Method for predicting wind power under short-term scale

Abstract

The invention discloses a method for predicting wind power in a short-term scale, which relates to the field of intelligent power grids, and comprises the following steps: according to a set sampling interval, collecting initial predicted wind speed, actually measured wind speed and actual power, preprocessing qualified input data, and according to the preprocessed initial predicted wind speed and actually measured wind speed, calculating the actual power of the wind power in the short-term scale; calculating a predicted wind speed deviation by using a first system deviation formula, and further correcting the initial predicted wind speed to obtain a corrected predicted wind speed; according to the corrected predicted wind speed, calculating initial predicted power by using an improved Richards model; calculating a power prediction deviation by using a second system deviation formula according to the initial prediction power and the actual power, and further correcting the initial prediction power to obtain a corrected prediction power; and detecting whether the corrected predicted power is qualified, and if yes, outputting the corrected predicted power according to a set interval. Through two-stage system deviation correction, prediction deviation expansion is inhibited, and short-term wind power prediction accuracy can be improved.

Description

[0001] This application is a divisional application. The application number of the original application is 202410475895.8, the application date is April 19, 2024, and the name of the invention is "A two-stage short-term wind power prediction method and device". Technical Field

[0002] The present invention relates to the technical field of smart grids, and more particularly to a method for predicting wind power in a short-term scale. Background Art

[0003] Wind power forecasting refers to the establishment of a wind farm output power forecasting model based on the historical power, historical wind speed, topography, numerical weather forecast, wind turbine operating status and other data of the wind farm. The wind speed, power or numerical weather forecast data is used as the input of the model, combined with the equipment status and operating conditions of the wind farm units, to obtain the future output power of the wind farm. Among them, the short-term scale (0-72 hours) wind power forecast is highly dependent on the predicted wind speed provided by the weather forecast.

[0004] Classic wind power curve fitting techniques include polynomial regression model, piecewise linear model, logistic function model and non-parametric regression model. However, polynomial regression model, piecewise linear model, logistic function model and so on cannot take into account the uncertainty of wind turbine operation status; non-parametric regression model such as kernel density estimation method has large estimation deviation in low wind speed section and near rated wind speed section, and is more sensitive to local outliers; while machine learning and other methods are more sensitive to data quality and parameter selection, and have higher requirements on computing resources. In terms of weather forecast uncertainty, signal decomposition method and prediction error probability distribution function are the most commonly used tools. However, signal decomposition method and probability distribution function also need to be based on a large amount of actual observation data, and are easily affected by outliers.

[0005] Therefore, how to improve the accuracy of short-term wind power prediction is an urgent problem that technical personnel in this field need to solve. Summary of the invention

[0006] In view of this, the present invention provides a method for predicting wind power in the short term. Firstly, an uncertainty term is introduced into the Richards curve, which characterizes the uncertainty of the unit output while characterizing the basic characteristics of the S-type of wind power; secondly, by estimating and correcting the system prediction deviation in the two links of wind speed and power prediction, the final prediction result is optimized, thereby improving the accuracy of short-term wind power prediction.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] The present invention discloses a two-stage short-term wind power forecasting method, such asFigure 1 As shown, the specific steps are as follows:

[0009] Data input: According to the set sampling interval, collect the initial predicted wind speed, measured wind speed, and actual power to obtain input data;

[0010] Data verification: Check whether the input data is qualified. If qualified, pre-process it, otherwise issue a warning;

[0011] Forecast wind speed correction: according to the preprocessed initial forecast wind speed and the measured wind speed, the forecast wind speed deviation is calculated using the first system deviation formula; the initial forecast wind speed is corrected according to the forecast wind speed deviation to obtain the corrected forecast wind speed;

[0012] Power prediction and correction: according to the corrected predicted wind speed, the initial predicted power is calculated using the improved Richards model; according to the initial predicted power and the actual power, the power prediction deviation is calculated using the second system deviation formula; according to the power prediction deviation, the initial predicted power is corrected to obtain the corrected predicted power;

[0013] Data output: Detect whether the corrected predicted power is qualified. If qualified, output the corrected predicted power according to the set interval.

[0014] Furthermore, the initial predicted wind speed and the measured wind speed are input from a weather forecast workstation located in safety zone III / IV via a reverse isolation device, and the actual power is input from a power forecast workstation located in safety zone II via a forward isolation device.

[0015] Furthermore, in the data verification, the formula for calculating the pass rate is:

[0016]

[0017] Among them, N 总 is the total number of sampling points in the detection time period, N 缺测点 is the total number of missing points in the detection period, N 离群点 is the total number of outliers in the detection period; when the qualified rate is greater than or equal to 90%, data preprocessing is performed, otherwise a warning is issued;

[0018] The outlier point in the detection time period is the sliding correlation coefficient r between the measured wind speed and the actual power. VP The sliding correlation coefficient r VP The calculation formula is:

[0019]

[0020] Where V i and P iare the values ​​of the time series of measured wind speed and actual power in a sliding window with a width of k, and is the average value of the points in the corresponding sliding window; in the process of sliding the detection window forward, the point where the sliding correlation coefficient is less than 0.9 is recorded as the mutation point.

[0021] Further, the preprocessing is: using a neighbor weighted average method to generate a replacement value for the unqualified data point, and replacing the unqualified data point;

[0022] The calculation formula of the replacement value is:

[0023]

[0024] In the formula, y is the replacement value, y -2 ,y -1 , y1 and y2 are the 4 points adjacent to the unqualified point y0, w -2 , w -1 , w1 and w2 are the weight values ​​of the corresponding neighboring points.

[0025] Furthermore, the first system deviation formula is:

[0026]

[0027] Where E is the systematic deviation between the initial predicted wind speed and the measured wind speed, n is the length of the wind speed time series, V i 、V' i are the measured wind speed and the initial predicted wind speed at time i respectively;

[0028] The correction of the initial predicted wind speed includes: using the ARIMA (p, d, q) model to estimate the wind speed prediction deviation E w The time series of E w The time series and the initial predicted wind speed are used to calculate the corrected predicted wind speed V' i ', the specific calculation formula is:

[0029]

[0030] In the formula, φ i is the coefficient of the autoregressive term, which indicates the influence of previous observations on the time series; p is the number of autoregressive terms; q is the number of sliding average terms; d is the number of differences required to stabilize the series; L is the lag operator L(y t )=y t-1 ,L k is the kth power of the lag operator, indicating that y t Lag k time steps, L k (y t )=y t-k θi is the coefficient of the sliding average term, which indicates the impact of the first q random shocks on the current value; ε t is a random error term.

[0031] Furthermore, the calculation formula of the initial predicted power is:

[0032]

[0033] Where P' i represents the initial predicted power at time i, P max is the rated power of the fan, a is the shape parameter, e is the natural constant, V″ i To correct the predicted wind speed, v0 is the cut-in wind speed of the wind turbine, and K is a scaling function used to characterize the uncertainty of the unit output under different wind speeds. Its expression is:

[0034]

[0035] In the formula, σ is the scale parameter, and x is the fluctuation range of wind speed;

[0036] The second system deviation formula is:

[0037]

[0038] Where E' is the system deviation between the initial predicted power and the measured power, n' is the length of the power time series, P i , P' i are the measured power and the initial predicted power at time i respectively;

[0039] The correcting of the initial predicted power includes: using the ARIMA (p, d, q) model to estimate the power prediction deviation E p The time series of E p The time series and the initial prediction power are used to calculate the corrected prediction power P″ i , the specific calculation formula is:

[0040]

[0041] In the formula, φ i is the coefficient of the autoregressive term, , which represents the influence of previous observations on the time series; p is the number of autoregressive terms; q is the number of sliding average terms; d is the number of differences required for the sequence to be stabilized; L is the lag operator L(y t )=y t-1 ,L k is the kth power of the lag operator, indicating that y t Lag k time steps, L k (y t )=y t-k θi is the coefficient of the sliding average term, which indicates the impact of the first q random shocks on the current value; ε t is a random error term.

[0042] Furthermore, in the data output, the output qualification rate calculation formula is:

[0043]

[0044] Among them, N' 总 is the total number of output time points within the prediction period, N' 缺测点 is the total number of missing points in the prediction time period, N' 离群点 is the total number of outliers in the prediction time period; when the qualified rate is less than 95%, a warning is issued; when it is greater than or equal to 90%, the power prediction result is output to the power prediction workstation through the reverse isolation device, and then sent to the dispatching center by the power prediction workstation;

[0045] The outlier point within the forecast time period is a mutation point in the sliding correlation coefficient r of the corrected forecast wind speed and the corrected forecast power. The calculation formula of the sliding correlation coefficient r is:

[0046]

[0047] Where V″ i and P″ i are the values ​​of the time series of the corrected predicted wind speed and the corrected predicted power in a sliding window with a width of k, and is the average value of the points in the corresponding sliding window; in the process of sliding the detection window forward, the point where the sliding correlation coefficient is less than 0.9 is recorded as the mutation point.

[0048] The present invention also discloses a two-stage short-term wind power prediction device, comprising:

[0049] Data input module: collects the initial predicted wind speed, measured wind speed, and actual power according to the set sampling interval to obtain input data;

[0050] Data verification module: detects whether the input data is qualified, performs data preprocessing if qualified, otherwise issues a warning;

[0051] Predicted wind speed correction module: according to the preprocessed initial predicted wind speed and the measured wind speed, using the first system deviation formula, calculate the predicted wind speed deviation; according to the predicted wind speed deviation, correct the initial predicted wind speed to obtain the corrected predicted wind speed;

[0052] Power prediction and correction module: according to the corrected predicted wind speed, the initial predicted power is calculated using the improved Richards model; according to the initial predicted power and the actual power, the power prediction deviation is calculated using the second system deviation formula; according to the power prediction deviation, the initial predicted power is corrected to obtain the corrected predicted power;

[0053] Data output module: detects whether the corrected predicted power is qualified, and if qualified, outputs the corrected predicted power at a set interval.

[0054] It can be known from the above technical solutions that compared with the prior art, the present invention discloses a two-stage short-term wind power forecasting method and device, which is based on the improved Richards model. While ensuring the basic S-shaped characteristics of the wind power curve, it takes into account the uncertainty of the unit operation status. At the same time, through the two-stage system deviation correction, the deviation propagation in the wind power forecasting process is suppressed. While ensuring the prediction accuracy, the stability and field applicability of the power prediction results are enhanced; and the device proposed by the present invention is an independent module, which is connected to the power forecasting system through the forward and reverse isolation devices, ensuring the security of key data and compatible with the existing forecasting system. The present invention is suitable for short-term wind power forecasting correction of both single machines and entire stations, which can improve the prediction accuracy and stability, increase the revenue from electricity sales of wind farms, and reduce the operating costs and operating risks of power grids. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0056] Figure 1 The figure is a schematic diagram of the overall flow of the prediction method according to an embodiment of the present invention.

[0057] Figure 2 Schematic diagram of the overall structure of the prediction device according to an embodiment of the present invention.

[0058] Figure 3 It is a schematic diagram of the implementation effect of an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] The embodiment of the present invention discloses a two-stage short-term wind power prediction method, such as Figure 1 As shown, the specific steps are as follows:

[0061] Data input: According to the set sampling interval, collect the initial predicted wind speed, measured wind speed, and actual power to obtain input data;

[0062] Data verification: Check whether the input data is qualified. If qualified, pre-process it, otherwise issue a warning;

[0063] Forecast wind speed correction: According to the pre-processed initial forecast wind speed and the measured wind speed, the forecast wind speed deviation is calculated using the first system deviation formula; the initial forecast wind speed is corrected according to the forecast wind speed deviation to obtain the corrected forecast wind speed;

[0064] Power prediction and correction: Based on the corrected predicted wind speed, the initial predicted power is calculated using the improved Richards model; based on the initial predicted power and the actual power, the power prediction deviation is calculated using the second system deviation formula; the initial predicted power is corrected based on the power prediction deviation to obtain the corrected predicted power;

[0065] Data output: Check whether the corrected predicted power is qualified. If qualified, output the corrected predicted power according to the set interval.

[0066] In a specific embodiment, Figure 2 As shown, the initial predicted wind speed and the measured wind speed are input from the weather forecast workstation located in the safety zone III / IV via the reverse isolation device, and the actual power is input from the power forecast workstation located in the safety zone II via the forward isolation device, thereby ensuring the safety of key data and being compatible with the existing forecast system.

[0067] In a specific embodiment, in data verification, the formula for calculating the pass rate is:

[0068]

[0069] Among them, N 总 is the total number of sampling points in the detection time period, N 缺测点 is the total number of missing points in the detection period, N 离群点 is the total number of outliers in the detection period; when the qualified rate is greater than or equal to 90%, data preprocessing is performed, otherwise a warning is issued;

[0070] The outliers in the detection period are the sliding correlation coefficient r between the measured wind speed and the actual power. VP The mutation point in the sliding correlation coefficient r VP The calculation formula is:

[0071]

[0072] Where V i and P i are the values ​​of the time series of measured wind speed and actual power in a sliding window with a width of k, and is the average value of the points in the corresponding sliding window; in the process of sliding the detection window forward, the point where the sliding correlation coefficient is less than 0.9 is recorded as the mutation point.

[0073] In a specific embodiment, the preprocessing is: using a neighbor weighted average method to generate replacement values ​​for unqualified data points, and replacing the unqualified data points;

[0074] The replacement value is calculated as:

[0075]

[0076] In the formula, y is the replacement value, y -2 ,y -1 , y1 and y2 are the 4 points adjacent to the unqualified point y0, w -2 , w -1 , w1 and w2 are the weight values ​​of the corresponding neighboring points. After data verification, the time series of the initial predicted wind speed V', the measured wind speed V and the actual power P are obtained.

[0077] In a specific embodiment, the first system deviation formula is:

[0078]

[0079] Where E is the systematic deviation between the initial predicted wind speed and the measured wind speed, n is the length of the wind speed time series, V i 、V' i are the measured wind speed and the initial predicted wind speed at time i respectively;

[0080] Correct the initial predicted wind speed, including: using the ARIMA (p, d, q) model to estimate the wind speed prediction deviation E w The time series of E w The time series and initial forecast wind speed are used to calculate the corrected forecast wind speed V″ i , the specific calculation formula is:

[0081]

[0082] In the formula, φ i is the coefficient of the autoregressive term, which indicates the influence of previous observations on the time series; p is the number of autoregressive terms; q is the number of sliding average terms; d is the number of differences required to stabilize the series; L is the lag operator L(y t )=y t-1 ,L k is the kth power of the lag operator, indicating that y t Lag k time steps, L k (y t )=y t-k θ i is the coefficient of the sliding average term, which indicates the impact of the first q random shocks on the current value; ε t is a random error term. Based on the input data historical wind speed prediction L k (E w ), and output L(E w ), use the automatic fitting method to determine the values ​​of model parameters p, d, q, and establish the ARIMA (p, d, q) model. k-1 (y t ) back into the model to get the predicted value For E w The value of the time series at time i.

[0083] Specifically, first, the stage of evaluating the stationarity of the data is to evaluate the wind speed prediction deviation E w Perform at least one unit root test on the time series data. If the sequence is not stable, perform successive difference operations on the original data until a stable sequence is obtained; the number of difference operations performed is determined as the difference number parameter d. Then, set the candidate range of the autoregressive term p and the moving average term q, which is generally a positive integer of [1,5]; search for parameter combinations, and estimate the p,q combination with the minimum residual in the process of fitting historical data. The residual calculation formula is:

[0084]

[0085] Where N is the number of points for fitting historical data.

[0086] In a specific embodiment, the calculation formula of the initial predicted power is:

[0087]

[0088] Where P' i represents the initial predicted power at time i, P max is the rated power of the fan, a is the shape parameter, e is the natural constant, V' i' is the corrected predicted wind speed, v0 is the cut-in wind speed of the wind turbine, and K is the scaling function used to characterize the uncertainty of the unit output under different wind speeds. Its expression is:

[0089]

[0090] In the formula, σ is the scale parameter, and x is the fluctuation range of wind speed;

[0091] Among them, the determination method of the unknown parameters σ and a is:

[0092] Set the candidate range of σ and a. Generally, the shape parameter a is between [0.5, 2] and σ is between [0, 3]. Search for the parameter combination, the σ and a combination with the smallest energy loss in the process of fitting historical data. The energy loss calculation formula is:

[0093]

[0094] It is stated that in the initial forecasting process, the forecast accuracy of the low wind speed segment was abandoned, and more emphasis was placed on the forecast accuracy of the medium and high wind speed segments and the period with more power output. In particular, for wind farms with different types of wind turbines, the wind turbines are first classified according to the cut-in wind speed, and then the initial power forecast is performed according to the category.

[0095] The second system deviation formula is:

[0096]

[0097] Where E' is the system deviation between the initial predicted power and the measured power, n' is the length of the power time series, P i , P' i are the measured power and the initial predicted power at time i respectively;

[0098] Correct the initial predicted power, including: using the ARIMA (p, d, q) model to estimate the power prediction deviation E p The time series of E p The time series and the initial forecast power are used to calculate the corrected forecast power P″ i , the specific calculation formula is:

[0099]

[0100] In the formula, φ i is the coefficient of the autoregressive term, which indicates the influence of previous observations on the time series; p is the number of autoregressive terms; q is the number of sliding average terms; d is the number of differences required to stabilize the series; L is the lag operator L(y t )=y t-1 ,L k is the kth power of the lag operator, indicating that y t Lag k time steps, Lk (y t )=y t-k θ i is the coefficient of the sliding average term, which indicates the impact of the first q random shocks on the current value; ε t is a random error term. According to the historical power prediction of input data L k (E p ), and output L(E p ), use the automatic fitting method to determine the values ​​of model parameters p, d, q, and establish the ARIMA (p, d, q) model. k-1 (y t ) back into the model to get the predicted value For E p The value of the time series at time i.

[0101] Specifically, the values ​​of the model parameters p, d, and q are determined by automatic fitting. First, in the stage of evaluating the stationarity of the data, the power prediction deviation E p Perform at least one unit root test on the time series data. If the sequence is not stable, perform successive difference operations on the original data until a stable sequence is obtained; the number of difference operations performed is determined as the difference number parameter d. Then, set the candidate range of the autoregressive term p and the moving average term q, which is generally a positive integer of [1,5]; search for parameter combinations, and estimate the p,q combination with the minimum residual in the process of fitting historical data. The residual calculation formula is:

[0102]

[0103] Where N' is the number of points for fitting historical data.

[0104] In a specific embodiment, in data output, the output qualification rate calculation formula is:

[0105]

[0106] Among them, N' 总 is the total number of output time points within the prediction period, N' 缺测点 is the total number of missing points in the prediction time period, N' 离群点 is the total number of outliers in the prediction time period; when the qualified rate is less than 95%, a warning is issued; when it is greater than or equal to 90%, the power prediction result is output to the power prediction workstation through the reverse isolation device, and then sent to the dispatching center by the power prediction workstation;

[0107] The outliers in the forecast period are mutation points in the sliding correlation coefficient r of the corrected forecast wind speed and the corrected forecast power. The calculation formula of the sliding correlation coefficient r is:

[0108]

[0109] Where V' i 'and P' i ' are the values ​​of the time series of corrected predicted wind speed and corrected predicted power in a sliding window with a width of k, and is the average value of the points in the corresponding sliding window; in the process of sliding the detection window forward, the point where the sliding correlation coefficient is less than 0.9 is recorded as the mutation point.

[0110] In a specific embodiment, a two-stage short-term wind power forecasting method is implemented as follows: Figure 3 As shown, after the optimization prediction by the method of the present invention, the RMSE error decreased by 6.3%; the systematic deviation was 3.10MW before the optimization, and -0.13MW after the optimization.

[0111] The present invention also discloses a two-stage short-term wind power prediction device, comprising:

[0112] Data input module: collects the initial predicted wind speed, measured wind speed, and actual power according to the set sampling interval to obtain input data;

[0113] Data verification module: detects whether the input data is qualified. If qualified, data preprocessing is performed, otherwise a warning is issued;

[0114] Predicted wind speed correction module: according to the pre-processed initial predicted wind speed and the measured wind speed, the predicted wind speed deviation is calculated using the first system deviation formula; the initial predicted wind speed is corrected according to the predicted wind speed deviation to obtain the corrected predicted wind speed;

[0115] Power prediction and correction module: according to the corrected predicted wind speed, the initial predicted power is calculated using the improved Richards model; according to the initial predicted power and the actual power, the power prediction deviation is calculated using the second system deviation formula; according to the power prediction deviation, the initial predicted power is corrected to obtain the corrected predicted power;

[0116] Data output module: detects whether the corrected predicted power is qualified. If qualified, the corrected predicted power is output according to the set interval.

[0117] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0118] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting wind power in a short-term scale, characterized in that: The specific steps are as follows: Data input: According to the set sampling interval, collect the initial predicted wind speed, measured wind speed, and actual power to obtain input data; Data verification: Check whether the input data is qualified. If qualified, pre-process it, otherwise issue a warning; Forecast wind speed correction: According to the pre-processed initial forecast wind speed and the measured wind speed, the forecast wind speed deviation is calculated using the first system deviation formula; the initial forecast wind speed is corrected according to the forecast wind speed deviation to obtain the corrected forecast wind speed; Power prediction and correction: Based on the corrected predicted wind speed, the initial predicted power is calculated using the improved Richards model; based on the initial predicted power and the actual power, the power prediction deviation is calculated using the second system deviation formula; the initial predicted power is corrected based on the power prediction deviation to obtain the corrected predicted power; Data output: Check whether the corrected predicted power is qualified. If qualified, output the corrected predicted power at the set interval; Among them, the first system deviation formula is: Where E is the systematic deviation between the initial predicted wind speed and the measured wind speed, n is the length of the wind speed time series, V i 、V' i are the measured wind speed and the initial predicted wind speed at time i respectively; Correct the initial predicted wind speed, including: using the ARIMA (p, d, q) model to estimate the wind speed prediction deviation E w The time series of wind speed prediction deviation E w Perform at least one unit root test on the time series data. If the time series is not stationary, perform successive difference operations on the original data until a stationary series is obtained. The number of difference operations performed is determined as the number of differences d required for the series to be stationary. Set the number of autoregressive terms p and the number of sliding average terms q to be positive integers [1,5]. Search for parameter combinations and estimate the p and q combinations with the minimum residuals in the process of fitting historical data. The residual calculation formula is: Where N is the number of points of fitting historical data; According to E w The time series and initial forecast wind speed are used to calculate the corrected forecast wind speed V″ i , the specific calculation formula is: In the formula, φ i is the coefficient of the autoregressive term, which indicates the influence of previous observations on the time series; p is the number of autoregressive terms; q is the number of sliding average terms; d is the number of differences required to stabilize the series; L is the lag operator L(y t )=y t-1 , L k is the kth power of the lag operator, indicating that y t Lag k time steps, L k (y t )=y t-k θ i is the coefficient of the sliding average term, which indicates the impact of the first q random shocks on the current value; ε t is a random error term.

2. A method for predicting wind power in a short-term scale according to claim 1, characterized in that: The initial predicted wind speed and measured wind speed are input from the weather forecast workstation located in safety zone III / IV via the reverse isolation device, and the actual power is input from the power forecast workstation located in safety zone II via the forward isolation device.

3. The method for predicting wind power in a short-term scale according to claim 1, characterized in that: In data verification, the formula for calculating the pass rate is: Among them, N 总 is the total number of sampling points in the detection time period, N 缺测点 is the total number of missing points in the detection time period, N 离群点 is the total number of outliers in the detection period; when the qualified rate is greater than or equal to 90%, data preprocessing is performed, otherwise a warning is issued; The outliers in the detection period are the sliding correlation coefficient r between the measured wind speed and the actual power. VP The mutation point in the sliding correlation coefficient r VP The calculation formula is: Where V i and P i are the values ​​of the time series of measured wind speed and actual power in a sliding window with a width of k, and is the average value of the points in the corresponding sliding window; in the process of sliding the detection window forward, the point where the sliding correlation coefficient is less than 0.9 is recorded as the mutation point.

4. The method for predicting wind power in a short-term scale according to claim 1, characterized in that: The preprocessing is as follows: using the neighbor weighted average method to generate replacement values ​​for unqualified data points and replace the unqualified data points; The replacement value is calculated as: In the formula, y is the replacement value, y -2 ,y -1 , y1 and y2 are the 4 points adjacent to the unqualified point y0, w -2 , w -1 , w1 and w2 are the weight values ​​of the corresponding neighboring points.

5. The method for predicting wind power in a short-term scale according to claim 1, characterized in that: The calculation formula for the initial predicted power is: Where P' i represents the initial predicted power at time i, P max is the rated power of the fan, a is the shape parameter, e is the natural constant, v″ u To correct the predicted wind speed, v0 is the cut-in wind speed of the wind turbine, and K is a scaling function used to characterize the uncertainty of the unit output under different wind speeds. Its expression is: In the formula, σ is the scale parameter, and x is the fluctuation range of wind speed; The second system deviation formula is: Where E' is the system deviation between the initial predicted power and the measured power, n' is the length of the power time series, P i , P' i are the measured power and the initial predicted power at time i respectively; Correct the initial predicted power, including: using the ARIMA (p, d, q) model to estimate the power prediction deviation E p The time series of E p The time series and the initial forecast power are used to calculate the corrected forecast power P″ i , the specific calculation formula is: In the formula, φ i is the coefficient of the autoregressive term, which indicates the influence of previous observations on the time series; p is the number of autoregressive terms; q is the number of sliding average terms; d is the number of differences required to stabilize the series; L is the lag operator L(y t )=y t-1 , L k is the kth power of the lag operator, indicating that y t Lag k time steps, L k (y t )=y t-k θ i is the coefficient of the sliding average term, which indicates the impact of the first q random shocks on the current value; ε t is a random error term.

6. A method for predicting wind power in a short-term scale according to claim 5, characterized in that: The scale parameter σ and shape parameter a are determined as follows: Set the candidate range of σ and a, the shape parameter a is between [0.5, 2], and the scale parameter σ is between [0, 3]; search for parameter combinations, the combination of σ and a with the smallest energy loss in the process of fitting historical data, and the energy loss calculation formula is: Among them, for wind farms with different models, the wind turbines are first classified according to the cut-in wind speed, and then initial power prediction is performed according to the category.

7. The method for predicting wind power in a short-term scale according to claim 5, characterized in that: The values ​​of the number of autoregressive terms p, the number of moving average terms q, and the number of differences d required for sequence stabilization are determined by automatic fitting, specifically: Power prediction deviation E p Perform at least one unit root test on the time series data. If the series is not stable, perform successive difference operations on the original data until a stable series is obtained. The number of difference operations performed is determined as the number of differences d required for the series to be stable. The candidate range of the number of autoregressive terms p and the number of moving average terms q is set to a positive integer of [1,5]; the search parameter combination is to estimate the p,q combination with the minimum residual in the process of fitting historical data. The residual calculation formula is: Where N' is the number of points for fitting historical data.

8. The method for predicting wind power in a short-term scale according to claim 1, characterized in that: In data output, the output qualified rate calculation formula is: Among them, N' 总 is the total number of output time points within the prediction period, N' 缺测点 is the total number of missing points in the prediction time period, N' 离群点 is the total number of outliers in the prediction time period; when the qualified rate is less than 95%, a warning is issued; when it is greater than or equal to 90%, the power prediction result is output to the power prediction workstation through the reverse isolation device, and then sent to the dispatching center by the power prediction workstation; The outliers in the forecast period are mutation points in the sliding correlation coefficient r of the corrected forecast wind speed and the corrected forecast power. The calculation formula of the sliding correlation coefficient r is: Where V' i 'and P' i ' are the values ​​of the time series of corrected predicted wind speed and corrected predicted power in a sliding window with a width of k, and is the average value of the points in the corresponding sliding window; in the process of sliding the detection window forward, the point where the sliding correlation coefficient is less than 0.9 is recorded as the mutation point.

Images