Design wind speed prediction method, device and equipment of photovoltaic power station and storage medium

Abstract

The embodiment of the application provides a photovoltaic power station design wind speed prediction method, device, equipment and storage medium, determines target wind speed data corresponding to actual stress center height of a photovoltaic module; according to the data sequence length of the target wind speed data, a target extreme value sample extraction algorithm is determined, and an extreme value sample corresponding to the target wind speed data is extracted; a target extreme value model is determined from at least two preset candidate extreme value models; for the windward edge area of the photovoltaic power station, a preset target return period and an extreme value sample are input into the target extreme value model to obtain a first design wind speed; for the internal sheltered area, the target return period, the extreme value sample and a wind speed attenuation coefficient are input into the target extreme value model to obtain a second design wind speed; the first design wind speed and the second design wind speed are respectively introduced into a surface average wind speed reduction to obtain a target design wind speed of the photovoltaic power station. The embodiment of the application can provide more reliable design wind speed basis for fine wind resistance design of the photovoltaic power station.

Description

Technical Field

[0001] This application relates to the field of engineering meteorology technology, and in particular to a method and apparatus for predicting the design wind speed of a photovoltaic power station, an electronic device, and a storage medium. Background Technology

[0002] Currently, in the engineering design of photovoltaic power plants, the derivation of the design wind speed for the return period mainly relies on the conventional extreme value analysis method recommended by current standards. However, this method has the following limitations in practical engineering applications: (1) When converting wind speed data to height, the relevant technologies directly apply the standard 10m high-altitude power law formula for calculation. However, the actual force center height of photovoltaic modules is usually only 1.5m to 5m. The wind profile at this height is greatly affected by the surface vegetation and the roughness of the photovoltaic panel itself. The wind speed conversion according to the standard 10m high-altitude power law formula will produce a large profile conversion error.

[0003] (2) The on-site measured data of photovoltaic power plants are usually only short-series wind measurement data of two or three years, which cannot obtain effective extreme value samples. Moreover, the relevant schemes generally use extreme value type I or P-III distribution as extreme value model to derive the return period design wind speed. Among them, the shape parameter of extreme value type I is fixed and cannot effectively fit the thick tail characteristics of wind speed distribution in areas with frequent strong winds. The P-III distribution is also not flexible enough in handling the thick tail characteristics of wind speed distribution in areas with frequent strong winds.

[0004] (3) The relevant technologies assume that the entire photovoltaic power station is designed with the same return period design wind speed. For the rear photovoltaic modules that actually withstand lower wind speeds, unreasonable wind resistance redundancy will be generated.

[0005] (4) Photovoltaic power plants occupy a large area, while the design wind speed calculated by conventional extreme value theory is only the maximum wind speed at a single point. However, in actual extreme weather, the gust front will not cover the entire photovoltaic power plant at the same time. If the entire photovoltaic power plant is designed according to the maximum wind speed at a single point, unreasonable wind resistance redundancy will be generated.

[0006] In summary, how to provide a more reliable basis for the design wind speed for the refined wind-resistant design of photovoltaic power plants is an urgent problem to be solved in this field. Summary of the Invention

[0007] This application provides a method for predicting the design wind speed of a photovoltaic power plant to solve or at least partially solve the above-mentioned problems.

[0008] Accordingly, embodiments of this application also provide a design wind speed prediction device for a photovoltaic power station, an electronic device, and a storage medium to ensure the implementation and application of the above methods.

[0009] To address the aforementioned problems, this application discloses a method for predicting the design wind speed of a photovoltaic power plant, the method comprising: Obtain the actual stress center height and initial wind speed data of the photovoltaic modules of the photovoltaic power station; Determine the target wind speed data corresponding to the actual stress center height of the photovoltaic module based on the initial wind speed data; Based on the data sequence length of the target wind speed data, a target extreme value sample extraction algorithm is determined, and the target extreme value sample extraction algorithm is used to extract the extreme value samples corresponding to the target wind speed data; Based on the extreme value samples, a target extreme value model is determined from at least two preset candidate extreme value models; Determine the windward edge zone and the internal shading zone of the photovoltaic power station; For the windward edge region, the preset target return period and the extreme value sample are input into the target extreme value model to obtain the first design wind speed; and for the internal shading region, the wind speed attenuation coefficient corresponding to the internal shading region is determined according to the preset wake attenuation function, and the target return period, the extreme value sample and the wind speed attenuation coefficient are input into the target extreme value model to obtain the second design wind speed. The first design wind speed and the second design wind speed are respectively converted into the first face average design wind speed and the second face average design wind speed; The average design wind speed of the first face and the average design wind speed of the second face are used as the target design wind speed of the photovoltaic power station, and the target design wind speed is used in the engineering design of the photovoltaic power station.

[0010] Optionally, the initial wind speed data includes meteorological station observation data, field measurement data, and mesoscale reanalysis data. Determining the target wind speed data corresponding to the actual stress center height of the photovoltaic module based on the initial wind speed data includes: The initial wind speed data is subjected to a uniformity test to obtain the tested initial wind speed data; Acquire photovoltaic array data of the photovoltaic power station, including average height of the photovoltaic array, tilt angle of photovoltaic modules, and data on the land cover corresponding to the photovoltaic power station; The windward area index of the photovoltaic power station is determined based on the tilt angle of the photovoltaic modules; The zero-plane displacement of the photovoltaic power station is calculated based on the average height of the photovoltaic array and the windward surface area index. The background roughness of the photovoltaic power station is determined based on the surface cover data; The aerodynamic roughness of the photovoltaic power station is determined based on the windward surface area index and the background roughness. Based on the zero plane displacement, the aerodynamic roughness, and the actual stress center height of the photovoltaic module, the verified initial wind speed data is converted into the initial wind speed data corresponding to the actual stress center height of the photovoltaic module; The initial wind speed data corresponding to the actual stress center height of the photovoltaic module are fused to obtain the target wind speed data corresponding to the actual stress center height of the photovoltaic module.

[0011] Optionally, the step of determining the target extreme value sample extraction algorithm based on the data sequence length of the target wind speed data includes: If the data sequence length of the target wind speed data is greater than or equal to a preset sequence length threshold, then the annual maximum value method is determined as the target extreme value sample extraction algorithm; If the data sequence length of the target wind speed data is less than the sequence length threshold, then the peak value exceeding the threshold method is determined as the target extreme value sample extraction algorithm.

[0012] Optionally, the target wind speed data includes target wind speed data from multiple years, and the step of extracting extreme value samples corresponding to the target wind speed data using the target extreme value sample extraction algorithm includes: If the target extreme value sample extraction algorithm is the annual maximum value method, then the maximum wind speed extreme point in the target wind speed data for each year is determined, and the maximum wind speed extreme point in the target wind speed data for each year is integrated into the extreme value sample corresponding to the target wind speed data.

[0013] Optionally, the step of extracting extreme value samples corresponding to the target wind speed data using the target extreme value sample extraction algorithm includes: If the target extreme value sample extraction algorithm is the over-threshold peak method, then the average residual lifetime map method is used to determine the extreme value threshold corresponding to the target wind speed data; Get the storm time window; Based on the extreme value threshold and the storm time window, the target wind speed data is de-clustered to obtain the extreme value samples corresponding to the target wind speed data.

[0014] Optionally, the candidate extreme value model includes a generalized extreme value distribution model and a generalized Pareto distribution model, and the step of determining the target extreme value model from at least two preset candidate extreme value models based on the extreme value samples includes: The maximum likelihood estimation method is used to determine the first model parameters and the first maximum likelihood function value corresponding to the generalized extreme value distribution model based on the extreme value samples. Calculate the first KS test statistic, the first Akaike information criterion, and the first Bayesian information criterion corresponding to the generalized extreme value distribution model based on the first model parameters and the first maximum likelihood function value; and calculate the first score corresponding to the generalized extreme value distribution model based on the first KS test statistic, the first Akaike information criterion, and the first Bayesian information criterion. Using the maximum likelihood estimation method, the second model parameters and the second maximum likelihood function values ​​corresponding to the generalized Pareto distribution model are determined based on the extreme value samples. The second KS test statistic, the second Akaike information criterion, and the second Bayesian information criterion corresponding to the generalized Pareto distribution model are calculated based on the second model parameters and the second maximum likelihood function value. The second score corresponding to the generalized Pareto distribution model is then calculated based on the second KS test statistic, the second Akaike information criterion, and the second Bayesian information criterion. If the first score is greater than the second score, then the generalized extreme value distribution model is used as the target extreme value model; If the second score is greater than the first score, then the generalized Pareto distribution model is used as the target extreme value model; If the first score equals the second score, then the generalized extreme value distribution model or the generalized Pareto distribution model shall be used as the target extreme value model.

[0015] Optionally, the photovoltaic power station is equipped with at least one row of photovoltaic modules, and determining the windward edge area and the internal shading area of ​​the photovoltaic power station includes: Obtain historical extreme wind meteorological data corresponding to the photovoltaic power station, including historical wind direction and historical wind speed; A strong wind rose diagram was drawn based on the aforementioned historical extreme wind meteorological data. Based on the strong wind rose diagram, determine the historical frequency of occurrence and the corresponding historical wind speed for the historical wind direction; The contribution of the historical wind direction is determined based on the historical frequency of occurrence and the historical wind speed. The historical wind direction with the highest contribution is identified as the dominant strong wind direction; A predetermined number of photovoltaic modules located in the direction of the prevailing strong wind are identified as target photovoltaic modules. The area where the target photovoltaic module is located is designated as the windward edge zone of the photovoltaic power station, and the area of ​​the photovoltaic power station other than the windward edge zone is designated as the internal shading zone of the photovoltaic power station.

[0016] Optionally, determining the wind speed attenuation coefficient corresponding to the internal blocking area based on a preset wake attenuation function includes: Obtain the photovoltaic array spacing, vertical wind direction row spacing, effective shading height of the photovoltaic array, wake diffusion coefficient, and number of shading rows in the internal shading area of ​​the photovoltaic power station; Based on the tilt angle of the photovoltaic module, the spacing of the photovoltaic array, the vertical wind direction column spacing, the effective shading height of the photovoltaic array, and the wake diffusion coefficient, the initial single-row attenuation coefficient of the internal shading area is determined using the wake attenuation function; The initial attenuation coefficient of a single row is superimposed and corrected based on the number of rows of shielding to obtain the wind speed attenuation coefficient corresponding to the inner shielding area.

[0017] Optionally, converting the first design wind speed and the second design wind speed into a first face-average design wind speed and a second face-average design wind speed respectively includes: Obtain the target length of the photovoltaic power station along the dominant strong wind direction and the target width perpendicular to the dominant strong wind direction, as well as the target relationship between the variance of the surface average wind speed and the variance of the single-point wind speed. The area reduction factor calculation formula for the photovoltaic power station is determined based on the target relationship, the target length, the target width, and the Davenport spatial coherence empirical function. Solve the formula for calculating the area reduction factor to obtain the area reduction factor of the photovoltaic power station; The product of the first design wind speed and the area reduction factor is taken as the first average design wind speed of the face, and the product of the second design wind speed and the area reduction factor is taken as the second average design wind speed of the face.

[0018] This application also discloses a design wind speed prediction device for a photovoltaic power station, the device comprising: The wind speed data acquisition module is used to acquire the actual stress center height and initial wind speed data of the photovoltaic modules of the photovoltaic power station. The wind speed data conversion module is used to determine the target wind speed data corresponding to the actual stress center height of the photovoltaic module based on the initial wind speed data. The extreme value sample extraction module is used to determine the target extreme value sample extraction algorithm based on the data sequence length of the target wind speed data, and to use the target extreme value sample extraction algorithm to extract the extreme value samples corresponding to the target wind speed data. The target model determination module is used to determine a target extreme value model from at least two preset candidate extreme value models based on the extreme value samples. The site area division module is used to determine the windward edge area and the internal shading area of ​​the photovoltaic power station; The wind speed calculation module is designed to obtain a first design wind speed by inputting a preset target return period and the extreme value sample into the target extreme value model for the windward edge area; and to obtain a second design wind speed by inputting the target return period, the extreme value sample, and the wind speed attenuation coefficient into the target extreme value model for the internal shading area according to a preset wake attenuation function. The wind speed conversion module is designed to convert the first design wind speed and the second design wind speed into the first average design wind speed and the second average design wind speed, respectively. The design wind speed determination module is used to take the average design wind speed of the first face and the average design wind speed of the second face as the target design wind speed of the photovoltaic power station.

[0019] This application also discloses an electronic device, including: a processor; and a memory storing executable code thereon, which, when executed, causes the processor to perform one or more of the photovoltaic power plant design wind speed prediction methods described in this application.

[0020] This application also discloses a machine-readable medium storing executable code, which, when executed, causes a processor to perform one or more of the photovoltaic power plant design wind speed prediction methods described in this application.

[0021] Compared with the prior art, the embodiments of this application have the following advantages: This application converts the initial wind speed data into target wind speed data corresponding to the actual stress center height of the photovoltaic module. This ensures that the target wind speed data fully considers the roughness influence at the actual stress center height of the photovoltaic module, avoiding the errors caused by directly applying the standard 10m height power law formula. Based on the data sequence length of the target wind speed data, this application selects a suitable target extreme value sample extraction algorithm, thereby ensuring that effective extreme value samples can be obtained even for short-sequence wind measurement data. On this basis, based on the extracted extreme value samples, a target extreme value model is determined from at least two preset candidate extreme value models. This overcomes the limitations of traditional methods, which are limited to using a Type I extreme value distribution with fixed shape parameters, resulting in an inability to effectively fit the thick-tailed characteristics of wind speed distribution in areas with frequent strong winds, and the inflexibility of the P-III type distribution in handling the thick-tailed characteristics of such areas. This significantly improves the fitting accuracy of short-sequence data under thick-tailed distribution conditions. This application's embodiments divide the photovoltaic power station into a windward edge zone and an internal shading zone. For the internal shading zone, which experiences lower actual wind speeds, a wind speed attenuation coefficient is introduced to calculate the design wind speed. This avoids the unreasonable wind resistance redundancy caused by using the same return-time design wind speed for the entire photovoltaic power station. Simultaneously, both the first and second design wind speeds are reduced by surface-average wind speed, transforming them into first-surface-average design wind speeds and second-surface-average design wind speeds. This corrects the unreasonable wind resistance redundancy caused by directly applying the maximum wind speed at a single point to the design of a large-area photovoltaic array in the traditional method. Thus, while ensuring structural safety, the amount of steel used and the cost of the support structure are effectively controlled. Therefore, this application's embodiments can provide a more reliable design wind speed basis for the refined wind resistance design of photovoltaic power stations. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating the steps of an embodiment of a method for predicting the design wind speed of a photovoltaic power station according to this application; Figure 2 This is a structural block diagram of an embodiment of a design wind speed prediction device for a photovoltaic power station according to this application; Figure 3 This is a schematic diagram of the structure of a device provided in an embodiment of this application. Detailed Implementation

[0023] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0024] Reference Figure 1 This is a flowchart illustrating the steps of an embodiment of a photovoltaic power plant design wind speed prediction method according to this application, including the following steps: Step 101: Obtain the actual stress center height and initial wind speed data of the photovoltaic modules of the photovoltaic power station.

[0025] The method for predicting the design wind speed of a photovoltaic power station provided in this application embodiment can be applied to the fields of engineering meteorological parameter estimation and structural wind-resistant design technology. The target design wind speed with the predicted target return period can be specifically applied to the engineering design of photovoltaic power stations (including ground-mounted, fixed adjustable and single-axis tracking types) such as array structure wind-resistant design, micro-site selection and support control strategy optimization.

[0026] In this embodiment, the return period refers to the statistically significant time interval between two consecutive occurrences of a wind speed greater than the design wind speed. The return period design wind speed is the number of years at which a specific wind speed might be encountered, based on a long-term statistical average. For example, using a maximum wind speed with a return period of 50 years as the design wind speed means that the average time interval between occurrences of this maximum wind speed over a long time series is 50 years. The photovoltaic power plant design wind speed prediction method provided in this embodiment predicts the target design wind speed for the target return period, thereby obtaining the target design wind speed for the target return period required for the engineering design of the photovoltaic power plant.

[0027] In step 101, the actual stress center height and initial wind speed data of the photovoltaic modules of the photovoltaic power station are obtained.

[0028] The actual stress center height of the photovoltaic module is obtained based on the actual situation of the photovoltaic module in the photovoltaic power station, and is usually 1.5m to 5m.

[0029] The initial wind speed data can be single-source or multi-source. In one embodiment, the initial wind speed data includes at least one or more wind speed data from meteorological station observations, on-site measured data from photovoltaic power plants, and mesoscale reanalysis data. The mesoscale reanalysis data is obtained by extracting mesoscale meteorological reanalysis datasets such as ERA5 or MERRA2. These datasets have long coverage periods but low spatial resolution (typically tens of kilometers of grid space), and subsequent algorithms perform local downscaling to obtain the mesoscale reanalysis data.

[0030] If the initial wind speed data is single-source wind speed data, then one of the following should be used as the initial wind speed data: meteorological station observation data, on-site measured data of photovoltaic power station, and mesoscale reanalysis data; if the initial wind speed data is multi-source wind speed data, then at least two of the following should be used as the initial wind speed data: meteorological station observation data, on-site measured data of photovoltaic power station, and mesoscale reanalysis data.

[0031] Step 102: Determine the target wind speed data corresponding to the actual stress center height of the photovoltaic module based on the initial wind speed data.

[0032] In step 102, the initial wind speed data needs to be converted to the actual stress center height of the photovoltaic module to obtain the target wind speed data corresponding to the actual stress center height of the photovoltaic module.

[0033] Optionally, when the initial wind speed data includes meteorological station observation data, field measurement data, and mesoscale reanalysis data, i.e., the initial wind speed data is multi-source wind speed data, the step of determining the target wind speed data corresponding to the actual stress center height of the photovoltaic module based on the initial wind speed data includes: The initial wind speed data is subjected to a uniformity test to obtain the tested initial wind speed data; Acquire photovoltaic array data of the photovoltaic power station, including average height of the photovoltaic array, tilt angle of photovoltaic modules, and data on the land cover corresponding to the photovoltaic power station; The windward area index of the photovoltaic power station is determined based on the tilt angle of the photovoltaic modules; The zero-plane displacement of the photovoltaic power station is calculated based on the average height of the photovoltaic array and the windward surface area index. The background roughness of the photovoltaic power station is determined based on the surface cover data; The aerodynamic roughness of the photovoltaic power station is determined based on the windward surface area index and the background roughness. Based on the zero plane displacement, the aerodynamic roughness, and the actual stress center height of the photovoltaic module, the verified initial wind speed data is converted into the initial wind speed data corresponding to the actual stress center height of the photovoltaic module; The initial wind speed data corresponding to the actual stress center height of the photovoltaic module are fused to obtain the target wind speed data corresponding to the actual stress center height of the photovoltaic module.

[0034] In this embodiment, after acquiring meteorological station observation data, field measurement data, and mesoscale reanalysis data, a uniformity test is performed. Considering the low-profile characteristics of photovoltaic power stations, the zero-plane displacement and aerodynamic roughness of the photovoltaic power station are dynamically calculated based on the photovoltaic array data. This replaces the conventional power-law formula to uniformly convert the initial wind speed data from multiple sources to the target wind speed data corresponding to the actual stress center height of the photovoltaic module.

[0035] Specifically, the uniformity of meteorological station observation data, field measurement data, and mesoscale reanalysis data was tested to obtain the initial wind speed data after testing.

[0036] In one example, the Standard Normal Homogeneity Test (SNHT) or the Petitt mutation test can be used to check homogeneity. Algorithms identify abrupt changes in the initial wind speed data time series due to unnatural climate changes (e.g., step changes caused by the relocation of wind towers or obstruction by surrounding buildings), and a mean smoothing model is used to correct the data segments before and after the abrupt changes, ensuring the consistency of the long-series statistical characteristics.

[0037] Specifically, the SNHT test divides the sequence into a first subsequence S1 and a second subsequence S2 at each possible time t, calculates the standardized mean difference between the two subsequences, and finds the time t that maximizes this difference, which is the most likely mutation point. The Pettitt mutation test, on the other hand, compares the overall positional differences between the first and second halves of the sequence based on rank to determine whether there is a step jump at the distribution center.

[0038] In addition to the SNHT test and Petitt's mutation test used in this example, other processing methods that can effectively control data quality can also be used to perform homogeneity tests on initial wind speed data.

[0039] Acquire photovoltaic array data for photovoltaic power plants. The photovoltaic array data includes the average height of the photovoltaic array, the tilt angle of the photovoltaic modules, and the data of the corresponding land cover of the photovoltaic power plant.

[0040] The average height of the photovoltaic array is the average physical height of the photovoltaic array (unit: m), which is the average value of all photovoltaic modules in the photovoltaic power station from the ground to the highest point of the photovoltaic module.

[0041] The tilt angle of a photovoltaic module is a key geometric parameter that determines the aerodynamic characteristics of a photovoltaic array. Specifically, based on the tilt angle θ of the photovoltaic module, the windward area index λ is calculated using formula (1). f Windward surface area index λ f (Frontal area index, dimensionless) refers to the photovoltaic array density, which is defined as the sum of the projected windbreak areas of photovoltaic modules per unit surface area in the prevailing wind direction.

[0042] (1) Where N is the number of photovoltaic modules per unit area; A is the surface area of ​​a single photovoltaic module (unit: m²). 2 ); θ is the tilt angle of the photovoltaic module, that is, the angle between the photovoltaic module and the horizontal plane (unit: °); S total Total surface area of ​​the unit region (unit: m²) 2 Therefore, the larger the tilt angle θ of the photovoltaic module, the larger the windward area index λ. f The larger the diameter, the stronger its obstruction of airflow.

[0043] The drag coefficient c can be dynamically adjusted based on the tilt angle of the photovoltaic module. The drag coefficient c is an empirical constant obtained from the aerodynamic characteristics of the photovoltaic module and wind tunnel tests, used for subsequent calculations of zero-plane displacement. Specifically, according to wind tunnel calibration, the airflow separation characteristics on the module surface differ significantly at different photovoltaic module tilt angles. The drag coefficient c can be determined according to the following rules: when the photovoltaic module tilt angle is 0°-15° (the normal operating angle of the photovoltaic module's single axis), c=0.7; when the photovoltaic module tilt angle is 15°-30° (the normal angle of the photovoltaic module's fixed support), c=0.8; when the photovoltaic module tilt angle is 30°-45° (the fixed support for high-latitude photovoltaic modules), c=0.9; and when the photovoltaic module tilt angle is 90° (the photovoltaic module's single-axis high-wind defense posture), c=1.2.

[0044] In one embodiment, for the single-axis tracking bracket of the photovoltaic module, the actual tilt angle of the photovoltaic module can be read hourly, and the corresponding windward area index λ can be updated in real time. f And the drag coefficient c, to avoid the calculation error caused by using the traditional fixed tilt angle method when the photovoltaic module is in a wind-defense posture.

[0045] Zero-plane displacement (unit: m) refers to the effective theoretical height at which the average wind speed tends to zero due to the presence of ground roughness elements (i.e., photovoltaic arrays). The zero-plane displacement of a photovoltaic power station is calculated based on the average height of the photovoltaic array and the windward surface area index according to the following formula (2).

[0046] (2) Where d is the zero-plane displacement (unit: m), H is the average height of the photovoltaic array (unit: m), c is the drag coefficient (dimensionless), and λ is the mean height of the photovoltaic array. f The windward surface area index (dimensionless).

[0047] The photovoltaic array data obtained above includes surface cover data corresponding to the photovoltaic power station, reflecting the type of surface cover. Surface cover is a fundamental drag source for near-surface wind fields, used to determine the background roughness of the photovoltaic power station and correct for the contribution of additional roughness from the photovoltaic array.

[0048] Specifically, the background roughness z 0b This refers to the roughness of the surface cover itself when there is no photovoltaic array. The empirical values ​​based on the type of surface cover are shown in Table 1 below: Table 1 Correspondence between background roughness and surface cover type

[0049] In one embodiment, for seasonal land surfaces such as crops and grasslands, the phenological calendar can be updated quarterly based on the region where the photovoltaic power station is located. 0b Taking cornfields in North China as an example, the background roughness can be set to the corn seedling stage (April-May). 0b =0.02m, corn maturity period (July-August) z 0b =0.2m.

[0050] When the surface roughness of the land where a photovoltaic power station is located is high, the additional obstruction effect of the photovoltaic array will be weakened. Therefore, it is necessary to introduce a surface cover correction factor k. b As shown in the following formula (3): (3) Among them, z 0b Let be the background roughness, and H be the average height of the photovoltaic array. In one example, when z 0b When k > 0.1H, b A value of 0.5 can be used to avoid over-correction.

[0051] Aerodynamic roughness length (unit: m) is a drag scale parameter characterizing the effect of the ground surface (including vegetation and photovoltaic panels) on the attenuation of near-surface wind speed. The aerodynamic roughness of the photovoltaic power station is determined based on the windward surface area index and the background roughness. As shown in formula (4), the aerodynamic roughness is dynamically calculated using a superposition model combining the background roughness with the additional roughness of the photovoltaic array, combined with the drag balance principle and wind tunnel test data: (4) in, The aerodynamic roughness of the photovoltaic array (unit: m). It refers to the background roughness of the ground surface where the photovoltaic power station is located. It is the land cover correction factor. This refers to the additional aerodynamic roughness (unit: m) of the photovoltaic array. The calculation is performed using the optimized MacDonald empirical formula, as shown in formula (5) below: (5) Where H is the average height of the photovoltaic array; κ is the von Kármán constant, a dimensionless empirical constant in fluid dynamics, typically taking a value of approximately 0.4; λ is the drag coefficient of a single photovoltaic module, and its value is consistent with the drag coefficient c in the zero-plane displacement calculation formula (2), thereby ensuring the physical consistency between the zero-plane displacement d and the aerodynamic roughness z0 calculation; fThis refers to the area index of the windward side.

[0052] Based on the zero-plane displacement, aerodynamic roughness, and the actual center height of the photovoltaic module, the initial wind speed data after verification is converted into the initial wind speed data corresponding to the actual center height of the photovoltaic module according to the following formula (6): (6) Wherein, U(z) is the average wind speed at height z (unit: m / s), and in this embodiment, it refers to the initial wind speed data corresponding to the actual force center height of the photovoltaic module after conversion; z is the actual force center height of the photovoltaic module (unit: m), that is, the centroid height of the photovoltaic module to be calculated. For example, if the centroid height of a photovoltaic module is 2.5m, then z=2.5m. Friction velocity (unit: m / s) is a physical quantity characterizing the intensity of shear stress on the Earth's surface, obtained by inverse calculation from meteorological stations or measured wind speeds; κ is the von Kármán constant; d is the zero-plane displacement; and z0 is the aerodynamic roughness.

[0053] Multi-source initial wind speed data converted to the actual stress center height of the photovoltaic module are fused to obtain target wind speed data corresponding to the actual stress center height of the photovoltaic module. Specifically, after completing the above-mentioned multi-source data preprocessing, uniformity verification, and low-level wind profile dynamic correction operations, the three types of data—meteorological station observation data, field measured data, and mesoscale reanalysis data—are fused into a unified long-series wind speed data set for subsequent extreme value analysis. In one embodiment, the fusion quality of the fused target wind speed data can also be verified.

[0054] In one example, the data fusion method can include two approaches: simplified mathematical statistical fusion and physical constraint variational data assimilation fusion. Simplified mathematical statistical fusion uses the actual measured data from the weather station as a benchmark, corrects the deviation between the weather station observation data and the mesoscale reanalysis data through linear regression, and then dynamically weights the data according to the reliability of the data source to generate a unified wind speed sequence. Physical constraint variational data assimilation fusion uses 3DVAR / 4DVAR variational assimilation technology, embeds the photovoltaic low-profile wind profile correction formula (i.e., the above formula (6)) into the observation operator as a physical constraint, minimizes the cost function, and obtains high-precision wind field data that is spatiotemporally continuous and physically self-consistent after fusion.

[0055] If the initial wind speed data is multi-source data from other data sources, the above method can also be used to determine the target wind speed data corresponding to the actual stress center height of the photovoltaic module; if the initial wind speed data is single-source data, after converting the verified initial wind speed data into the initial wind speed data corresponding to the actual stress center height of the photovoltaic module, there is no need to perform the data fusion step.

[0056] This application embodiment converts the initial wind speed data into target wind speed data corresponding to the actual stress center height of the photovoltaic module, so that the target wind speed data fully considers the roughness influence at the actual stress center height of the photovoltaic module, avoiding the error caused by directly applying the standard 10m height power law formula for calculation.

[0057] Step 103: Determine the target extreme value sample extraction algorithm based on the data sequence length of the target wind speed data, and use the target extreme value sample extraction algorithm to extract the extreme value samples corresponding to the target wind speed data.

[0058] In step 103, a suitable target extreme value sample extraction algorithm is matched based on the data sequence length of the target wind speed data. In one embodiment, the optional extreme value sample extraction algorithms include the annual maximum value method and the peak value exceeding the threshold method.

[0059] Optionally, the step of determining the target extreme value sample extraction algorithm based on the data sequence length of the target wind speed data includes: If the data sequence length of the target wind speed data is greater than or equal to a preset sequence length threshold, then the annual maximum value method is determined as the target extreme value sample extraction algorithm; If the data sequence length of the target wind speed data is less than the sequence length threshold, then the peak value exceeding the threshold method is determined as the target extreme value sample extraction algorithm.

[0060] In this embodiment, for long-sequence data with a sequence length greater than or equal to a preset sequence length threshold, the Annual Maximum Method (AM) is used as the target extreme value sample extraction algorithm, such as long-sequence data like meteorological station observation data and mesoscale reanalysis data. For short-sequence data with a sequence length less than the preset sequence length threshold, the Peaks-Over-Threshold (POT) method is used as the target extreme value sample extraction algorithm. This extracts extreme wind speed samples through independence testing (i.e., declustering) to expand the effective data volume, such as short-sequence data like field-measured data. If the target wind speed data is obtained by fusing multi-source data, the Annual Maximum Method or the Peaks-Over-Threshold method is selected for extreme value extraction based on the sequence length of the fused target wind speed data.

[0061] The preset sequence length threshold can be set according to actual conditions. For example, data with a sequence length greater than or equal to 20 years can be considered long sequence data, and data with a sequence length less than 20 years can be considered short sequence data. This application does not impose any restrictions on the specific value of the preset sequence length threshold.

[0062] The embodiments of this application select a suitable target extreme value sample extraction algorithm based on the data sequence length of the target wind speed data, thereby ensuring that effective extreme value samples can be obtained even for short-sequence wind measurement data.

[0063] After determining the target extreme value sample extraction algorithm, the target extreme value sample extraction algorithm is used to extract the extreme value samples corresponding to the target wind speed data.

[0064] Optionally, the target wind speed data includes target wind speed data from multiple years, and the step of extracting extreme value samples corresponding to the target wind speed data using the target extreme value sample extraction algorithm includes: If the target extreme value sample extraction algorithm is the annual maximum value method, then the maximum wind speed extreme point in the target wind speed data for each year is determined, and the maximum wind speed extreme point in the target wind speed data for each year is integrated into the extreme value sample corresponding to the target wind speed data.

[0065] In this embodiment, the annual maximum value method is used to extract extreme value samples of long-sequence target wind speed data. Specifically, using a complete natural year (or a specific wind season) as a time window, the maximum wind speed extreme point within the target wind speed data of each year is extracted. The maximum wind speed extreme points extracted from the target wind speed data of each year are integrated into the extreme value samples corresponding to the target wind speed data, thereby obtaining an independent extreme value data sequence with the same sample size as the annual time of the target wind speed data.

[0066] The embodiments of this application use the annual maximum value method to extract extreme value samples of target wind speed data in long series, which ensures the reliability of extreme value samples.

[0067] Optionally, the step of extracting extreme value samples corresponding to the target wind speed data using the target extreme value sample extraction algorithm includes: If the target extreme value sample extraction algorithm is the over-threshold peak method, then the average residual lifetime map method is used to determine the extreme value threshold corresponding to the target wind speed data; Get the storm time window; Based on the extreme value threshold and the storm time window, the target wind speed data is de-clustered to obtain the extreme value samples corresponding to the target wind speed data.

[0068] In this embodiment, the extreme value samples of the target wind speed data in the short sequence are extracted using the over-threshold peak method.

[0069] Specifically, the parameters of the average residual lifetime map are set based on experience, the target wind speed data is processed using the average residual lifetime map method, and the wind speed corresponding to the starting point of the approximately linear trend of the average residual lifetime map is selected as the extreme value threshold.

[0070] Set a storm time window (e.g., 48 hours) and perform declustering on the target wind speed data to obtain the extreme value samples corresponding to the target wind speed data. The declustering steps are as follows: If the wind speed of the target wind speed data exceeds the extreme value threshold multiple times consecutively, only the storm peak value that is both locally maximum and globally maximum is extracted as the extreme value sample, that is, the highest point in the wind speed sequence exceeding the threshold is extracted as the extreme value sample; if the time interval between two peaks is less than the storm time window, or the wind speed between two peaks has not fallen below the local minimum, it is considered as the same meteorological event, and only the maximum value in the same meteorological event is extracted as an independent extreme value sample. Here, the local minimum refers to the trough node in the wind speed sequence where both the preceding and following times are greater than the current time. If the wind speed between two peaks has not fallen below this local minimum, it indicates that an independent storm gap has not formed, and it is still considered as the same meteorological event, and only its maximum value is extracted as an independent extreme value sample.

[0071] The embodiments of this application use the over-threshold peak method to extract extreme value samples of target wind speed data in short sequences, ensuring that effective extreme value samples can be extracted from short sequence data.

[0072] Step 104: Based on the extreme value samples, determine the target extreme value model from at least two preset candidate extreme value models.

[0073] In step 104, a target extreme value model is determined from at least two pre-defined candidate extreme value models based on the extreme value samples. In one embodiment, the candidate extreme value models include a generalized extreme value distribution (GEV) model and a generalized Pareto distribution (GPD) model.

[0074] Optionally, determining the target extreme value model from at least two preset candidate extreme value models based on the extreme value samples includes: The maximum likelihood estimation method is used to determine the first model parameters and the first maximum likelihood function value corresponding to the generalized extreme value distribution model based on the extreme value samples. Calculate the first KS test statistic, the first Akaike information criterion, and the first Bayesian information criterion corresponding to the generalized extreme value distribution model based on the first model parameters and the first maximum likelihood function value; and calculate the first score corresponding to the generalized extreme value distribution model based on the first KS test statistic, the first Akaike information criterion, and the first Bayesian information criterion. Using the maximum likelihood estimation method, the second model parameters and the second maximum likelihood function values ​​corresponding to the generalized Pareto distribution model are determined based on the extreme value samples. The second KS test statistic, the second Akaike information criterion, and the second Bayesian information criterion corresponding to the generalized Pareto distribution model are calculated based on the second model parameters and the second maximum likelihood function value. The second score corresponding to the generalized Pareto distribution model is then calculated based on the second KS test statistic, the second Akaike information criterion, and the second Bayesian information criterion. If the first score is greater than the second score, then the generalized extreme value distribution model is used as the target extreme value model; If the second score is greater than the first score, then the generalized Pareto distribution model is used as the target extreme value model; If the first score equals the second score, then the generalized extreme value distribution model or the generalized Pareto distribution model shall be used as the target extreme value model.

[0075] In this embodiment, after estimating the model parameters of the generalized extreme value distribution model and the generalized Pareto distribution model respectively, the KS test is used to evaluate the global goodness of fit of the model. At the same time, the scores of the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC) are calculated to penalize overfitting. Based on the weighted total score of the KS test statistic, the Akaike Information Criterion, and the Bayesian Information Criterion, the extreme value model with the best fit is selected from the generalized extreme value distribution model and the generalized Pareto distribution model.

[0076] Specifically, for the generalized extreme value distribution model, the shape, scale, and location parameters of the generalized extreme value distribution model are calculated using the maximum likelihood estimation (MLE) method as the first model parameters, and the maximum likelihood function value that appears in the model parameter estimation process is the first maximum likelihood function value.

[0077] The first KS test statistic of the first model parameters is calculated to evaluate the global goodness of fit, and the first Akaike information criterion and the first Bayesian information criterion are calculated based on the first model parameters and the first maximum likelihood function value to penalize overfitting.

[0078] The First Akaike Information Criterion (AIC) is calculated according to formula (7): AIC=2k 2ln(L)(7) The first Bayesian information criterion (BIC) is calculated according to formula (8): BIC = kln(n) 2ln(L)(8) Where AIC / BIC are the information content criterion scores (dimensionless), the smaller the value, the better the overall performance of the model in balancing fitting accuracy and model complexity; k is the number of independent parameters in the target extreme value model. The generalized extreme value distribution model contains three parameters: shape, scale, and location, so k=3; L is the first maximum likelihood function value, which characterizes the goodness of fit of the model to the current data; n is the total number of extreme wind speed samples participating in the model fitting, i.e., the number of extreme value samples; ln is the natural logarithm function.

[0079] The first KS test statistic, the first Akaike information criterion, and the first Bayesian information criterion are normalized and weighted by weight matrix to obtain the first score corresponding to the generalized extreme value distribution model.

[0080] For the generalized Pareto distribution model, the shape and scale parameters of the generalized Pareto distribution model are calculated using the maximum likelihood estimation method as the second model parameters. The maximum likelihood function value that appears in the model parameter estimation process is the second maximum likelihood function value.

[0081] The second KS test statistic of the second model parameters is calculated to evaluate the global goodness of fit, and the second Akaike information criterion and the second Bayesian information criterion are calculated based on the second model parameters and the second maximum likelihood function value to penalize overfitting. The second Akaike information criterion AIC is calculated according to the above formula (7), and the second Bayesian information criterion BIC is calculated according to the above formula (8). However, since the generalized Pareto distribution model contains two parameters, shape and scale, k=2 is taken in the calculation.

[0082] The second KS test statistic, the second Akaike information criterion, and the second Bayesian information criterion are normalized and weighted by the weight matrix to obtain the second score corresponding to the generalized Pareto distribution model.

[0083] If the first score is greater than the second score, the generalized extreme value distribution model is used as the target extreme value model; if the second score is greater than the first score, the generalized Pareto distribution model is used as the target extreme value model; if the first score is equal to the second score, one of the generalized extreme value distribution model and the generalized Pareto distribution model is selected as the target extreme value model.

[0084] In one example, in the model parameter estimation stage, in addition to MLE, when faced with field measured data with short sequence lengths, the Bayesian Markov Chain Monte Carlo (MCMC) method can be used as an alternative. By introducing long-sequence meteorological data of the region as a prior distribution, the stability of parameter estimation in the case of extremely small samples can be further enhanced.

[0085] Specifically, using a Bayesian inference framework, the distribution of extreme statistical parameters of long-sequence regional meteorological stations is used as the prior distribution, and the measured data from stations with shorter sequence lengths are used as the likelihood function to construct the posterior distribution of the extreme value model parameters. Since the posterior distribution has no analytical solution, a Markov chain is constructed using the Metropolis-Hastings sampling algorithm. The Markov chain is used to form a stationary walk path in the parameter space. Hastings' algorithm guarantees chain convergence to the true posterior distribution through the acceptance-rejection criterion. After the Markov chain converges, samples are extracted and the expected value of the posterior distribution of the parameters is calculated, which is then used as the final model parameters for the extreme value model. When faced with field-measured data with short sequence lengths, this method can replace the traditional MLE and effectively solve the problem of parameter estimation divergence under extremely small sample conditions.

[0086] Based on the extracted extreme value samples, this application's embodiments determine a better target extreme value model from the generalized extreme value distribution model and the generalized Pareto distribution model according to the scores. This overcomes the limitations of traditional methods, which are limited to using the extreme value type I distribution with fixed shape parameters, resulting in an inability to effectively fit the thick-tailed characteristics of wind speed distribution in areas with frequent strong winds, and the problem that the P-III type distribution is not flexible enough in handling the thick-tailed characteristics of such areas. This significantly improves the fitting accuracy of short sequence data under the condition of thick-tailed distribution.

[0087] Step 105: Determine the windward edge area and the internal shading area of ​​the photovoltaic power station.

[0088] In step 105, the photovoltaic power station is divided into the windward edge zone, which is usually directly subjected to the impact of strong winds, and the internal shading zone, which is actually subjected to lower wind speeds.

[0089] Optionally, the photovoltaic power station is equipped with at least one row of photovoltaic modules, and determining the windward edge area and the internal shading area of ​​the photovoltaic power station includes: Obtain historical extreme wind meteorological data corresponding to the photovoltaic power station, including historical wind direction and historical wind speed; A strong wind rose diagram was drawn based on the aforementioned historical extreme wind meteorological data. Based on the strong wind rose diagram, determine the historical frequency of occurrence and the corresponding historical wind speed for the historical wind direction; The contribution of the historical wind direction is determined based on the historical frequency of occurrence and the historical wind speed. The historical wind direction with the highest contribution is identified as the dominant strong wind direction; A predetermined number of photovoltaic modules located in the direction of the prevailing strong wind are identified as target photovoltaic modules. The area where the target photovoltaic module is located is designated as the windward edge zone of the photovoltaic power station, and the area of ​​the photovoltaic power station other than the windward edge zone is designated as the internal shading zone of the photovoltaic power station.

[0090] This application embodiment achieves the division of the windward edge zone and the internal shading zone of a photovoltaic power station by determining the dominant strong wind direction based on historical data.

[0091] Obtain historical extreme wind weather data for the area where the photovoltaic power station is located. This data includes historical wind direction and speed. Based on this historical extreme wind weather data, construct a strong wind rose diagram, a statistical chart specifically designed to illustrate the characteristics of strong wind direction.

[0092] In historical extreme wind events, for long-series data, the wind direction with the highest frequency and significant contribution to wind speed can be identified as the dominant strong wind direction; for short-series data, the wind direction with a significant contribution to wind speed can be identified as the dominant strong wind direction. In this embodiment, based on a strong wind direction rose diagram, the historical frequency and corresponding historical wind speed of each historical wind direction are determined. The contribution of each historical wind direction is then determined based on its historical frequency and historical wind speed, and the historical wind direction with the highest contribution is identified as the dominant strong wind direction. Specifically, determining the contribution of a historical wind direction based on its historical frequency and historical wind speed can be achieved by weighting the historical frequency and historical wind speed.

[0093] Along the direction of the prevailing strong wind, a predetermined number of photovoltaic (PV) modules located in that direction are identified as target PV modules. For example, the outermost first to third rows of PV modules directly facing the incoming wind can be designated as target PV modules. The area where the target PV modules are located is designated as the windward edge zone of the PV power plant, and the area of ​​the PV power plant excluding the windward edge zone, i.e., the area behind the target PV modules that is shaded by the wake, is designated as the internal shading zone of the PV power plant.

[0094] This application embodiment determines the dominant strong wind direction based on historical extreme wind meteorological data, and then divides the photovoltaic power station into a windward edge zone and an internal shading zone according to the dominant strong wind direction. This is used to introduce a wind speed attenuation coefficient for the internal shading zone, which has a lower actual wind speed, to calculate the design wind speed. This avoids the unreasonable wind resistance redundancy caused by using the same return period design wind speed for the entire photovoltaic power station.

[0095] Step 106: For the windward edge region, input the preset target return period and the extreme value sample into the target extreme value model to obtain the first design wind speed; and for the internal shading region, determine the wind speed attenuation coefficient corresponding to the internal shading region according to the preset wake attenuation function, and input the target return period, the extreme value sample and the wind speed attenuation coefficient into the target extreme value model to obtain the second design wind speed.

[0096] In step 106, the design wind speeds for the windward edge region and the internal shading region under the target return period are calculated separately. In one example, the target return period can be a typical return period such as 50 years or 100 years.

[0097] For the windward edge region, the preset target return period and extreme value samples are input into the target extreme value model to obtain the first design wind speed corresponding to the windward edge region. The windward edge region does not need to incorporate the wind speed attenuation coefficient α for calculation, or α can be set to 1 in the calculation.

[0098] For the internal obstruction zone, the wind speed attenuation coefficient α corresponding to the internal obstruction zone needs to be determined based on a preset wake attenuation function. The preset wake attenuation function can be the Jensen wake empirical formula. The target return period, extreme value samples, and wind speed attenuation coefficient α are input into the target extreme value model. In the model calculation, α is used as a product factor and directly multiplied by the model parameters of the target extreme value model to correct the model parameters, thereby obtaining the second design wind speed corresponding to the internal obstruction zone. The wind speed attenuation coefficient α is typically less than 1.

[0099] This application's embodiments introduce a wind speed attenuation coefficient for internal shading areas with relatively low actual wind speeds to calculate the design wind speed, avoiding unreasonable wind resistance redundancy caused by using the same return period design wind speed throughout the photovoltaic power station.

[0100] Optionally, determining the wind speed attenuation coefficient corresponding to the internal blocking area based on a preset wake attenuation function includes: The photovoltaic array spacing, vertical wind direction spacing, effective shading height of the photovoltaic array, wake diffusion coefficient, and number of shading rows in the internal shading area of ​​the photovoltaic power station are obtained. Based on the tilt angle of the photovoltaic module, the spacing of the photovoltaic array, the vertical wind direction column spacing, the effective shading height of the photovoltaic array, and the wake diffusion coefficient, the initial single-row attenuation coefficient of the internal shading area is determined using the wake attenuation function; The initial attenuation coefficient of a single row is superimposed and corrected based on the number of rows of shielding to obtain the wind speed attenuation coefficient corresponding to the internal shielding area.

[0101] This application's embodiments introduce the Jensen wake empirical formula adapted to the characteristics of photovoltaic arrays to calculate the wind speed attenuation coefficient corresponding to the internal shading area.

[0102] The following parameters are determined: the photovoltaic array spacing S along the prevailing strong wind direction, the vertical row spacing T perpendicular to the prevailing strong wind direction, the photovoltaic module tilt angle θ, the effective shading height H of the photovoltaic array, the wake diffusion coefficient k, and the number of shading rows in the internal shading zone. The wake diffusion coefficient k is selected based on the surface roughness of the photovoltaic power station. The number of shading rows is determined according to the number of rows of target photovoltaic modules. For example, if the target photovoltaic modules are the first to third rows on the outermost side facing the incoming flow, then the number of shading rows is 3.

[0103] Using the Jensen wake empirical formula adapted to the characteristics of photovoltaic arrays, the wake diffusion radius r(S) along the row spacing is calculated based on the photovoltaic array arrangement spacing S along the dominant strong wind direction, the vertical wind direction row spacing T perpendicular to the dominant strong wind direction, the photovoltaic module tilt angle θ, the effective shading height H of the photovoltaic array, and the wake diffusion coefficient k. The wake shading ratio C of a single row array is calculated based on the wake diffusion radius r(S), thereby obtaining the initial attenuation coefficient α1 of a single row in the internal shading area.

[0104] The initial attenuation coefficient of a single row is superimposed and corrected based on the number of rows of obstruction to obtain the wind speed attenuation coefficient corresponding to the inner obstruction area. For example, the power of the initial attenuation coefficient of a single row can be calculated for superposition and correction. If the number of rows of obstruction is 3, the cube of the initial attenuation coefficient of a single row is calculated as the final wind speed attenuation coefficient.

[0105] This application embodiment calculates the wind speed attenuation coefficient corresponding to the internal shading area and corrects the model parameters of the target extreme value model when calculating the second design wind speed of the internal shading area. For the rear photovoltaic modules that actually withstand lower wind speeds, unreasonable wind resistance redundancy is avoided.

[0106] Step 107: Convert the first design wind speed and the second design wind speed into the first average design wind speed and the second average design wind speed respectively.

[0107] In step 107, in order to solve the problem of unreasonable wind resistance redundancy caused by directly applying the maximum wind speed at a single point to the design of a large-area photovoltaic array in the photovoltaic power station in the traditional method, the first design wind speed and the second design wind speed are both reduced by the surface average wind speed and transformed into the first surface average design wind speed and the second surface average design wind speed.

[0108] Optionally, converting the first design wind speed and the second design wind speed into a first face-average design wind speed and a second face-average design wind speed respectively includes: Obtain the target length of the photovoltaic power station along the dominant strong wind direction and the target width perpendicular to the dominant strong wind direction, as well as the target relationship between the variance of the surface average wind speed and the variance of the single-point wind speed. The area reduction factor calculation formula for the photovoltaic power station is determined based on the target relationship, the target length, the target width, and the Davenport spatial coherence empirical function. Solve the formula for calculating the area reduction factor to obtain the area reduction factor of the photovoltaic power station; The product of the first design wind speed and the area reduction factor is taken as the first average design wind speed of the face, and the product of the second design wind speed and the area reduction factor is taken as the second average design wind speed of the face.

[0109] This application utilizes the Davenport spatial coherence empirical function to calculate the lateral and longitudinal correlation coefficients of the wind field under strong wind events in wind tunnel testing. Based on the total area occupied by the photovoltaic power station, the area reduction factor (ARF) of the photovoltaic power station is derived. The extreme wind speed at a single point (i.e., the first design wind speed and the second design wind speed) is then multiplied by the ARF to obtain the area-averaged design wind speed, thereby eliminating the overall load redundancy caused by the maximum wind speed at a single point in a long-span area.

[0110] Specifically, spatial coherence describes the correlation between wind speed fluctuations at different spatial locations during a strong wind event at the same moment. For large-area photovoltaic power plants, extreme gust fronts are moving systems with a finite spatial scale and will not reach peak wind speeds simultaneously at all locations across the entire site. Therefore, the maximum wind speed measured at a single point cannot directly represent the area-average maximum wind speed of the entire site, and spatial coherence reduction is needed to eliminate unreasonable wind resistance redundancy. Based on their relative position with wind direction, spatial correlation coefficients are divided into two categories: longitudinal correlation coefficients and lateral correlation coefficients, both of which jointly determine the spatial attenuation characteristics of the wind field.

[0111] The general Davenport spatial coherence empirical function is shown in the following formula (9). For any two points in space, the coherence coefficient (i.e., the square of the correlation coefficient) of their wind speed fluctuations is calculated using formula (9): (9) Wherein, γ is the spatial correlation coefficient of wind speed between two points (range 0~1, 1 indicates complete correlation, 0 indicates no correlation), Δx is the distance between two points along the dominant strong wind direction (longitudinal distance, unit: m), Δy is the distance between two points perpendicular to the dominant strong wind direction (lateral distance, unit: m), and z is the height above the ground (in this embodiment, the actual force center height of the photovoltaic module is taken). The longitudinal attenuation coefficient is dimensionless. L(z) is the transverse attenuation coefficient (dimensionless), and L(z) is the gust integral scale at height z (unit: m, representing the average spatial size of the gust vortex).

[0112] The gust integral scale L(z) is a core parameter determining the rate of spatial coherence decay, and its value increases with altitude. In one example, for the low-lying layer (1.5~5m) of a photovoltaic power station, the value of L(z) can be set as follows: if the photovoltaic power station is in a flat and open area (such as the Gobi Desert or grassland), then L(z) = 80-120m; if the photovoltaic power station is in a rural area (with a few trees and buildings), then L(z) = 60-80m; if the photovoltaic power station is in the suburbs of a city, then L(z) = 40-60m; if the photovoltaic power station is in a mountainous and complex terrain, then L(z) = 20-40m (mountainous terrain is fragmented, and the gust scale is smaller). In the embodiments of this application, L(z) will be dynamically valued according to the topography and surface roughness of the area where the photovoltaic power station is located, further improving the reduction accuracy.

[0113] In one embodiment, the longitudinal correlation coefficient (longitudinal refers to the direction along the prevailing strong wind) describes the degree of wind speed correlation between two measuring points arranged along the wind direction. As an example, when calculating the longitudinal correlation coefficient based on the Davenport spatial coherence empirical function, the longitudinal attenuation coefficient C is relatively slow due to the slower decay of wind speed correlation along the wind direction, as the scale of gust vortices along the wind direction is large. x The value is typically set to 10, so that, given a ground clearance of z = 2.5m and a gust integral scale L(z) = 100m, the longitudinal distance Δx and the longitudinal correlation coefficient γ can be calculated. x The relationship is shown in Table 2.

[0114] Table 2 Relationship between longitudinal distance and longitudinal correlation coefficient

[0115] In one embodiment, the lateral correlation coefficient (lateral refers to the direction perpendicular to the prevailing strong wind) describes the degree of wind speed correlation between two measuring points aligned perpendicular to the wind direction. As an example, when calculating the lateral correlation coefficient based on the Davenport spatial coherence empirical function, the lateral attenuation coefficient C is higher because the wind speed correlation in the vertical direction decays more rapidly, since the lateral scale of the gust vortex is smaller than its longitudinal scale. y The value is usually taken as 16, so that, given a ground height z = 2.5m and a gust integral scale L(z) = 100m, the lateral distance Δy and the lateral correlation coefficient γ can be calculated. y The relationship is shown in Table 3.

[0116] Table 3. Relationship between lateral distance and lateral correlation coefficient

[0117] In this embodiment, the area reduction factor is defined as the ratio of the maximum average wind speed over the entire photovoltaic power station to the maximum wind speed at a single point. Essentially, it is obtained by integrating the spatial correlation coefficient between point pairs across the entire power station to obtain the ratio of the variance of the average wind speed over the average wind speed to the variance of the wind speed at a single point, and then taking the square root to obtain the wind speed reduction factor. The detailed derivation process of the area reduction factor is as follows: For any region of area A, the variance of its surface-mean wind speed The variance σ of the single-point wind speed 2 The target relationship between the variance of the surface-average wind speed and the variance of the single-point wind speed is satisfied by the following formula (10): (10) Where γ is the wind speed correlation coefficient between the two points, Δx is the distance between the two points along the prevailing strong wind direction, and Δy is the distance between the two points perpendicular to the prevailing strong wind direction. and Let A be the coordinates of the first point within region A. and Let be the coordinates of the second point within region A.

[0118] Divide both sides of formula (10) by σ 2 Thus, we obtain formula (11): (11) Since extreme wind speed is proportional to the standard deviation of wind speed σ, specifically, in extreme value theory, the design wind speed UT for the return period is expressed as UT=μ+k·σ, where μ is the mean and k is the extreme value quantile coefficient. Therefore, the area reduction coefficient can be approximated by formula (12): (12) When applied to photovoltaic power plants, which are typically rectangular in layout, the target length L of the photovoltaic power plant along the prevailing strong wind direction is obtained. x and the target width L perpendicular to the prevailing strong wind direction y The total area of ​​the photovoltaic power station is A=L x ·L y Substituting the Davenport spatial coherence empirical function into the double integral of the above formula (12), we can obtain the formula (13) for calculating the area reduction factor of the rectangular area of ​​the photovoltaic power station: (13) Where x and y represent the coordinates of points within the region, The longitudinal attenuation coefficient is... denoted as the lateral attenuation coefficient, and L as the gust integral scale.

[0119] Since the integral of formula (13) has no analytical solution, the Gaussian numerical integration method is used to quickly solve formula (13) to obtain the area reduction coefficient of photovoltaic power station.

[0120] To facilitate engineering applications, this application also provides an approximate formula (14) based on a large number of numerical calculation results for calculating the area reduction factor of photovoltaic power stations, which is applicable to most photovoltaic power stations on flat terrain. The scope of application of formula (14) is photovoltaic power stations whose actual land area meets the requirement of 0.01km²≤A≤10km², which covers the vast majority of photovoltaic power station scales: (14) Where A is the actual area occupied by the station (unit: km²), and A0 is the reference area (which can be taken as 0.01 km², i.e. 1 hectare, corresponding to the equivalent area of ​​a single point extreme value).

[0121] In one example, based on the above derivation process, typical values ​​of the area reduction factor for photovoltaic power stations of different sizes on flat terrain are shown in Table 4 below. Table 4 shows typical values ​​for flat and open areas. Due to the smaller gust integral scale, the ARF value for mountainous terrain will be lower (the reduction ratio will be larger). In the embodiments of this application, this difference can be reflected by dynamically adjusting the gust integral scale during calculation.

[0122] Table 4 Typical values ​​of area reduction factor for photovoltaic power stations of different sizes on flat terrain

[0123] In one embodiment, for an area exceeding 1 km² 2 For extra-large stations, the station can be divided into multiple 1km sections. 2 The ARF of each of the left and right sub-regions is calculated separately to achieve a more refined partitioning design.

[0124] After obtaining the area reduction factor, the product of the first design wind speed and the area reduction factor is taken as the first average design wind speed, and the product of the second design wind speed and the area reduction factor is taken as the second average design wind speed.

[0125] In this embodiment, the design wind speeds corresponding to the windward edge area and the internal shading area need to be reduced by spatial coherence area. The first design wind speed and the second design wind speed are converted into the first average design wind speed and the second average design wind speed of the second face. This corrects the unreasonable wind resistance redundancy caused by directly applying the maximum wind speed at a single point to the design of a large-area photovoltaic array in the entire photovoltaic power station in the traditional method. Thus, while ensuring structural safety, the amount of steel used and the cost of the support structure are effectively controlled.

[0126] Step 108: The average design wind speed of the first face and the average design wind speed of the second face are used as the target design wind speed of the photovoltaic power station. The target design wind speed is used in the engineering design of the photovoltaic power station.

[0127] In step 108, the average design wind speed of the first face corresponding to the windward edge area and the average design wind speed of the second face corresponding to the internal shading area are taken as the target design wind speed of the photovoltaic power station. The target design wind speed is used for the engineering design of the photovoltaic power station.

[0128] This application converts the initial wind speed data into target wind speed data corresponding to the actual stress center height of the photovoltaic module. This ensures that the target wind speed data fully considers the roughness influence at the actual stress center height of the photovoltaic module, avoiding the errors caused by directly applying the standard 10m height power law formula. Based on the data sequence length of the target wind speed data, this application selects a suitable target extreme value sample extraction algorithm, thereby ensuring that effective extreme value samples can be obtained even for short-sequence wind measurement data. On this basis, based on the extracted extreme value samples, a target extreme value model is determined from at least two preset candidate extreme value models. This overcomes the limitations of traditional methods, which are limited to using a Type I extreme value distribution with fixed shape parameters, resulting in an inability to effectively fit the thick-tailed characteristics of wind speed distribution in areas with frequent strong winds, and the inflexibility of the P-III type distribution in handling the thick-tailed characteristics of such areas. This significantly improves the fitting accuracy of short-sequence data under thick-tailed distribution conditions. This application's embodiments divide the photovoltaic power station into a windward edge zone and an internal shading zone. For the internal shading zone, which experiences lower actual wind speeds, a wind speed attenuation coefficient is introduced to calculate the design wind speed. This avoids the unreasonable wind resistance redundancy caused by using the same return-time design wind speed for the entire photovoltaic power station. Simultaneously, both the first and second design wind speeds are reduced by surface-average wind speed, transforming them into first-surface-average design wind speeds and second-surface-average design wind speeds. This corrects the unreasonable wind resistance redundancy caused by directly applying the maximum wind speed at a single point to the design of a large-area photovoltaic array in the traditional method. Thus, while ensuring structural safety, the amount of steel used and the cost of the support structure are effectively controlled. Therefore, this application's embodiments can provide a more reliable design wind speed basis for the refined wind resistance design of photovoltaic power stations.

[0129] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of this application.

[0130] Based on the above embodiments, this embodiment also provides a design wind speed prediction device for photovoltaic power plants, which can be applied to electronic devices such as terminal equipment and servers.

[0131] Reference Figure 2The diagram shows a structural block diagram of an embodiment of a design wind speed prediction device for a photovoltaic power plant, which may specifically include the following modules: The wind speed data acquisition module 201 is used to acquire the actual stress center height and initial wind speed data of the photovoltaic modules of the photovoltaic power station. The wind speed data conversion module 202 is used to determine the target wind speed data corresponding to the actual stress center height of the photovoltaic module based on the initial wind speed data. The extreme value sample extraction module 203 is used to determine the target extreme value sample extraction algorithm based on the data sequence length of the target wind speed data, and to use the target extreme value sample extraction algorithm to extract the extreme value samples corresponding to the target wind speed data. The target model determination module 204 is used to determine a target extreme value model from at least two preset candidate extreme value models based on the extreme value samples. The site area division module 205 is used to determine the windward edge area and the internal shading area of ​​the photovoltaic power station; The wind speed calculation module 206 is designed to, for the windward edge area, input a preset target return period and the extreme value sample into the target extreme value model to obtain a first design wind speed; and, for the internal shading area, determine the wind speed attenuation coefficient corresponding to the internal shading area according to a preset wake attenuation function, and input the target return period, the extreme value sample and the wind speed attenuation coefficient into the target extreme value model to obtain a second design wind speed; Design wind speed conversion module 207 is used to convert the first design wind speed and the second design wind speed into the first average design wind speed and the second average design wind speed respectively. The design wind speed determination module 208 is used to take the average design wind speed of the first face and the average design wind speed of the second face as the target design wind speed of the photovoltaic power station, and the target design wind speed is used for the engineering design of the photovoltaic power station.

[0132] Optionally, the initial wind speed data includes meteorological station observation data, field measurement data, and mesoscale reanalysis data. The wind speed data conversion module 202 includes: The uniformity test submodule is used to perform a uniformity test on the initial wind speed data to obtain the tested initial wind speed data. The photovoltaic array data acquisition submodule is used to acquire the photovoltaic array data of the photovoltaic power station. The photovoltaic array data includes the average height of the photovoltaic array, the tilt angle of the photovoltaic modules, and the data of the ground cover corresponding to the photovoltaic power station. The windward area index determination submodule is used to determine the windward area index of the photovoltaic power station based on the tilt angle of the photovoltaic modules; The zero-plane displacement calculation submodule is used to calculate the zero-plane displacement of the photovoltaic power station based on the average height of the photovoltaic array and the windward surface area index. The background roughness determination submodule is used to determine the background roughness of the photovoltaic power station based on the surface cover data. An aerodynamic roughness calculation submodule is used to determine the aerodynamic roughness of the photovoltaic power station based on the windward surface area index and the background roughness. The wind speed data height conversion submodule is used to convert the verified initial wind speed data into the initial wind speed data corresponding to the actual force center height of the photovoltaic module based on the zero plane displacement, the aerodynamic roughness, and the actual force center height of the photovoltaic module. The data fusion submodule is used to fuse the initial wind speed data corresponding to the actual stress center height of the photovoltaic module to obtain the target wind speed data corresponding to the actual stress center height of the photovoltaic module.

[0133] Optionally, the extreme value sample extraction module 203 includes: The first algorithm determination submodule is used to determine the annual maximum value method as the target extreme value sample extraction algorithm if the data sequence length of the target wind speed data is greater than or equal to a preset sequence length threshold. The second algorithm determination submodule is used to determine the over-threshold peak method as the target extreme value sample extraction algorithm if the data sequence length of the target wind speed data is less than the sequence length threshold.

[0134] Optionally, the target wind speed data includes target wind speed data from multiple years, and the extreme value sample extraction module 203 includes: The first sample extraction submodule is used to determine the maximum wind speed extreme point in the target wind speed data for each year if the target extreme value sample extraction algorithm is the annual maximum value method, and to integrate the maximum wind speed extreme point in the target wind speed data for each year into the extreme value sample corresponding to the target wind speed data.

[0135] Optionally, the extreme value sample extraction module 203 includes: The extreme value threshold determination submodule is used to determine the extreme value threshold corresponding to the target wind speed data by using the average residual lifetime map method if the target extreme value sample extraction algorithm is the over-threshold peak method. The Storm Time Window Acquisition submodule is used to acquire storm time windows; The second sample extraction submodule is used to perform de-clustering processing on the target wind speed data according to the extreme value threshold and the storm time window to obtain the extreme value samples corresponding to the target wind speed data.

[0136] Optionally, the candidate extreme value model includes a generalized extreme value distribution model and a generalized Pareto distribution model, and the target model determination module 204 includes: The first parameter estimation submodule is used to determine the first model parameters and the first maximum likelihood function value corresponding to the generalized extreme value distribution model based on the extreme value samples using the maximum likelihood estimation method. The first score determination submodule is used to calculate the first KS test statistic, the first Akaike information criterion, and the first Bayesian information criterion corresponding to the generalized extreme value distribution model based on the first model parameters and the first maximum likelihood function value, and to calculate the first score corresponding to the generalized extreme value distribution model based on the first KS test statistic, the first Akaike information criterion, and the first Bayesian information criterion. The second parameter estimation submodule is used to determine the second model parameters and the second maximum likelihood function values ​​corresponding to the generalized Pareto distribution model based on the extreme value samples using the maximum likelihood estimation method. The second scoring determination submodule is used to calculate the second KS test statistic, the second Akaike information criterion, and the second Bayesian information criterion corresponding to the generalized Pareto distribution model based on the second model parameters and the second maximum likelihood function value, and to calculate the second score corresponding to the generalized Pareto distribution model based on the second KS test statistic, the second Akaike information criterion, and the second Bayesian information criterion. The first target model determination submodule is used to determine the target extreme value model if the first score is greater than the second score. The second target model determination submodule is used to determine the target extreme value model if the second score is greater than the first score; The third target model determination submodule is used to determine the target extreme value model if the first score is equal to the second score.

[0137] Optionally, the photovoltaic power station is equipped with at least one row of photovoltaic modules, and the station area division module 205 includes: The historical extreme data acquisition submodule is used to acquire historical extreme wind meteorological data corresponding to the photovoltaic power station. The historical extreme wind meteorological data includes historical wind direction and historical wind speed. The strong wind rose diagram determination submodule is used to draw a strong wind rose diagram based on the historical extreme wind meteorological data. The historical data determination submodule is used to determine the historical frequency of occurrence and the corresponding historical wind speed based on the strong wind rose diagram. The wind direction contribution determination submodule is used to determine the contribution of the historical wind direction based on the historical occurrence frequency and the historical wind speed corresponding to the historical wind direction. The dominant strong wind direction determination submodule is used to determine the historical wind direction with the highest contribution as the dominant strong wind direction; The target photovoltaic module determination submodule is used to determine the photovoltaic modules in a preset number of rows located in the direction of the prevailing strong wind as the target photovoltaic modules; The area division submodule is used to divide the area where the target photovoltaic module is located into the windward edge zone of the photovoltaic power station, and divide the area of ​​the photovoltaic power station other than the windward edge zone into the internal shading zone of the photovoltaic power station.

[0138] Optionally, the design wind speed calculation module 206 includes: The shading data acquisition submodule is used to acquire the photovoltaic array spacing, vertical wind direction column spacing, effective shading height of the photovoltaic array, wake diffusion coefficient, and number of shading rows in the internal shading area of ​​the photovoltaic power station. The initial attenuation coefficient determination submodule is used to determine the single-row initial attenuation coefficient of the internal shading area based on the photovoltaic module tilt angle, the photovoltaic array spacing, the vertical wind direction column spacing, the effective shading height of the photovoltaic array, and the wake diffusion coefficient, using the wake attenuation function. The wind speed attenuation coefficient determination submodule is used to superimpose and correct the initial attenuation coefficient of the single row according to the number of shading rows to obtain the wind speed attenuation coefficient corresponding to the internal shading area.

[0139] Optionally, the wind speed conversion module 207 includes: The data acquisition submodule is used to acquire the target length of the photovoltaic power station along the dominant strong wind direction and the target width perpendicular to the dominant strong wind direction, as well as the target relationship between the variance of the surface average wind speed and the variance of the single-point wind speed. The area reduction factor calculation formula determination submodule is used to determine the area reduction factor calculation formula of the photovoltaic power station based on the target relationship, the target length, the target width and the Davenport spatial coherence empirical function; The area reduction factor determination submodule is used to solve the area reduction factor calculation formula to obtain the area reduction factor of the photovoltaic power station; The area average wind speed reduction submodule is used to take the product of the first design wind speed and the area reduction coefficient as the first area average design wind speed, and to take the product of the second design wind speed and the area reduction coefficient as the second area average design wind speed.

[0140] This application also provides a non-volatile readable storage medium storing one or more modules (programs). When these modules are applied to a device, they enable the device to execute the instructions for the method steps in this application.

[0141] This application provides one or more machine-readable media storing instructions that, when executed by one or more processors, cause an electronic device to perform one or more of the methods described in the above embodiments. In this application, the electronic device includes various types of devices such as terminal devices and servers (clusters).

[0142] The embodiments of this disclosure can be implemented as an apparatus configured as desired using any suitable hardware, firmware, software, or any combination thereof, including electronic devices such as terminal devices, servers (clusters), etc. Figure 3 An exemplary apparatus 300 is schematically shown that can be used to implement the various embodiments described in this application.

[0143] In one embodiment, Figure 3 An exemplary device 300 is shown, which includes one or more processors 302, a control module (chipset) 304 coupled to at least one of the processors 302, a memory 306 coupled to the control module 304, a non-volatile memory (NVM) / storage device 308 coupled to the control module 304, one or more input / output devices 310 coupled to the control module 304, and a network interface 312 coupled to the control module 304.

[0144] Processor 302 may include one or more single-core or multi-core processors, and processor 302 may include any combination of general-purpose processors or special-purpose processors (e.g., graphics processors, application processors, baseband processors, etc.). In some embodiments, device 300 can serve as a terminal device, server (cluster), or other device as described in the embodiments of this application.

[0145] In some embodiments, the apparatus 300 may include one or more computer-readable media (e.g., memory 306 or NVM / storage device 308) having instructions 314 and one or more processors 302 that are combined with the one or more computer-readable media and configured to execute the instructions 314 to implement the module and thus perform the actions described in this disclosure.

[0146] In one embodiment, the control module 304 may include any suitable interface controller to provide any suitable interface to at least one of the processors 302 and / or any suitable device or component communicating with the control module 304.

[0147] The control module 304 may include a memory controller module to provide an interface to the memory 306. The memory controller module may be a hardware module, a software module, and / or a firmware module.

[0148] Memory 306 may be used, for example, to load and store data and / or instructions 314 for device 300. In one embodiment, memory 306 may include any suitable volatile memory, such as suitable DRAM. In some embodiments, memory 306 may include double data rate type quad synchronous dynamic random access memory (DDR4 SDRAM).

[0149] In one embodiment, the control module 304 may include one or more input / output controllers to provide an interface to the NVM / storage device 308 and (one or more) input / output devices 310.

[0150] For example, NVM / storage device 308 may be used to store data and / or instructions 314. NVM / storage device 308 may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (one or more) non-volatile storage devices (e.g., one or more hard disk drive (HDD), one or more optical disc (CD) drives, and / or one or more digital universal optical disc (DVD) drives).

[0151] NVM / storage device 308 may include storage resources that are physically part of a device on which device 300 is mounted, or that are accessible to the device but do not necessarily have to be part of the device. For example, NVM / storage device 308 may be accessed via a network via one or more input / output devices 310.

[0152] One or more input / output devices 310 may provide an interface for device 300 to communicate with any other suitable device. Input / output devices 310 may include communication components, audio components, sensor components, etc. A network interface 312 may provide an interface for device 300 to communicate via one or more networks. Device 300 may wirelessly communicate with one or more components of a wireless network according to any of one or more wireless network standards and / or protocols, such as accessing a wireless network based on a communication standard, such as WiFi, 2G, 3G, 4G, 5G, etc., or a combination thereof.

[0153] In one embodiment, at least one of the processors 302 may be logically packaged with one or more controllers (e.g., memory controller modules) of the control module 304. In one embodiment, at least one of the processors 302 may be logically packaged with one or more controllers of the control module 304 to form a system-in-package (SiP). In one embodiment, at least one of the processors 302 may be integrated with the logic of one or more controllers of the control module 304 on the same die. In one embodiment, at least one of the processors 302 may be integrated with the logic of one or more controllers of the control module 304 on the same die to form a system-on-a-chip (SoC).

[0154] In various embodiments, device 300 may be, but is not limited to, a server, desktop computing device, or mobile computing device (e.g., laptop computing device, handheld computing device, tablet computer, netbook, etc.). In various embodiments, device 300 may have more or fewer components and / or different architectures. For example, in some embodiments, device 300 includes one or more cameras, a keyboard, a liquid crystal display (LCD) screen (including a touchscreen display), a non-volatile memory port, multiple antennas, a graphics chip, an application-specific integrated circuit (ASIC), and a speaker.

[0155] The detection device can use a main control chip as a processor or control module, and sensor data, position information, etc. can be stored in a memory or NVM / storage device. The sensor group can be used as an input / output device, and the communication interface can include a network interface.

[0156] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0157] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0158] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should 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 photovoltaic power plant design wind speed forecasting terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable photovoltaic power plant design wind speed forecasting terminal device, 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.

[0159] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable photovoltaic power plant design wind speed prediction terminal device to operate in a specific manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including an instruction means, which is implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0160] These computer program instructions can also be loaded onto a computer or other programmable photovoltaic power plant design wind speed forecasting terminal device, causing a series of operational steps to be executed on the computer or other programmable terminal device to produce computer-implemented processing, thereby providing instructions that execute on the computer or other programmable terminal device 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.

[0161] Although preferred embodiments of the present application 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 the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0162] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0163] The foregoing has provided a detailed description of a method and apparatus for predicting the design wind speed of a photovoltaic power station, an electronic device, and a storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and its core ideas. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for predicting the design wind speed of a photovoltaic power station, characterized in that, The method includes: Obtain the actual stress center height and initial wind speed data of the photovoltaic modules of the photovoltaic power station; Determine the target wind speed data corresponding to the actual stress center height of the photovoltaic module based on the initial wind speed data; Based on the data sequence length of the target wind speed data, a target extreme value sample extraction algorithm is determined, and the target extreme value sample extraction algorithm is used to extract the extreme value samples corresponding to the target wind speed data; Based on the extreme value samples, a target extreme value model is determined from at least two preset candidate extreme value models; Determine the windward edge zone and the internal shading zone of the photovoltaic power station; For the windward edge region, the preset target return period and the extreme value sample are input into the target extreme value model to obtain the first design wind speed; and for the internal shading region, the wind speed attenuation coefficient corresponding to the internal shading region is determined according to the preset wake attenuation function, and the target return period, the extreme value sample and the wind speed attenuation coefficient are input into the target extreme value model to obtain the second design wind speed. The first design wind speed and the second design wind speed are respectively converted into the first face average design wind speed and the second face average design wind speed; The average design wind speed of the first face and the average design wind speed of the second face are used as the target design wind speed of the photovoltaic power station, and the target design wind speed is used in the engineering design of the photovoltaic power station.

2. The method according to claim 1, characterized in that, The initial wind speed data includes meteorological station observation data, field measurement data, and mesoscale reanalysis data. Determining the target wind speed data corresponding to the actual stress center height of the photovoltaic module based on the initial wind speed data includes: The initial wind speed data is subjected to a uniformity test to obtain the tested initial wind speed data; Acquire photovoltaic array data of the photovoltaic power station, including average height of the photovoltaic array, tilt angle of photovoltaic modules, and data on the land cover corresponding to the photovoltaic power station; The windward area index of the photovoltaic power station is determined based on the tilt angle of the photovoltaic modules; The zero-plane displacement of the photovoltaic power station is calculated based on the average height of the photovoltaic array and the windward surface area index. The background roughness of the photovoltaic power station is determined based on the surface cover data; The aerodynamic roughness of the photovoltaic power station is determined based on the windward surface area index and the background roughness. Based on the zero plane displacement, the aerodynamic roughness, and the actual stress center height of the photovoltaic module, the verified initial wind speed data is converted into the initial wind speed data corresponding to the actual stress center height of the photovoltaic module; The initial wind speed data corresponding to the actual stress center height of the photovoltaic module are fused to obtain the target wind speed data corresponding to the actual stress center height of the photovoltaic module.

3. The method according to claim 1, characterized in that, The algorithm for determining the target extreme value sample extraction based on the data sequence length of the target wind speed data includes: If the data sequence length of the target wind speed data is greater than or equal to a preset sequence length threshold, then the annual maximum value method is determined as the target extreme value sample extraction algorithm; If the data sequence length of the target wind speed data is less than the sequence length threshold, then the peak value exceeding the threshold method is determined as the target extreme value sample extraction algorithm.

4. The method according to claim 3, characterized in that, The target wind speed data includes target wind speed data from multiple years. The step of extracting extreme value samples corresponding to the target wind speed data using the target extreme value sample extraction algorithm includes: If the target extreme value sample extraction algorithm is the annual maximum value method, then the maximum wind speed extreme point in the target wind speed data for each year is determined, and the maximum wind speed extreme point in the target wind speed data for each year is integrated into the extreme value sample corresponding to the target wind speed data.

5. The method according to claim 3, characterized in that, The step of extracting extreme value samples corresponding to the target wind speed data using the target extreme value sample extraction algorithm includes: If the target extreme value sample extraction algorithm is the over-threshold peak method, then the average residual lifetime map method is used to determine the extreme value threshold corresponding to the target wind speed data; Get the storm time window; Based on the extreme value threshold and the storm time window, the target wind speed data is de-clustered to obtain the extreme value samples corresponding to the target wind speed data.

6. The method according to claim 1, characterized in that, The candidate extreme value models include a generalized extreme value distribution model and a generalized Pareto distribution model. The step of determining the target extreme value model from at least two pre-defined candidate extreme value models based on the extreme value samples includes: The maximum likelihood estimation method is used to determine the first model parameters and the first maximum likelihood function value corresponding to the generalized extreme value distribution model based on the extreme value samples. Calculate the first KS test statistic, the first Akaike information criterion, and the first Bayesian information criterion corresponding to the generalized extreme value distribution model based on the first model parameters and the first maximum likelihood function value; and calculate the first score corresponding to the generalized extreme value distribution model based on the first KS test statistic, the first Akaike information criterion, and the first Bayesian information criterion. Using the maximum likelihood estimation method, the second model parameters and the second maximum likelihood function values ​​corresponding to the generalized Pareto distribution model are determined based on the extreme value samples. The second KS test statistic, the second Akaike information criterion, and the second Bayesian information criterion corresponding to the generalized Pareto distribution model are calculated based on the second model parameters and the second maximum likelihood function value. The second score corresponding to the generalized Pareto distribution model is then calculated based on the second KS test statistic, the second Akaike information criterion, and the second Bayesian information criterion. If the first score is greater than the second score, then the generalized extreme value distribution model is used as the target extreme value model; If the second score is greater than the first score, then the generalized Pareto distribution model is used as the target extreme value model; If the first score equals the second score, then the generalized extreme value distribution model or the generalized Pareto distribution model shall be used as the target extreme value model.

7. The method according to claim 2, characterized in that, The photovoltaic power station is equipped with at least one row of photovoltaic modules. Determining the windward edge area and the internal shading area of ​​the photovoltaic power station includes: Obtain historical extreme wind meteorological data corresponding to the photovoltaic power station, including historical wind direction and historical wind speed; A strong wind rose diagram was drawn based on the aforementioned historical extreme wind meteorological data. Based on the strong wind rose diagram, determine the historical frequency of occurrence and the corresponding historical wind speed for the historical wind direction; The contribution of the historical wind direction is determined based on the historical frequency of occurrence and the historical wind speed. The historical wind direction with the highest contribution is identified as the dominant strong wind direction; A predetermined number of photovoltaic modules located in the direction of the prevailing strong wind are identified as target photovoltaic modules. The area where the target photovoltaic module is located is designated as the windward edge zone of the photovoltaic power station, and the area of ​​the photovoltaic power station other than the windward edge zone is designated as the internal shading zone of the photovoltaic power station.

8. The method according to claim 7, characterized in that, The step of determining the wind speed attenuation coefficient corresponding to the internal shielding area according to the preset wake attenuation function includes: Obtain the photovoltaic array spacing, vertical wind direction row spacing, effective shading height of the photovoltaic array, wake diffusion coefficient, and number of shading rows in the internal shading area of ​​the photovoltaic power station; Based on the tilt angle of the photovoltaic module, the spacing of the photovoltaic array, the vertical wind direction column spacing, the effective shading height of the photovoltaic array, and the wake diffusion coefficient, the initial single-row attenuation coefficient of the internal shading area is determined using the wake attenuation function; The initial attenuation coefficient of a single row is superimposed and corrected based on the number of rows of shielding to obtain the wind speed attenuation coefficient corresponding to the inner shielding area.

9. The method according to claim 7, characterized in that, The step of converting the first design wind speed and the second design wind speed into a first face average design wind speed and a second face average design wind speed, respectively, includes: Obtain the target length of the photovoltaic power station along the dominant strong wind direction and the target width perpendicular to the dominant strong wind direction, as well as the target relationship between the variance of the surface average wind speed and the variance of the single-point wind speed. The area reduction factor calculation formula for the photovoltaic power station is determined based on the target relationship, the target length, the target width, and the Davenport spatial coherence empirical function. Solve the formula for calculating the area reduction factor to obtain the area reduction factor of the photovoltaic power station; The product of the first design wind speed and the area reduction factor is taken as the first average design wind speed of the face, and the product of the second design wind speed and the area reduction factor is taken as the second average design wind speed of the face.

10. A design wind speed prediction device for a photovoltaic power station, characterized in that, The device includes: The wind speed data acquisition module is used to acquire the actual stress center height and initial wind speed data of the photovoltaic modules of the photovoltaic power station. The wind speed data conversion module is used to determine the target wind speed data corresponding to the actual stress center height of the photovoltaic module based on the initial wind speed data. The extreme value sample extraction module is used to determine the target extreme value sample extraction algorithm based on the data sequence length of the target wind speed data, and to use the target extreme value sample extraction algorithm to extract the extreme value samples corresponding to the target wind speed data. The target model determination module is used to determine a target extreme value model from at least two preset candidate extreme value models based on the extreme value samples. The site area division module is used to determine the windward edge area and the internal shading area of ​​the photovoltaic power station; The wind speed calculation module is designed to obtain a first design wind speed by inputting a preset target return period and the extreme value sample into the target extreme value model for the windward edge area; and to obtain a second design wind speed by inputting the target return period, the extreme value sample, and the wind speed attenuation coefficient into the target extreme value model for the internal shading area according to a preset wake attenuation function. The wind speed conversion module is designed to convert the first design wind speed and the second design wind speed into the first average design wind speed and the second average design wind speed, respectively. The design wind speed determination module is used to take the average design wind speed of the first face and the average design wind speed of the second face as the target design wind speed of the photovoltaic power station, and the target design wind speed is used for the engineering design of the photovoltaic power station.

11. An electronic device, characterized in that, include: processor; and A memory having executable code stored thereon, which, when executed, causes the processor to perform the design wind speed prediction method for a photovoltaic power plant as described in any one of claims 1-9.

12. A machine-readable medium having executable code stored thereon, which, when executed, causes a processor to perform the design wind speed prediction method for a photovoltaic power plant as claimed in any one of claims 1-9.

Images