Data-experience fusion method and system for power prediction of wind farm under low temperature condition

Abstract

The application provides a data-experience fusion wind farm low-temperature condition power prediction method and system, and relates to the wind farm power prediction technical field. The method comprises the following steps: establishing an experience-based wind farm engineering wake field model; based on the model, establishing a wind farm low-temperature condition power prediction method based on the wind farm engineering wake field model; using historical data to optimize the coefficient in the wind farm low-temperature condition power prediction method, and establishing a data-experience fusion wind farm low-temperature condition power prediction method; and using the established data-experience fusion wind farm low-temperature condition power prediction method to predict the wind farm power. The application considers the influence of temperature on the wind farm engineering wake field model, and corrects the prediction deviation of the wind farm engineering wake field model under the influence of temperature according to the actual data, thereby effectively reducing the power prediction deviation of the wind farm under the low-temperature condition and improving the wind farm power prediction precision.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wind farm power prediction, in particular to a data-experience fusion wind farm power prediction method under low temperature conditions. BACKGROUND

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

[0003] The advantage of data-driven method is based on machine learning and deep learning technology, which can process a large amount of meteorological data, including wind speed, temperature, humidity and other multi-dimensional information, to capture complex wind speed patterns. It can automatically learn and adapt to wind speed prediction patterns in different environments, including seasonal and geographical differences. It provides high prediction accuracy, especially in long-term and short-term prediction. However, its shortcomings are also obvious, it needs a large amount of historical meteorological data for training, and there may be a problem of data scarcity in some areas and time periods. Complex data-driven models require a large amount 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 empirical formula method is based on physical principles or empirical laws, which can usually provide fast and simple wind speed prediction without the need for a large amount of training data. In some specific situations, such as stable weather conditions, empirical formula method may perform well. It is usually more suitable for short-term wind speed prediction. The disadvantage is that empirical formula method may not be accurate for complex weather and geographical conditions, as it ignores the interaction between multi-dimensional data. Moreover, it lacks adaptability and cannot automatically learn or adapt to changes in different environments. Its accuracy is low in long-term and special conditions.

[0005] In summary, there is an urgent need for a wind power prediction method with high response speed and high accuracy. SUMMARY

[0006] In view of the problems and deficiencies of the above-mentioned two existing technologies, the purpose of the present application is to provide a data-experience fusion wind farm power prediction method and system under low temperature conditions, which can retain the advantages of data-driven method and empirical formula method, make the best of both worlds, improve prediction accuracy and make it applicable to most time resolution conditions.

[0007] To achieve the above-mentioned purpose of the application, the embodiments of the present application provide a data-experience fusion wind farm power prediction method under low temperature conditions, which comprises:

[0008] Step S1: establishing an empirical wind farm engineering wake field model;

[0009] Step S2: based on the model, establishing a wind farm low-temperature condition power prediction method based on the wind farm engineering wake field model;

[0010] Step S3: using historical data, optimizing the coefficients in the wind farm low-temperature condition power prediction method, and establishing a data-experience fusion-based wind farm low-temperature condition power prediction method, specifically including:

[0011] S31: collecting wind farm actual data, including wind farm temperature, wind speed, wind direction, and power data, and pre-processing and feature engineering processing the actual data;

[0012] S32: based on the actual data after pre-processing and feature engineering processing, establishing a data-based wind farm low-temperature condition temperature, wind speed, and wind direction prediction method;

[0013] S33: initializing undetermined coefficients, and using different undetermined coefficients to establish an empirical wind farm low-temperature condition power prediction method;

[0014] S34: according to historical data, selecting the optimal coefficient combination to establish a data-experience fusion-based wind farm low-temperature condition power prediction method;

[0015] S4: using the established data-experience fusion-based wind farm low-temperature condition power prediction method to predict the wind farm power.

[0016] Preferably, in step S1, the empirical wind farm engineering wake field model is established, specifically including:

[0017] Establish a single-machine wake model:

[0018]

[0019] 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, C T is the thrust coefficient of the wind turbine, is the initial wake diameter, is the wake growth rate, D is the wind turbine rotor diameter, z h is the wind turbine hub height, and β is the ratio of the initial wake area to the wind wheel swept area;

[0020] Establish a wake superposition model:

[0021]

[0022] wherein u n+1 represents the wind speed at the target wind turbine, u0 represents the inflow wind speed of the wind farm, and n represents the number of upstream wind turbines.

[0023] Preferably, in step S2, based on the model, a wind farm power prediction method under low temperature conditions based on a wind farm engineering wake field model is established, specifically including:

[0024] The single-machine wake model is corrected according to the ambient temperature:

[0025] Under low temperature conditions, the wake of the wind turbine will change, and the wake growth rate is used for correction:

[0026] k * = a0+ a1T+ a2T 2 +…+ a n-1 T n-1 +a n T n

[0027] wherein k * is the wake growth rate, T is the ambient Celsius temperature, a0, a1, …, a n are undetermined coefficients, and n is the highest order of the ambient Celsius temperature;

[0028] A wind speed-thrust coefficient curve is established:

[0029] C T =f(v)

[0030] wherein C T is the thrust coefficient, and f(v) is a function of the thrust coefficient with respect to the wind speed;

[0031] A wind speed-power curve is established:

[0032] P=g(v)

[0033] wherein P is the output power of the wind turbine, and g(v) is a function of the output power of the wind turbine with respect to the wind speed.

[0034] Preferably, in step S31, the initial actual data is preprocessed and feature engineered, specifically including:

[0035] First, the initial actual data is preprocessed: a series of intrinsic mode decomposition functions are formed by hierarchical decomposition for different characteristic scales, and the original signal is reconstructed by linear superposition after analyzing each component, wherein the data preprocessing relationship is as follows:

[0036]

[0037] where x(t) is an original data sequence, q i (t) is the i-th eigenmode component, r n (t) is a residual component.

[0038] Then, the preprocessed actual data is subjected to low-temperature feature engineering processing: designing and creating features related to power prediction under low-temperature conditions, specifically including:

[0039] Creating temperature-related features, including average temperature, minimum temperature, and maximum temperature.

[0040] Considering time features, including time of day and day of week.

[0041] Considering time lag features, including power data at past time points, and considering the lagging effect of temperature changes on power.

[0042] Preferably, in step S32, a data-based temperature, wind speed, and wind direction prediction method under low-temperature conditions for a wind farm is established, specifically including:

[0043] Using an LSTM recurrent neural network, a data-driven power prediction model under low-temperature conditions for a wind farm is established, where the relationship of the prediction model is as follows:

[0044] i t =σ(W xi x t +W hi h t-1 +W ci c t-1 +b i )

[0045] f t =σ(W xf x t +W hf h t-1 +W cf c t-1 +b f )

[0046] c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c )

[0047] o t =σ(W xo x t +W ho ht-1 +W co c t +b o )

[0048] h t =o t tanh(c t )

[0049] Where: i is the input gate, f is the forget gate, c is the cell state, o is the output gate, W is the weight coefficient matrix, x is the given sequence, h is the hidden layer sequence, t is the time, b is the corresponding bias term, σ is the sigmoid activation function, and tanh is the hyperbolic tangent activation function.

[0050] Preferably, in step S33, initializing the undetermined coefficients specifically includes:

[0051] Value range: wake growth rate k * The value range of is determined based on experience, and the value range of the ambient temperature T is determined based on the meteorological conditions of the area where the wind farm is located;

[0052] Change rule: According to the reality that wind turbine power increases as the ambient temperature decreases, and the wake growth rate formula, the wake growth rate should increase when the ambient temperature decreases.

[0053] According to the above value range and change rules, the initial value of the unknown coefficient is determined:

[0054]

[0055] Where a0, a1, ..., a n is the unknown coefficient, n is the highest order term of the ambient temperature in the wake growth rate formula, and in its subscript, the first parameter 0, 1, 2, …, n represents the different order terms in the wake growth rate formula, and the second parameter 1, 2, …, m represents different schemes, where m is the total number of schemes.

[0056] Preferably, in step S34, based on historical data, an optimal coefficient combination is selected to establish a wind farm power prediction method under low temperature conditions based on data-experience fusion, specifically including:

[0057] 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. The output is a wind farm power time series with the same length as the prediction set and obtained from different undetermined coefficient schemes. The result is compared with the wind farm power of 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.

[0058] The embodiment of the present application also provides a data-experience fusion wind farm low-temperature condition power prediction system, which is used for realizing the data-experience fusion wind farm low-temperature condition power prediction method and specifically comprises the following steps:

[0059] A wind farm engineering wake field model establishing module is used for establishing an experience-based wind farm engineering wake field model.

[0060] A prediction method establishing module based on the wind farm engineering wake field model is used for establishing a wind farm low-temperature condition power prediction method based on the wind farm engineering wake field model based on the model.

[0061] A data-experience-based prediction method establishing module is used for optimizing coefficients in the wind farm low-temperature condition power prediction method by using historical data, and establishing a data-experience fusion wind farm low-temperature condition power prediction method, and specifically comprises the following steps:

[0062] A data collection preprocessing unit is used for collecting wind farm actual data including wind farm temperature, wind speed, wind direction and power data, and pre-processing and feature engineering processing the actual data.

[0063] A first prediction method establishing unit is used for establishing a data-based wind farm low-temperature condition temperature, wind speed and wind direction prediction method based on the actual data after pre-processing and feature engineering processing.

[0064] A second prediction method establishing unit is used for initializing undetermined coefficients and establishing an experience-based wind farm low-temperature condition power prediction method by using different undetermined coefficients.

[0065] A third prediction method establishing unit is used for selecting an optimal coefficient combination according to historical data, and establishing a data-experience fusion wind farm low-temperature condition power prediction method.

[0066] A wind farm power prediction module is used for predicting wind farm power by using the established data-experience fusion wind farm low-temperature condition power prediction method.

[0067] The embodiment of the present application also provides an electronic device, which comprises a processor, a memory and a bus system, the processor and the memory are connected through the bus system, the memory is used for storing instructions, and the processor is used for executing the instructions stored in the memory to realize the data-experience fusion wind farm low-temperature condition power prediction method.

[0068] The embodiment of the present application also provides a computer storage medium, which stores a computer software product, and the computer software product comprises a plurality of instructions to make a computer device execute the data-experience fusion wind farm power prediction method under low temperature conditions.

[0069] From the above technical solutions, the present application has the following advantages:

[0070] The embodiment of the present application provides a data-experience fusion wind farm power prediction method under low temperature conditions and a system thereof, the data-driven method can capture complex patterns and long-term trends to provide more accurate input data for the experience formula method; the experience formula method can provide fast short-term prediction to make up for the shortcomings of the data-driven method in some cases; the present application combines the data-driven method and the experience formula method, can retain the advantages of the data-driven method and the experience formula method, and avoid their shortcomings, can improve the comprehensive wind speed prediction performance, especially in complex environments. The present application considers the influence of temperature on the wind farm engineering wake field model, and corrects the prediction deviation of the wind farm engineering wake field model under the influence of temperature according to actual data, thereby effectively reducing the power prediction deviation of the wind farm under low temperature conditions and improving the wind farm power prediction accuracy. The present application determines the optimal fusion strategy and coefficient through in-depth research and experiment, which is of great significance for improving the efficiency and reliability of the wind farm and the accuracy of the weather prediction. BRIEF DESCRIPTION OF DRAWINGS

[0071] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly described below. The features and advantages of the present application can be more clearly understood by referring to the drawings. The drawings are schematic and should not be understood as any limitation on the present application. Those skilled in the art can obtain other drawings without creative labor on the basis of these drawings. Among them:

[0072] Figure 1 A flowchart of the data-experience fusion wind farm power prediction method provided in the embodiment;

[0073] Figure 2 A wind turbine thrust coefficient curve of the present application;

[0074] Figure 3 A wind turbine power curve of the present application;

[0075] Figure 4 A temperature prediction data graph of the wind farm under low temperature conditions based on data of the present application;

[0076] Figure 5A data-based wind speed prediction data graph under low-temperature conditions of a wind farm according to the present application;

[0077] Figure 6 A data-based wind direction prediction data graph under low-temperature conditions of a wind farm according to the present application;

[0078] Figure 7 A root mean square error graph of the experience-based wind farm power prediction method according to the various schemes of the present application;

[0079] Figure 8 A comparison graph of the power prediction data and the actual data under low-temperature conditions of a wind farm according to the optimal scheme of the present application;

[0080] Figure 9 A block diagram of a data-experience fusion wind farm power prediction system under low-temperature conditions according to an embodiment. DETAILED DESCRIPTION

[0081] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in conjunction with the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0082] Embodiment One

[0083] As shown in Figure 1 , the present application proposes a data-experience fusion wind farm power prediction method under low-temperature conditions, which specifically includes:

[0084] Step S1: Establishing an experience-based wind farm engineering wake field model;

[0085] Step S2: Based on the model, establishing a wind farm power prediction method under low-temperature conditions based on the wind farm engineering wake field model;

[0086] Step S3: Using historical data to optimize the coefficients in the wind farm power prediction method under low-temperature conditions, and establishing a data-experience fusion wind farm power prediction method under low-temperature conditions, which specifically includes:

[0087] S31: Collecting wind farm actual data, including temperature, wind speed, wind direction and power data in the wind farm, and pre-processing and feature engineering processing the actual data;

[0088] S32: Based on the actual data after pre-processing and feature engineering processing, establishing a data-based temperature, wind speed and wind direction prediction method under low-temperature conditions of a wind farm;

[0089] S33: initialize the pending coefficient, and use different pending coefficients to establish an empirical-based power prediction method for wind farms under low temperature conditions;

[0090] S34: according to historical data, select the optimal coefficient combination, and establish a data-experience fusion-based power prediction method for wind farms under low temperature conditions;

[0091] S4: using the established data-experience fusion-based power prediction method for wind farms under low temperature conditions, the power of the wind farm is predicted.

[0092] From the above technical solution, the present application provides a data-experience fusion-based power prediction method for wind farms under low temperature conditions, which establishes an empirical-based wind farm engineering wake field model; establishes a wind farm low temperature condition power prediction method based on the wind farm engineering wake field model; uses historical data to optimize the coefficients in the wind farm low temperature condition power prediction method, and establishes a data-experience fusion-based wind farm low temperature condition power prediction method. The present application considers the influence of temperature on the wind farm engineering wake field model, and corrects the prediction deviation of the wind farm engineering wake field model under the influence of temperature according to the actual data, thereby effectively reducing the power prediction deviation of the wind farm under low temperature conditions and improving the wind farm power prediction accuracy.

[0093] In this embodiment, in step S1, an empirical-based wind farm engineering wake field model is established.

[0094] (1) Establish a single machine wake model:

[0095]

[0096] In the formula, u0 is the wind speed at the hub center of the wind turbine, Δu(x, y, z) is the speed reduction at the position (x, y, z) behind the wind turbine, C T is the thrust coefficient of the wind turbine, is the initial wake diameter, is the wake growth rate, 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 wind wheel swept area;

[0097] (2) Establish a wake superposition model:

[0098] In a wind farm, the downwind wind turbine is affected by the wake of the upstream wind turbines. The wake of the wind turbines is evaluated to obtain the flow field model of the wind farm. There are four main wake superposition models, namely, square and superposition model, geometric and superposition model, linear superposition model, and energy conservation model. Using any one of the wake superposition models, the linear superposition model is taken as an example in the present application:

[0099]

[0100] In the formula, u n+1 represents the wind speed at the target wind turbine, u0 represents the inflow wind speed of the wind farm, and n represents the number of upstream wind turbines.

[0101] It should be understood that the linear superposition model is only used to illustrate the present application and is not used to limit the scope of the present application. The use of other forms of wake superposition model or various equivalent modifications of the wake superposition model falls within the scope defined by the claims.

[0102] In the present embodiment, in step S2, based on the model, a wind farm power prediction method under low temperature conditions based on the wake field model of the wind farm project is established.

[0103] (1) The above single-machine wake model is modified according to the ambient temperature:

[0104] Under low temperature conditions, the wake of the wind turbine will change, and the wake growth rate is used for correction:

[0105] k * =a0+a1T+a2T 2 +…+a n-1 T n-1 +a n T n

[0106] In the formula, k * is the wake growth rate, T is the ambient temperature in Celsius, a0, a1, …, a n are undetermined coefficients, and n is the highest order of the ambient temperature in Celsius.

[0107] (2) Establish the wind speed-thrust coefficient curve:

[0108] In the single-machine wake model, the thrust coefficient C T can be determined by the wind speed at the hub center of the wind turbine.

[0109] The wind turbine power curve is closely related to the selection of the wind turbine. For a specific model, the wind turbine manufacturer will provide the thrust coefficient of the model corresponding to each wind speed between the cut-in wind speed and the cut-out wind speed:

[0110] CT = f(v)

[0111] where C T is the thrust coefficient and f(v) is a function of the thrust coefficient with respect to the wind speed, usually provided by the wind turbine manufacturer. The results are shown in Fig. 2. Figure 2

[0112] (3) Establish the wind speed-power curve:

[0113] The output power of each wind turbine in the wind farm can be determined by the wind speed obtained from the single-turbine wake model.

[0114] For a specific model, the wind turbine manufacturer will provide the power of the model corresponding to each wind speed between the cut-in wind speed and the cut-out wind speed:

[0115] P = g(v)

[0116] where P is the output power of the wind turbine and g(v) is a function of the output power of the wind turbine with respect to the wind speed, usually provided by the wind turbine manufacturer. The results are shown in Fig. 3. Figure 3

[0117] In this embodiment, in step S3, the coefficients in the power prediction method under low temperature conditions of the wind farm are optimized using historical data, and a data-experience fusion-based power prediction method under low temperature conditions of the wind farm is established, which specifically includes:

[0118] S31: Collect actual data of the wind farm, including temperature, wind speed, wind direction, and power data in the wind farm, and perform preprocessing and feature engineering processing on the actual data.

[0119] Specifically, historical power data includes power output of the wind farm at different time points, which is usually actual operation data. Meteorological data, including temperature, wind speed, wind direction, etc. are obtained from meteorological stations, satellite data or other sources.

[0120] First, the initial actual data is preprocessed: for different characteristic scales, a series of intrinsic mode decomposition functions (IMF) are formed by step-by-step decomposition, and after analyzing each component, the original signal can be reconstructed by linear superposition. The sequence after decomposition has stronger regularity compared to the original power sequence, which is beneficial to improve the prediction accuracy. The data preprocessing relationship is as follows:

[0121]

[0122] where x(t) is an original data sequence, q i (t) is the i-th intrinsic mode component, and r n (t) is a residual component.

[0123] ​​Then the pre-processed actual data is processed by low temperature feature engineering. Feature engineering is to extract useful features from raw data for the prediction model: design and create features related to power prediction under low temperature conditions, including:

[0124] Create temperature-related features, including average temperature, minimum temperature, maximum temperature;

[0125] Consider time features, including time of day, day of week;

[0126] Consider time lag features, including power data at past time points, and consider the lagging effect of temperature change on power.

[0127] S32: Based on the pre-processed and feature-engineered actual data, a data-based temperature, wind speed, and wind direction prediction method under low temperature conditions in a wind farm is established.

[0128] Specifically, a data-driven power prediction model under low temperature conditions in a wind farm is established using an LSTM recurrent neural network, where the relationship of the prediction model is as follows:

[0129] i t =σ(W xi x t +W hi h t-1 +W ci c t-1 +b i )

[0130] f t =σ(W xf x t +W hf h t-1 +W cf c t-1 +b f )

[0131] c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c )

[0132] o t =σ(W xo x t +W ho h t-1 +W co c t +b o)

[0133] h t =o t tanh(c t )

[0134] Where: i is the input gate, f is the forget gate, c is the cell state, o is the output gate, W is the weight coefficient matrix, x is the given sequence, h is the hidden layer sequence, t is the time, b is the corresponding bias term, σ is the sigmoid activation function, and tanh is the hyperbolic tangent activation function. The prediction results of temperature, wind speed, and wind direction are shown in the attached figure. Figure 4 , Attachment Figure 5 , Attachment Figure 6 shown.

[0135] S33: Initialize the unknown coefficients, and use different unknown coefficients to establish an experience-based power prediction method for wind farms under low temperature conditions.

[0136] Furthermore, the undetermined coefficients are initialized, specifically including:

[0137] Value range: In the single-machine wake model, the wake growth rate k * The value range of is determined based on experience; in the wake growth rate formula, the value range of the ambient temperature T in Celsius is determined based on the meteorological conditions of the area where the wind farm is located.

[0138] Change rule: According to the reality that the power of wind turbines increases as the ambient temperature decreases, and the wake growth rate formula, the wake growth rate should increase when the ambient temperature decreases.

[0139] According to the above value range and change rules, the initial value of the unknown coefficient is determined:

[0140]

[0141] Where a0, a1, ..., a n is the unknown coefficient, n is the highest order term of the ambient temperature in the wake growth rate formula, and in its subscript, the first parameter 0, 1, 2, …, n represents the different order terms in the wake growth rate formula, and the second parameter 1, 2, …, m represents different schemes, where m is the total number of schemes.

[0142] Specifically, taking n=1 as an example, 30 schemes are established as follows:

[0143]

[0144] Where p01, p02,…, p30 are the scheme numbers.

[0145] It should be understood that the described initial value scheme of the pending coefficient is only used to illustrate the present application and is not used to limit the scope of the present application, and other initial value schemes of the pending coefficient or various equivalent modifications of the initial value scheme of the pending coefficient all fall within the scope defined by the claims.

[0146] Further, using different pending coefficients, an empirical-based wind farm power prediction method is established:

[0147] The different pending coefficient schemes are input into the wind farm low-temperature power prediction method based on the wind farm engineering wake field model to obtain the predicted wind speed time series:

[0148]

[0149] In the formula, u0(t) is the wind speed time series at the hub center of the wind turbine, Δu(x, y, z, t) is the speed reduction time series at the position (x, y, z) behind the wind turbine, C T is the thrust coefficient of the wind turbine, is the initial wake diameter, a0, a1, …, a n is the pending 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.

[0150] Further, in the wind farm, the downwind wind turbine is affected by the wake of multiple upstream wind turbines, and the wake effect of multiple wind turbines is evaluated, that is, the wind farm flow field model is obtained. There are currently four main wake superposition models, namely square and superposition model, geometric and superposition model, linear superposition model, and energy conservation model. Using any one of the wake superposition models, taking the linear superposition model as an example:

[0151]

[0152] In the formula, u n+1 (t) represents the wind speed at the target wind turbine, u0(t) represents the inflow wind speed of the wind farm, n 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.

[0153] It should be understood that the described linear superposition model is only used to illustrate the present application and is not used to limit the scope of the present application, and other forms of wake superposition model or various equivalent modifications of the wake superposition model all fall within the scope defined by the claims.

[0154] S34: According to the historical data, the optimal coefficient combination is selected to establish a data-empirical fusion-based wind farm low-temperature power prediction method.

[0155] Specifically, 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 of 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.

[0156] To evaluate the approximation of the model, the root mean square error (RMSE) can be used to measure it.

[0157]

[0158] Where x t y t The approximation coefficient of .

[0159] The RMSE of each solution is shown in the attached Figure 7 As shown in Figure 2, the RMSE of Scheme 21 is the lowest, and Scheme 21 is selected as the optimal coefficient combination. According to Scheme 21, a power prediction method for wind farms under low temperature conditions based on data-experience fusion is established. The comparison between the wind farm power prediction data and the actual data is shown in the attached figure. Figure 8 .

[0160] In this embodiment, in step S4, the wind farm power is predicted using the established wind farm power prediction method under low temperature conditions based on data-experience fusion.

[0161] Example 2

[0162] like Figure 9 As shown, the present invention provides a data-experience fusion wind farm power prediction system under low temperature conditions. The system is used to implement the data-experience fusion wind farm power prediction method under low temperature conditions described in Example 1, specifically including:

[0163] A wind farm engineering tail flow field model establishment module 10 is used to establish a wind farm engineering tail flow field model based on experience;

[0164] A prediction method establishment module 20 based on a wind farm engineering tail flow field model is used to establish a wind farm power prediction method under low temperature conditions based on the wind farm engineering tail flow field model based on the model;

[0165] A data-experience-based prediction method establishment module 30 is established to utilize historical data to optimize the coefficients in the wind farm power prediction method under low temperature conditions, and to establish a wind farm power prediction method under low temperature conditions based on data-experience fusion, specifically including:

[0166] The data collection preprocessing unit 31 is configured to collect actual data of the wind farm, including temperature, wind speed, wind direction and power data in the wind farm, and to pre-process and perform feature engineering on the actual data.

[0167] The first prediction method establishment unit 32 is configured to establish a data-based temperature, wind speed and wind direction prediction method for the wind farm under low-temperature conditions based on the pre-processed and feature-engineered actual data.

[0168] The second prediction method establishment unit 33 is configured to initialize undetermined coefficients and establish an experience-based power prediction method for the wind farm under low-temperature conditions using different undetermined coefficients.

[0169] The third prediction method establishment unit 34 is configured to select an optimal coefficient combination based on historical data and establish a data-experience fusion-based power prediction method for the wind farm under low-temperature conditions.

[0170] The wind farm power prediction module 40 is configured to predict the power of the wind farm using the established data-experience fusion-based power prediction method for the wind farm under low-temperature conditions.

[0171] The data-experience fusion-based power prediction system for the wind farm under low-temperature conditions of the present embodiment is used to implement the aforementioned data-experience fusion-based power prediction method for the wind farm under low-temperature conditions, and thus the specific embodiments of the data-experience fusion-based power prediction system for the wind farm under low-temperature conditions can be seen from the aforementioned embodiment part of the data-experience fusion-based power prediction method for the wind farm under low-temperature conditions. For example, the wind farm engineering wake field model establishment module 10, the prediction method establishment module 20 based on the wind farm engineering wake field model, the prediction method establishment module 30 based on data-experience, and the wind farm power prediction module 40 are respectively used to implement steps S1, S2, S3 and S4 of the aforementioned data-experience fusion-based power prediction method for the wind farm under low-temperature conditions. Therefore, the specific embodiments can refer to the description of the corresponding respective part embodiments, and in order to avoid redundancy, they will not be described here.

[0172] Embodiment three

[0173] The embodiment of the present application also provides an electronic device, which comprises a processor, a memory and a bus system, the processor and the memory are connected through the bus system, the memory is used to store instructions, and the processor is used to execute the instructions stored in the memory to realize the data-experience fusion-based power prediction method for the wind farm under low-temperature conditions.

[0174] Embodiment four

[0175] The embodiment of the present application further provides a computer storage medium, which stores a computer software product, and the computer software product comprises a plurality of instructions to enable a computer device to execute the data-experience fusion wind farm power prediction method under low-temperature conditions.

[0176] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a 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.) containing computer usable program codes.

[0177] The present application is described with reference to flowcharts and / or block diagrams of the method, device (system), and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts 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 devices to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that carries out the functions specified in one or more flows and / or blocks.

[0178] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing devices to work in a specific manner, so that the instructions stored in the computer readable memory produce a product including an instruction apparatus that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that carries out the functions specified in one or more flows and / or blocks. These computer program instructions can also be loaded into a computer or other programmable data processing device to cause a series of operation steps to be performed on the computer or other programmable data processing device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing device provide a process for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that carries out the functions specified in one or more flows and / or blocks.

[0179] Obviously, the above embodiments are merely example for clearly illustrating, and are not limitation to the embodiments. For ordinary skilled in the art, other different forms of changes or variations can be made on the basis of the above description. Here, all the embodiments need not and can not be exhausted. The obvious changes or variations derived therefrom are still within the scope of the present invention.

Claims

1. A data-experience fusion method for predicting wind farm power under low temperature conditions, characterized in that: include: Step S1: Establishing a wind farm engineering wake field model based on experience; Step S2: Based on the model, a method for predicting wind farm power under low temperature conditions based on a wind farm engineering tail flow field model is established, specifically comprising: 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 Where k * is the wake growth rate, T is the ambient temperature in Celsius, a0, a1, …, a n is the unknown coefficient, n is the highest order term of the ambient temperature in Celsius; Create a wind speed-thrust coefficient curve: C T =f(v) Where C T is the thrust coefficient, f(v) is the function of the thrust coefficient with respect to wind speed; Create a wind speed-power curve: P=g(v) Where P is the output power of the wind turbine, g(v) is the function of the output power of the wind turbine with respect to wind speed; Step S3: Utilizing historical data, optimize the coefficients in the wind farm power prediction method under low temperature conditions, and establish a wind farm power prediction method under low temperature conditions based on data-experience fusion, specifically including: S31: Collect actual data of the wind farm, including temperature, wind speed, wind direction, and power data within the wind farm, and perform preprocessing and feature engineering on the actual data; S32: Based on the actual data after preprocessing and feature engineering, establish a data-based prediction method for temperature, wind speed, and wind direction under low temperature conditions in wind farms; S33: Initializing the undetermined coefficients and using different undetermined coefficients to establish an experience-based power prediction method for wind farms under low temperature conditions, wherein the initialized undetermined coefficients specifically include: Value range: wake growth rate k * The value range of is determined based on experience, and the value range of the ambient temperature T is determined based on the meteorological conditions of the area where the wind farm is located; Change rule: According to the reality that wind turbine power increases as the ambient temperature decreases, and the wake growth rate formula, the wake growth rate should increase when the ambient temperature decreases. According to the above value range and change rules, the initial value of the unknown coefficient is determined: Where a0, a1, ..., a n is the unknown coefficient, n is the highest order term of the ambient temperature in the wake growth rate formula, and in its subscript, the first parameter 0, 1, 2, ..., n represents the different order terms in the wake growth rate formula, and the second parameter 1, 2, ..., m represents different schemes, where m is the total number of schemes; S34: Based on historical data, select the optimal coefficient combination and establish a wind farm power prediction method under low temperature conditions based on data-experience fusion; S4: Use the established wind farm power prediction method under low temperature conditions based on data-experience fusion to predict the wind farm power.

2. The data-experience fusion method for predicting wind farm power under low temperature conditions according to claim 1 is characterized in that: In step S1, an experience-based wind farm engineering wake field model is established, specifically including: To build 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 (x,y,z) position behind the wind turbine, and C T is the thrust coefficient of the wind turbine, 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 x, 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; Build a wake superposition model: Where u n+1 represents the wind speed at the target wind turbine, u0 represents the inflow wind speed of the wind farm, and n represents the number of upstream wind turbines.

3. The data-experience fusion method for predicting wind farm power under low temperature conditions according to claim 1 is characterized in that: In step S31, the initial actual data is preprocessed and feature-engineered, specifically including: First, the initial actual data is preprocessed: it is decomposed step by step according to different characteristic scales to form a series of intrinsic mode decomposition functions. After analyzing each component, it can be linearly superimposed and reconstructed to form the original signal. The data preprocessing relationship is as follows: In the formula, x(t) is an original data sequence, q i (t) is the i-th eigenmode component, r n (t) is a residual component; The pre-processed data is then subjected to low-temperature feature engineering: features related to power prediction under low-temperature conditions are designed and created, including: Create temperature-related features, including average temperature, minimum temperature, and maximum temperature; Consider temporal features, including time of day and day of the week; Consider time-lag characteristics, including power data at past time points, and consider the lagged effects of temperature changes on power.

4. The data-experience fusion method for predicting wind farm power under low temperature conditions according to claim 1 is characterized in that: In step S32, a method for predicting temperature, wind speed, and wind direction under low temperature conditions of a wind farm based on data is established, specifically including: Using the LSTM recurrent neural network, a data-driven power prediction model for wind farms under low temperature conditions is established. The relationship between the prediction model is as follows: i t =σ(W xi x t +W hi h t-1 +W ci c t-1 +b i ) f t =σ(W xf x t +W hf h t-1 +W cf c t-1 +b f ) c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c ) o t =σ(W xo x t +W ho h t-1 +W co c t +b o ) h t =o t fishy t ) Where: i is the input gate, f is the forget gate, c is the cell state, o is the output gate, W is the weight coefficient matrix, x is the given sequence, h is the hidden layer sequence, t is the time, b is the corresponding bias term, σ is the sigmoid activation function, and tanh is the hyperbolic tangent activation function.

5. The data-experience fusion method for predicting wind farm power under low temperature conditions according to claim 1 is characterized in that: In step S34, based on historical data, an optimal coefficient combination is selected to establish a wind farm power prediction method under low temperature conditions based on data-experience fusion, specifically including: 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. The output is a wind farm power time series with the same length as the prediction set and obtained from different undetermined coefficient schemes. The result is compared with the wind farm power of 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.

6. A data-experience fusion wind farm power prediction system under low temperature conditions, characterized by: The system is used to implement the data-experience fusion power prediction method for wind farms under low temperature conditions as described in any one of claims 1 to 5, specifically comprising: Wind farm engineering wake field model establishment module, used to establish a wind farm engineering wake field model based on experience; A prediction method establishment module based on a wind farm engineering tail flow field model is used to establish a wind farm power prediction method under low temperature conditions based on the wind farm engineering tail flow field model based on the model; A data-experience-based prediction method establishment module is established to use historical data to optimize the coefficients in the wind farm power prediction method under low temperature conditions, and to establish a wind farm power prediction method under low temperature conditions based on data-experience fusion, specifically including: The data collection and preprocessing unit is used to collect actual data of the wind farm, including temperature, wind speed, wind direction, and power data within the wind farm, and perform preprocessing and feature engineering on the actual data; A first prediction method establishment unit is used to establish a data-based prediction method for temperature, wind speed, and wind direction under low temperature conditions in a wind farm based on actual data that has been preprocessed and feature-engineered; The second prediction method establishment unit is used to initialize the undetermined coefficients and use different undetermined coefficients to establish an experience-based power prediction method for wind farms under low temperature conditions; The third prediction method establishment unit is used to select the optimal coefficient combination based on historical data and establish a power prediction method for wind farms under low temperature conditions based on data-experience fusion; The wind farm power prediction module is used to predict the wind farm power by using the established wind farm power prediction method under low temperature conditions based on data-experience fusion.

7. An electronic device, characterized in that: The electronic device includes a processor, a memory and a bus system, the processor and the memory are connected through the bus system, the memory is used to store instructions, and the processor is used to execute the instructions stored in the memory to implement the data-experience fusion power prediction method for wind farms under low temperature conditions as described in any one of claims 1 to 5.

8. A computer storage medium, characterized in that The computer storage medium stores a computer software product, which includes several instructions for enabling a computer device to execute the data-experience fusion power prediction method for wind farms under low temperature conditions as described in any one of claims 1 to 5.

Images