Short-term wind speed prediction method, system and storage medium based on time and space dimensions

Abstract

The present invention discloses a short-term wind speed prediction method, system and storage medium based on time and space dimensions, including: S1, obtaining a wind speed time series, and using a re-scaled range analysis method to obtain the Hurst index of the wind speed time series, and judging the predictability of the wind speed time series; S2, calculating the parameters in the affine transformation, performing fractal interpolation and extrapolation prediction for each moment, and obtaining a fractal interpolation and extrapolation prediction result; S3, defining the last moment before the acquired wind speed time series as an interpolation point, using the atmospheric parameters of all heights of the interpolation point to obtain the Richardson number Ri, judging the atmospheric stability, according to the atmospheric stability state, substituting into the wind shear index extrapolation model, extrapolating the data at this moment, and obtaining the wind shear index extrapolation result; S4, fusing the fractal interpolation and extrapolation prediction result and the wind shear index extrapolation result, and evaluating the result. The present invention combines the fractal interpolation and extrapolation algorithm with the wind shear index, so that the prediction result is more consistent with the nature of the wind profile.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind speed prediction, and in particular relates to a short-term wind speed prediction method based on time and space dimensions. Background Art

[0002] For wind power generation, factors such as the topography, unit installation location, temperature and humidity, and wind energy formation will have a certain impact on wind power generation, and wind speed is one of the biggest factors affecting the wind power generation system. Since wind energy has strong randomness, intermittency and volatility, the active power and reactive power of wind power have obvious volatility and intermittency characteristics with the change of wind speed. When the wind power access rate exceeds 8%, it will have an impact on the normal operation of the power system. Large-scale wind power integration into the power grid will bring huge challenges to the safe and stable operation and economic optimization of the power system. Wind speed prediction is an important basis and prerequisite for realizing wind power prediction. Obtaining the prediction uncertainty information of wind speed in advance can provide data reference for wind power enterprise management and power grid dispatching and operation personnel, assist in making optimized decisions, help wind power stations operate scientifically, reduce wind power generation costs, optimize power grid operation and dispatching, ensure the quality of power supply, reduce the adverse effects of large-scale wind power grid connection on the power system, and improve the power system's carrying capacity and operation stability for wind power, thereby further improving wind energy utilization and reliability. Summary of the invention

[0003] The purpose of the present invention is to provide a short-term wind speed prediction method based on time and space dimensions, so as to realize post-processing of prediction completion of incomplete data in time and layered height measurement, and improve the completeness of radar wind measurement data.

[0004] In order to solve the above problems, the present invention discloses a short-term wind speed prediction method based on time and space dimensions, comprising the following steps:

[0005] S1. Obtain wind speed time series;

[0006] S2. Use the rescaled range analysis (R / S analysis) method to process the wind speed time series data and obtain the Hurst index of the wind speed time series to determine the predictability of the series;

[0007] S21. Calculate the Hurst index of the wind speed forecast sequence using the R / S analysis method:

[0008] Arrange the wind speed data in chronological order to form a wind speed time series data {ξ t =x 1 ,x 2 ,...,x n}. For any positive integer k ≥ 1, its mean sequence, cumulative deviation F tk , range R k and standard deviation S k They are:

[0009] 1≤t≤k;

[0010]

[0011] R k =max F tk -min F tk 1≤t≤k;

[0012]

[0013] S22. Determine the predictability of the sequence

[0014] If there is R k / S k ∝k H This indicates that the time series has the Hurst phenomenon, where H is the Hurst index. H = 0.5 indicates that the time series is completely independent, 0 < H < 0.5 indicates that the time series has anti-persistence, and 0.5 < H < 1 indicates that the time series has persistence; the closer H is to 0, the stronger the anti-persistence; the closer H is to 1, the stronger the persistence.

[0015] S3. Calculate the parameters in the affine transformation, including the vertical scale factor d i , the parameters of the affine transformation iterated function system (IFS) of the interval and its attractor; after obtaining the parameters, fractal interpolation and extrapolation prediction are performed at each moment;

[0016] S31. Calculation method of attractors in fractal interpolation

[0017] Suppose that a set of mappings with compression factors 0≤si<1 defined on a complete distance space (X, d) is {ω i :X→R(X), i=1,2…N}, which constitutes a (hyperbolic) iterated function system IFS. For this IFS, W:(X)→R(X) defined by the following formula is a complete distance space (R(X), h(d)) with s=max{s i :i=1,2…N} is the compression mapping of the compression factor:

[0018]

[0019] where B∈R(X); the only fixed point of this mapping A∈R(X) satisfies For any initial point B∈R(X), A can be obtained by W iteration, that is,

[0020]

[0021] Where B∈R(X); the fixed point A∈R(X) is called the attractor of IFS. Usually the distance function can be of Hausdorff type.

[0022] S32. Calculation method of IFS parameters in fractal interpolation

[0023] Fractal interpolation constructs an IFS in the entire interval, and iterates using IFS from any initial point in this function space. Its attractor is the function point graph passing through this set of interpolation points, which is determined by IFS instead of the usual analytical function.

[0024] Generally speaking, the following affine transformation group can be used as the IFS to be constructed for fractal interpolation:

[0025]

[0026] Where: (x, y) is the coordinate of a point; ω i is the affine transformation relationship; a i 、c i ,d i is the transformation matrix element; e i 、f i is the constant component after the transformation of (x,y).

[0027] If the attractor of this IFS is required to pass through the given interpolation point {(x i ,y i ):i=1,2…N}, each transformation must satisfy the following endpoint conditions:

[0028]

[0029] This means that the left endpoint of the large interval (x 0 ,y 0 ) is mapped to the left endpoint of the subinterval (x i-1 ,y i-1 ), the right endpoint of the large interval (x N ,y N ) is mapped to the right endpoint of the subinterval (x i ,y i ).

[0030] Substitute the transformation conditions into the affine transformation group and take d i is a free parameter (0 <d i <1), the other parameters of the affine transformation are:

[0031]

[0032] After obtaining the parameters, the attractor of IFS can be obtained by deterministic or random iterative algorithms. As the number of iterations increases, the degree of fit between the interpolated curve and the original curve continues to improve. After multiple iterations, a stable invariant interpolation curve will be formed, which not only passes through the sampled interpolation points, but also fully approximates the original sampling curve.

[0033] S33, fractal interpolation extrapolation model

[0034] (1) Select an initial point set from known data to obtain wind speed curve characteristics.

[0035] (2) Obtain the vertical scale factor d of the affine transformation by geometric or analytical methods i .

[0036] (3) Calculate other parameters in the affine transformation to obtain the affine transformation type IFS of the inner interval system.

[0037] (4) Obtain the attractor of IFS using a deterministic or random iterative algorithm.

[0038] (5) Determine the horizontal coordinate x of the point to be extrapolated or predicted B , and assign an initial value to its ordinate (The initial value should be close to the true solution.) At the same time, set the extrapolation search times variable l = 0, and set y C The perturbation value ε y Assign an appropriate value (such as 10 to a negative power).

[0039] (6) From point Start iterating an appropriate number of times (K times) to obtain the point set B of the Kth iteration 1 (excluding the initial point and the points generated before the Kth iteration), compare them with the attractor, and obtain the mean square deviation E from the attractor 1 .

[0040] (7) From point Start iterating an appropriate number of times (K times) to obtain the point set B of the Kth iteration 2 (excluding the initial point and the points generated before the Kth iteration), compare them with the attractor, and obtain the mean square deviation E from the attractor 2 .

[0041] (8) Calculate the gradient change of the mean square deviation

[0042] (9) Convergence judgment: If (ε is the gradient convergence limit), then go to step (10) to continue; if Then The result is output and the search process ends.

[0043] (10) According to the revised formula Find the new y C value.

[0044] S4. Select atmospheric parameters at all heights from the interpolation point, obtain the Richardson number Ri, and judge the atmospheric stability. According to the atmospheric stability state, obtain the wind speed condition under this state. Substitute it into the wind shear index extrapolation model and extrapolate the data at this moment;

[0045] S41. According to the atmospheric stability state, obtain the wind speed condition under this state

[0046] S411. Calculation of Richardson number Ri

[0047] Atmospheric stability has the greatest impact on the wind shear index. Different stability will cause significant changes in the wind profile. The atmospheric stability state parameter is often expressed by the gradient Richardson number Ri, which is expressed as:

[0048]

[0049] Where: g is the acceleration due to gravity, which is 9.8 m / s; T is the average absolute temperature of the two layers of air; △T is the height z 2 With z 1 The temperature difference at △v = v 2 -v 1 , v 2 and v 1 is the height z 2 and z 1 Wind speed at

[0050] Mau-scale computational model

[0051] S412. Determine the state of atmospheric stability

[0052] If -0.6≤Ri<0.2, the atmosphere is neutral and the airflow does not undergo obvious vertical mixing or stratification; if Ri≥0.2, the atmosphere is stable and the vertical movement of the air is weak; if Ri<-0.6, it is unstable and the air is prone to mixing and turbulence, usually accompanied by strong vertical shear of the wind speed.

[0053] S42, Wind Shear Index Extrapolation Model

[0054] S421. Calculation of wind shear index based on atmospheric thermal stability

[0055] (1) Using the formula Calculate the Richardson number. If -0.6≤Ri<0.2, the atmosphere is neutral; if Ri≥0.2, the atmosphere is stable; if Ri<-0.6, it is unstable.

[0056] (2) Using the formula Obtain the Obukhov length (L), which is determined by factors such as wind speed, temperature gradient and turbulent kinetic energy.

[0057] (3) Using the PD algorithm, the wind shear index is calculated using atmospheric stability and roughness length. The formula is:

[0058]

[0059] in,

[0060] S422. Wind shear index extrapolation method

[0061] The law of average wind speed changing with height is called wind speed profile or wind shear. Wind speed profile can be distributed in a logarithmic law as follows:

[0062] Power law formula:

[0063]

[0064] Index formula:

[0065]

[0066] Where: z 1 , z 2 is the vertical height in meters; v 1 , v 2 The vertical height z 1 , z 2 The wind speed at the point is in m / s; α is the wind shear exponent. When calculating the wind speed profile in flat and open terrain, the initial value of the wind shear exponent is usually 0.143, which is called the 1 / 7 power law. On surfaces with more vegetation and when the wind speed is close to medium (less than 7m / s), the exponent is usually larger.

[0067] S5. The results of the horizontal prediction and the vertical prediction are fused using the weighted average method, and the model results are evaluated.

[0068] On the other hand, the present application proposes a system for implementing a short-term wind speed prediction method based on time and space dimensions, comprising:

[0069] An acquisition module, configured to acquire a wind speed time series;

[0070] a judgment module configured to analyze the predictability of the wind speed time series;

[0071] A fractal interpolation and extrapolation module is configured to calculate a fractal interpolation and extrapolation prediction result;

[0072] A wind shear index extrapolation module is configured to calculate a wind shear index extrapolation result;

[0073] The result fusion module is configured to fuse the prediction results of the type interpolation extrapolation and the wind shear index extrapolation results to obtain the results and evaluate the results.

[0074] On the other hand, the present application also proposes an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the short-term wind speed prediction method based on time and space dimensions when executing the computer program.

[0075] On the other hand, the present application further proposes a storage medium, wherein the computer-readable storage medium stores a computer program, and when the program is executed by a processor, the short-term wind speed prediction method based on time and space dimensions is implemented.

[0076] Compared with the prior art, the present invention has the following beneficial effects:

[0077] (1) The present invention extends the fractal characteristics of the inner interval to the outer side of the interval by using an affine transformation type IFS, and thereby constructs a fractal interpolation extrapolation algorithm with an extrapolation function.

[0078] (2) The present invention combines the fractal interpolation extrapolation algorithm with the wind shear index to make the prediction results more consistent with the properties of the wind profile, further improve the accuracy of the algorithm, and achieve the completion of incomplete data in time and altitude. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 It is a general flow chart of the short-term wind speed prediction method based on time and space dimensions provided by the present invention;

[0080] Figure 2 is a radar raw data graph provided by an embodiment of the present invention;

[0081] Figure 3 is a 180-minute sample data curve diagram of an embodiment of the present invention;

[0082] Figure 4 It is a simulation result diagram of the short-term wind speed prediction method based on time and space dimensions provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0083] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0084] In the description of the present invention, it should be noted that the terms "upper", "lower", "inner", "outer", "front end", "rear end", "two ends", "one end", "the other end" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0085] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "provided with", "connected", etc. should be understood in a broad sense. For example, "connected" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0086] 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.

[0087] Example 1

[0088] This embodiment discloses a short-term wind speed prediction method based on time and space dimensions, including:

[0089] S1. Obtain a wind speed time series, and use a re-standardized range analysis method to obtain the Hurst index of the wind speed time series to determine the predictability of the wind speed time series; including:

[0090] For the wind speed time series {x i ,t=0,1,...,T}, let e T is the mean of the sequence, then the cumulative deviation X(t,T), range R(T) and standard deviation S(T) of the sequence are:

[0091]

[0092] The quotient R(T) / S(T) is called the re-calibrated range, denoted as (R / S) T :

[0093] (R / S) T =cTH ;

[0094] Among them, H is the Hurst exponent, c is a positive constant independent of the parameter T, and the relationship between the Hurst exponent H of the time series and the fractal dimension D satisfies: H + D = 2;

[0095] If H = 0.5, it indicates that the time series is random or non-autocorrelated and is unpredictable at this time; if H ≠ 0.5, it indicates that the time series is a non-random sequence with fractal characteristics. At this time, it is divided into two cases: if 0 ≤ H < 0.5, the original sequence has anti-persistence, and the time series shows anti-correlation at any scale. The closer the H value is to 0, the greater the intensity of the anti-persistence of the time series; if 0.5 < H ≤ 1, it indicates that the sequence has positive correlation, that is, the trend of the sequence in the latter period is the same as that in the previous period, indicating that the time series has predictability. The closer the H value is to 1, the greater the intensity of the persistence of the time series. When H = 1, the time series is completely predictable.

[0096] S2. Calculate the parameters in the affine transformation, perform fractal interpolation extrapolation prediction for each moment, and obtain the fractal interpolation extrapolation prediction result; including:

[0097] S201. Select the initial point set from the known data;

[0098] S202. Obtain the vertical scale factor d of the affine transformation by the empirical method, geometric method or analytical method i ;

[0099] S203. Calculate other parameters in the affine transformation to obtain the affine transformation type IFS of the inner interval system;

[0100] S204. Obtain the attractor of the IFS by the deterministic or stochastic iterative algorithm;

[0101] S205. Determine its abscissa x from the point to be extrapolated or predicted B , and assign an initial value to its ordinate At the same time, assign the extrapolation search times variable l = 0, and assign a value to the micro-perturbation value ε of y C ; y ;

[0102] S206. Start iterating K times from the point to obtain the point set B of the Kth iteration 1 , compare them with the attractor, and obtain the mean square deviation E from the attractor 1 ;

[0103] S207. Start iterating K times from the point to obtain the point set B of the Kth iteration 2 , compare them with the attractor, and obtain the mean square deviation E from the attractor2 ;

[0104] S208, calculate the gradient change of the mean square deviation And make convergence judgment,

[0105] like Then Output as the result, and then end the search process;

[0106] like ε is the gradient convergence limit, then proceed to the next step;

[0107] S209, based on the revised formula Find the new y C value;

[0108] S211, set l=l+1, and then return to step S206 to continue execution.

[0109] S3. Define the time before the last moment of the acquired wind speed time series as the interpolation point, use the atmospheric parameters at all heights of the interpolation point to obtain the Richardson number Ri, judge the atmospheric stability, substitute the atmospheric stability state into the wind shear index extrapolation model, extrapolate the data at that moment, and obtain the wind shear index extrapolation result; including:

[0110] Richardson number (Ri):

[0111]

[0112] Monin-Obukhov length (L):

[0113]

[0114] Where: g is the acceleration due to gravity, which is 9.8 m / s 2 ; z2 and z1 represent the vertical air height, T is the average absolute temperature of the two layers of air; △T is the temperature difference between the two layers of vertical air height z2 and z1; △u=u2-u1, u2 and u1 are the wind speeds at the two layers of vertical air height z2 and z1, is the geometric mean height,

[0115] u * is the friction velocity, k is the von Karman constant, is the temperature gradient;

[0116] Friction speed u * :

[0117]

[0118] τ 0is the surface shear stress (unit: Pa), ρ is the air density, usually ρ = 1.225 kg / m 3 ;

[0119] When -0.6≤Ri<0.2, the atmospheric stability is neutral, the wind speed does not change much in the vertical direction, and turbulence exists;

[0120] When Ri ≥ 0.2, the atmospheric stability is stable, the vertical wind speed is small, and it is characterized by a steady wind speed;

[0121] When Ri<-0.6, the atmospheric stability is unstable, the vertical wind speed is strong, there is strong convection and turbulence, the wind speed changes greatly, and there are obvious fluctuations in height.

[0122] According to the atmospheric stability state, the wind shear index extrapolation model is substituted to extrapolate the data at that moment to obtain the wind shear index extrapolation results, including:

[0123] Calculate the ground roughness z 0 And the stability function Φ m ,

[0124]

[0125] When the atmospheric stability is stable,

[0126]

[0127] When the atmospheric stability is neutral,

[0128]

[0129] When the atmospheric stability is unstable,

[0130]

[0131]

[0132] δ 0 =[1-15(z 0 / L)] -0.25 ;

[0133] Where: Φ m and is the Monin-Obukhov stability function; z 0 is the roughness length; h 1 and h 2 are 2 different heights; δ and δ 0is the Monin-Obukhov scaling function;

[0134] Calculate the wind shear exponent α:

[0135]

[0136] Calculate the wind speed based on the wind shear exponent power law formula:

[0137]

[0138] Where: z1, z2 are vertical heights; v1, v2 are wind speeds at vertical heights z1, z2 respectively; α is the wind shear index.

[0139] S4. Fractal interpolation prediction results and wind shear index extrapolation results are integrated to evaluate the results. Including:

[0140] Considering the relative importance of different result data, the weighted average method is used to fuse the fractal interpolation extrapolation prediction results and the wind shear index extrapolation results. By assigning a weight to each data, its contribution to the final result is affected. The expression is:

[0141]

[0142] Where: n is the dimension of the data, which is 2 in this paper; y i (t) is the result obtained by the i-th method at time t; w i (t) is the weighting factor of the ith method.

[0143] Example 2

[0144] This embodiment discloses a short-term wind speed prediction system based on time and space dimensions, including:

[0145] An acquisition module, configured to acquire a wind speed time series;

[0146] a judgment module configured to analyze the predictability of the wind speed time series;

[0147] A fractal interpolation and extrapolation module is configured to predict the results of fractal interpolation and extrapolation;

[0148] A wind shear index extrapolation module is configured to calculate a wind shear index extrapolation result;

[0149] The result fusion module is configured to fuse the prediction results of the type interpolation extrapolation and the wind shear index extrapolation results to obtain the results and evaluate the results.

[0150] Specific examples:

[0151] The radar data (20230720) of a domestic wind farm is used as the feature extraction object of the hybrid prediction algorithm. All data are valid values. The original data are as follows: Figure 2 This specific example takes radar data time series as an example. If there are other uses, you can choose indicators according to your own needs. The specific steps include:

[0152] (1) Obtain the wind speed time series. In this example, 180 minutes of continuous data are extracted as samples. The sample curve is as follows: Figure 3 shown.

[0153] (2) Import the radar data 20230720 in the file into MATLAB software to generate the file "WSarr.m", put this file and the "R_S.m" code in the same folder, and use MATLAB to open the "R_S.m" program.

[0154] (3) Modify the file input path according to your personal situation.

[0155] (4) Run the code to obtain the calculation result graph of R / S analysis and the Hurst index. By calculating the Hurst index, we get H = 0.79561>0.5, which means that the time series has a strong long-memory positive correlation and can be analyzed using the fractal method.

[0156] (5) Calculate the parameters a, c, e, and f in the affine transformation using the formula in S32. Four groups of 1×180 one-dimensional arrays are obtained through calculation.

[0157] (6) Place the "WSarr.m" and "iteration.m" codes in the same folder and open the "iteration.m" program using MATLAB software.

[0158] (7) Modify the file input path according to your personal situation.

[0159] (8) Wherein “WSarr.m” is the original data, and “WSarr.m” is the input data of the fractal interpolation model which takes the original data as the interpolation point with 1 as the step interval.

[0160] (9) Since the only influencing parameter of the fractal interpolation model is the vertical scale factor d i , respectively take d i =0.3, 0.5 and 0.7 for calculation.

[0161] (10) Click Run to obtain the calculation results of the fractal interpolation model parameters. The results show that when the absolute value of the vertical scale factor is larger, the local fluctuation of the fractal interpolation curve is more obvious.

[0162] (11) Select atmospheric parameters at all altitudes from the interpolation points, calculate the Richardson number, and determine the atmospheric stability.

[0163] (12) Place the files “WSarr.m” and “Monin_Obukhov.m” code in the same folder and open the “Monin_Obukhov.m” program using MATLAB.

[0164] (13) Modify the file input path according to your personal situation.

[0165] (14) Run the code to obtain the calculation result of the Richardson number.

[0166] (15) The atmospheric stability is judged. At a certain moment, by calculating the Richardson number, it is obtained that Ri = -0.6 < 9.3590e-04 < -0.2, indicating that the atmospheric stability state at this moment is neutral.

[0167] (16) After obtaining the parameters, the hybrid algorithm is used to make predictions at each moment.

[0168] (17) Place the “completed_fenxing.m”, “predict.m”, “Monin_Obukhov.m”, “windshear.m” and “iteration.m” codes in the same folder, and open the “completed_fenxing.m” program using MATLAB software.

[0169] (18) Click Run to simulate the fractal interpolation model.

[0170] (19) The prediction results at a height of 20 meters are as follows Figure 4 As shown, the results show that there is a small difference between the original data curve and the simulation result curve. The simulation results can not only reflect the fluctuations in most periods of the time series, but also perform well at short-term mutation data points without large deviations.

[0171] (20) Model results evaluation.

[0172] The error between the simulation results of the fractal interpolation model and the true value of each point is calculated. The results show that the average error rate (average error / average true value) is 0.0337, and the overall error rate is small, which proves that (the short-term wind speed prediction algorithm based on the time and space dimensions) has a good prediction effect and is qualified as a wind speed time series prediction model.

[0173] The above is only an embodiment of the present invention, and the common sense such as the known specific structure and characteristics in the scheme is not described in detail here. For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive, and the scope of the present invention is defined by the attached claims rather than the above description, so it is intended to include all changes within the meaning and scope of the equivalent elements of the claims. Any figure mark in the claims should not be regarded as limiting the claims involved.

Claims

1. A short-term wind speed prediction method based on time and space dimensions, characterized in that: Including: S1. Obtain the wind speed time series, and use the rescaled range analysis method to obtain the Hurst exponent of the wind speed time series, and judge the predictability of the wind speed time series; S2. Calculate the parameters in the affine transformation, perform fractal interpolation extrapolation prediction for each moment, and obtain the fractal interpolation extrapolation prediction result; S3. Define the points before the last moment of the obtained wind speed time series as interpolation points, obtain the Richardson number Ri using the atmospheric parameters at all heights of the interpolation points, judge the atmospheric stability, and according to the atmospheric stability state, substitute it into the wind shear exponent extrapolation model to extrapolate the data at this moment to obtain the wind shear exponent extrapolation result; S4. Integrate the fractal interpolation extrapolation prediction result and the wind shear exponent extrapolation result, and conduct result evaluation.

2. The short-term wind speed prediction method based on time and space dimensions according to claim 1 is characterized in that: The obtaining of the wind speed time series, and using the rescaled range analysis method to obtain the Hurst exponent of the wind speed time series, and judging the predictability of the wind speed time series includes: For the wind speed time series {x i ,t=0,1,...,T}, let e T is the mean of the sequence, then the cumulative deviation X(t,T), range R(T) and standard deviation S(T) of the sequence are: The quotient R(T) / S(T) is called the re-calibrated range, denoted as (R / S) T : (R / S) T =cT H ; Among them, H is the Hurst exponent, c is a positive constant independent of the parameter T, and the relationship between the Hurst exponent H of the time series and the fractal dimension D satisfies: H + D = 2; When H = 0.5, it indicates that the time series is random or non-autocorrelated, and it is not predictable at this time; when H ≠ 0.5, it indicates that the time series is a non-random sequence with fractal characteristics. At this time, it is divided into two cases: if 0 ≤ H < 0.5, the original sequence has anti-persistence, and the time series shows anti-correlation at any scale. The closer the H value is to 0, the stronger the anti-persistence intensity of the time series; if 0.5 < H ≤ 1, it indicates that the sequence has positive correlation, that is, the trend of the sequence in the latter period is the same as that in the previous period, indicating that the time series has predictability. The closer the H value is to 1, the stronger the persistence intensity of the time series. When H = 1, the time series is completely predictable.

3. The short-term wind speed prediction method based on time and space dimensions according to claim 1 or 2 is characterized in that: The calculating of the parameters in the affine transformation, performing fractal interpolation extrapolation prediction for each moment, and obtaining the fractal interpolation extrapolation prediction result includes: S201. Select the initial point set from the known data; S202, obtain the vertical scale factor d of the affine transformation by empirical method, geometric method or analytical method i ; S203. Calculate other parameters in the affine transformation to obtain the affine transformation type IFS of the inner interval system; S204. Use the deterministic or stochastic iterative algorithm to obtain the attractor of the IFS; S205, determine the horizontal coordinate x of the point to be extrapolated or predicted B , and assign an initial value to its ordinate At the same time, assign the extrapolated search times variable l = 0, and set y C The perturbation value ε y Assignment; S206, from point Start iterating K times, get the point set B1 of the Kth iteration, compare them with the attractor, and find the mean square deviation E1 from the attractor; S207, from point Start iterating K times, get the point set B2 of the Kth iteration, compare them with the attractor, and find the mean square deviation E2 from the attractor; S208, calculate the gradient change of the mean square deviation And make convergence judgment, like Then Output as the result, and then end the search process; like ε is the gradient convergence limit, then proceed to the next step; S209, based on the revised formula Find the new y C value; S211. Let l = l + 1, and then return to step S206 to continue execution.

4. The short-term wind speed prediction method based on time and space dimensions according to claim 3 is characterized in that: The obtaining of the Richardson number Ri and judging the atmospheric stability includes: Richardson number (Ri): Monin-Obukhov length (L): Where: g is the acceleration due to gravity, which is 9.8 m / s 2 ; z1, z2 are vertical heights, T is the average absolute temperature of the two layers at vertical heights; △T is the temperature difference between the two layers of air at vertical heights z2 and z1; △v = v2-v1, v2 and v1 are the wind speeds at the two layers of air at vertical heights z2 and z1, is the geometric mean height, u * is the friction velocity, k is the von Karman constant, is the temperature gradient; Friction speed u * : When -0.6 ≤ Ri < 0.2, the atmospheric stability is neutral, the wind speed changes little in the vertical direction, and there is turbulence; When Ri ≥ 0.2, the atmospheric stability is stable, the vertical wind speed is small, and the wind speed shows a steady state; When Ri < -0.6, the atmospheric stability is unstable, the vertical wind speed is strong, there is strong convection and turbulence, the change of the wind speed is large, and there are obvious fluctuations in height.

5. The short-term wind speed prediction method based on time and space dimensions according to claim 4 is characterized in that: According to the atmospheric stability state, substitute it into the wind shear exponent extrapolation model to extrapolate the data at this moment to obtain the wind shear exponent extrapolation result, including: Calculate the ground roughness z0 and stability function When the atmospheric stability is stable, When the atmospheric stability is neutral, When the atmospheric stability is unstable, δ0 = [1 - 15(z0 / L)]; Where: Φ m and is the Monin-Obukhov stability function; z0 is the roughness length; h1 and h2 are two different heights; δ and δ0 are the Monin-Obukhov scaling functions; Obtain the wind shear exponent α: Estimate the wind speed according to the wind shear exponent power law formula: Where: v1, v2 are the wind speeds at vertical heights z1 and z2 respectively; α is the wind shear index.

6. The method according to claim 5, characterized in that The fractal interpolation extrapolation prediction result and the wind shear index extrapolation result are integrated to evaluate the result, including: Considering the relative importance of different result data, the weighted average method is used to fuse the fractal interpolation extrapolation prediction results and the wind shear index extrapolation results. By assigning a weight to each data, its contribution to the final result is affected. The expression is: Where: n is the dimension of the data, which is 2 in this paper; y i (t) is the result obtained by the i-th method at time t; w i (t) is the weighting factor of the ith method.

7. A system for implementing the short-term wind speed prediction method based on time and space dimensions as described in any one of claims 1 to 6, characterized in that: include: An acquisition module, configured to acquire a wind speed time series; a judgment module configured to analyze the predictability of the wind speed time series; A fractal interpolation and extrapolation module is configured to calculate a fractal interpolation and extrapolation prediction result; A wind shear index extrapolation module is configured to calculate a wind shear index extrapolation result; The result fusion module is configured to fuse the prediction results of the type interpolation extrapolation and the wind shear index extrapolation results to obtain the results and evaluate the results.

8. An electronic device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 6 when executing the computer program.

9. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images