Method, device and electronic equipment for generating wind power fluctuation interval prediction model

Abstract

The present application relates to a method, device and electronic equipment for generating a wind power fluctuation interval prediction model. The specific scheme is: obtaining historical wind power data of the wind farm; determining the power generation prediction function of the wind power fluctuation interval prediction model based on the historical wind power data; determining the power generation fluctuation interval prediction function based on the power generation prediction function and the preset confidence level; determining the power generation prediction function and the power generation fluctuation interval prediction function as the wind power fluctuation interval prediction model; inputting wind speed sample data and wind direction sample data into the wind power fluctuation interval prediction model to obtain power generation prediction data and power generation fluctuation interval prediction data; determining the target index value based on the power generation measured sample data, power generation prediction data and power generation fluctuation interval prediction data, and adjusting the parameters of the wind power fluctuation interval prediction model based on the target index value. The present application can obtain the wind power mean and interval prediction results.

Description

Technical Field

[0001] The present application relates to the technical field of new energy wind power generation prediction, and in particular to a method, device and electronic equipment for generating a wind power fluctuation interval prediction model. Background Art

[0002] In related technologies, the single-point prediction technology for wind farm output power is relatively mature. However, the uncertainty of wind energy and the inherent defects of the prediction model make it inevitable that there will be errors in the single-point prediction of wind power, and the prediction results cannot quantitatively describe the uncertainty of wind power. From the application level of wind power, the planning and safe and stable operation of power grids containing wind power require accurate estimation of the fluctuation range of wind power. In this case, a single-point fixed value prediction cannot meet the application requirements, and the significance of wind power fluctuation range prediction is highlighted. Mainstream wind power fluctuation range prediction methods such as quantile regression, kernel density estimation, and resampling methods are mostly based on existing wind power single-point prediction models superimposed on uncertainty analysis of the prediction results. They are mainly for post-evaluation of the model and cannot be directly used for model optimization. Summary of the Invention

[0003] To this end, the present application provides a method, device, and electronic device for generating a wind power fluctuation range prediction model. The technical solution of the present application is as follows:

[0004] According to a first aspect of an embodiment of the present application, a method for generating a wind power fluctuation interval prediction model is provided, the method comprising:

[0005] Acquire historical wind power data of the wind farm; wherein the historical wind power data includes wind speed sample data, wind direction sample data and generated power measured sample data;

[0006] Based on the historical wind power data, determining the power generation prediction function of the wind power fluctuation range prediction model according to the Bayesian rule;

[0007] Determining a power generation fluctuation interval prediction function based on the power generation prediction function and a preset confidence level;

[0008] Determining the generated power prediction function and the generated power fluctuation interval prediction function as a wind power fluctuation interval prediction model;

[0009] Inputting the wind speed sample data and the wind direction sample data into the wind power fluctuation interval prediction model to obtain the generated power prediction data and the generated power fluctuation interval prediction data output by the wind power fluctuation interval prediction model;

[0010] A target index value is determined based on the generated power measured sample data, the generated power forecast data and the generated power fluctuation interval forecast data, and the wind power fluctuation interval forecast model is adjusted based on the target index value.

[0011] According to one embodiment of the present application, determining the power generation prediction function of the wind power fluctuation interval prediction model based on the historical wind power data and according to the Bayesian rule includes:

[0012] Determining a first function based on the historical wind power data; wherein the first function is used to characterize the probability distribution of the historical wind power data of the wind farm;

[0013] Determine the covariance kernel function and mean function respectively;

[0014] Determining a Gaussian process function based on the covariance kernel function and the mean function;

[0015] According to the first function and the Gaussian process function, a power generation prediction function of the wind power fluctuation interval prediction model is determined based on the Bayesian rule.

[0016] According to one embodiment of the present application, the covariance kernel function is determined by the following method:

[0017] Determine a first kernel function, a second kernel function, and a third kernel function respectively; wherein the first kernel function is used to characterize the overall fluctuation characteristics of wind power, the second kernel function is used to characterize the change of wind power with wind speed variables, and the third kernel function is used to characterize the change of wind power with wind direction variables;

[0018] Adding the second kernel function to the third kernel function to obtain an intermediate function;

[0019] The intermediate function is multiplied by the first kernel function to obtain the covariance kernel function.

[0020] According to one embodiment of the present application, the second kernel function is obtained by adding a linear kernel function and a Brownian kernel function; and the first kernel function and the third kernel function are both Matérn kernel functions.

[0021] According to one embodiment of the present application, determining a target index value based on the measured power generation sample data, the power generation prediction data, and the power generation fluctuation interval prediction data, and adjusting parameters of the wind power fluctuation interval prediction model based on the target index value includes:

[0022] Based on the generated power measured sample data and the generated power predicted data, a root mean square error of the generated power predicted data is determined. The root mean square error is calculated using the following formula:

[0023]

[0024] Wherein, RMSE is the root mean square error of the power generation prediction data, P Mi is the actual power at the i-th moment, P Pi is the predicted power at the i-th moment, C i is the wind farm startup capacity at the i-th moment, and N is the number of moments;

[0025] Comparing the root mean square error with a first preset threshold to obtain a first comparison result;

[0026] In response to the first comparison result being that the root mean square error is greater than the first preset threshold, the wind power fluctuation interval prediction model is adjusted.

[0027] According to one embodiment of the present application, determining a target index value based on the measured power generation sample data, the power generation prediction data, and the power generation fluctuation interval prediction data, and adjusting parameters of the wind power fluctuation interval prediction model based on the target index value further includes:

[0028] Based on the power generation fluctuation interval value and the preset confidence level, a coverage index is determined. The coverage index value is calculated using the following formula:

[0029]

[0030]

[0031] in, is a 0-1 variable, 1-α is the confidence level, I i α is the predicted power fluctuation interval I at the i-th moment i α =[L i α ,U i α ], L i α is the lower limit of the power fluctuation range predicted at the i-th moment, U i α is the upper limit of the power fluctuation range predicted at the i-th moment, R c is the coverage index of the fluctuation interval prediction, and N is the number of moments;

[0032] Based on the power generation fluctuation interval value and the preset confidence level, an average relative width index value of the power generation fluctuation interval value is determined. The average relative width index value is calculated using the following formula:

[0033]

[0034] Among them, W mean α is the average relative width indicator of the fluctuation range prediction, 1-α is the confidence level, W i α is the predicted width of the fluctuation range at time i, W i α =U t α -L t α , C i is the wind farm operating capacity at time i, and N is the number of time moments;

[0035] In response to the average relative width index value not meeting a first preset requirement and / or the coverage index value not meeting a second preset requirement, the wind power fluctuation interval prediction model is adjusted.

[0036] According to one embodiment of the present application, determining the power generation fluctuation interval prediction function based on the power generation prediction function and a preset confidence level includes:

[0037] Determining a standard deviation function of the power generation prediction data based on the wind speed sample data, the wind direction sample data, the covariance kernel function, and the mean function;

[0038] Based on the standard deviation function, the mean function, and the power generation prediction data, a lower limit and an upper limit of the power generation fluctuation range are respectively determined to obtain the power generation fluctuation range prediction function; wherein the power generation fluctuation range prediction function is calculated by the following formula:

[0039] L i α =P Pi -zσ i

[0040] U i α =P Pi +zσ i

[0041] I i α =[L i α , U i α ] Among them, 1-α is the confidence level, the coefficient z is obtained from the normal distribution table, L i α is the lower limit of the power fluctuation range predicted at the i-th moment, U iα is the upper limit of the power fluctuation range predicted at the i-th moment, P Pi is the predicted power at the i-th moment, σ i is the standard deviation at the i-th moment, I i α is the power prediction interval at time i.

[0042] According to a second aspect of an embodiment of the present application, a device for generating a wind power fluctuation interval prediction model is provided, the device comprising:

[0043] An acquisition module is used to acquire historical wind power data of a wind farm; wherein the historical wind power data includes wind speed sample data, wind direction sample data and generated power measured sample data;

[0044] A first determining module is configured to determine, based on the historical wind power data and in accordance with the Bayesian rule, a power generation prediction function of the wind power fluctuation interval prediction model;

[0045] A second determining module is configured to determine a power generation fluctuation interval prediction function based on the power generation prediction function and a preset confidence level;

[0046] A third determining module is configured to determine the generated power prediction function and the generated power fluctuation interval prediction function as a wind power fluctuation interval prediction model;

[0047] an input module, configured to input the wind speed sample data and the wind direction sample data into the wind power fluctuation interval prediction model, and obtain the generated power prediction data and the generated power fluctuation interval prediction data output by the wind power fluctuation interval prediction model;

[0048] A parameter adjustment module is used to determine a target index value based on the measured sample data of the generated power, the generated power prediction data and the generated power fluctuation interval prediction data, and to adjust the parameters of the wind power fluctuation interval prediction model based on the target index value.

[0049] According to a third aspect of the embodiments of the present application, there is provided an electronic device, characterized in that it includes: a processor, and a memory communicatively connected to the processor;

[0050] The memory stores computer-executable instructions;

[0051] The processor executes the computer-executable instructions stored in the memory to implement the method as described in any one of the first aspects.

[0052] According to a fourth aspect of an embodiment of the present application, the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to implement the method as described in any one of the first aspects when executed by a processor.

[0053] The technical solutions provided by the embodiments of this application bring at least the following beneficial effects:

[0054] By obtaining historical wind power data from the wind farm; determining the power generation prediction function of the wind power fluctuation interval prediction model based on the historical wind power data according to the Bayesian rule; determining the power generation fluctuation interval prediction function based on the power generation prediction function and a preset confidence level; determining the power generation prediction function and the power generation fluctuation interval prediction function as the wind power fluctuation interval prediction model; inputting wind speed sample data and wind direction sample data into the wind power fluctuation interval prediction model to obtain power generation prediction data and power generation fluctuation interval prediction data output by the wind power fluctuation interval prediction model; determining target index values ​​based on the measured power generation sample data, power generation prediction data, and power generation fluctuation interval prediction data, and adjusting the parameters of the wind power fluctuation interval prediction model based on the target index values. Thus, a wind power fluctuation interval prediction model based on a combination of characteristic kernel functions is established. While improving the accuracy of single-point wind power prediction, wind power interval prediction results at different confidence levels can be obtained.

[0055] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application, and do not constitute an improper limitation on the present application.

[0057] Figure 1 This is a flow chart of a method for generating a wind power fluctuation interval prediction model in an embodiment of the present application;

[0058] Figure 2 This is a structural block diagram of a device for generating a wind power fluctuation interval prediction model in an embodiment of the present application;

[0059] Figure 3 is a block diagram of an electronic device in an embodiment of the present application;

[0060] Figure 4 This is a trend chart of wind power interval prediction results in the embodiment of this application;

[0061] Figure 5This is a comparison trend diagram of the monthly power forecast error change curve of the traditional artificial neural network forecasting method in the embodiment of the present application and the monthly power forecast error change curve of a wind power fluctuation interval forecasting model. DETAILED DESCRIPTION

[0062] In order to enable ordinary people in the art to better understand the technical solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0063] It should be noted that the terms "first," "second," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0064] It should be noted that accurately predicting wind farm output power, turning wind power into a dispatchable and friendly power source, is an effective way to mitigate the impact of large-scale wind power grid integration on the power system. Currently, single-point prediction technology for wind farm output power is relatively mature, but the uncertainty of wind energy and the inherent flaws of the prediction model inevitably lead to errors in single-point wind power predictions, and the prediction results cannot quantitatively describe the uncertainty of wind power. From the application level of wind power, the planning and safe and stable operation of wind power-inclusive grids require accurate estimation of the fluctuation range of wind power. In this case, a single-point fixed value prediction cannot meet the application requirements, and the significance of wind power fluctuation range prediction is highlighted.

[0065] Mainstream wind power fluctuation range forecasting methods (such as quantile regression, kernel density estimation, and resampling) are mostly based on existing single-point wind power prediction models and superimposed with uncertainty analysis of the prediction results. These methods primarily serve as post-evaluation of the model and cannot be directly used for model optimization. However, the Gaussian process method adaptively determines the "hyperparameters" in the prior covariance function based on learning from training data. This outputs the variance of the prediction results while providing the model's mean prediction, directly guiding the optimization of the prediction model.

[0066] Based on the above problems, the present application proposes a method, device, and electronic device for generating a wind power fluctuation interval prediction model, which can be achieved by obtaining historical wind power data of a wind farm; determining the power generation prediction function of the wind power fluctuation interval prediction model based on the historical wind power data according to the Bayesian rule; determining the power generation fluctuation interval prediction function based on the power generation prediction function and a preset confidence level; determining the power generation prediction function and the power generation fluctuation interval prediction function as a wind power fluctuation interval prediction model; inputting wind speed sample data and wind direction sample data into the wind power fluctuation interval prediction model to obtain power generation prediction data and power generation fluctuation interval prediction data output by the wind power fluctuation interval prediction model; determining a target index value based on the measured power generation sample data, the power generation prediction data, and the power generation fluctuation interval prediction data, and adjusting the parameters of the wind power fluctuation interval prediction model based on the target index value. Thus, a wind power fluctuation interval prediction model based on a combination of characteristic kernel functions is established, which can obtain wind power interval prediction results under different confidence levels while improving the accuracy of single-point wind power prediction.

[0067] Figure 1 This is a flow chart of a method for generating a wind power fluctuation interval prediction model in an embodiment of the present application.

[0068] like Figure 1 As shown, the method for generating the wind power fluctuation range prediction model includes:

[0069] Step 101: Acquire historical wind power data of a wind farm.

[0070] In the embodiment of the present application, the historical wind power data includes wind speed sample data, wind direction sample data and generated power measured sample data.

[0071] As an example of a possible implementation, the historical wind power data may be wind speed sample data, wind direction sample data, and actual power generation sample data of the entire wind farm from a wind tower for a consecutive year.

[0072] Step 102 : Based on historical wind power data and according to the Bayesian rule, a power generation prediction function of a wind power fluctuation range prediction model is determined.

[0073] In some embodiments of the present application, step 102 includes:

[0074] Step a1: Determine a first function based on historical wind power data.

[0075] In this embodiment of the present application, the first function is used to characterize the probability distribution of historical wind power data of the wind farm.

[0076] As an example of a possible implementation, for historical wind power data D = {(xi ,y i )|i=1,2,...,n;x i ∈X,y i =f(x i )∈R},x i The input variables are wind speed sample data and wind direction sample data, y i The first function is p(D), which is the probability distribution function of the historical wind power data of the wind farm, and the prior function of the wind power fluctuation range prediction model.

[0077] Step a2: determine the covariance kernel function and the mean function respectively.

[0078] In this embodiment of the present application, the mean function is μ(x)=E[f(x)].

[0079] In some embodiments of the present application, the covariance kernel function is determined by the following method:

[0080] Step b1, respectively determine the first kernel function, the second kernel function and the third kernel function; wherein the first kernel function is used to characterize the overall fluctuation characteristics of wind power, the second kernel function is used to characterize the change of wind power with wind speed variable, and the third kernel function is used to characterize the change of wind power with wind direction variable.

[0081] In some embodiments of the present application, the second kernel function is obtained by adding a linear kernel function and a Brownian kernel function; the first kernel function and the third kernel function are both Matérn kernel functions.

[0082] Step b2: Add the second kernel function and the third kernel function to obtain an intermediate function.

[0083] Step b3: multiply the intermediate function by the first kernel function to obtain the covariance kernel function.

[0084] As an example of a possible implementation, the covariance kernel function k GP The calculation formula is:

[0085]

[0086] in, is the second kernel function. Since wind power increases with the increase of incoming wind speed and remains constant after reaching the rated power, its change is similar to the wind turbine power curve, so the linear kernel function is used. and Brownian kernel function The addition method describes the change of wind power with wind speed variable; is the third kernel function. Since the wind direction is symmetrical, the periodic Matérn series kernel function is used. Describe the variation of wind power with wind direction variables; is the first kernel function, and the Matérn series kernel function with a smoothness between the exponential function and the square exponential function is used to describe the overall fluctuation characteristics of wind power.

[0087] The calculation principle of the Matérn kernel function k(x, x') is shown in the following formula:

[0088]

[0089] Where r is the distance between two elements, l is the length scale parameter, K v This is a modified Basel function; v represents the smoothness of the curve, and its value can be predetermined based on actual needs. As the value of v increases, the smoothness of the curve represented by the Matérn function increases. When modeling the full variables of wind speed and direction, v is set to 5 / 2, corresponding to the Matérn52 kernel function; when modeling the wind direction variable, v is set to 3 / 2, corresponding to the Matérn32 kernel function.

[0090] The calculation principle of the Brownian kernel function k(x, x') is shown in the following formula:

[0091] When 0≤x≤1 and 0≤x'≤1, the mean of the kernel function is 0 according to the following formula.

[0092] k(x,x')=min(x,x')

[0093] k(x,x')=min(x,x')-xx'(0≤x,x'≤1)

[0094] Step a3: Determine a Gaussian process function based on the covariance kernel function and the mean function.

[0095] Step a4: determining the power generation prediction function of the wind power fluctuation interval prediction model based on the first function and the Gaussian process function and the Bayesian theorem.

[0096] As an example of a possible implementation, the Gaussian process function f(x) is calculated as follows:

[0097] f(x)~GP(μ(x),k(x,x′))

[0098] μ(x)=E[f(x)]

[0099] k(x,x′)=E[(f(x)-μ(x))(f(x′)-μ(x′))]

[0100] Where μ(x) is the mean function. The Gaussian process function f(x) is characterized by its mean function μ(x) and covariance kernel function k(x,x′), where x and x′ are two different input variables x i , and both can be multidimensional variables.

[0101] Based on the Bayesian theorem, the power generation prediction function p(f|D) of the wind power fluctuation interval prediction model is determined by the following formula, that is, the posterior function of the wind power fluctuation interval prediction model:

[0102]

[0103] Where f is the Gaussian process function f(x).

[0104] Step 103: Determine a power generation fluctuation range prediction function based on the power generation prediction function and a preset confidence level.

[0105] Step c1: determining the standard deviation function of the power generation prediction data based on the wind speed sample data, the wind direction sample data, the covariance kernel function and the mean function.

[0106] In step c2, based on the standard deviation function, the mean function, and the power generation prediction data, the lower limit and the upper limit of the power generation fluctuation range are determined to obtain a power generation fluctuation range prediction function. The power generation fluctuation range prediction function is calculated using the following formula:

[0107] L i α =P Pi -zσ i

[0108] L i α =P Pi -zσ i

[0109] I i α =[L i α , U i α ]

[0110] Among them, 1-α is the confidence level, L i α is the lower limit of the power fluctuation range predicted at the i-th moment, U i α is the upper limit of the power fluctuation range predicted at the i-th moment, P Pi is the predicted power at the i-th moment, σ i is the standard deviation at the i-th moment, I iα is the power prediction interval at time i.

[0111] It should be noted that the coefficient z is obtained through the normal distribution table. For example, when the confidence level 1-α=90%, z=1.645; when the confidence level 1-α=95%, z=1.96; when the confidence level 1-α=99%, z=2.576.

[0112] Step 104 : Determine the generated power prediction function and the generated power fluctuation interval prediction function as a wind power fluctuation interval prediction model.

[0113] For example, Figure 4 This is a trend chart of wind power interval prediction results in the embodiment of this application. Figure 4 As shown, each test time point corresponds to predicted power (generated power prediction data), measured power (generated power measured sample data) and predicted interval (generated power fluctuation interval prediction data).

[0114] Figure 5 This is a comparison trend chart of the monthly power prediction error change curve of the traditional artificial neural network prediction method in the embodiment of the present application and the monthly power prediction error change curve of a wind power fluctuation range prediction model. Compared with the monthly power prediction error change curve of the traditional artificial neural network prediction method, the monthly power prediction error change curve of a wind power fluctuation range prediction model can better perform single-point prediction of the wind power mean, and the error is relatively small.

[0115] Step 105 : Input the wind speed sample data and the wind direction sample data into the wind power fluctuation interval prediction model to obtain the generated power prediction data and the generated power fluctuation interval prediction data output by the wind power fluctuation interval prediction model.

[0116] Step 106 : determining a target index value based on the measured sample data of generated power, the generated power prediction data, and the generated power fluctuation interval prediction data, and adjusting the parameters of the wind power fluctuation interval prediction model based on the target index value.

[0117] In some embodiments of the present application, step 106 includes:

[0118] Step d1: Based on the measured sample data of generated power and the generated power predicted data, the root mean square error of the generated power predicted data is determined. The root mean square error is calculated by the following formula:

[0119]

[0120] Among them, RMSE is the root mean square error of the power generation prediction data, P Mi is the actual power at the i-th moment, P Piis the predicted power at the i-th moment, C i is the wind farm startup capacity at the i-th moment, and N is the number of moments;

[0121] Step d2, comparing the root mean square error with a first preset threshold to obtain a first comparison result;

[0122] Step d3: In response to the first comparison result that the root mean square error is greater than the first preset threshold, the wind power fluctuation interval prediction model is adjusted.

[0123] In some embodiments of the present application, step 106 further includes:

[0124] In step e1, based on the power generation fluctuation interval value and the preset confidence level, a coverage index is determined. The coverage index value is calculated using the following formula:

[0125]

[0126]

[0127] in, is a 0-1 variable, 1-α is the confidence level, I i α is the predicted power fluctuation interval I at the i-th moment i α =[L i α ,U i α ], L i α is the lower limit of the power fluctuation range predicted at the i-th moment, U i α is the upper limit of the power fluctuation range predicted at the i-th moment, R c is the coverage index of the fluctuation interval prediction, and N is the number of moments;

[0128] Step e2: Based on the power generation fluctuation interval value and the preset confidence level, determine the average relative width index value of the power generation fluctuation interval value. The average relative width index value is calculated using the following formula:

[0129]

[0130] Among them, W mean α is the average relative width indicator of the fluctuation range prediction, 1-α is the confidence level, W i α is the predicted width of the fluctuation range at time i, W i α =U t α -Lt α , C i is the wind farm startup capacity at time i, and N is the number of time moments.

[0131] Step e3: in response to the average relative width index value not meeting the first preset requirement and / or the coverage index value not meeting the second preset requirement, adjusting the parameters of the wind power fluctuation interval prediction model.

[0132] Understandably, R c The closer to the confidence level 1-α, the better the prediction effect. Given the same prediction information, the average relative width index W of the interval prediction is mean α The smaller it is, the better the prediction effect is.

[0133] In some embodiments of the present application, the winter months (December, January, February) and spring months (March, April, May) with more severe wind speed fluctuations can be classified into one category based on the month. Based on the historical wind power data of winter and spring, a monthly power prediction model is established using a combined kernel function, and the value of the length scale parameter l| is taken as 100; the summer months (June, July, August) and autumn months (September, October, November) with relatively gentle wind speed fluctuations are classified into one category. Based on the historical wind power data of summer and autumn, a monthly power prediction model is established using a combined kernel function, and the value of the length scale parameter l| is taken as 280, which is higher than that of the winter and spring seasons. The same parameterization method can also be extended to different months for use separately. For months with severe power fluctuations, a lower length scale parameter value is used, and for months with gentle power fluctuations, a higher length scale parameter value is used. By constructing power prediction models for different seasons, the generated power can be more accurately predicted based on the wind speed fluctuation characteristics of different months.

[0134] According to the method for generating a wind power fluctuation interval prediction model according to an embodiment of the present application, the following steps are performed: obtaining historical wind power data of a wind farm; determining a power generation prediction function of the wind power fluctuation interval prediction model based on the historical wind power data according to the Bayesian rule; determining a power generation fluctuation interval prediction function based on the power generation prediction function and a preset confidence level; determining the power generation prediction function and the power generation fluctuation interval prediction function as a wind power fluctuation interval prediction model; inputting wind speed sample data and wind direction sample data into the wind power fluctuation interval prediction model to obtain power generation prediction data and power generation fluctuation interval prediction data output by the wind power fluctuation interval prediction model; determining a target index value based on the measured power generation sample data, the power generation prediction data, and the power generation fluctuation interval prediction data; and adjusting the parameters of the wind power fluctuation interval prediction model based on the target index value. Thus, a wind power fluctuation interval prediction model based on a combination of characteristic kernel functions is established, which improves the accuracy of single-point wind power prediction and can obtain wind power interval prediction results under different confidence levels.

[0135] Figure 2 This is a structural block diagram of a device for generating a wind power fluctuation interval prediction model in an embodiment of the present application.

[0136] like Figure 2 As shown, the generating device of the wind power fluctuation range prediction model includes:

[0137] An acquisition module 201 is configured to acquire historical wind power data of a wind farm; wherein the historical wind power data includes wind speed sample data, wind direction sample data, and generated power measured sample data;

[0138] A first determination module 202 is configured to determine a power generation prediction function of a wind power fluctuation range prediction model based on historical wind power data and in accordance with the Bayesian rule;

[0139] The second determining module 203 is configured to determine a power generation fluctuation range prediction function based on the power generation prediction function and a preset confidence level;

[0140] A third determining module 204 is configured to determine the generated power prediction function and the generated power fluctuation interval prediction function as a wind power fluctuation interval prediction model;

[0141] An input module 205 is configured to input wind speed sample data and wind direction sample data into a wind power fluctuation interval prediction model to obtain generated power prediction data and generated power fluctuation interval prediction data output by the wind power fluctuation interval prediction model;

[0142] The parameter adjustment module 206 is used to determine the target index value based on the measured sample data of generated power, the generated power prediction data and the generated power fluctuation interval prediction data, and adjust the parameters of the wind power fluctuation interval prediction model based on the target index value.

[0143] According to the device for generating a wind power fluctuation interval prediction model according to the embodiment of the present application, the method obtains historical wind power data of a wind farm; determines the power generation prediction function of the wind power fluctuation interval prediction model according to the Bayesian rule based on the historical wind power data; determines the power generation fluctuation interval prediction function based on the power generation prediction function and a preset confidence level; determines the power generation prediction function and the power generation fluctuation interval prediction function as a wind power fluctuation interval prediction model; inputs wind speed sample data and wind direction sample data into the wind power fluctuation interval prediction model to obtain power generation prediction data and power generation fluctuation interval prediction data output by the wind power fluctuation interval prediction model; determines a target index value based on the measured power generation sample data, the power generation prediction data, and the power generation fluctuation interval prediction data; and adjusts the parameters of the wind power fluctuation interval prediction model based on the target index value. Thus, a wind power fluctuation interval prediction model based on a combination of characteristic kernel functions is established, which improves the accuracy of single-point wind power prediction and can obtain wind power interval prediction results at different confidence levels.

[0144] Figure 3 FIG. 1 is a block diagram of an electronic device according to an embodiment of the present application. Figure 3 As shown, the electronic device may include: a transceiver 31 , a processor 32 , and a memory 33 .

[0145] The processor 32 executes the computer-executable instructions stored in the memory, so that the processor 32 implements the solutions in the above embodiments. The processor 32 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0146] The memory 33 is connected to the processor 32 via a system bus and communicates with the processor 32. The memory 33 is used to store computer program instructions.

[0147] The transceiver 31 may be used to obtain tasks to be executed and configuration information of the tasks to be executed.

[0148] The system bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, among others. The system bus can be divided into an address bus, a data bus, a control bus, and so on. For ease of illustration, the figure shows only one thick line, but this does not mean that there is only one bus or one type of bus. Transceivers are used to enable communication between the database access device and other computers (e.g., clients, read-write libraries, and read-only libraries). Memory may include random access memory (RAM) and may also include non-volatile memory.

[0149] The electronic device provided in the embodiment of the present application may be the terminal device of the above embodiment.

[0150] An embodiment of the present application also provides a chip for executing instructions, which is used to execute the technical solution of the message processing method in the above embodiment.

[0151] An embodiment of the present application further provides a computer-readable storage medium, in which computer instructions are stored. When the computer instructions are executed on a computer, the computer executes the technical solution of the message processing method of the above embodiment.

[0152] An embodiment of the present application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium, and when at least one processor executes the computer program, it can implement the technical solution of the message processing method in the above embodiment.

[0153] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0154] It should be understood that the present application is not limited to the exact structure described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for generating a wind power fluctuation interval prediction model, characterized in that: The method comprises: Acquire historical wind power data of the wind farm; wherein the historical wind power data includes wind speed sample data, wind direction sample data and generated power measured sample data; Based on the historical wind power data, determining the power generation prediction function of the wind power fluctuation range prediction model according to the Bayesian rule; Determining a power generation fluctuation interval prediction function based on the power generation prediction function and a preset confidence level; Determining the generated power prediction function and the generated power fluctuation interval prediction function as a wind power fluctuation interval prediction model; Inputting the wind speed sample data and the wind direction sample data into the wind power fluctuation interval prediction model to obtain the generated power prediction data and the generated power fluctuation interval prediction data output by the wind power fluctuation interval prediction model; Determining a target index value based on the generated power measured sample data, the generated power forecast data, and the generated power fluctuation interval forecast data, and adjusting parameters of the wind power fluctuation interval forecast model based on the target index value; The step of determining the power generation fluctuation range prediction function based on the power generation prediction function and the preset confidence level includes: Determining a standard deviation function of the power generation prediction data based on the wind speed sample data, the wind direction sample data, the covariance kernel function, and the mean function; Based on the standard deviation function and the power generation prediction function, the lower limit and the upper limit of the power generation fluctuation range are respectively determined to obtain the power generation fluctuation range prediction function; wherein the power generation fluctuation range prediction function is calculated by the following formula: Among them, 1- α is the confidence level, and the coefficient z is obtained through the normal distribution table. L i α For the i The lower limit of the power fluctuation range predicted at each moment, U i α For the i The upper limit of the power fluctuation range predicted at each moment, P Pi For the i The predicted power at each moment, σ i For the i The standard deviation at each moment, I i α for i The power prediction interval at the time.

2. The method according to claim 1, characterized in that The determining, based on the historical wind power data and according to the Bayesian rule, of the power generation prediction function of the wind power fluctuation interval prediction model comprises: Determining a first function based on the historical wind power data; wherein the first function is used to characterize the probability distribution of the historical wind power data of the wind farm; Determine the covariance kernel function and mean function respectively; Determining a Gaussian process function based on the covariance kernel function and the mean function; According to the first function and the Gaussian process function, a power generation prediction function of the wind power fluctuation interval prediction model is determined based on the Bayesian rule.

3. The method according to claim 2, characterized in that The covariance kernel function is determined by the following method: Determine a first kernel function, a second kernel function, and a third kernel function respectively; wherein the first kernel function is used to characterize the overall fluctuation characteristics of wind power, the second kernel function is used to characterize the change of wind power with wind speed variables, and the third kernel function is used to characterize the change of wind power with wind direction variables; Adding the second kernel function to the third kernel function to obtain an intermediate function; The intermediate function is multiplied by the first kernel function to obtain the covariance kernel function.

4. The method according to claim 3, characterized in that The second kernel function is obtained by adding a linear kernel function and a Brownian kernel function; the first kernel function and the third kernel function are both Matern kernel functions.

5. The method according to claim 1, wherein The determining of a target index value based on the measured sample data of generated power, the generated power prediction data, and the generated power fluctuation interval prediction data, and adjusting parameters of the wind power fluctuation interval prediction model based on the target index value, includes: Based on the generated power measured sample data and the generated power predicted data, a root mean square error of the generated power predicted data is determined. The root mean square error is calculated using the following formula: Wherein, RMSE is the root mean square error of the power generation prediction data, P Mi For the i The actual power at a moment, P Pi For the i The predicted power at each moment, C i For the i The wind farm operating capacity at a given moment, N is the number of moments; Comparing the root mean square error with a first preset threshold to obtain a first comparison result; In response to the first comparison result being that the root mean square error is greater than the first preset threshold, the wind power fluctuation interval prediction model is adjusted.

6. The method according to claim 1, characterized in that The step of determining a target index value based on the generated power measured sample data, the generated power forecast data, and the generated power fluctuation interval forecast data, and adjusting parameters of the wind power fluctuation interval forecast model based on the target index value further includes: Based on the power generation fluctuation range prediction data and the preset confidence level, a coverage index is determined. The coverage index value is calculated using the following formula: in, is a 0-1 variable, 1- α is the confidence level, I i α For the i Predicted power fluctuation range at each moment I i α =[ L i α ,U i α ], L i α For the i The lower limit of the power fluctuation range predicted at each moment, U i α For the i The upper limit of the power fluctuation range predicted at each moment, R c is the coverage index of the fluctuation range forecast, N is the number of moments; Based on the power generation fluctuation interval prediction data and the preset confidence level, an average relative width index value of the power generation fluctuation interval prediction data is determined. The average relative width index value is calculated using the following formula: in, W mean α is the average relative width indicator of the fluctuation range forecast, 1- α is the confidence level, W i α For the moment i The predicted width of the fluctuation range, W i α = U t α - L t α , C i for i The wind farm operating capacity at the time, N is the number of moments; In response to the average relative width index value not meeting a first preset requirement and / or the coverage index value not meeting a second preset requirement, the wind power fluctuation interval prediction model is adjusted.

7. A device for generating a wind power fluctuation range prediction model, characterized in that: The device comprises: An acquisition module is used to acquire historical wind power data of a wind farm; wherein the historical wind power data includes wind speed sample data, wind direction sample data and generated power measured sample data; A first determining module is configured to determine, based on the historical wind power data and in accordance with the Bayesian rule, a power generation prediction function of the wind power fluctuation interval prediction model; A second determining module is configured to determine a power generation fluctuation interval prediction function based on the power generation prediction function and a preset confidence level; A third determining module is configured to determine the generated power prediction function and the generated power fluctuation interval prediction function as a wind power fluctuation interval prediction model; an input module, configured to input the wind speed sample data and the wind direction sample data into the wind power fluctuation interval prediction model, and obtain the generated power prediction data and the generated power fluctuation interval prediction data output by the wind power fluctuation interval prediction model; a parameter adjustment module, configured to determine a target index value based on the generated power measured sample data, the generated power forecast data, and the generated power fluctuation interval forecast data, and adjust the parameters of the wind power fluctuation interval forecast model based on the target index value; The second determining module is specifically configured to: Determining a standard deviation function of the power generation prediction data based on the wind speed sample data, the wind direction sample data, the covariance kernel function, and the mean function; Based on the standard deviation function and the power generation prediction function, the lower limit and the upper limit of the power generation fluctuation range are respectively determined to obtain the power generation fluctuation range prediction function; wherein the power generation fluctuation range prediction function is calculated by the following formula: Among them, 1- α is the confidence level, and the coefficient z is obtained through the normal distribution table. L i α For the i The lower limit of the power fluctuation range predicted at each moment, U i α For the i The upper limit of the power fluctuation range predicted at each moment, P Pi For the i The predicted power at each moment, σ i For the i The standard deviation at each moment, I i α for i The power prediction interval at the time.

8. An electronic device, characterized in that: include: a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 6 when executed by a processor.

Images