake area to the rotor swept area; ② Establish a wake superposition model: In a wind farm, the downwind wind turbines are affected by the wakes of multiple upstream wind turbines. The wake impact of multiple wind turbines is evaluated to obtain the wind farm engineering wake field model; the wake superposition model is selected from at least one of the square sum superposition model, geometric sum superposition model, linear superposition model, and energy conservation model; using any wake superposition model, take the linear superposition model as an example: Where u e+1 represents the wind speed at the target wind turbine, u0 represents the inflow wind speed of the wind farm, e represents the number of upstream wind turbines, Δu i Represents the speed reduction at the i-th wind turbine.

3. The method for predicting power of a wind farm under low temperature conditions according to claim 2, characterized in that: The step 2) specifically includes: ① Modify the single-machine wake model according to the ambient temperature: Under low temperature conditions, the wake of the wind turbine will change, and the wake growth rate correction is used: k * =a0+a1T+a2T 2 +…+a n-1 T n-1 +a n T n (4) Where k * is the wake growth rate, T is the ambient temperature in Celsius, a0, a1, …, a n is the coefficient to be determined, n is the highest order term of the ambient temperature in Celsius in equation (4); ② Wind speed-thrust coefficient curve: Thrust coefficient C of wind turbine T Determined by the wind speed at the center of the wind turbine hub; C T =f(u) (5) Where C T is the thrust coefficient of the wind turbine, f(u) is the function of the thrust coefficient of the wind turbine with respect to the wind speed u; ③ Wind speed-power curve: The output power of each wind turbine in the wind farm is determined by the wind speed obtained by equation (1); P=g(u) (6) Where P is the output power of the wind turbine, and g(u) is the function of the output power of the wind turbine with respect to the wind speed u.

4. The method for predicting power of a wind farm under low temperature conditions according to claim 1, characterized in that: The step 3) specifically includes: Acquire historical power data and corresponding historical meteorological data, wherein the historical power data includes the real power output of the wind farm at different time points; the meteorological data includes the temperature, wind speed, and wind direction at the corresponding time points; Perform eigenmode decomposition on temperature, wind speed and wind direction respectively to obtain a series of corresponding eigenmode components and residual components; According to a series of intrinsic mode components and residual components of temperature, wind speed and wind direction in the previous time period, they are input into the LSTM recurrent neural network respectively to obtain a series of intrinsic mode components and residual components of the predicted values ​​of temperature, wind speed and wind direction in the next time period; According to a series of intrinsic mode components and residual components of the temperature, wind speed and wind direction prediction values, linear superposition and reconstruction are performed to obtain the temperature, wind speed and wind direction prediction values ​​for the next time period; According to the predicted values ​​of temperature, wind speed and wind direction, the power prediction model of the wind farm under low temperature conditions with different undetermined coefficient schemes is used to obtain the predicted power value of the wind farm at the corresponding time point; According to the calculation error between the wind farm power prediction value and the corresponding real power output, the undetermined coefficient scheme with the smallest error is selected as the determination coefficient, and the power prediction model of the wind farm under low temperature conditions based on data-experience fusion is obtained.

5. The method for predicting power of a wind farm under low temperature conditions according to claim 1 or 4, characterized in that: The eigenmode decomposition and linear superposition reconstruction are performed using the following formula: In the formula, x(t) is an original data sequence, q j (t) is the jth eigenmode component, there are d eigenmode components in total, r d (t) is a residual component. ② Establish a data-based method for predicting temperature, wind speed and wind direction under low temperature conditions in wind farms.

6. The method for predicting power of a wind farm under low temperature conditions according to claim 1 or 4, characterized in that: The LSTM recurrent neural network includes: i t =σ(W xi x t +W hi h t-1 +W ci c t-1 +b i ) (8) f t =σ(W xf x t +W hf h t-1 +W cf c t-1 +b f ) (9) c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c ) (10) o t =σ(W xo x t +W ho h t-1 +W co c t +b o ) (11) h t =o t fishy t ) (12) Where: i t is the input gate, f t is the forget gate, c t is the cell state, o t is the output gate, W xi , W hi , W ci , W xf , W hf , W cf , W xc , W hc , W xo , W ho , W co is the weight coefficient matrix, x t For a given sequence, h t is the hidden layer sequence, t is the time, b i 、b f 、b c 、b o is the corresponding bias term, σ is the sigmoid activation function, and tanh is the hyperbolic tangent activation function.

7. The method for predicting power of a wind farm under low temperature conditions according to claim 4, characterized in that: According to the predicted values ​​of temperature, wind speed and wind direction, the wind farm power prediction model under low temperature conditions with different undetermined coefficient schemes is used to obtain the wind farm power prediction value at the corresponding time point, including: Initialize the unknown coefficient: wake growth rate k * There is a range of values ​​determined based on experience. The range of values ​​considered for the ambient temperature T is determined according to the meteorological conditions of the area where the wind farm is located. According to the actual situation that the power of wind turbines increases as the ambient temperature decreases, the wake growth rate should increase when the ambient temperature decreases. Based on the above range of values ​​and the law of change, the initial value of the coefficient to be determined is determined as follows: Where a0, a1, ..., a n is the coefficient to be determined, n is the highest order of the ambient temperature in Celsius, in its subscript, the first parameter 0, 1, 2, ..., n represents different order terms, the second parameter 1, 2, ..., m represents different schemes, and m is the total number of coefficient schemes to be determined; According to the predicted values ​​of temperature, wind speed and wind direction, different undetermined coefficient schemes are input into the power prediction of the wind farm under low temperature conditions to obtain the predicted wind speed time series: Where u0(t) is the wind speed time series at the center of the wind turbine hub, Δu(x,y,z,t) is the speed reduction time series at the position (x,y,z) behind the wind turbine, and C T is the thrust coefficient of the wind turbine, ε is the initial wake diameter, a0, a1, …, a n is the unknown coefficient, D is the wind turbine rotor diameter, z h is the hub height of the wind turbine, β is the ratio of the initial wake area to the rotor swept area; According to the wind farm flow field model, the wind speed at each wind turbine is obtained: In the formula, +1 (t) represents the wind speed at the target wind turbine, u0(t) represents the inflow wind speed of the wind farm, e represents the number of upstream wind turbines, and Δu(x, y, z, t) is the speed reduction time series at the position (x, y, z) behind the wind turbine; Then, the power prediction value of each wind turbine is obtained according to the wind speed-power curve, thereby obtaining the power prediction value of the wind farm.

8. A power prediction device for wind farms under low temperature conditions based on data-experience fusion, characterized in that: including processor and storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the method according to any one of claims 1 to 7.

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

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images