Wind power output prediction method and device considering micrometeorology and microtopography cooperative influence

Abstract

The invention discloses a wind power output prediction method and device considering micrometeorology and microtopography cooperative influence. The method comprises the following steps: acquiring real-time and historical first data and a historical wind power output curve of a target area; the first data comprises micrometeorological and microtopographic parameters; dividing the target area into a preset number of sub-areas based on a first condition and a second condition according to the real-time first data and a historical wind power output curve; obtaining a wind power output prediction model of each sub-region according to the micro-meteorological parameters of the real-time first data in the sub-region, the wind turbine generator set parameters and the effective wind speed model; the effective wind speed model considers the influence of turbulent flow and wake flow; after the wind power output prediction model is optimized according to historical first data, the wind power output prediction models of all the sub-regions are subjected to weighted fusion, and a total wind power output prediction model is obtained; and realizing wind power output prediction of the target area according to the wind power output prediction total model.

Description

Technical Field

[0001] The present invention relates to the field of wind power output prediction, and in particular, to a wind power output prediction method and device considering the collaborative influence of micro-meteorology and micro-topography. Background Art

[0002] As an efficient, clean and sustainable energy form, wind power is playing an increasingly important role in the power structure. The output level of a wind farm not only depends on the large-scale numerical weather forecast field and the power curve of the wind turbine equipment itself, but is also deeply affected by the micro-meteorological factors and micro-topographical factors within the farm.

[0003] Micro-meteorological factors usually show significant spatial heterogeneity in a wind farm, resulting in a deviation between the actual output of a single wind turbine and the power predicted based on the existing average meteorological conditions; at the same time, the regulating effect of micro-topographical factors on micro-meteorological conditions will also cause fluctuations in the meteorological parameters within the farm. On the one hand, most of the current mainstream wind power output prediction methods follow the average meteorological parameters of the whole farm or rough terrain influence models, ignoring the distribution characteristics of the micro-scale meteorological field and its complex regulating effect on the effective wind speed, thus making it difficult to meet the requirements of modern power grids for high-precision and low-fluctuation prediction. On the other hand, the terrain channel effect will accelerate or weaken the local wind speed flowing through the wind turbine, resulting in a significant deviation of the wind turbine output power from the calibrated power curve; the terrain occlusion or the interaction between wind turbines will form complex wake and turbulence structures, further enhancing or attenuating the turbulence intensity and wind speed gradient of the downstream wind turbines; the local reshaping of temperature by micro-topographical factors will change the air density, thereby affecting the mechanical capture efficiency and power generation performance of the wind turbines. Although some studies have tried to compensate for local effects through empirical zoning or statistical correction, the existing methods often lack a continuous evaluation and dynamic adjustment mechanism for the system, making it difficult to cope with the uncertainties brought by instantaneous or seasonal micro-meteorological factor fluctuations, and thus resulting in large errors and significant fluctuations in the existing wind power output prediction methods in high-precision scenarios, making it difficult to meet the requirements of the power grid for rapid response and intelligent scheduling of wind power.

[0004] Therefore, there is an urgent need for a new technical solution to solve the technical problem of how to improve the accuracy of wind power output prediction. Summary of the Invention

[0005] The present invention provides a wind power output prediction method and device considering the collaborative influence of micro-meteorology and micro-topography, so as to solve the technical problem of how to improve the accuracy of wind power output prediction.

[0006] To achieve the above object, the present invention provides a wind power output prediction method considering the collaborative influence of micro-meteorology and micro-topography, including: Obtain the real-time first data, historical first data, and historical wind power output curve of the target area; the first data includes preset micro-meteorological parameters and micro-topographical parameters; divide the target area into a preset number of sub-areas based on the real-time first data and the historical wind power output curve according to the first condition and the second condition.

[0007] The first condition includes that the wind turbines in the same sub-area have first data within a preset similarity range; the second condition includes that the maximum difference between the characteristic center vectors of the sub-areas on each coordinate component is less than or equal to a preset threshold.

[0008] Obtain the wind power output prediction model for each sub-area according to the micro-meteorological parameters, wind turbine parameters, and effective wind speed model of the real-time first data in the sub-area; the effective wind speed model takes into account the influence of turbulence and wake.

[0009] Optimize the wind power output prediction model according to the historical first data, and then perform weighted fusion on the wind power output prediction models of each sub-area to obtain the total wind power output prediction model.

[0010] Realize the wind power output prediction of the target area according to the total wind power output prediction model.

[0011] Preferably, dividing the target area into a preset number of sub-areas based on the real-time first data and the historical wind power output curve according to the first condition and the second condition includes: Divide the target area into a preset number of sub-areas based on the real-time first data and the historical wind power output curve according to the first condition; obtain the first index; the first index includes the maximum difference between the characteristic center vectors of the sub-areas on each coordinate component; if the first index is greater than the preset threshold, adjust the partition based on the first condition and obtain the first index again until the first index meets the second condition to complete the division of the target area, including: ; Among them, represents the set of sub-areas that are geographically adjacent to sub-area k; and both represent sub-areas; represents sub-area and the maximum difference between the characteristic centers of on each coordinate component; represents the preset threshold; represents the output power of the hth unit at the historical input wind speed ; and represent the polynomial coefficients of the hth unit; represents the historical input wind speed; represents the The multi-dimensional feature vectors extracted from the sampling points, where the multi-dimensional feature vectors include micro-meteorological parameters, micro-topographical parameters, and the predicted wind power values considering only the wind speed; Denotes the set of all sub-region indicator variables; Denotes the sub-region Characteristic center vector; Denotes the total number of sub-regions; Denotes the total number of sampling points in all sub-regions; Denotes the sampling point; Denotes the Whether the -th sampling point is assigned to the sub-region k. If Denotes the -th sampling point is assigned to the k-th sub-region. If Denotes the -th sampling point is not assigned to the k-th sub-region; Denotes the sub-region Characteristic center vector.

[0012] Preferably, the effective wind speed model includes: Obtain the wake influence model of the sub-region according to the real-time first data of the sub-region combined with the influence model of the nearby wake on the target wind turbine; obtain the turbulence influence model of the sub-region according to the micro-topographical data in the real-time first data of the sub-region combined with the influence model of the nearby turbulence on the target wind turbine; obtain the effective wind speed model of the sub-region according to the wake influence model and the turbulence influence model.

[0013] Preferably, obtaining the effective wind speed model of the sub-region according to the wake influence model and the turbulence influence model includes: The effective wind speed model includes: ; ; ; ; Wherein, Denotes the effective wind speed; And Respectively denote the weight coefficients of the wake and the turbulence in the sub-region ; Denotes the local turbulence intensity of the coordinate , that is, the turbulence influence model; Denotes the wake wind speed at the coordinate , that is, the wake influence model; Denotes the local slope angle of the wind turbine position; Denotes taking the partial derivative; Denotes the non-affected turbulence intensity; Denotes the turbulence correction model; Indicates the distance extending from the wind turbine location along the wind direction; represents the offset relative to the wake centerline; Indicates the height deviation relative to the turbine hub in the vertical direction; A model representing the impact of nearby turbulence on the target wind turbine; represents the initial turbulence intensity amplification factor; represents the turbulence attenuation scale; represents the micro-topography influence coefficient; Microtopography correction factor representing the deviation between local terrain height and average terrain height and slope effect; represents the lateral diffusion coefficient; Indicates the reference altitude; represents the diffusion shape index; Represents the overall weight coefficient of the surrounding wind turbine interference; express The first The surrounding wind turbines that cause disturbances to the target wind turbine; Indicates the total number of surrounding fans that disturb the turbulence of the target fan; Indicates the target fan With the The center distance between typhoon turbines; represents the relative bearing attenuation factor; represents the angle attenuation index; Indicates the The angle between the line connecting the typhoon turbine and the target turbine and the main wind direction; 、 、 Indicates the The three-dimensional coordinates of the typhoon; Indicates the absolute height of the local terrain; represents the average terrain height of the region; Indicates the wind speed when there is no disturbance; represents the density correction factor; represents the wake attenuation model; The model represents the impact of nearby wake on the target wind turbine; represents the humidity correction factor; represents the initial wake attenuation coefficient; represents the wake recovery length under flat terrain; represents the micro-topography influence coefficient; represents the micro-topography correction factor; represents the wake diffusion parameter; Indicates the fan impeller diameter; express The first Surrounding wind turbines that generate disturbances to the target wind turbine; Indicates the total number of surrounding wind turbines that generate disturbances to the wake of the target wind turbine; , , Indicates the three-dimensional coordinates of the Indicates the target wind turbine and the central distance between the Indicates the interference intensity factor of the h-th surrounding wind turbine; Indicates the characteristic decay length of the interference effect of the h-th wind turbine; Indicates the interference decay exponent; Indicates the sensitivity of wake decay to density change; Indicates the air density; Indicates the reference density; Indicates the influence factor of humidity on wake recovery; Indicates the relative humidity; Indicates the reference humidity.

[0014] Preferably, the wind power output prediction model for each sub-region obtained according to the micro-meteorological parameters, wind turbine parameters and effective wind speed model in the real-time first data in the sub-region includes: The wind power output prediction model includes: ; Wherein, Indicates the predicted wind power output value of the sub-region ; Indicates the air density within the sub-region , which is the micro-meteorological parameter from the real-time first data; Indicates the area swept by the wind turbine within the sub-region ; Indicates the power coefficient of the wind turbine within the sub-region .

[0015] Preferably, after optimizing the wind power output prediction model according to the historical first data, the wind power output prediction models of each sub-region are weighted and fused to obtain the total wind power output prediction model, including: Obtain the prediction results of each sub-region according to the wind power output prediction model of each sub-region; obtain the prediction error of the wind power output prediction model of each sub-region according to the prediction results of each sub-region and the historical wind power output curve.

[0016] Iteratively optimize the wind power output prediction model with a prediction error greater than the preset error threshold according to the historical first data, and obtain the prediction error of the wind power output prediction model after each optimization; when the iterative optimization meets the third condition or the fourth condition and the prediction error is less than or equal to the preset error threshold, complete the optimization of the wind power output prediction model.

[0017] The third condition includes that the number of iterations reaches the preset upper limit; the fourth condition includes that the relative change rate of the prediction error after two consecutive optimizations of the wind power output prediction model is less than the preset change rate threshold.

[0018] When the prediction errors of the wind power output prediction models of each sub-region are all less than or equal to the preset error threshold, weight and fuse the wind power output prediction models of each sub-region to obtain the total wind power output prediction model.

[0019] Preferably, obtaining the prediction error of the wind power output prediction model of each sub-region according to the prediction result of each sub-region and the historical wind power output curve includes: Compare the prediction result of the sub-region with the historical wind power output curve to obtain the prediction error: ; wherein, represents the prediction error of the k-th prediction in sub-region o; represents the prediction result of the k-th prediction in sub-region o; represents the historical actual wind power output value of the k-th prediction in sub-region o.

[0020] Preferably, iteratively optimizing the wind power output prediction model with a prediction error greater than the preset error threshold according to the historical first data includes: Calculate the deviation of the micro-meteorological parameters and the deviation of the micro-topographic parameters between the historical first data and the real-time first data; normalize the deviation; take the average within the first data set for the normalized deviation to obtain the correction direction; obtain the correction step size through the exponential decay formula according to the number of samples within the first data set; use the correction step size to update the parameters of the wind power output prediction model with a prediction error greater than the preset error threshold along the correction direction, including: ; wherein, represents the historical data set of the -th type of environmental characteristics in sub-region represents the environmental characteristic category number; represents the environmental characteristic classification function; represents the historical sample; represents the -th topographic feature vector corresponding to the historical sample in sub-region Represents the micro-meteorological feature vector corresponding to the th historical sample of the sub-region; Represents the parameter error vector of the th sample of the sub-region; Represents the original model parameter vector of the th sample of the sub-region; Represents the optimal parameter obtained by offline fitting for the th sample of the sub-region; Represents the average correction direction of the sub-region ; Represents the update step size of the sub-region ; Represents the step size factor; Represents the natural base; Represents the attenuation rate coefficient; Represents the updated parameter vector of the sub-region ; Represents the parameter vector before update of the sub-region ; Represents the wind power output prediction model after updating the parameters of the sub-region ; Represents the wind power output prediction model before updating the parameters of the sub-region .

[0021] The relative change rate of the prediction error in the fourth condition includes: ; wherein, represents the relative change rate of the k-th prediction error; represents the prediction error after the (k + 1)-th iteration; represents the prediction error after the k-th iteration.

[0022] Preferably, the wind power output prediction models of each sub-region are weighted and fused to obtain the total wind power output prediction model, including: ; wherein, represents the total wind power output prediction model; represents the weight of sub-region o.

[0023] The present invention also provides a wind power output prediction device considering the collaborative influence of micro-meteorology and micro-topography, for the method of the present invention. The device includes a first module, a second module, a third module, and a fourth module.

[0024] The first module is used to obtain real-time first data, historical first data, and historical wind power output curves of the target area; the first data includes preset micro-meteorological parameters and micro-topographical parameters.

[0025] The second module is used to divide the target area into a preset number of sub-areas based on the real-time first data and the historical wind power output curve according to the first condition and the second condition.

[0026] The first condition includes that the wind turbines in the same sub-area have first data within a preset similarity range; the second condition includes that the maximum difference between the characteristic center vectors of the sub-areas on each coordinate component is less than or equal to a preset threshold.

[0027] The third module is used to obtain a wind power output prediction model for each sub-area according to the micro-meteorological parameters, wind turbine parameters, and effective wind speed model of the real-time first data in the sub-area; the effective wind speed model takes into account the influence of turbulence and wake.

[0028] The fourth module is used to optimize the wind power output prediction model according to the historical first data and then perform weighted fusion on the wind power output prediction models of each sub-area to obtain a total wind power output prediction model; the wind power output of the target area is predicted according to the total wind power output prediction model.

[0029] The present invention has the following beneficial effects: The wind power output prediction method considering the collaborative influence of micro-meteorology and micro-topography according to the present invention divides a target area into a preset number of sub-areas based on a first condition and a second condition according to real-time first data and a historical wind power output curve. The method partitions the target area considering real-time micro-meteorological parameters, micro-topographical parameters, and the historical wind power output curve. That is, the partitioning method of the present invention considers more factors that affect the partitioning, making the partitioning have a higher accuracy. At the same time, when partitioning, it is ensured that the wind turbines within the same sub-area have first data within a preset similarity range, enabling the partitioning to divide similar areas into the same sub-area and further improving the accuracy of the partitioning. Further, when partitioning, it is also ensured that the maximum difference between the characteristic center vectors of different sub-areas on each coordinate component is less than or equal to a preset threshold, enabling the method of the present invention to further consider the accuracy of the partitioning. The method of the present invention considers multiple factors and two conditions to divide the target area, making the partitioning of the method of the present invention have a higher accuracy and providing a basis for subsequent wind power output prediction. When constructing the wind power output prediction model for the sub-area, the method of the present invention considers the influence of turbulence and wake, enabling the wind power output prediction model of the present invention to consider more factors that affect wind power generation, thereby improving the prediction accuracy of the wind power output prediction model. In the method of the present invention, the wind power output prediction model is optimized according to historical first data, enabling the wind power output prediction model of the sub-area to have good prediction performance and avoiding prediction errors. In the method of the present invention, the wind power output prediction models of each sub-area are weighted and fused to obtain a total wind power output prediction model, enabling the total wind power output prediction model to consider the weights of each sub-area, improving the attention to key sub-areas, and further making a more accurate wind power output prediction for the entire target area. The method of the present invention can realize dynamic adjustment of partitioning, iterative correction of sub-area prediction models, and continuous evaluation, thereby improving the accuracy of wind power output prediction in the target area, reducing the warning response delay, and providing reliable support for the safe and stable operation of the power grid.

[0030] The wind power output prediction device considering the collaborative influence of micro-meteorology and micro-topography according to the present invention is used for the method of the present invention and has the same beneficial effects as the method of the present invention.

[0031] In addition to the purposes, features, and advantages described above, the present invention has other purposes, features, and advantages. The following will refer to the accompanying drawings for a further detailed description of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1 is a schematic flow chart of the method of the preferred embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, the present invention can be implemented in many different ways defined and covered by the claims.

[0034] See Figure 1 , in a preferred embodiment of the present invention, a wind power output prediction method considering the collaborative influence of micro-meteorology and micro-topography is provided, including: S1. Obtain the real-time first data, historical first data, and historical wind power output curve of the target area; the first data includes preset micro-meteorological parameters and micro-topographical parameters.

[0035] In a preferred embodiment of the present invention, the influence of micro-topographical parameters on micro-meteorological parameters mainly considers the interference with wind speed. Different micro-topographical parameters will cause changes in wind speed. For example, changes in depressions and slopes will change the direction and velocity of the air flow, thereby forming complex local meteorology. For example, areas with higher wind speeds may cause rapid mixing of local air, making the temperature distribution tend to be uniform, and thus directly affecting the wind capture efficiency and power generation output of wind turbines.

[0036] In a preferred embodiment of the present invention, the influence of changes in temperature and humidity in micro-meteorological parameters on wind power output is considered. For example, changes in air temperature and humidity will directly affect air density, that is, changes in local temperature and humidity affect air density, thereby affecting wind power output. At the same time, local temperature differences will change air density, and humidity changes will affect surface vegetation and soil moisture, and thus play a regulatory role in surface roughness.

[0037] In a preferred embodiment of the present invention, the micro-meteorological parameters and micro-topographical parameters are preset as needed. For example, the micro-meteorological parameters can be set to include, but are not limited to, parameters such as temperature, wind speed, turbulence intensity, humidity, and air density; the micro-topographical parameters can be set to include, but are not limited to, parameters such as slope, terrain height, absolute terrain height, steep slope and gentle slope classification, etc.

[0038] In a preferred embodiment of the present invention, preset local airflow characteristics can be obtained according to the micro-meteorological parameters, such as turbulence intensity, airflow acceleration area, and deceleration area, etc.

[0039] S2. Divide the target area into a preset number of sub-areas based on the real-time first data and the historical wind power output curve according to the first condition and the second condition.

[0040] The first condition includes that the wind turbines in the same sub-area have first data within a preset similarity range; the second condition includes that the maximum difference between the characteristic center vectors of the sub-areas on each coordinate component is less than or equal to a preset threshold.

[0041] In a preferred embodiment of the present invention, S2 specifically includes: Divide the target area into a preset number of sub-areas based on the real-time first data and the historical wind power output curve according to the first condition; obtain the first index; the first index includes the maximum difference of the characteristic center vectors between sub-areas on each coordinate component; if the first index is greater than the preset threshold, adjust the partition based on the first condition and obtain the first index again until the first index meets the second condition to complete the division of the target area, including: ; Among them, represents the set of sub-areas geographically adjacent to the sub-area; and both represent sub-areas; represents the maximum difference of the characteristic center of the sub-area and on each coordinate component; represents the preset threshold; represents the output power of the h-th unit at the historical input wind speed ; and represent the polynomial coefficients of the h-th unit; represents the historical input wind speed; represents the multi-dimensional feature vector extracted from the -th sampling point, and the multi-dimensional feature vector includes micro-meteorological parameters, micro-topographic parameters, and wind power prediction values considering only wind speed; represents the set of all sub-area indicator variables; represents the characteristic center vector of the sub-area ; represents the total number of sub-areas; represents the total number of sampling points of all sub-areas; represents the sampling point; represents whether the -th sampling point is assigned to the sub-area k. If represents that the -th sampling point is assigned to the k-th sub-area. If represents that the -th sampling point is not assigned to the k-th sub-area; represents the characteristic center vector of the sub-area ;

[0042] When performing area division, it specifically includes: first, fitting the h-th unit using the historical data set to construct the power model . Then, using as the feature vector of the -th point. During the clustering process, define the binary indicator variable . If the point The Euclidean distance from the k-th cluster center is the smallest, then , otherwise it is 0. When all clusters are stable, calculate the infinity norm difference and between adjacent clusters . If its maximum value exceeds the preset threshold , it is considered that the sub-region feature difference is too large and needs to be returned for re-partitioning; otherwise, the partitioning is completed.

[0043] In a preferred embodiment of the present invention, when adjusting the partition based on the first condition and obtaining the first index again until the first index meets the second condition, including a preset partition adjustment times threshold, if the number of times of adjusting the partition exceeds the preset partition adjustment times threshold, a severe warning is output, indicating that manual intervention is triggered.

[0044] S3. Obtain the wind power output prediction model of each sub-region according to the micro-meteorological parameters, wind turbine parameters and effective wind speed model of the real-time first data in the sub-region; the effective wind speed model considers the influence of turbulence and wake.

[0045] In a preferred embodiment of the present invention, the effective wind speed model includes: Obtain the wake influence model of the sub-region according to the real-time first data of the sub-region combined with the influence model of the nearby wake on the target wind turbine; obtain the turbulence influence model of the sub-region according to the micro-topography data in the real-time first data of the sub-region combined with the influence model of the nearby turbulence on the target wind turbine; obtain the effective wind speed model of the sub-region according to the wake influence model and the turbulence influence model.

[0046] In a preferred embodiment of the present invention, obtaining the effective wind speed model of the sub-region according to the wake influence model and the turbulence influence model includes: The effective wind speed model includes: ; ; ; ; wherein, represents the effective wind speed; and respectively represent the weight coefficients of the wake and turbulence in the sub-region ; represents the local turbulence intensity of the coordinate , that is, the turbulence influence model; represents the wake wind speed at the coordinate , that is, the wake influence model; represents the local slope angle of the wind turbine position; represents taking the partial derivative; Indicates the unaffected turbulence intensity; Indicates the turbulence correction model; Indicates the distance extending along the wind direction from the fan position; Indicates the offset relative to the wake centerline; Indicates the height deviation relative to the turbine hub in the vertical direction; Indicates the influence model of the nearby turbulence on the target fan; Indicates the initial turbulence intensity amplification factor; Indicates the turbulence decay scale; Indicates the microtopography influence coefficient; Indicates the microtopography correction factor of the deviation between the local terrain height and the average terrain height and the slope effect; Indicates the lateral diffusion coefficient; Indicates the reference height; Indicates the diffusion shape index; Indicates the overall weight coefficient of the surrounding fan interference; Indicates the th surrounding fan that disturbs the target fan within; Indicates the total number of surrounding fans that disturb the turbulence of the target fan; Indicates the target fan and the th fan; Indicates the relative azimuth attenuation factor; Indicates the angular attenuation index; Indicates the angle between the line connecting the hth fan and the target fan and the main wind direction; Indicates the th fan's three-dimensional coordinates; Indicates the absolute height of the local terrain; Indicates the regional average terrain height; Indicates the wind speed when not disturbed; Indicates the density correction factor; Indicates the wake attenuation model; Indicates the influence model of the nearby wake on the target fan; Indicates the humidity correction factor; Indicates the initial wake attenuation coefficient; Indicates the wake recovery length under flat terrain; Indicates the microtopography influence coefficient; Indicates the microtopography correction factor; Indicates the wake diffusion parameter; Indicates the fan impeller diameter; Indicates the Surrounding wind turbines that generate disturbances to the target wind turbine; Indicates the total number of surrounding wind turbines that generate disturbances to the wake of the target wind turbine; , , Indicates the three-dimensional coordinates of the th wind turbine; Indicates the target wind turbine and the th wind turbine; Indicates the interference intensity factor of the hth surrounding wind turbine; Indicates the characteristic decay length of the interference effect of the hth wind turbine; Indicates the interference decay exponent; Indicates the sensitivity of density change to wake decay; Indicates the air density; Indicates the reference density; Indicates the influence factor of humidity on wake recovery; Indicates the relative humidity; Indicates the reference humidity.

[0047] In a preferred embodiment of the present invention, the wind power output prediction model includes: ; Wherein, Indicates the predicted value of wind power output in the sub-region ; Indicates the air density in the sub-region , which is a micro-meteorological parameter from real-time first data; Indicates the area swept by the wind turbine in the sub-region ; Indicates the power coefficient of the wind turbine in the sub-region .

[0048] S4. After optimizing the wind power output prediction model according to historical first data, the wind power output prediction models of each sub-region are weighted and fused to obtain a total wind power output prediction model.

[0049] In a preferred embodiment of the present invention, S4 specifically includes: Obtain the prediction result of each sub-region according to the wind power output prediction model of each sub-region; obtain the prediction error of the wind power output prediction model of each sub-region according to the prediction result of each sub-region and the historical wind power output curve.

[0050] Iteratively optimize the wind power output prediction model whose prediction error based on historical first data is greater than the preset error threshold, and obtain the prediction error of the wind power output prediction model after each optimization; when the iterative optimization meets the third condition or the fourth condition and the prediction error is less than or equal to the preset error threshold, complete the optimization of the wind power output prediction model.

[0051] The third condition includes that the number of iterations reaches the preset upper limit; the fourth condition includes that the relative change rate of the prediction error after two consecutive optimizations of the wind power output prediction model is less than the preset change rate threshold.

[0052] When the prediction errors of the wind power output prediction models of each sub-region are all less than or equal to the preset error threshold, weight and fuse the wind power output prediction models of each sub-region to obtain the total wind power output prediction model.

[0053] In a preferred embodiment of the present invention, obtaining the prediction error of the wind power output prediction model of each sub-region according to the prediction result of each sub-region and the historical wind power output curve includes: Compare the prediction result of the sub-region with the historical wind power output curve to obtain the prediction error: ; wherein, represents the prediction error of the k-th prediction in sub-region o; represents the prediction result of the k-th prediction in sub-region o; represents the historical actual wind power output value of the k-th prediction in sub-region o.

[0054] In a preferred embodiment of the present invention, iteratively optimizing the wind power output prediction model whose prediction error based on historical first data is greater than the preset error threshold includes: Calculate the micro-meteorological parameter deviation and micro-topographical parameter deviation between the historical first data and the real-time first data; normalize the deviation; take the average within the first data group of the normalized deviation to obtain the correction direction; obtain the correction step size through the exponential decay formula according to the number of samples within the first data group; update the parameters of the wind power output prediction model whose prediction error is greater than the preset error threshold along the correction direction by using the correction step size, including: ; wherein, represents the historical data set of the th type of environmental characteristics in sub-region represents the environmental characteristic category number; represents the environmental characteristic classification function; represents the historical sample; represents the th Topographic feature vectors corresponding to historical samples; Denote the sub-region The th micro-meteorological feature vector corresponding to historical samples; Denote the sub-region The th parameter error vector of the sample; Denote the sub-region The th original model parameter vector of the sample; Denote the sub-region The th optimal parameter obtained by offline fitting of the sample; Denote the sub-region Average correction direction; Denote the sub-region Update step size; Denote the step size factor; Denote the natural base; Denote the decay rate coefficient; Denote the sub-region Updated parameter vector; Denote the sub-region Parameter vector before update; Denote the sub-region Wind power output prediction model after parameter update; Denote the sub-region Wind power output prediction model before parameter update.

[0055] The relative change rate of prediction error in the fourth condition includes: ; Wherein, Denote the relative change rate of the kth prediction error; Denote the prediction error after the (k + 1)th iteration; Denote the prediction error after the kth iteration.

[0056] In the preferred embodiment of the present invention, iteratively optimizing the wind power output prediction model with a prediction error greater than the preset error threshold according to historical first data further includes: When the iterative optimization satisfies the third condition or the fourth condition but the prediction error is still greater than the preset error threshold, the correction stops and a warning level signal is output , including: ; Wherein, Denote the current number of wind power output predictions; Denote the current number of overall region divisions; represents the maximum allowable number of wind power output predictions; represents the maximum allowable number of regional division times; and represents the sub-region in the error thresholds for each level of early warning of historical measured wind power in the nth time; and [[ID=]14] represents the change rate threshold.

[0057] If = 1, output a mild early warning, indicating that the current error and change rate slightly exceed the threshold, and the early warning prompt needs attention.

[0058] If 3] = 2, output a moderate early warning, indicating that the prediction error fluctuates greatly and needs to be adjusted.

[0059] If = 3, output a severe early warning, indicating that the prediction error is abnormal and triggers manual intervention. And record the data in the prediction, such as errors, etc.

[0060] In the preferred embodiment of the present invention, the wind power output prediction models of each sub-region are weighted and fused to obtain the total wind power output prediction model, including: ; Among them, represents the total wind power output prediction model; represents the sub-region weight.

[0061] In the preferred embodiment of the present invention, there are differences in resource conditions such as wind speed and light in different sub-regions. If simply added equally, it will underestimate the contribution of high-quality partitions or overestimate the contribution of low-quality partitions, thus resulting in deviations in the total amount prediction; at the same time, the contribution degrees of different-sized sub-regions or sub-regions with different installed capacities are different. Therefore, the weights of each region are set, and the wind power output prediction models of each sub-region are weighted and fused.

[0062] S5. Implement the wind power output prediction of the target area according to the total wind power output prediction model.

[0063] The wind power output prediction method considering the collaborative influence of micro-meteorology and micro-topography according to the present invention divides the target area into a preset number of sub-areas based on the first condition and the second condition according to the real-time first data and the historical wind power output curve. When partitioning, it considers the real-time micro-meteorological parameters, micro-topographical parameters and the historical wind power output curve of the target area. That is, the partitioning method of the present invention considers more factors that affect the partitioning, making the partitioning have a higher accuracy. At the same time, when partitioning, it is ensured that the wind turbines in the same sub-area have the first data within a preset similarity range, so that the partitioning can divide similar areas into the same sub-area, further improving the accuracy of the partitioning. Further, when partitioning, it is also ensured that the maximum difference between the characteristic center vectors of different sub-areas on each coordinate component is less than or equal to a preset threshold, so that the method of the present invention further considers the accuracy of the partitioning. The method of the present invention divides the target area considering multiple factors and two conditions, making the partitioning of the method of the present invention have a higher accuracy, providing a basis for subsequent wind power output prediction. When constructing the wind power output prediction model for the sub-area, the method of the present invention considers the influence of turbulence and wake, making the wind power output prediction model of the present invention consider more factors that affect wind power generation, thereby improving the prediction accuracy of the wind power output prediction model. In the method of the present invention, the wind power output prediction model is optimized according to the historical first data, so that the wind power output prediction model of the sub-area has better prediction performance and avoids prediction errors. In the method of the present invention, the wind power output prediction models of each sub-area are weighted and fused to obtain the total wind power output prediction model, so that the total wind power output prediction model considers the weights of each sub-area, can improve the attention to key sub-areas, and further perform more accurate wind power output prediction on the whole target area. The method of the present invention can realize dynamic adjustment of partitioning, iterative correction of sub-area prediction models and continuous evaluation, thereby improving the accuracy of wind power output prediction in the target area, reducing the warning response delay, and providing reliable support for the safe and stable operation of the power grid.

[0064] In a preferred embodiment of the present invention, there is also provided a wind power output prediction device considering the collaborative influence of micro-meteorology and micro-topography for the method of the present invention. The device includes a first module, a second module, a third module and a fourth module.

[0065] The first module is used to obtain the real-time first data, historical first data and historical wind power output curve of the target area; the first data includes preset micro-meteorological parameters and micro-topographical parameters.

[0066] The second module is used to divide the target area into a preset number of sub-areas based on the first condition and the second condition according to the real-time first data and the historical wind power output curve.

[0067] The first condition includes that the wind turbines in the same sub-region have first data within a preset similarity range; the second condition includes that the maximum difference between the characteristic center vectors of different sub-regions in each coordinate component is less than or equal to a preset threshold.

[0068] The third module is used to obtain the wind power output prediction model of each sub-region according to the micro-meteorological parameters, wind turbine parameters and effective wind speed model of the real-time first data in the sub-region; the effective wind speed model takes into account the influence of turbulence and wake.

[0069] The fourth module is used to optimize the wind power output prediction model according to the historical first data, and then perform weighted fusion on the wind power output prediction models of each sub-region to obtain the total wind power output prediction model; the wind power output of the target region is predicted according to the total wind power output prediction model.

[0070] The wind power output prediction device considering the collaborative influence of micro-meteorology and micro-topography of the present invention is used for the method of the present invention and has the same beneficial effects as the method of the present invention.

[0071] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for predicting wind power output considering the collaborative influence of micrometeorology and microtopography, characterized in that, Including: Obtain the real-time first data, historical first data, and historical wind power output curve of the target area; the first data includes preset micro-meteorological parameters and micro-topographical parameters; divide the target area into a preset number of sub-areas based on the first condition and the second condition according to the real-time first data and the historical wind power output curve; The first condition includes that the wind turbines in the same sub-area have first data within a preset similarity range; the second condition includes that the maximum difference between the characteristic center vectors of different sub-areas on each coordinate component is less than or equal to a preset threshold; Obtain the wind power output prediction model of each sub-area according to the micro-meteorological parameters, wind turbine parameters, and effective wind speed model of the real-time first data in the sub-area; the effective wind speed model takes into account the effects of turbulence and wake; Optimize the wind power output prediction model according to the historical first data, and then perform weighted fusion on the wind power output prediction models of each sub-area to obtain the total wind power output prediction model; Realize the wind power output prediction of the target area according to the total wind power output prediction model.

2. The wind power output prediction method considering the combined influence of micro-meteorology and micro-topography according to claim 1, characterized in that Dividing the target area into a preset number of sub-areas based on the first condition and the second condition according to the real-time first data and the historical wind power output curve includes: Dividing the target area into a preset number of sub-areas based on the first condition according to the real-time first data and the historical wind power output curve; obtain the first index; the first index includes the maximum difference between the characteristic center vectors of different sub-areas on each coordinate component; if the first index is greater than the preset threshold, adjust the partition based on the first condition and obtain the first index again until the first index meets the second condition to complete the division of the target area, including: ； Among them, represents the set of sub-regions that are geographically adjacent to sub-region k; and both represent sub-regions; represents sub-region and the maximum difference in each coordinate component of the characteristic centers; represents the preset threshold; represents the output power of the h-th unit under the historical input wind speed ; and represent the polynomial coefficients of the h-th unit; represents the historical input wind speed; represents the multi-dimensional feature vector extracted from the -th sampling point. The multi-dimensional feature vector includes micro-meteorological parameters, micro-topographical parameters, and a wind power prediction value considering only the wind speed; represents the set of all sub-region indicator variables; represents the characteristic center vector of sub-region ; represents the total number of sub-regions; represents the total number of sampling points of all sub-regions; represents a sampling point; represents whether the -th sampling point is assigned to sub-region . If represents that the -th sampling point is assigned to the -th sub-region, and if represents that the -th sampling point is not assigned to the -th sub-region; represents the characteristic center vector of sub-region ; 3. The wind power output prediction method considering the collaborative influence of micro-meteorology and micro-topography according to claim 2, characterized in that The effective wind speed model includes: Obtain the wake influence model of the sub-area according to the real-time first data of the sub-area combined with the influence model of the nearby wake on the target wind turbine; obtain the turbulence influence model of the sub-area according to the micro-topographical data in the real-time first data of the sub-area combined with the influence model of the nearby turbulence on the target wind turbine; obtain the effective wind speed model of the sub-area according to the wake influence model and the turbulence influence model.

4. The wind power output prediction method considering the collaborative influence of micro-meteorology and micro-topography according to claim 3, wherein, Obtaining the effective wind speed model of the sub-area according to the wake influence model and the turbulence influence model includes: The effective wind speed model includes: ； ； ； ； Among them, represents the effective wind speed; and respectively represent the weight coefficients of wake and turbulence in the sub-region ; represents the local turbulence intensity of the coordinate , that is, the turbulence influence model; represents the wake wind speed at the coordinate , that is, the wake influence model; represents the local slope angle of the fan position; represents taking the partial derivative; represents the unperturbed turbulence intensity; represents the turbulence correction model; represents the distance extending along the wind direction from the fan position; represents the offset relative to the wake centerline; represents the height deviation relative to the turbine hub in the vertical direction; represents the influence model of the nearby turbulence on the target fan; represents the initial turbulence intensity amplification factor; represents the turbulence decay scale; represents the micro-topography influence coefficient; represents the micro-topography correction factor of the deviation between the local terrain height and the average terrain height and the slope effect; represents the transverse diffusion coefficient; represents the reference height; represents the diffusion shape index; represents the overall weight coefficient of the surrounding fan interference; represents the th surrounding fan that disturbs the target fan within; represents the total number of surrounding fans that disturb the turbulence of the target fan; represents the target fan and the th fan represents the relative azimuth attenuation factor; represents the angular attenuation index; represents the th fan and the included angle between the line connecting the target fan and the th fan and the main wind direction; represents the three-dimensional coordinates of the th fan; represents the wind speed without interference; represents the density correction factor; represents the wake attenuation model; Represents the influence model of the nearby wake on the target wind turbine; Represents the humidity correction factor; Represents the initial wake decay coefficient; Represents the wake recovery length under flat terrain; Represents the micro-topography influence coefficient; Represents the micro-topography correction factor; Represents the wake diffusion parameter; Represents the diameter of the wind turbine impeller; Represents The th surrounding wind turbine that disturbs the target wind turbine; Represents the total number of surrounding wind turbines that disturb the wake of the target wind turbine; Represents the three-dimensional coordinates of the th wind turbine; Represents the target wind turbine and the th wind turbine The center distance between them; Represents the interference intensity factor of the hth surrounding wind turbine; Represents the characteristic decay length of the interference effect of the hth wind turbine; Represents the interference attenuation exponent; Represents the sensitivity of density change to wake decay; Represents the air density; Represents the reference density; Represents the influence factor of humidity on wake recovery; Represents the relative humidity; Represents the reference humidity.

5. The wind power output prediction method considering the combined influence of micrometeorology and microtopography according to claim 4, wherein Obtaining the wind power output prediction model of each sub-area according to the micro-meteorological parameters, wind turbine parameters, and effective wind speed model of the real-time first data in the sub-area includes: The wind power output prediction model includes: ； Among them, represents the predicted value of wind power output in the sub-region ; represents the air density within the sub-region , which is a micrometeorological parameter from real-time first data; represents the area swept by the wind turbine within the sub-region ; represents the power coefficient of the wind turbine within the sub-region .

6. The wind power output prediction method considering the combined influence of micro-meteorology and micro-topography according to claim 5, characterized in that Optimizing the wind power output prediction model according to the historical first data, and then performing weighted fusion on the wind power output prediction models of each sub-area to obtain the total wind power output prediction model includes: Obtain the prediction result of each sub-area according to the wind power output prediction model of each sub-area; obtain the prediction error of the wind power output prediction model of each sub-area according to the prediction result of each sub-area and the historical wind power output curve; Iteratively optimize the wind power output prediction model with a prediction error greater than the preset error threshold according to the historical first data, and obtain the prediction error of the wind power output prediction model after each optimization; when the iterative optimization meets the third condition or the fourth condition and the prediction error is less than or equal to the preset error threshold, complete the optimization of the wind power output prediction model. The third condition includes that the number of iterations reaches the preset upper limit; the fourth condition includes that the relative change rate of the prediction error after two consecutive optimizations of the wind power output prediction model is less than the preset change rate threshold. When the prediction errors of the wind power output prediction models for each sub-region are all less than or equal to the preset error threshold, weight and fuse the wind power output prediction models for each sub-region to obtain the total wind power output prediction model.

7. The wind power output prediction method considering the combined influence of micro-meteorology and micro-topography according to claim 6, characterized in that, Obtaining the prediction error of the wind power output prediction model for each sub-region according to the prediction result of each sub-region and the historical wind power output curve includes: Compare the prediction result of the sub-region with the historical wind power output curve to obtain the prediction error: ； Among them, represents the prediction error of the k-th prediction in sub-region o; represents the prediction result of the k-th prediction in sub-region o; represents the historical actual wind power output value of the k-th prediction in sub-region o.

8. The wind power output prediction method considering the combined influence of micrometeorology and microtopography according to claim 7, wherein Iteratively optimizing the wind power output prediction model with a prediction error greater than the preset error threshold according to the historical first data includes: Calculate the deviation of the micro-meteorological parameters and the micro-topographical parameters between the historical first data and the real-time first data; normalize the deviation; take the average within the first data group for the normalized deviation to obtain the correction direction; obtain the correction step size through the exponential decay formula according to the number of samples within the first data group; use the correction step size to update the parameters of the wind power output prediction model with a prediction error greater than the preset error threshold along the correction direction, including: ； Among them, represents the historical data set of the th category of environmental features in the sub-region; represents the environmental feature category number; represents the environmental feature classification function; represents the historical sample; represents the th topographic feature vector corresponding to the historical sample in the sub-region; represents the th micrometeorological feature vector corresponding to the historical sample in the sub-region; represents the th parameter error vector of the sample in the sub-region; represents the th original model parameter vector of the sample in the sub-region; represents the th optimal parameter obtained by offline fitting of the sample in the sub-region; represents the average correction direction of the sub-region; represents the update step size of the sub-region; represents the step size factor; represents the natural base; represents the decay rate coefficient; represents the updated parameter vector of the sub-region; represents the parameter vector before update of the sub-region; represents the wind power output prediction model after updating the parameters of the sub-region; represents the wind power output prediction model before updating the parameters of the sub-region; The relative change rate of the prediction error in the fourth condition includes: ； Among them, represents the relative change rate of the k-th prediction error; represents the prediction error after the (k + 1)-th iteration; represents the prediction error after the k-th iteration.

9. The wind power output prediction method considering the collaborative influence of micro-meteorology and micro-topography according to claim 8, characterized in that, The process of weighting and fusing the wind power output prediction models for each sub-region to obtain the total wind power output prediction model includes: ； Among them, represents the total model for wind power output prediction; represents the sub-region weight.

10. A wind power output prediction device considering the collaborative influence of micro-meteorology and micro-topography, for use in the method according to any one of claims 1 to 9, characterized in that, The device includes a first module, a second module, a third module, and a fourth module. The first module is used to obtain the real-time first data, historical first data, and historical wind power output curve of the target region; the first data includes preset micro-meteorological parameters and micro-topographical parameters. The second module is used to divide the target region into a preset number of sub-regions based on the real-time first data and the historical wind power output curve according to the first condition and the second condition. The first condition includes that the wind turbines within the same sub-region have first data within a preset similarity range; the second condition includes that the maximum difference of the characteristic center vectors between sub-regions on each coordinate component is less than or equal to the preset threshold. The third module is used to obtain the wind power output prediction model for each sub-region according to the micro-meteorological parameters, wind turbine parameters, and effective wind speed model of the real-time first data within the sub-region; the effective wind speed model takes into account the influence of turbulence and wake. The fourth module is used to optimize the wind power output prediction model according to the historical first data, then weight and fuse the wind power output prediction models for each sub-region to obtain the total wind power output prediction model; realize the wind power output prediction of the target region according to the total wind power output prediction model.

Images