A weather forecast-based wind power output regional composite calculation method and system

Abstract

The application discloses a wind power output regional composite calculation method and system based on weather forecast, wherein the method comprises: wind field data processing and regional composite model initialization, including: processing national near-surface wind field high-resolution gridded wind field data to obtain wind field data at a specified height; generating a national prevailing wind direction network based on the wind field data at the specified height; setting an internal wind turbine arrangement structure of a wind farm; calculating the output power of a single wind farm considering the wake effect; configuring the distribution sites of wind farms in a region; establishing a regional composite model, including: calculating the regional average wind power output based on the boundary conditions; calculating the confidence interval of the wind power output based on the historical probability weather forecast and the regional average wind power output; and calculating the wind power output and its confidence interval based on the future weather forecast through the regional composite model.

Description

Technical Field

[0001] This invention relates to the field of wind power prediction and output characteristic analysis technology, and more specifically, to a method and system for regional composite calculation of wind power output based on weather forecast. Background Technology

[0002] China boasts abundant wind energy resources, ranking second in both installed wind power capacity and electricity generation among clean energy sources (after hydropower). At the provincial level, 13 provincial-level administrative regions lead in either installed wind power capacity or electricity generation among renewable energy sources. However, wind power generation is subject to weather conditions, resulting in unstable output. Different weather conditions have varying impacts on wind power output, with regional extreme weather events often causing widespread fluctuations. As my country's wind power share further increases in the future, the range and uncertainty of wind power output fluctuations will also increase. Therefore, predicting inter-provincial wind power output changes under different weather conditions is crucial.

[0003] The theoretical basis of wind power generation is to convert the kinetic energy of air movement into electrical energy. The wind energy per unit volume is directly proportional to the air density and the cube of the wind speed. However, air density and wind speed are also affected by a variety of weather factors. The actual wind energy converted into electricity is also related to the wind turbine parameters, wind farm structure, etc.

[0004] Currently, there are two main methods for studying the impact of weather on wind power output: First, by comparing and analyzing the changes in wind power output under different weather types using historical meteorological data and measured output data. For example, Existing Technology 1 (Chinese Patent Announcement No. CN115994624A) discloses a wind power prediction system and method based on meteorological data. Existing Technology 1 determines the similarity between real-time meteorological data and historical meteorological data, and selects the wind power corresponding to historically similar weather types to predict wind power output in the future. Second, by using short-term weather forecasts to predict changes in meteorological conditions, a response model of wind power to meteorological elements is established. Modeling methods include neural network optimization based on genetic algorithms (Chinese Patent Announcement No. CN105447572A) and two-stage dynamic optimization combining Gaussian regression algorithm and differential delay operator (Chinese Patent Announcement No. CN109190845A).

[0005] However, current methods for studying the impact of weather on wind power output primarily focus on specific wind farms or small regions, emphasizing optimized layout within a single wind farm. These methods fail to address the comprehensive assessment of weather-related wind power impacts across provincial borders in China and the issue of regional differences. Applying independent wind power response models to meteorological elements to larger areas requires considering not only the actual location of wind farms but also the prevailing wind directions (i.e., the most frequent wind directions in each region) when determining the orientation of wind farm layouts. This is a missing element in existing technologies. Furthermore, current numerical weather prediction employs multi-model ensemble forecasting, resulting in probabilistic weather forecasts where the forecast is not a single deterministic quantity but rather a probability value within a certain range. However, the relationship between the probability range of the weather forecast and the confidence interval for wind power output remains unclear. Therefore, a technique is needed to construct a regional composite model of wind farm changes affecting wind power output, considering the uncertainty range of probabilistic weather forecasts, to predict the provincial average wind power output and confidence level. Summary of the Invention

[0006] The technical solution of this invention provides a method and system for calculating wind power output in a region based on weather forecasts. This method addresses how to consider the prevailing wind direction and probabilistic weather forecasts in the regional composite model of wind power output. It takes into account the orientation of wind farms in the region according to the characteristics of prevailing wind direction in different areas, and predicts the regional wind power output. It can also be used to study regional differences in the relationship between complex weather and wind power output.

[0007] To address the aforementioned problems, this invention provides a method for comprehensive calculation of wind power output regions based on weather forecasts, the method comprising:

[0008] Wind field data processing and regional composite model initialization include: processing high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at specified altitudes; generating a nationwide prevailing wind direction network based on the wind field data at specified altitudes; setting the internal wind turbine layout structure of wind farms; calculating the output power of a single wind farm considering wake effects; configuring the distribution locations of wind farms within the region; and establishing a regional composite model, including: calculating the configured regional average wind power output based on boundary conditions; and calculating the confidence interval of wind power output based on historical probabilistic weather forecasts and the regional average wind power output.

[0009] Based on future weather forecasts, wind power output and its confidence interval are calculated using the regional composite model. Preferably, the processing of high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified altitude includes:

[0010] The high-resolution gridded wind field data of near-surface wind fields across the country were processed according to the logarithmic law:

[0011]

[0012] Where Z is the specified height, u(Z) is the wind speed at the specified height Z, u* is the friction velocity constant, K is the Karman constant, and Z0 is the ground roughness length.

[0013] Preferably, the process of processing the high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified altitude includes:

[0014] High-resolution gridded wind field data of near-surface wind fields across the country were processed according to the exponential law:

[0015]

[0016] Where Z is the specified height at which the desired wind speed is obtained. A Given the height of the wind speed, u(Z) and u(Z) A () are heights Z and Z respectively. A The wind speed at point n is the Hermann exponent.

[0017] Preferably, the process of processing the high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified altitude includes:

[0018] For simulated wind field data, corrections are made to the actual wind field data using a variance matching method:

[0019] Calculate the average value of the measured data O within the selected region and calibration period. ave And variance O var And the average value M of the simulated data M ave and variance M var Finally, the correction data M is calculated according to the following formula. adj :

[0020]

[0021] Based on the correction data M adj Correcting the simulated wind field data.

[0022] Preferably, generating a nationwide prevailing wind direction network based on wind field data at a specified altitude includes:

[0023] Missing values ​​in the wind field data are supplemented using interpolation:

[0024] Nearest neighbor interpolation is used to assign the value of the input grid point closest to the output grid point to the output grid point, thus filling in the missing values ​​of the output grid point; or

[0025] By using bilinear interpolation, the weighted average of the grid points in the nearest 2×2 neighbors to the output grid point is used to supplement the missing values ​​of the output grid point; or

[0026] Bicubic interpolation is used to supplement the missing values ​​of the output grid point by averaging the weighted average of the grid points in the nearest 4×4 neighboring points.

[0027] Preferably, generating a nationwide prevailing wind direction network based on wind field data at a specified altitude includes:

[0028] Calculate the wind direction frequency distribution map for each grid point; based on the wind direction frequency distribution map, using a polar coordinate system, divide the circumference into equal parts according to a preset number of parts, and the prevailing wind direction is the direction in which the wind direction appears most frequently during the statistical data period. Traverse all grid points to generate a national prevailing wind direction network.

[0029] Preferably, it further includes: calculating the regional average wind power output after configuration based on the double-weighted averaging method.

[0030]

[0031]

[0032] Where P i w is the normalized output value of the i-th wind farm in the region. i Here, c is the corresponding weighting coefficient, S is the absolute median deviation, and S is a robust measure of sample bias.

[0033] S = median{|P i -P * |}

[0034] The median wind power output of the region is used to calculate P* as the initial value to start iterative calculation until the preset convergence condition is reached.

[0035] Based on another aspect of the present invention, the present invention provides a wind power output regional composite calculation system based on weather forecast, the system comprising:

[0036] The initial unit is used to: process high-resolution gridded wind field data of near-surface wind fields across the country to obtain wind field data at a specified altitude; generate a nationwide prevailing wind direction network based on the wind field data at the specified altitude; set the internal wind turbine layout structure of wind farms; calculate the output power of a single wind farm considering the wake effect; and configure the distribution locations of wind farms within the region.

[0037] The regional composite model building unit is used to calculate the configured regional average wind power output based on boundary conditions; and to calculate the confidence interval of wind power output based on historical probabilistic weather forecasts and regional average wind power output.

[0038] The output unit is used to calculate the wind power output and its confidence interval based on future weather forecasts using the regional composite model.

[0039] Preferably, the boundary setting unit is used to process high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified height, and is also used for:

[0040] The high-resolution gridded wind field data of near-surface wind fields across the country were processed according to the logarithmic law:

[0041]

[0042] Where Z is the specified height, u(Z) is the wind speed at the specified height Z, and u * Let be the friction velocity constant, K be the Karman constant, and Z0 be the surface roughness length.

[0043] Preferably, the boundary setting unit is used to process high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified height, and is also used for:

[0044] High-resolution gridded wind field data of near-surface wind fields across the country were processed according to the exponential law:

[0045]

[0046] Where Z is the specified height at which the desired wind speed is obtained. A Given the height of the wind speed, u(Z) and u(Z) A () are heights Z and Z respectively. A The wind speed at point n is the Hermann exponent.

[0047] Preferably, the boundary setting unit processes high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified height, and is also used for:

[0048] For simulated wind field data, corrections are made to the actual wind field data using a variance matching method:

[0049] Calculate the average value of the measured data O within the selected region and calibration period. ave And variance O var And the average value M of the simulated data M ave and variance M var Finally, the correction data M is calculated according to the following formula. adj :

[0050]

[0051] Based on the correction data M adj Correcting the simulated wind field data.

[0052] Preferably, the boundary setting unit is used to generate a nationwide prevailing wind direction network based on wind field data at a specified altitude, and is also used to:

[0053] Missing values ​​in the wind field data are supplemented using interpolation:

[0054] Nearest neighbor interpolation is used to assign the value of the input grid point closest to the output grid point to the output grid point, thus filling in the missing values ​​of the output grid point; or

[0055] By using bilinear interpolation, the weighted average of the grid points in the nearest 2×2 neighbors to the output grid point is used to supplement the missing values ​​of the output grid point; or

[0056] Bicubic interpolation is used to supplement the missing values ​​of the output grid point by averaging the weighted average of the grid points in the nearest 4×4 neighboring points.

[0057] Preferably, the boundary setting unit is used to generate a nationwide prevailing wind direction network based on wind field data at a specified altitude, and is also used to:

[0058] Calculate the wind direction frequency distribution map for each grid point; based on the wind direction frequency distribution map, using a polar coordinate system, divide the circumference into equal parts according to a preset number of parts, and the prevailing wind direction is the direction in which the wind direction appears most frequently during the statistical data period. Traverse all grid points to generate a national prevailing wind direction network.

[0059] Preferably, the regional composite model establishment unit is further used to calculate the regional average wind power output after configuration based on the double-weighted averaging method:

[0060]

[0061]

[0062] Where P i w is the normalized output value of the i-th wind farm in the region. i Here, c is the corresponding weighting coefficient, S is the absolute median deviation, and S is a robust measure of sample bias.

[0063] S = median{|P i -P * |}

[0064] The median wind power output of the region is used to calculate P* as the initial value to start iterative calculation until the preset convergence condition is reached.

[0065] According to another aspect of the present invention, the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program for executing a composite calculation method for wind power output area based on weather forecast.

[0066] According to another aspect of the present invention, the present invention provides an electronic device, characterized in that the electronic device comprises: a processor and a memory; wherein,

[0067] The memory is used to store the processor-executable instructions;

[0068] The processor is configured to read the executable instructions from the memory and execute the instructions to implement a composite calculation method for wind power output area based on weather forecast.

[0069] This invention provides a method and system for regional composite calculation of wind power output based on weather forecasts. The method includes: wind field data processing and regional composite model initialization, including: processing high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified altitude; generating a nationwide prevailing wind direction network based on the wind field data at the specified altitude; setting the internal wind turbine layout structure of wind farms; calculating the output power of a single wind farm considering wake effects; configuring the distribution locations of wind farms within the region; and establishing a regional composite model, including: calculating the configured regional average wind power output based on boundary conditions; calculating the confidence interval of wind power output based on historical probabilistic weather forecasts and the regional average wind power output; and calculating the wind power output and its confidence interval based on future weather forecasts using the regional composite model. This invention considers the range of probabilistic weather forecasts and proposes a regional composite model that addresses the impact of wind field changes on wind power output. This model can be used to estimate the provincial average wind power output and confidence level, and can also be used to study regional differences in the relationship between complex weather and wind power output. Attached Figure Description

[0070] Exemplary embodiments of the present invention can be more fully understood by referring to the following figures:

[0071] Figure 1 This is a flowchart of a wind power output area composite calculation method based on weather forecast according to a preferred embodiment of the present invention;

[0072] Figure 2 This is a flowchart of a wind power output area composite calculation method based on weather forecast according to a preferred embodiment of the present invention;

[0073] Figure 3 This is a schematic diagram showing the predicted output curve of the Chaganhot Wind Farm on April 20th with a 3-hour resolution, one week in advance, according to a preferred embodiment of the present invention, and its comparison with the actual output curve.

[0074] Figure 4 This is a diagram showing the predicted average power output of wind farms in Jilin Province on April 20th, one week in advance, according to a preferred embodiment of the present invention, and its comparison with the actual power output curve; and

[0075] Figure 5 This is a structural diagram of a wind power output regional composite calculation system based on weather forecast according to a preferred embodiment of the present invention. Detailed Implementation

[0076] Exemplary embodiments of the invention will now be described with reference to the accompanying drawings. However, the invention may be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to fully and completely disclose the invention and to fully convey its scope to those skilled in the art. The terminology used in the exemplary embodiments illustrated in the drawings is not intended to limit the invention. In the drawings, the same units / elements are referred to by the same reference numerals.

[0077] Unless otherwise stated, the terms used herein (including technical terms) have their common meaning as understood by one of ordinary skill in the art. Furthermore, it is understood that terms defined in commonly used dictionaries should be understood to have a meaning consistent with the context of their relevant field, and not to be interpreted as having an idealized or overly formal meaning.

[0078] Figure 1 This is a flowchart of a wind power output regional composite calculation method based on weather forecast according to a preferred embodiment of the present invention. Figure 1 As shown, this invention provides a method for comprehensive calculation of wind power output regions based on weather forecasts, the method comprising:

[0079] Step 101: Wind field data processing and regional composite model initialization, including: processing high-resolution gridded wind field data of near-surface wind fields across the country to obtain wind field data at a specified height; generating a nationwide prevailing wind direction network based on the wind field data at the specified height; setting the internal wind turbine layout structure of wind farms; calculating the output power of a single wind farm considering the wake effect; and configuring the distribution locations of wind farms within the region.

[0080] Preferably, high-resolution gridded wind field data of near-surface wind fields nationwide are processed to obtain wind field data at a specified altitude, including:

[0081] The high-resolution gridded wind field data of near-surface wind fields across the country were processed according to the logarithmic law:

[0082]

[0083] Where Z is the specified height, u(Z) is the wind speed at the specified height Z, and u * Let be the friction velocity constant, K be the Karman constant, and Z0 be the surface roughness length.

[0084] Preferably, high-resolution gridded wind field data of near-surface wind fields nationwide are processed to obtain wind field data at a specified altitude, including:

[0085] High-resolution gridded wind field data of near-surface wind fields across the country were processed according to the exponential law:

[0086]

[0087] Where Z is the specified height at which the desired wind speed is obtained. A Given the height of the wind speed, u(Z) and u(Z) A () are heights Z and Z respectively. A The wind speed at point n is the Hermann exponent.

[0088] Preferably, high-resolution gridded wind field data of near-surface wind fields nationwide are processed to obtain wind field data at a specified altitude, including:

[0089] For simulated wind field data, corrections are made to the actual wind field data using a variance matching method:

[0090] Calculate the average value of the measured data O within the selected region and calibration period. ave And variance O var And the average value M of the simulated data M ave and variance M var Finally, the correction data M is calculated according to the following formula. adj :

[0091]

[0092] Based on correction data M adj Correcting the simulated wind field data.

[0093] Preferably, based on wind field data at a specified altitude, a nationwide prevailing wind direction network is generated, including:

[0094] Missing values ​​in the wind field data are supplemented using interpolation:

[0095] Nearest neighbor interpolation is used to assign the value of the input grid point closest to the output grid point to the output grid point, thus filling in the missing values ​​of the output grid point; or

[0096] By using bilinear interpolation, the weighted average of the grid points in the nearest 2×2 neighbors to the output grid point is used to supplement the missing values ​​of the output grid point; or

[0097] Bicubic interpolation is used to supplement the missing values ​​of the output grid point by averaging the weighted average of the grid points in the nearest 4×4 neighboring points.

[0098] Preferably, based on wind field data at a specified altitude, a nationwide prevailing wind direction network is generated, including:

[0099] Calculate the wind direction frequency distribution map for each grid point; based on the wind direction frequency distribution map, using a polar coordinate system, divide the circumference into equal parts according to a preset number of parts, and the prevailing wind direction is the direction in which the wind direction appears most frequently during the statistical data period. Traverse all grid points to generate a national prevailing wind direction network.

[0100] This invention first processes high-resolution gridded near-surface wind field data nationwide. The data is in the form of two-dimensional wind vectors, i.e., zonal wind (westerly) and meridional wind (southern), or wind speed scalars and wind direction data, with a temporal resolution of no less than 6 hours and a coverage period of no less than 5 years. The installation height of wind turbine hubs is generally 80-120m, and some meteorological data only provide wind fields at a height of 10m, which can be converted to a specified height using the logarithmic law or exponential law. The logarithmic law is derived from Prandtl's mixing length theory, which describes the near-surface surface-level wind speed profile for neutral stratification:

[0101]

[0102] Where Z is the height (in m), u(Z) is the wind speed at height Z (in m / s), u* is the friction velocity constant (in m / s), K is the Karman constant (dimensionless, usually taken as 0.4), Z0 is the ground roughness length (in m), and the typical value and range of Z0 are 0.03 (0.01 to 0.1) in grassland, 0.1 (0.1 to 0.3) in cultivated land, 0.3 (0.2 to 1) in forest, 1 (1 to 4) in densely populated residential areas, and 4 or above in densely populated residential areas.

[0103] The exponential law is a wind speed profile pattern derived from Schmidt's semi-empirical theory:

[0104]

[0105] Where Z and Z A Let u(Z) and u(Z) be the heights (in meters) where the wind speed is to be found and the wind speed is known, respectively. A Let be the wind speed (in m / s) at heights Z and ZA, and n be the Hermann exponent, which is affected by underlying surface conditions and atmospheric stability, generally ranging from 0.1 (extremely unstable) to 0.3 (stable). If simulated wind field data is used, it is necessary to correct it to the actual observed wind field using a variance matching method to ensure that the mean and variance of the simulated data are consistent with the observed data within the calibration period. The specific algorithm is to calculate the average value O of the measured data O within the selected area and calibration period (10 years or more). ave And variance O var And the average value M of the simulated data M ave and variance M var, and finally calculate the correction data M according to the following formula. adj .

[0106]

[0107] This invention generates a nationwide prevailing wind direction network. First, the wind field data is spatially resampled to the required spatial resolution. An optimized algorithm of first filling gaps and then interpolating is adopted to avoid the offset or amplification of blank or missing grid points at the coastline or regional boundary during the interpolation operation. The model provides three interpolation methods: (1) nearest neighbor interpolation, that is, the value assigned to the output grid point is the value of the nearest input grid point, without considering other pixels; (2) bilinear interpolation, that is, the output grid point value is the weighted average of the grid points in the nearest 2×2 neighboring points; (3) bicubic interpolation, that is, the output grid point value is the weighted average of the grid points in the nearest 4×4 neighboring points. Based on this, the wind direction frequency distribution map of each grid point is calculated. Using a polar coordinate system, the circumference is divided into 36 equal parts with 10° intervals. The direction with the highest frequency of wind direction in the statistical data period is taken as the prevailing wind direction of that point. By traversing all grid points, a nationwide prevailing wind direction network is generated.

[0108] This invention designs the internal wind turbine layout structure of a wind farm. It supports importing actual wind farm turbine layout and parameter data. If no corresponding data is input, default data is used based on general medium-sized wind farm planning principles. The default turbine model is Vitas V90-1.8MW, with a rotor diameter of 90.0 meters, a swept rotor area of ​​6362.0 square meters, a cut-in wind speed of 3.5 m / s, a rated wind speed of 12.0 m / s, a cut-out wind speed of 25.0 m / s, and a thrust coefficient of 0.88. In a single wind farm, the turbines are arranged in a 5×5 differential configuration. The longitudinal spacing between turbines is 7 times the rotor diameter, while the transverse spacing is 4.5 times the rotor diameter. The longitudinal direction of the wind farm turbines is consistent with the local prevailing wind direction to ensure maximum annual output.

[0109] This invention considers the wake effect when calculating the output power of a single wind farm. Wind experiences energy attenuation after passing through the rotor, leading to a decrease in downstream wind speed, known as the wake effect. This effect causes uneven wind speed distribution within a multi-row wind farm. This invention uses the Jensen wake model to estimate this effect, which assumes that the wind decelerates after passing through the rotor, and the diameter of the deceleration region initially matches the rotor diameter, subsequently increasing linearly with distance from the rotor. Combining this with the law of conservation of momentum, we can obtain:

[0110]

[0111] D(x) = D0 + 2kx

[0112] Where V0 is the wind speed before it decays at the wind turbine, V(x) is the wind speed after passing the wind turbine a distance x, and C TThe thrust coefficient is given by D0, the rotor diameter is given by D(x), and the diameter of the deceleration region after passing through the rotor by a distance x is given by D(x). The diameter is calculated using the wake region expansion coefficient k, which is determined by the rotor hub height Z and the ground roughness length Z0.

[0113]

[0114] The fans are arranged in a differential configuration, so only a portion of the swept area of ​​the downstream fan is affected by the upstream fan. Let d represent the distance between the downstream and upstream fans in the direction perpendicular to the wind direction, then the affected swept area A can be expressed as:

[0115]

[0116] This invention configures the distribution locations of wind farms within a specified area. Considering the number of wind farms and their corresponding geographical locations (latitude and longitude coordinates) within the area, the model calculates the optimal orientation of the wind farms based on the prevailing wind direction at the grid points where the wind farms are located.

[0117] Step 102: Establish a regional composite model, including: calculating the regional average wind power output after configuration based on boundary conditions; calculating the confidence interval of wind power output based on historical probabilistic weather forecasts and regional average wind power output;

[0118] Preferably, it further includes: calculating the regional average wind power output after configuration based on the double-weighted averaging method.

[0119]

[0120]

[0121] Where P i w is the normalized output value of the i-th wind farm in the region. i Here, c is the corresponding weighting coefficient, S is the absolute median deviation, and S is a robust measure of sample bias.

[0122] S = median{|P i -P * |}

[0123] The median wind power output of the region is used to calculate P* as the initial value to start iterative calculation until the preset convergence condition is reached.

[0124] This invention calculates the regional average wind power output using a double-weighted averaging method to reduce extreme errors.

[0125] This invention calculates the confidence interval for wind power output based on probabilistic weather forecasts, uses kernel distribution to fit the deviation between historical predicted output and actual output, and calculates the confidence interval for wind power output predicted 1 day, 3 days, or 7 days in advance at a given level (50%, 70%, or 90%).

[0126] This invention calculates the regional average wind power output. Since the power output of each wind farm within a region at a given time does not follow an ideal normal distribution, using the arithmetic mean to calculate the regional average wind power output is not an unbiased estimate. This invention employs a double-weighted averaging method to calculate the regional average wind power output to reduce extreme errors. The average wind power output P* of a region at time t is calculated through the following iterative process:

[0127]

[0128]

[0129] Where P i w is the normalized output value of the i-th wind farm in the region. i Here, c is the corresponding weighting coefficient, c is a constant, usually taken as 6-9 (the model defaults to 7.5), and S is the absolute median deviation, which is a robust measure of sample bias.

[0130] S = median{|P i -P * |}

[0131] The median is used to calculate P* as the initial value to begin iterative calculations. During the iteration process, w... i It is not monotonically changing; it may not converge or may exhibit alternating cycles of the two values. However, after 3-4 iterations, the range of P* usually does not exceed 10. -3 The default maximum number of iterations for the model is 10.

[0132] This invention calculates the confidence interval for wind power output based on probabilistic weather forecasts. It obtains forecasted and actual wind power output values ​​using historical weather forecast data and observed weather data (at least one year of continuous data, with a resolution of at least 3 hours). The deviation between the two values ​​is calculated, and a kernel distribution is fitted to the deviation. Based on this, a confidence level for the predicted wind power output at a given level is calculated. The model supports calculating confidence levels of 50%, 70%, or 90%, and supports generating confidence levels for predicted wind power output 1 day (24 hours), 3 days (72 hours), and 7 days (168 hours) in advance.

[0133] Step 103: Based on future weather forecasts, calculate wind power output and its confidence interval using the regional composite model. This invention inputs weather forecast data and outputs the corresponding wind power output and a selected confidence level. It can be used to predict the average output of wind farms in a specific area or region in the future; it can also input historical weather data from different locations to study the spatial patterns of the relationship between extreme weather and wind power output.

[0134] like Figure 2 As shown, this invention provides a method for calculating wind power output regional composite based on weather forecasts, comprising the following steps:

[0135] S1. Process high-resolution gridded data of near-surface wind fields across the country. For wind speeds at different heights near the ground, convert them according to the logarithmic law or exponential law. For simulated wind field data, correct them to the actual wind field according to the variance matching method.

[0136] S2. Generate a nationwide prevailing wind direction network, spatially resample the wind field data to the required spatial resolution, and calculate the direction with the highest frequency of wind direction for each grid point as the prevailing wind direction for that point.

[0137] S3. Set the wind turbine layout structure inside the wind farm, and arrange them in a 5×5 differential arrangement. The longitudinal spacing of the wind turbines is 7 times the rotor diameter, and the transverse spacing is 4.5 times the rotor diameter.

[0138] S4. Considering the wake effect, calculate the output power of a single wind farm. Use the Jensen wake model to calculate the influence of the upstream wind turbine on the wind speed at the downstream wind turbine.

[0139] S5. Configure the location of wind farms within the area and calculate the optimal orientation of the wind farms based on the prevailing wind direction at the grid point where the wind farms are located.

[0140] S6. Calculate the regional average wind power output. Use the double-weighted averaging method to calculate the regional average wind power output to reduce extreme errors.

[0141] S7. Calculate the confidence interval for wind power output based on probabilistic weather forecasts, use kernel distribution to fit the deviation between historical predicted output and actual output, and calculate the confidence interval for wind power output 1 day, 3 days or 7 days in advance at a given level (50%, 70% or 90%).

[0142] S8. Input weather forecast data and output the corresponding wind power output and the confidence level of the selected level.

[0143] This invention calculates the average value O of measured data O within a selected region and calibration period (10 years or more). ave And variance O var And the average value M of the simulated data M ave and variance M varFinally, the correction data M is calculated according to the following formula. adj :

[0144]

[0145] This ensures that the mean and variance of the simulated data are consistent with the observed data during the calibration period.

[0146] This invention employs an optimized algorithm that first fills in the missing data and then interpolates. The interpolation can be performed using the nearest neighbor method (the value of the output grid point is the value of the nearest input grid point, without considering other pixels), the bilinear method (the value of the output grid point is the weighted average of the grid points in the nearest 2×2 neighbors), or the bicubic method (the value of the output grid point is the weighted average of the grid points in the nearest 4×4 neighbors) to avoid the offset or magnification of blank or missing grid points at the coastline or region boundary during the interpolation operation.

[0147] This invention uses a polar coordinate system, dividing the circumference into 36 equal parts at 10° intervals. The prevailing wind direction is the direction with the highest frequency of wind occurrence during the statistical data period. By traversing all grid points, a national prevailing wind direction network is generated.

[0148] This invention calculates through the following iterative process:

[0149]

[0150]

[0151] Where P i w is the normalized output value of the i-th wind farm in the region. i Here, c is the corresponding weighting coefficient, c is a constant, usually taken as 6-9 (the model defaults to 7.5), and S is the absolute median deviation, which is a robust measure of sample bias.

[0152] S = median{|P i -P * |}

[0153] This invention uses the median to calculate P* as the initial value to begin iterative calculations, with a maximum of 10 iterations, ensuring that the variation range of P* between adjacent iterations does not exceed 10. -3 .

[0154] This invention constructs a composite regional wind power output model that considers prevailing wind direction and probabilistic weather forecasts. It takes into account the prevailing wind direction characteristics of different regions and the orientation of wind farms within the region, combined with a single wind farm output model. This provides a solution for predicting inter-provincial regional wind power in China and offers technical support for comprehensively assessing the characteristics and spatial differences of wind power output in regions affected by weather. In processing simulated wind field data, the model uses variance matching to correct systematic deviations from measured wind field data. When resampling gridded meteorological data, an optimized algorithm of first filling gaps and then interpolating is adopted to avoid the offset or amplification of blank or missing grid points at coastlines or regional boundaries during interpolation. A double-weighted averaging method is used to reduce extreme errors when calculating the regional average wind power output. Based on weather forecasts 1 day to 1 week in advance, the model can predict the range of single wind farms or regional average wind power output at corresponding time points. Based on climate simulations of future socio-economic development paths, the model can predict the changing characteristics of regional wind power output under different scenarios.

[0155] This invention applies the prediction made one week in advance of the Chaganhot Wind Farm in Baicheng City, Jilin Province, and the large-scale cold wave and temperature drop that occurred across the province on April 20, 2020. Cold waves cause a sharp drop in regional temperature, which is usually accompanied by an increase in power load. Predicting the wind power output level in a province during extreme weather events such as cold waves helps to ensure the rational scheduling and operation of the power grid and guarantee residents' electricity supply. Information retrieval shows that the Baicheng Meteorological Bureau issued a yellow cold wave warning signal at 20:10 on the evening of April 19, 2020, predicting that the minimum temperature would drop by more than 10℃ in the next 24 hours. The Jilin Provincial Meteorological Observatory first issued a blue cold wave warning signal at 19:05 on the evening of April 19, predicting that the minimum temperature in northeastern Songyuan, Changchun, Siping, northern Liaoyuan, and western Jilin would drop by more than 8℃ in the next 24 hours, with some areas reaching around 10℃. Subsequently, at 8:45 on the morning of April 20, it upgraded the cold wave warning signal to a yellow warning signal, predicting that the minimum temperature in eastern Songyuan, Changchun, northeastern Siping, Liaoyuan, most of Jilin, northern Tonghua, and western Baishan would drop by more than 10℃ in the next 24 hours, with some areas reaching 12℃.

[0156] S1: Processing high-resolution gridded near-surface wind field data nationwide; historical wind field data comes from the European Centre for Medium-Range Weather Forecasts (ECMWF) next-generation reanalysis data ERA5. "Reanalysis" refers to the integration of conventional meteorological observation data and satellite remote sensing data from different regions and sources around the world, and optimal integration and re-fusion through a data assimilation system. The ERA5 data has a spatial resolution of 0.25° and a temporal resolution of 1 hour. Its surface analysis products include near-surface wind speed variables at two heights: 10m and 100m. The data format is a two-dimensional wind vector. In this example, the 100m height wind field data is selected.

[0157] S2: Generate a nationwide prevailing wind direction network; select data from 10 consecutive years (2012-2021) to calculate the most frequent wind direction at each grid point. The Northeast Plain and North China Plain are mainly controlled by northwest or southwest winds; while the middle and lower reaches of the Yangtze River are mainly controlled by east and southeast winds; the Inner Mongolia Plateau and the hilly and mountainous areas of southern China are mainly controlled by north and northwest winds; the Sichuan Basin and Tarim Basin are mainly controlled by east winds; and the Tibetan Plateau and the Yunnan-Guizhou Plateau have more south and southwest winds.

[0158] S3: Set the internal wind turbine layout of the wind farm; use the default value.

[0159] S4: Calculate the output power of a single wind farm considering the wake effect; no data input is required for this step.

[0160] S5: Configure the location of wind farms within the region; the geographical distribution data of wind farms in China comes from the Global Wind Energy Tracking Project (GWPT), which includes information on major wind farms with an installed capacity of over 10MW worldwide. According to statistics from the International Renewable Energy Agency (IRENA), China's total onshore wind power installed capacity in 2021 was approximately 302,583MW, of which GWPT recorded 293,116MW, exceeding 96% of the total, which can relatively completely reflect the spatial distribution of wind farms in China. GWPT has recorded 136 onshore wind farms and their latitude and longitude information in Jilin Province, and the location of wind farms within the province is configured accordingly.

[0161] S6: Calculate the regional average wind power output; no data input is required for this step.

[0162] S7: Calculate the confidence interval for wind power output forecast based on probabilistic weather forecasts; in this example, the forecast wind power output interval and actual wind power output value are obtained by using the 2021 national 0.25° resolution grid historical weather forecast dataset and observed weather data, and a 70% confidence level is selected.

[0163] S8: Input weather forecast data and output the corresponding wind power output and the confidence level of the selected level. Figure 3 The predicted output curve of the Chaganhot Wind Farm on April 20th with a 3-hour resolution was presented one week in advance, and its comparison with the actual output curve was shown. The results show that the daily fluctuation of wind power output can be predicted relatively accurately 7 days in advance, including the near full power output from 0:00 to 14:00, and the output dropping to below 50% of the installed capacity after 14:00. The average prediction deviation of instantaneous output is 3.2%, and the maximum deviation occurs at 20:00 in the evening, at 17.7%, but it is still within the 70% confidence range. The predicted total power generation on that day is 95.9% of the actual value. Figure 4The paper presents a one-week advance forecast curve for the average wind power output of Jilin Province's wind farms on April 20th, with a 3-hour resolution, and compares it with the actual output curve. The results show that forecasting the daily fluctuations in the province's average wind power output seven days in advance is relatively accurate, including a gradual increase in output from 0:00 to 14:00, nearing full capacity at 14:00, followed by a rapid decline to below 50% of installed capacity. The average prediction deviation for instantaneous output is 10.0%, with the largest deviation occurring at 5:00 AM at 28.4%, exceeding the 70% confidence level. The predicted total power generation for that day was 116.7% of the actual value.

[0164] The present invention provides a computer-readable storage medium storing a computer program for executing a composite calculation method for wind power output area based on weather forecast.

[0165] This invention provides an electronic device, which includes: a processor and a memory; wherein,

[0166] Memory, used to store processor-executable instructions;

[0167] A processor is used to read executable instructions from memory and execute the instructions to implement a composite calculation method for wind power output areas based on weather forecasts.

[0168] Figure 5 This is a structural diagram of a wind power output regional composite calculation system based on weather forecast according to a preferred embodiment of the present invention.

[0169] like Figure 5 As shown, this invention provides a wind power output regional composite calculation system based on weather forecasts. The system includes:

[0170] Initial unit 501 is used for: processing high-resolution gridded wind field data of near-surface wind fields across the country to obtain wind field data at a specified height; generating a nationwide prevailing wind direction network based on the wind field data at the specified height; setting the internal wind turbine layout structure of wind farms; calculating the output power of a single wind farm considering wake effects; and configuring the distribution locations of wind farms within the region.

[0171] The regional composite model building unit 502 is used to calculate the configured regional average wind power output based on boundary conditions; and to calculate the confidence interval of wind power output based on historical probabilistic weather forecasts and regional average wind power output.

[0172] Output unit 503 is used to calculate wind power output and its confidence interval based on future weather forecasts using the regional composite model.

[0173] Preferably, the boundary setting unit 501 is used to process high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified height, and is also used for:

[0174] The high-resolution gridded wind field data of near-surface wind fields across the country were processed according to the logarithmic law:

[0175]

[0176] Where Z is the specified height, u(Z) is the wind speed at the specified height Z, u* is the friction velocity constant, K is the Karman constant, and Z0 is the ground roughness length.

[0177] Preferably, the boundary setting unit 501 is used to process high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified height, and is also used for:

[0178] High-resolution gridded wind field data of near-surface wind fields across the country were processed according to the exponential law:

[0179]

[0180] Where Z is the specified height at which the desired wind speed is obtained. A Given the height of the wind speed, u(Z) and u(Z) A () are heights Z and Z respectively. A The wind speed at point n is the Hermann exponent.

[0181] Preferably, the boundary setting unit 501 processes the high-resolution gridded wind field data of the near-surface wind field across the country to obtain wind field data at a specified height, and is also used for:

[0182] For simulated wind field data, corrections are made to the actual wind field data using a variance matching method:

[0183] Calculate the average value of the measured data O within the selected region and calibration period. ave And variance O var And the average value M of the simulated data M ave and variance M var Finally, the correction data M is calculated according to the following formula. adj :

[0184]

[0185] Based on correction data M adj Correcting the simulated wind field data.

[0186] Preferably, the boundary setting unit 501 is used to generate a nationwide prevailing wind direction network based on wind field data at a specified height, and is also used for:

[0187] Missing values ​​in the wind field data are supplemented using interpolation:

[0188] Nearest neighbor interpolation is used to assign the value of the input grid point closest to the output grid point to the output grid point, thus filling in the missing values ​​of the output grid point; or

[0189] By using bilinear interpolation, the weighted average of the grid points in the nearest 2×2 neighbors to the output grid point is used to supplement the missing values ​​of the output grid point; or

[0190] Bicubic interpolation is used to supplement the missing values ​​of the output grid point by averaging the weighted average of the grid points in the nearest 4×4 neighboring points.

[0191] Preferably, the boundary setting unit 501 is used to generate a nationwide prevailing wind direction network based on wind field data at a specified height, and is also used for:

[0192] Calculate the wind direction frequency distribution map for each grid point; based on the wind direction frequency distribution map, using a polar coordinate system, divide the circumference into equal parts according to a preset number of parts, and the prevailing wind direction is the direction in which the wind direction appears most frequently during the statistical data period. Traverse all grid points to generate a national prevailing wind direction network.

[0193] Preferably, the regional composite model establishment unit 502 is also used to calculate the regional average wind power output after the calculation configuration based on the double-weighted averaging method:

[0194]

[0195]

[0196] Where P i w is the normalized output value of the i-th wind farm in the region. i Here, c is the corresponding weighting coefficient, S is the absolute median deviation, and S is a robust measure of sample bias.

[0197] S = median{|P i -P * |}

[0198] Calculate P using the median of regional wind power output * The calculation begins with the initial value and continues iteratively until the preset convergence condition is met.

[0199] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0200] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0201] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0202] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0203] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0204] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0205] The invention has been described with reference to a few embodiments. However, as will be known to those skilled in the art, and as defined in the appended claims, other embodiments besides those disclosed above fall equivalently within the scope of the invention.

[0206] Generally, all terms used in the claims are to be interpreted according to their ordinary meaning in the art, unless otherwise expressly defined herein. All references to “a / the / the [device, component, etc.]” ​​are openly interpreted as at least one instance of said device, component, etc., unless otherwise expressly stated. The steps of any method disclosed herein need not be performed in the exact order disclosed unless explicitly stated otherwise.

Claims

1. A method for calculating wind power output regional composite based on weather forecasts, the method comprising: Wind field data processing and regional composite model initialization include: processing high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified altitude; the processing of high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified altitude includes: For simulated wind field data, corrections are made to the actual wind field data using a variance matching method: Calculate the measured data within the selected area and calibration period. O average O ave and variance O var and simulation data M average M ave and variance M var Finally, the correction data is calculated according to the following formula. M adj : Based on the correction data M adj Correcting the simulated wind field data; Based on wind field data at a specified altitude, a nationwide prevailing wind direction network is generated; the generation of the nationwide prevailing wind direction network based on wind field data at a specified altitude includes: Calculate the wind direction frequency distribution map for each grid point; based on the wind direction frequency distribution map, use the polar coordinate system to divide the circumference into equal parts according to a preset number of parts, and the prevailing wind direction is the direction in which the wind direction appears most frequently during the statistical data period. Traverse all grid points to generate a national prevailing wind direction network. The internal wind turbine layout of the wind farm is designed; the output power of a single wind farm considering wake effects is calculated; the locations of wind farms within the region are configured; a regional composite model is established, including: calculating the regional average wind power output after configuration based on boundary conditions, and also: calculating the regional average wind power output after configuration based on the double-weighted averaging method. in P i For the first in the region i Normalized output value of a wind farm w i For the corresponding weighting coefficients, c It is a constant. S The absolute median is a robust measure of sample bias. Calculate using the median of regional wind power output P* The calculation is started with the initial value and iterated until the preset convergence condition is reached. The predicted output confidence interval of wind power is calculated based on historical probabilistic weather forecasts and regional average wind power output; based on future weather forecasts, the wind power output and its confidence interval are calculated using the aforementioned regional composite model.

2. The method according to claim 1, wherein processing the high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified altitude includes: The high-resolution gridded wind field data of near-surface wind fields across the country were processed according to the logarithmic law: in Z To specify the height, u ( Z () specifies the height Z Wind speed at the location, u * The friction velocity constant is K Kármán's constant, Z 0 represents the ground roughness length.

3. The method according to claim 1, wherein processing the high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified altitude includes: High-resolution gridded wind field data of near-surface wind fields nationwide were processed according to the exponential law: in Z At the specified altitude for the desired wind speed, Z A At a height where the wind speed is known, u ( Z )and u ( Z A ( ) are respectively height Z and Z A Wind speed at the location, n This is the Hermann index.

4. The method according to claim 1, wherein generating a national prevailing wind direction network based on wind field data at a specified altitude includes: Missing values ​​in the wind field data are supplemented using interpolation: Nearest neighbor interpolation is used to assign the value of the input grid point closest to the output grid point to the output grid point, thus filling in the missing values ​​of the output grid point; or By using bilinear interpolation, the weighted average of the grid points in the nearest 2×2 neighbors to the output grid point is used to supplement the missing values ​​of the output grid point; or By using bicubic interpolation, the weighted average of the grid points in the nearest 4×4 neighbors to the output grid point is used to supplement the missing values ​​of the output grid point.

5. A composite calculation system for wind power output area based on weather forecast, the system comprising: The initial unit is used to process high-resolution gridded wind field data of near-surface wind fields across the country to obtain wind field data at a specified height. Based on wind field data at a specified altitude, a nationwide prevailing wind direction network is generated; the internal wind turbine layout structure of the wind farm is set; the output power of a single wind farm considering the wake effect is calculated; and the distribution locations of wind farms within the region are configured. The boundary setting unit processes high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified height, and is also used for: For simulated wind field data, corrections are made to the actual wind field data using a variance matching method: Calculate the measured data within the selected area and calibration period. O average O ave and variance O var and simulation data M average M ave and variance M var Finally, the correction data is calculated according to the following formula. M adj : Based on the correction data M adj Correcting the simulated wind field data; The boundary setting unit is used to generate a nationwide prevailing wind direction network based on wind field data at a specified height, and is also used for: Calculate the wind direction frequency distribution map for each grid point; based on the wind direction frequency distribution map, use the polar coordinate system to divide the circumference into equal parts according to a preset number of parts, and the prevailing wind direction is the direction in which the wind direction appears most frequently during the statistical data period. Traverse all grid points to generate a national prevailing wind direction network. The regional composite model building unit is used to calculate the configured regional average wind power output based on boundary conditions; and to calculate the predicted output confidence interval of wind power output based on historical probabilistic weather forecasts and regional average wind power output. The regional composite model building unit is used to calculate the configured regional average wind power output based on boundary conditions, and also to calculate the configured regional average wind power output based on the double-weighted averaging method. in P i For the first in the region i Normalized output value of a wind farm w i For the corresponding weighting coefficients, c It is a constant. S The absolute median is a robust measure of sample bias. Calculate using the median of regional wind power output P* The calculation is started with the initial value and iterated until the preset convergence condition is reached. The output unit is used to calculate the wind power output and its confidence interval based on future weather forecasts using the regional composite model.

6. The system according to claim 5, wherein the boundary setting unit is used to process high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified height, and is further used to: The high-resolution gridded wind field data of near-surface wind fields across the country were processed according to the logarithmic law: in Z To specify the height, u ( Z () specifies the height Z Wind speed at the location, u * The friction velocity constant is K Kármán's constant, Z 0 represents the ground roughness length.

7. The system according to claim 5, wherein the boundary setting unit is used to process high-resolution gridded wind field data of near-surface wind fields nationwide to obtain wind field data at a specified height, and is further used to: High-resolution gridded wind field data of near-surface wind fields nationwide were processed according to the exponential law: in Z At the specified altitude for the desired wind speed, Z A At a height where the wind speed is known, u ( Z )and u ( Z A ( ) are respectively height Z and Z A Wind speed at the location, n This is the Hermann index.

8. The system according to claim 5, wherein the boundary setting unit is configured to generate a nationwide prevailing wind direction network based on wind field data at a specified altitude, and is further configured to: Missing values ​​in the wind field data are supplemented using interpolation: Nearest neighbor interpolation is used to assign the value of the input grid point closest to the output grid point to the output grid point, thus filling in the missing values ​​of the output grid point; or By using bilinear interpolation, the weighted average of the grid points in the nearest 2×2 neighbors to the output grid point is used to supplement the missing values ​​of the output grid point; or By using bicubic interpolation, the weighted average of the grid points in the nearest 4×4 neighbors to the output grid point is used to supplement the missing values ​​of the output grid point.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program for performing the method of any one of claims 1-4.

10. An electronic device, characterized in that, The electronic device includes: a processor and a memory; wherein, The memory is used to store the processor-executable instructions; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method of any one of claims 1-4.

Images