A design method for refined simulation of wind load based on photovoltaic array

Abstract

The present application relates to the technical field of photovoltaic structure wind engineering, in particular to a fine simulation design method for wind load based on photovoltaic array, comprising: obtaining terrain elevation data and ground roughness data, combining photovoltaic array layout drawing to construct a three-dimensional terrain grid model, cutting and extracting the inflow wind speed profile and the turbulence intensity profile data of the array area through Boolean operation, calibrating the aerodynamic interference factor and the wind pressure coefficient correction coefficient suitable for the site, reconstructing the high-resolution calculation grid containing the component gap and the support details, loading the two extreme wind field boundary conditions of typhoon and non-typhoon to carry out transient numerical simulation, collecting dynamic wind pressure time history data, extracting extreme wind pressure distribution atlas and wind vibration response amplification coefficient, importing the structure design checking module to automatically generate support rod section selection suggestion and connector spacing configuration parameters. The method can realize the site accurate adaptation of wind load simulation and the automatic output of support system design parameters.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic structure wind engineering technology, and in particular to a refined simulation design method for wind loads on photovoltaic arrays. Background Technology

[0002] The existing photovoltaic array wind load design uses the general wind parameters given in the standard for calculation, without combining the site topographic elevation data, surface roughness data and photovoltaic array layout drawings to construct a three-dimensional terrain mesh model, without performing Boolean operation cutting processing on the three-dimensional geometric solid model, and without being able to extract the specific incoming wind speed profile data and turbulence intensity profile data of the photovoltaic array location area. The wind load parameter calibration process is not based on the actual wind field data of the site, and the aerodynamic interference factor and wind pressure coefficient correction coefficient used are all general values ​​that do not match the actual wind field conditions of the photovoltaic array site.

[0003] Existing wind load simulation technologies mostly employ steady-state simulation modes, setting only a single wind field condition as the boundary condition. They fail to distinguish between typhoon and non-typhoon extreme wind field conditions, making it impossible to conduct transient wind load numerical simulations, collect dynamic wind pressure time history data on the photovoltaic array surface, and obtain extreme wind pressure distribution maps and wind vibration response amplification factors under different wind direction angles through statistical analysis. The selection of photovoltaic array support system member cross-sections and connector spacing configurations rely on manual experience, failing to achieve automatic generation of design parameters based on refined wind load simulation results. This invention requires site-adaptive wind load parameter calibration, transient wind load simulation under dual extreme wind fields, and automatic generation of support system design configuration parameters based on wind-induced response parameters. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a refined simulation design method for wind load on photovoltaic arrays.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a refined simulation design method for wind load on photovoltaic arrays, comprising: Topographic elevation data and surface roughness data were acquired, and a three-dimensional topographic mesh model was constructed in conjunction with the layout drawings of the photovoltaic array. Boolean operations were performed on the three-dimensional geometric solid model to cut it, and the incoming wind speed profile data and turbulence intensity profile data of the area where the photovoltaic array is located were extracted. The incoming wind speed profile data and turbulence intensity profile data are input into the wind load parameter calibration module to generate a set of aerodynamic interference factors and wind pressure coefficient correction coefficients applicable to the site conditions of the voltage array. Based on the aerodynamic interference factor and wind pressure coefficient correction coefficient, the three-dimensional geometric solid model of the photovoltaic array is reconstructed into a mesh to generate a high-resolution computational mesh that includes the module gaps and support details. Based on the high-resolution computing grid, boundary conditions for two extreme wind fields, namely typhoon and non-typhoon conditions, were loaded, and transient wind load numerical simulations were run to collect dynamic wind pressure time history data on the surface of the photovoltaic array. Statistical analysis was performed on the dynamic wind pressure time history data to extract the extreme wind pressure distribution map and wind vibration response amplification factor of the photovoltaic array under different wind direction angles; The extreme wind pressure distribution map and wind vibration response amplification factor are imported into the structural design verification module to automatically generate recommendations for the selection of rod cross sections and configuration parameters for the spacing of connectors in the photovoltaic array support system.

[0006] As a further aspect of the present invention, topographic elevation data and surface roughness data are obtained, and a three-dimensional topographic mesh model is constructed in conjunction with the layout drawings of the photovoltaic array. Boolean operations are performed on the three-dimensional geometric solid model to cut it, and incoming wind speed profile data and turbulence intensity profile data of the area where the photovoltaic array is located are extracted, including: Obtain the topographic elevation data and surface roughness data of the site where the photovoltaic array is located, and construct a three-dimensional terrain mesh model in combination with the layout drawings of the photovoltaic array; Import the three-dimensional geometric solid model of the photovoltaic module into the three-dimensional terrain mesh model, and perform Boolean operation to cut the three-dimensional geometric solid model so that its boundary fits the undulating surface of the three-dimensional terrain mesh model; The steady-state wind field initialization simulation was performed on the fitted three-dimensional terrain mesh model using computational fluid dynamics software, and the incoming wind speed profile data and turbulence intensity profile data of the area where the photovoltaic array is located were extracted. The process of acquiring terrain elevation and surface roughness data of the site where the photovoltaic array is located, and constructing a three-dimensional terrain mesh model in conjunction with the layout drawings of the photovoltaic array, includes: Digital elevation model data within the site selection area of ​​the photovoltaic array is retrieved from the geographic information system database and then interpolated and converted into regular terrain raster data. The vegetation cover type and building distribution around the photovoltaic array were surveyed and recorded on-site, and the surface roughness length parameter was quantified and generated. Read the layout drawings of the photovoltaic array and identify the row spacing, column spacing, and tilt and azimuth parameters of each module; The terrain raster data, the surface roughness length parameter, and the array arrangement parameters are imported into 3D modeling software to generate a base surface with terrain undulation features. The array layout is projected onto the base surface according to the array arrangement parameters to initially construct the three-dimensional terrain mesh model.

[0007] As a further aspect of the present invention, a three-dimensional geometric solid model of a photovoltaic module is imported into the three-dimensional terrain mesh model, and Boolean operations are performed on the three-dimensional geometric solid model to cut it, including: Construct a cuboid solid model with the same dimensions as the actual photovoltaic module, and set a small tilt angle on the long side of the cuboid solid model to simulate the installation slope; The cuboid solid model is copied in batches to the designated location of the three-dimensional terrain mesh model according to the array arrangement parameters; Traverse the interface between each cuboid solid model and the terrain mesh below, and calculate the cutting contour line at the interface. Using the cutting contour line, perform Boolean difference operation on the corresponding cuboid solid model to remove the model part embedded in the terrain and retain the exposed part that fits the terrain. All entity models processed by Boolean operations are merged into a single photovoltaic array assembly model.

[0008] As a further aspect of the present invention, a steady-state wind field initialization simulation is performed on the fitted three-dimensional terrain mesh model using computational fluid dynamics software, extracting incoming wind speed profile data and turbulence intensity profile data for the area where the photovoltaic array is located, including: A virtual wind tunnel inlet boundary is set on the windward side of the three-dimensional terrain mesh model, and a uniform incoming flow velocity is given. Set symmetrical or sliding wall boundaries on the top and sides of the model, and set non-slip wall boundaries on the leeward side and the ground. Select a turbulence model suitable for high Reynolds number flows and set the key parameters of the turbulence model, including the type of wall function and the analytical requirements of the near-wall mesh; Run steady-state calculations until the flow field residuals converge and the velocity changes at the monitoring points tend to stabilize; Several characteristic measurement points are selected within the height range of the photovoltaic array, and the average wind speed and turbulent kinetic energy data at each measurement point are recorded. The turbulent kinetic energy data is then converted into the turbulence intensity profile data.

[0009] As a further aspect of the present invention, the incoming wind speed profile data and turbulence intensity profile data are input into the wind load parameter calibration module to generate a set of aerodynamic interference factors and wind pressure coefficient correction coefficients suitable for the site conditions of the voltage array, including: The wind tunnel test database inside the wind load parameter calibration module is retrieved. The wind tunnel test database contains the aerodynamic interference factor reference value and wind pressure coefficient reference value of the photovoltaic array under different arrangement methods. The incoming wind speed profile data and the turbulence intensity profile data are used as environmental variables. Interpolation queries are performed in the wind tunnel test database to find the benchmark value corresponding to the closest environmental conditions. Based on the measured surface roughness length parameters of the current site and the actual spacing parameters of the photovoltaic array, the queried benchmark values ​​are linearly corrected. The corrected baseline values ​​are combined to form the aerodynamic interference factor and the wind pressure coefficient correction factor.

[0010] As a further aspect of the present invention, based on the aerodynamic interference factor and wind pressure coefficient correction coefficient, the three-dimensional geometric solid model of the photovoltaic array is reconstructed into a mesh to generate a high-resolution computational mesh containing details of component gaps and support structure, including: The aerodynamic interference factor is read, and the mesh density level at the component gap is determined according to the magnitude of the aerodynamic interference factor. The larger the interference factor, the higher the density level. Read the wind pressure coefficient correction factor, and adjust the mesh refinement at the edges and corners of the component according to the magnitude of the wind pressure coefficient correction factor. The larger the correction factor, the higher the refinement. Perform curvature analysis on all surfaces in the photovoltaic array assembly model and automatically insert additional mesh nodes in areas with drastic curvature changes. A hybrid meshing strategy is adopted, using structured quadrilateral meshes for the planar areas of the components and unstructured tetrahedral meshes for the connection between the support and the components; Check the quality parameters of the mesh, remove mesh cells with excessive distortion, and complete the generation of the high-resolution computational mesh.

[0011] As a further aspect of the present invention, boundary conditions for two extreme wind fields—typhoon and non-typhoon—are loaded onto the high-resolution computational grid, and transient wind load numerical simulations are run to collect dynamic wind pressure time-history data on the photovoltaic array surface, including: For typhoon conditions, an average wind speed profile with exponential decay characteristics and a high-intensity turbulence pulse signal are loaded onto the virtual wind tunnel inlet boundary. For non-typhoon conditions, an average wind speed profile with logarithmic law characteristics and a low-intensity turbulence pulse signal are loaded onto the virtual wind tunnel inlet boundary. Large eddy simulation mode was activated under both operating conditions to capture transient vortex shedding phenomena in the flow field. Several virtual pressure monitoring points are arranged on the surface of the photovoltaic array, and the pressure value change of each monitoring point within the simulation time step is recorded; The pressure data from each monitoring point are organized according to the time series to form the dynamic wind pressure time history data.

[0012] As a further aspect of the present invention, statistical analysis is performed on the dynamic wind pressure time history data to extract the extreme wind pressure distribution map and wind vibration response amplification coefficient of the photovoltaic array under different wind direction angles, including: The power spectral density of the wind pressure signal is obtained by performing a fast Fourier transform on the dynamic wind pressure time history data. The dominant frequency component is identified in the power spectral density, and the dominant frequency is compared with the first-order natural frequency of the photovoltaic array to be designed to calculate the frequency ratio. The amplification factor of the wind vibration response is calculated using random vibration theory based on the frequency ratio and damping ratio. By iterating through the simulation results for all wind direction angles, the peak positive pressure and peak negative pressure corresponding to each wind direction angle are found. The peak positive pressure and peak negative pressure corresponding to all wind direction angles are mapped onto the surface of the three-dimensional geometric solid model of the photovoltaic array, and the extreme wind pressure distribution map is drawn.

[0013] As a further aspect of the present invention, the extreme wind pressure distribution map and wind vibration response amplification factor are imported into the structural design verification module to automatically generate recommendations for the selection of member cross-sections and configuration parameters for the spacing of connectors in the photovoltaic array support system, including: Read the extreme wind pressure distribution map, multiply the wind pressure value on the map by the wind vibration response amplification factor, and obtain the design wind load value; The design wind load value is used as the input load and applied to the structural finite element model of the photovoltaic array. In the structural design verification module, the ratio of the material's yield strength to its allowable stress is set, and the axial force and bending moment of the members are automatically calculated. Based on the calculation results, an iterative search is performed in the member specification library to select the minimum cross-sectional specification that meets the strength and stiffness requirements as the member cross-section selection suggestion; Simultaneously, based on the distribution of wind load, the arrangement density of the support beams is optimized, and the spacing configuration parameters of the connectors are generated.

[0014] As a further aspect of the present invention, it also includes: Based on the recommended selection of the rod cross-section and the configuration parameters of the connector spacing, the three-dimensional geometric solid model of the photovoltaic array is structurally reinforced and modeled, and a modal analysis and static equilibrium verification are performed. Output reinforced modeling and verification photovoltaic array design drawings and wind load simulation analysis report; Based on the recommended selection of the rod cross-section and the configuration parameters of the connector spacing, a structural reinforcement model of the three-dimensional geometric solid model of the photovoltaic array is performed, followed by a modal analysis and static equilibrium verification, including: In the three-dimensional geometric solid model of the photovoltaic array, the original rod solids are replaced with solids having the specifications specified in the rod cross-section selection recommendations; Based on the adjusted connector spacing configuration parameters, rearrange the positions of bolts or welded nodes on the support beam; A free vibration modal analysis was performed on the reinforced structural model to extract its first few natural frequencies and mode shapes. Apply standard static loads to the reinforced structural model, perform static equilibrium solutions, and check for any abnormal displacements or stress concentrations. If no problems are found in modal analysis and static equilibrium verification, the current reinforcement modeling state will be confirmed as the final design scheme.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: A three-dimensional terrain mesh model is constructed based on topographic elevation data, surface roughness data, and photovoltaic array layout drawings. Boolean operations are performed on the three-dimensional geometric model to extract incoming wind speed and turbulence intensity profiles for the photovoltaic array area. This wind field profile data is then input into a wind load parameter calibration module to generate aerodynamic interference factors and wind pressure coefficient correction coefficients suitable for the site conditions of the photovoltaic array. The three-dimensional terrain mesh model closely matches the actual shape of the site and array layout. The Boolean operation cutting method accurately delineates the wind field parameter extraction range for the array area. The aerodynamic interference factors and wind pressure coefficient correction coefficients form a one-to-one correspondence with the site's topography and surface features. The wind field profile data extraction process is limited to the target area where the photovoltaic array is located.

[0016] Based on a high-resolution computational grid that includes details of component gaps and support structures, boundary conditions for both typhoon and non-typhoon extreme wind fields are applied. Transient wind load numerical simulations are run separately, collecting dynamic wind pressure time-history data on the photovoltaic array surface. Statistical analysis extracts extreme wind pressure distribution maps and wind-induced vibration response amplification factors at different wind angles. These parameters are then imported into the structural design verification module to generate recommendations for member cross-section selection and connector spacing configuration parameters for the photovoltaic array support system. The transient numerical simulation captures the dynamic characteristics of wind load changes over time. The dual-extreme wind field boundary conditions cover different types of extreme climate scenarios. The extreme wind pressure distribution maps and wind-induced vibration response amplification factors correspond to the wind-induced response states of the photovoltaic array at multiple wind angles. The structural design verification module directly generates the configuration parameters for the support system's members and connectors. Attached Figure Description

[0017] Figure 1 This is a flowchart of a refined simulation design method for wind load based on photovoltaic arrays, as described in this invention. Figure 2A flowchart for importing a solid model and performing Boolean operations to cut it; Figure 3 A flowchart for generating the aerodynamic interference factor and wind pressure coefficient correction factor; Figure 4 Comparison of wind speed profiles in the wind field of photovoltaic arrays; Figure 5 This is a map showing the distribution of extreme wind pressure. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] See Figure 1 The process involves acquiring topographic elevation and surface roughness data of the photovoltaic array site, constructing a 3D topographic mesh model based on the photovoltaic array layout drawings, and performing Boolean operations on the 3D geometric solid model of the photovoltaic modules to fit the terrain. Incoming wind speed and turbulence intensity profiles for the photovoltaic array area are then extracted. These extracted data are input into a wind load parameter calibration module to generate a set of aerodynamic interference factors and wind pressure coefficient correction coefficients suitable for the site conditions. Based on these coefficients, the 3D geometric solid model of the photovoltaic array is reconstructed to generate a high-resolution computational mesh that includes module gaps and support details. On this mesh, boundary conditions for both typhoon and non-typhoon extreme wind fields are applied, and transient wind load numerical simulations are run to collect dynamic wind pressure time history data on the photovoltaic array surface. Statistical analysis of the collected dynamic wind pressure time history data is performed to extract extreme wind pressure distribution maps and wind vibration response amplification coefficients for the photovoltaic array under different wind direction angles. The extreme wind pressure distribution map and wind vibration response amplification factor are imported into the structural design verification module to automatically generate recommendations for the selection of rod cross sections and configuration parameters for the spacing of connectors in the photovoltaic array support system.

[0021] In one embodiment of the present invention, digital elevation model data within the site selection area of ​​the photovoltaic array is retrieved from a geographic information system database and interpolated into regular terrain raster data. Vegetation cover type and building distribution around the photovoltaic array are recorded through on-site surveys, and surface roughness length parameters are quantified. The layout drawings of the photovoltaic array are read to identify the row spacing, column spacing, and tilt and azimuth angles of each module. The aforementioned terrain raster data, surface roughness length parameters, and array layout parameters are imported into 3D modeling software to generate a base surface with terrain undulations. The array layout is projected onto this base surface according to the array layout parameters to initially construct a 3D terrain mesh model. Next, the 3D geometric solid model of the photovoltaic module is imported into the 3D terrain mesh model, and Boolean operations are performed on the 3D geometric solid model to cut its boundaries to fit the undulating surface of the 3D terrain mesh model. Computational fluid dynamics software is used to perform a steady-state wind field initialization simulation on the fitted 3D terrain mesh model to extract incoming wind speed profile data and turbulence intensity profile data for the area where the photovoltaic array is located.

[0022] In specific implementation, the method involves a photovoltaic array site located at 30 degrees north latitude and 120 degrees east longitude. Digital elevation model (DEM) data within the selected photovoltaic array site area is retrieved from a geographic information system (GIS) database. This DEM data has a resolution of 5 meters and is interpolated into regular terrain raster data with a grid spacing of 1 meter. In some embodiments, the DEM data originates from publicly available satellite remote sensing data or airborne lidar scanning data. The elevation difference between the two data sources is less than 0.5 meters in flat areas, but can reach 2 meters in steep slope areas. The data source with the smaller difference is selected for subsequent modeling. On-site surveys are conducted and the vegetation cover type and building distribution around the photovoltaic array are recorded. Vegetation cover types include low shrubs and tall grass, and building distribution includes the height and outline of nearby warehouses. A surface roughness length parameter is quantified. Optionally, the quantification of the surface roughness length parameter is calculated based on vegetation height and coverage using the following formula:

[0023] in: This represents the length parameter of surface roughness, in meters. This indicates the average height of the vegetation, in meters. This represents vegetation coverage, with a value ranging from 0 to 1. and These are empirical coefficients, with values ​​of 0.15 and 2.0 respectively. In practice, the average height of low shrubs recorded in the field survey was 0.8 meters, with a coverage rate of 0.6, while the average height of tall grass was 0.5 meters, with a coverage rate of 0.3. The surface roughness length parameter for the low shrub area was calculated to be 0.12 meters, and for the tall grass area, it was 0.075 meters. For the building area, a fixed value of 1.0 meter was directly assigned as the surface roughness length parameter. The photovoltaic array layout drawing, in CAD format, was read, revealing a row spacing of 5 meters, a column spacing of 3 meters, and a tilt angle of 20 degrees and an azimuth angle of due south for each module. It can be understood that the row spacing parameter defines the center-to-center distance between rows of photovoltaic modules, the column spacing parameter defines the center-to-center distance between columns of photovoltaic modules, the tilt angle parameter defines the angle between the module plane and the horizontal plane, and the azimuth angle parameter defines the projection direction of the module plane normal onto the horizontal plane. The terrain raster data, surface roughness length parameters, and array layout parameters are imported into the 3D modeling software, Rhino, to generate a base surface with terrain undulation features. The grid resolution of the base surface is consistent with that of the terrain raster data.

[0024] In some embodiments, the base surface is generated by fitting a non-uniform rational B-spline surface or by triangulation. The maximum elevation error of the 3D terrain mesh model generated by both methods is less than 0.1 meters in the slope variation area. An array layout projection is performed on the base surface according to the array arrangement parameters to initially construct a 3D terrain mesh model. The 3D terrain mesh model includes terrain surface and component position information. In a specific implementation, the array layout projection is achieved by mapping the coordinates of the component center point to the corresponding elevation point on the base surface. The component position information includes the 3D coordinates of the center point of each component and the direction of the normal vector. A 3D geometric solid model of the photovoltaic module is imported into the 3D terrain mesh model. The 3D geometric solid model of the photovoltaic module is a cuboid with dimensions of 1.6 meters by 0.9 meters by 0.05 meters. Boolean operations are performed on the 3D geometric solid model to cut its boundaries to fit the undulating surface of the 3D terrain mesh model. It can be understood that the Boolean operation cutting is achieved by calculating the intersection of the component model and the terrain mesh and removing the embedded parts. After cutting, the component model retains the portion exposed on the terrain. Steady-state wind field initialization simulation was performed on the fitted 3D terrain mesh model using computational fluid dynamics software, ANSYS Fluent. Incoming wind speed profiles and turbulence intensity profiles for the photovoltaic array area were extracted. In the specific implementation, the steady-state wind field initialization simulation was set with an inlet wind speed of 10 m / s, and the standard k-epsilon model was selected as the turbulence model. The simulation domain size was 5 times the terrain area. Measuring points were set at 1-meter heights above the centerline of the photovoltaic array, and the average wind speed and turbulent kinetic energy were recorded at each point. Turbulence intensity profile data were calculated from turbulent kinetic energy and average wind speed. The turbulence intensity profile data showed that the turbulence intensity was 15% at a height of 10 meters and 25% at a height of 2 meters. Optionally, the extraction of incoming wind speed profiles and turbulence intensity profile data was also validated by comparing the results of different turbulence models, such as comparing the output differences between the standard k-epsilon model and the SST k-omega model. When the difference was less than 5%, the result of the standard k-epsilon model was used.

[0025] In one embodiment of the present invention, see [reference] Figure 2A rectangular solid model with dimensions identical to the actual photovoltaic module is constructed, and a small tilt angle is set on the long side of this rectangular solid model to simulate the installation slope. The rectangular solid models are then batch-copyed to specified locations on the 3D terrain mesh model according to the array layout parameters. The interfaces between each rectangular solid model and the underlying terrain mesh are traversed, and the cutting contour lines at the interfaces are calculated. Boolean difference operations are performed on the corresponding rectangular solid models using the calculated cutting contour lines to remove the model parts embedded within the terrain, retaining the exposed parts that conform to the terrain. After completing the Boolean operation processing for all models, all processed solid models are merged into a single photovoltaic array assembly model.

[0026] In practical implementation, a rectangular solid model with the same dimensions as the actual photovoltaic module is constructed, where the actual photovoltaic module's dimensions are its length. Meter, width meters, thickness Meters, and a small tilt angle is set along the long side of the cuboid solid model to simulate the installation slope, tilt angle The degree and tilt angle are set through rotation transformation, and the rotation matrix formula is:

[0027] in: It is the original coordinate vector of the vertices of the cuboid solid model. It is the rotated coordinate vector. It is a rotation matrix about the y-axis. It is the tilt angle in radians. In some embodiments, the cuboid solid model is constructed using the original solid creation function of the 3D modeling software. After creation, the model size has an error of less than 0.001 meters compared with the design drawings. The cuboid solid model is imported from the standard component library, and the model automatically matches the unit system to meters after import. The processing time difference between the two methods of creating cuboid solid models in subsequent Boolean operations is less than 10%. In specific implementations, the cuboid solid models are batch copied to specified positions in the 3D terrain mesh model according to the array layout parameters, including row spacing. meters and column spacing The coordinates of the center point of the copied component model are calculated from the layout parameters using the formula. ,in and These are row indexes and column indexes. This is a terrain elevation function. Batch copying is automated via a script. The script reads the layout parameters and generates a transformation matrix, instantiating the cuboid solid model at each location. After instantiation, the total number of component models is 100, and the matching error between the location coordinates and the terrain elevation data is less than 0.05 meters. It iterates through the interface between each cuboid solid model and the terrain mesh below, calculating the cutting contour line at the interface. The cutting contour line is defined by the sequence of intersection points between the terrain mesh and the bottom surface of the component model. In essence, the calculation of the cutting contour line is achieved by solving for the intersection points of the triangular facets of the terrain mesh and the rectangular bottom surface of the component model. Each intersection point is obtained through linear interpolation, and the coordinates of the intersection points satisfy the equation... ,in It is the elevation value of the bottom surface of the component model at the current point.

[0028] In specific implementation, Boolean difference operations are performed on the corresponding cuboid solid models using the cutting contour lines to remove the model parts embedded in the terrain, retaining the exposed parts that fit the terrain. In some embodiments, the Boolean difference operation is performed using the Boolean operation tool of the 3D modeling software, with a tolerance of 0.001 meters. The Boolean difference operation is implemented through a custom algorithm based on the voxelization method. The difference between the two methods in complex terrain areas is manifested in the different smoothness of the cutting boundary, with the smoothness difference at the millimeter level. All solid models processed by Boolean operations are merged into a single photovoltaic array assembly model. Optionally, the merging operation is achieved by stitching all component models to the terrain base surface, with a stitching tolerance of 0.01 meters. The merged photovoltaic array assembly model file format is STEP. Optionally, the integrity of each component model is checked before merging to ensure that there are no isolated faces or edges. It is understood that the photovoltaic array assembly model is used for subsequent fluid dynamics simulation, and its geometric integrity affects the simulation accuracy. Geometric integrity is verified through model closure checks.

[0029] In one embodiment of the invention, a virtual wind tunnel inlet boundary is set on the windward side of the three-dimensional terrain mesh model, and a uniform incoming flow velocity is given. Symmetrical boundaries or slip wall boundaries are set on the top and sides of the model, and non-slip wall boundaries are set on the leeward side and the ground. A turbulence model suitable for high Reynolds number flows is selected, and the key parameters of the turbulence model are set, including the type of wall function and the analytical requirements of the near-wall mesh. Steady-state calculations are run until the flow field residuals converge and the velocity changes at the monitoring points tend to stabilize. Several characteristic measuring points are selected within the height range of the photovoltaic array, and the average wind speed and turbulent kinetic energy data at each measuring point are recorded. The turbulent kinetic energy data are then converted into turbulence intensity profile data. (See also...) Figure 3The wind tunnel test database within the wind load parameter calibration module is retrieved. This database contains baseline values ​​for the aerodynamic interference factor and wind pressure coefficient of photovoltaic arrays under different arrangement configurations. The obtained incoming wind speed profile data and turbulence intensity profile data are used as environmental variables, and interpolation queries are performed in the wind tunnel test database to find the baseline values ​​corresponding to the closest environmental conditions. Based on the measured surface roughness length parameter and the actual spacing parameter of the photovoltaic arrays at the current site, the retrieved baseline values ​​are linearly corrected. The corrected baseline values ​​are then combined to form the aerodynamic interference factor and wind pressure coefficient correction coefficients.

[0030] In practice, a virtual wind tunnel inlet boundary is set on the windward side of the 3D terrain mesh model, with the windward side being the western boundary, and a uniform inflow velocity is given. The speed was set to 10 m / s. Symmetrical or slip wall boundaries were set on the top and sides of the model. The top boundary and the three side boundaries (east, south, and north) were set as pressure far-field boundary conditions to simulate the infinite far field. No-slip wall boundaries were set on the leeward side and the ground surface. The ground no-slip wall boundary was treated using standard wall functions. A turbulence model suitable for high Reynolds number flows was selected, specifically the standard k-epsilon model. Key parameters of the standard k-epsilon model were set, including the type of wall function and the analytical requirements of the near-wall mesh. The standard wall function was selected, and the height of the first layer of the near-wall mesh satisfied that the y+ value was between 30 and 300. Steady-state calculations were run until the flow field residuals converged. The residual convergence criterion was set to the residuals of all equations decreasing to a certain value. Below, the velocity changes at the monitoring points tend to stabilize, with the rate of change of the monitoring point velocity being less than 0.1% over 1000 consecutive iterations. Several characteristic measuring points are selected within the height range of the photovoltaic array, located above the center line of the photovoltaic array at heights of 1 meter, 2 meters, 5 meters, 10 meters, and 20 meters. The average wind speed and turbulent kinetic energy data at each measuring point are recorded, and the turbulent kinetic energy data is converted into turbulence intensity profile data using the following formula:

[0031] in: These are turbulence intensity profile data at height z. These are turbulent kinetic energy data at height z, expressed in square meters per second squared. The average wind speed at height z is expressed in meters per second. The calculated turbulence intensity profiles at heights of 1 meter, 2 meters, 5 meters, 10 meters, and 20 meters are 0.25, 0.23, 0.18, 0.15, and 0.12, respectively. Optionally, the extraction of turbulence intensity profile data can be validated by comparing results under different inlet turbulence settings. For example, comparing the differences between profiles obtained from a uniform inlet and those based on empirical formulas; if the difference is within 5%, the current result is adopted.

[0032] In practical implementation, the wind tunnel test database within the wind load parameter calibration module is retrieved. This database contains baseline values ​​for aerodynamic interference factors and wind pressure coefficients of photovoltaic arrays under different arrangement configurations. The database is stored in tabular form, with index parameters including row spacing, column spacing, tilt angle, incoming wind speed profile index, and inlet turbulence intensity. The extracted incoming wind speed profile data and turbulence intensity profile data are used as environmental variables. Interpolation queries are performed in the wind tunnel test database to find the baseline values ​​corresponding to the closest environmental conditions. The interpolation query uses a bilinear interpolation algorithm, with the input parameter being the incoming wind speed profile index for the current site. and inlet turbulence intensity The obtained aerodynamic interference factor baseline value The wind pressure coefficient is 1.25, which is the baseline value. The value is 0.85. In some embodiments, the interpolation query is performed on two dimensions: wind speed profile index and turbulence intensity. The interpolation query also incorporates the wind direction angle dimension. A comparison of the multidimensional interpolation results and the two-dimensional interpolation results shows that when the wind direction angle is 0 degrees, the difference in the aerodynamic interference factor baseline value is less than 3%. Based on the currently measured surface roughness length parameter and the actual spacing parameter of the photovoltaic array, the surface roughness length parameter... Meters, the actual row spacing parameter of the photovoltaic array Meter, column spacing parameters The metric is linearly corrected to the queried baseline value using the following formula:

[0033]

[0034] in: It is the corrected aerodynamic interference factor. This is the corrected wind pressure coefficient. , , These are the surface roughness length reference parameters, row spacing reference parameters, and column spacing reference parameters corresponding to the current query record in the wind tunnel test database, with values ​​of 0.05 meters, 4 meters, and 2.5 meters, respectively. These are linear correction coefficients, whose values ​​are obtained by fitting historical data, and are 0.5, 0.02, 0.01, 0.3, 0.015, and 0.008 respectively. The purpose of linear correction is to map the standard conditions of the wind tunnel test to actual site conditions. The calculated corrected aerodynamic disturbance factor is 1.28, and the wind pressure coefficient is 0.87. The corrected baseline values ​​are combined to form the aerodynamic disturbance factor and wind pressure coefficient correction coefficients, resulting in an aerodynamic disturbance factor of 1.28 and a wind pressure coefficient correction coefficient of 0.87. In some embodiments, the coefficients used for linear correction are derived from regression analysis of multiple sets of historical wind tunnel test and field measurement data, and from parameterized studies of computational fluid dynamics. The coefficients from both sources show basically consistent calculation results under typical flat terrain, but the difference is approximately 5% under mountainous terrain. Optionally, after correction, the generated aerodynamic disturbance factor and wind pressure coefficient correction coefficients are written to a specified parameter file for subsequent mesh generation module reading. The parameter file format is JSON, containing parameter names and numerical key-value pairs. It is understandable that the aerodynamic interference factor is used to quantify the flow obstruction effect between arrays, and the wind pressure coefficient correction factor is used to adjust the pressure distribution on individual components.

[0035] In one embodiment of the invention, an aerodynamic interference factor is read, and the mesh refinement level at the component gaps is determined based on the magnitude of the aerodynamic interference factor; the larger the interference factor, the higher the refinement level. A wind pressure coefficient correction factor is read, and the mesh refinement level at the component edges and corners is adjusted based on the magnitude of the wind pressure coefficient correction factor; the larger the correction factor, the higher the refinement level. Curvature analysis is performed on all surfaces in the photovoltaic array assembly model, and additional mesh nodes are automatically inserted in areas with drastic curvature changes. A hybrid mesh generation strategy is adopted, using structured quadrilateral meshes for the component planar areas and unstructured tetrahedral meshes for the connections between the support and the component. The quality parameters of the mesh are checked, and mesh cells with excessively high distortion are removed to complete the generation of a high-resolution computational mesh. Based on this high-resolution computational mesh, for typhoon conditions, an average wind speed profile with exponential decay characteristics and a high-intensity turbulent fluctuation signal are loaded onto the virtual wind tunnel inlet boundary. For non-typhoon conditions, an average wind speed profile with logarithmic law characteristics and a low-intensity turbulent fluctuation signal are loaded onto the virtual wind tunnel inlet boundary. Large eddy simulation (LES) was activated under both operating conditions to capture transient vortex shedding phenomena in the flow field. Several virtual pressure monitoring points were deployed on the surface of the photovoltaic array, and the pressure changes at each monitoring point were recorded within the simulation time step. The pressure data from each monitoring point were then organized into a time series to form dynamic wind pressure time history data.

[0036] In practice, the aerodynamic interference factor and wind pressure coefficient correction factor are read. The aerodynamic interference factor is 1.28 and the wind pressure coefficient correction factor is 0.87. The mesh density level at the component gap is determined according to the magnitude of the aerodynamic interference factor. The larger the interference factor, the higher the density level. For the specific relationship between the density level and the mesh base size, please refer to Table 1.

[0037] Table 1: Correspondence between specific encryption levels and grid base size

[0038] In practical implementation, the current aerodynamic interference factor is 1.28, corresponding to Level 2 mesh refinement, and the basic mesh size at the component gaps is set to 0.10 meters. The wind pressure coefficient correction factor is read, and the mesh refinement at the component edges and corners is adjusted according to the magnitude of the correction factor. The larger the correction factor, the higher the refinement. The final mesh size at the component edges and corners... for:

[0039] in: It is the final grid size at the edges and corners of the component. This is the basic grid dimension, which is 0.10 meters in this case. This is the refinement factor, with a value of 2.0. It is the wind pressure coefficient correction factor, which is substituted into the numerical values ​​to obtain the result. Meters. Curvature analysis is performed on all surfaces in the photovoltaic array assembly model. Additional mesh nodes are automatically inserted in areas of drastic curvature change. The threshold for determining curvature change is 0.1 millimeters. Nodes were automatically inserted in the rounded corner areas of the component frame and the weld seam areas at the bracket connection. In some embodiments, curvature analysis was achieved by calculating the Gaussian curvature of the surface, and curvature analysis was achieved by calculating the angle between the normal vectors of adjacent surfaces. Both methods yielded consistent results when identifying sharp edges. A hybrid meshing strategy was adopted, using structured quadrilateral meshes for the planar areas of the component, with the aspect ratio of the structured quadrilateral mesh controlled within 1:5. Unstructured tetrahedral meshes were used at the connection between the bracket and the component, with the maximum skewness of the unstructured tetrahedral mesh controlled within 0.8. The quality parameters of the mesh were checked, and mesh cells with excessive distortion were removed. Mesh quality parameters included Jacobian ratio and warpage. The lower limit of Jacobian ratio was set to 0.6, and the upper limit of warpage was set to 15 degrees. This completed the generation of the high-resolution computational mesh, which ultimately contained approximately 8.5 million mesh cells.

[0040] In practical implementation, boundary conditions for both typhoon and non-typhoon extreme wind fields are loaded onto a high-resolution computational grid. Transient wind load numerical simulations are then run to collect dynamic wind pressure time-history data on the photovoltaic array surface. For the typhoon condition, an average wind speed profile with exponential decay characteristics and a high-intensity turbulent fluctuation signal are loaded onto the virtual wind tunnel inlet boundary. The formula for the average wind speed profile with exponential decay characteristics is:

[0041] in: It is the average wind speed at height z. This is the reference wind speed at a height of 10 meters, taken as 42 meters per second. The attenuation index is 0.12. The high-intensity turbulent fluctuation signal is generated from a random fluctuating wind speed spectrum, and the turbulence intensity is set to 0.25. In some embodiments, the high-intensity turbulent fluctuation signal is generated based on the Kaimal spectrum, and the high-intensity turbulent fluctuation signal is generated based on the Von Karman spectrum. The energy distribution difference between the two spectra in the main frequency range is less than 5%. For non-typhoon conditions, a logarithmic-law-based average wind speed profile and a low-intensity turbulent fluctuation signal are loaded at the virtual wind tunnel inlet boundary. The logarithmic-law-based average wind speed profile uses the logarithmic-law formula:

[0042] in: This is the friction speed, taken as 1.5 meters per second. It is the Kármán constant of 0.4. The surface roughness length is 0.1 meters, and the turbulence intensity of the low-intensity turbulent fluctuation signal is set to 0.15. Large eddy simulation (LES) mode is enabled in both operating conditions to capture transient vortex shedding phenomena in the flow field. The subgrid-scale model for LES is the Smagorinsky-Lilly model, with the Smagorinsky constant set to 0.1. A second-order implicit scheme is used for time progression, and the time step is set to 0.001 seconds. It is understood that the time step setting must satisfy the stability condition of a Coulomb number less than 1. Several virtual pressure monitoring points are arranged on the surface of the photovoltaic array to record the pressure change of each monitoring point within the simulation time step. A total of 120 virtual pressure monitoring points are arranged evenly on the surface of 9 modules in three rows and three columns in the middle of the array. Each module surface has 5 monitoring points: 4 corner points and 4 center points. Optionally, the same number of pressure monitoring points are also symmetrically arranged on the leeward side of the modules. In the specific implementation, the total simulation time is 10 seconds (physical time), and pressure data for each monitoring point is recorded for a total of 10,000 time steps. The pressure data of each monitoring point is organized into a time series to form dynamic wind pressure time history data. The dynamic wind pressure time history data is stored in text file format, with each column representing the pressure change sequence of a monitoring point over time. Optionally, the dynamic wind pressure time history data is also stored in binary format to save space. The size of the binary format file is about 30% of that of the text format. It can be understood that the dynamic wind pressure time history data is the basis for subsequent time-domain statistical analysis.

[0043] See Figure 4 This is a comparative chart of wind speed profiles in a photovoltaic array, showing the difference in average wind speed with altitude under typhoon and non-typhoon conditions. It is a visualization of the core boundary conditions for refined simulation of wind load on photovoltaic arrays. Under typhoon conditions, wind speed decreases significantly with altitude, reaching approximately 42 m / s near the ground (10 meters) and dropping to approximately 16 m / s at 20 meters, reflecting the strong shear characteristics of the typhoon boundary layer. Under non-typhoon conditions, wind speed increases slowly with altitude, reaching approximately 8 m / s near the ground (1 meter) and increasing to approximately 20 m / s at 20 meters, reflecting the stable growth characteristics of the conventional atmospheric boundary layer. The wind speed gradient is gentler, resulting in weaker instantaneous impact on structures compared to typhoon conditions. During typhoon conditions, high wind speeds at low altitudes significantly amplify the wind pressure on the surface of photovoltaic modules, requiring close attention to the wind-resistant design of module gaps and support structures; in non-typhoon conditions, the focus is more on the fatigue and stability of the structure under sustained medium to high wind speeds.

[0044] In one embodiment of the present invention, a fast Fourier transform is performed on the dynamic wind pressure time history data to obtain the power spectral density of the wind pressure signal. The dominant frequency component is identified in the power spectral density, and this dominant frequency is compared with the first-order natural frequency of the photovoltaic array to be designed to calculate the frequency ratio. Based on the frequency ratio and damping ratio, the wind vibration response amplification factor is calculated using random vibration theory. The simulation results for all wind direction angles are iterated to find the peak positive pressure and peak negative pressure corresponding to each wind direction angle. The peak positive pressure and peak negative pressure corresponding to all wind direction angles are mapped onto the surface of the three-dimensional geometric solid model of the photovoltaic array to draw an extreme wind pressure distribution map. The extreme wind pressure distribution map is read, and the wind pressure value on the map is multiplied by the wind vibration response amplification factor to obtain the design wind load value. The design wind load value is used as the input load and applied to the structural finite element model of the photovoltaic array. In the structural design verification module, the ratio of the material's yield strength to the allowable stress is set, and the axial force and bending moment of the members are automatically calculated. Based on the calculation results, an iterative search is performed in the member specification library to select the minimum cross-sectional specification that meets the strength and stiffness requirements as the member cross-section selection suggestion. Simultaneously, based on the wind load distribution, the arrangement density of the support beams is optimized, generating connection spacing configuration parameters. Based on the member cross-section selection suggestion and connection spacing configuration parameters, a structural reinforcement model is created for the 3D geometric solid model of the photovoltaic array, followed by a modal analysis and static equilibrium verification. In the 3D geometric solid model of the photovoltaic array, the original member entities are replaced with entities having the specifications specified in the member cross-section selection suggestion. According to the adjusted connection spacing configuration parameters, the positions of bolts or welded joints on the support beams are rearranged. A free vibration modal analysis is performed on the reinforced structural model to extract its first few natural frequencies and mode shapes. A standard static load is applied to the reinforced structural model, and a static equilibrium solution is performed to check for any abnormal displacements or stress concentrations. If no problems are found in the modal analysis and static equilibrium verification, the current reinforced modeling state is confirmed as the final design scheme. Output the reinforced modeling and verification of the photovoltaic array design drawings and the wind load simulation analysis report.

[0045] In the specific implementation, a Fast Fourier Transform (FFT) was performed on the dynamic wind pressure time history data to obtain the power spectral density of the wind pressure signal. The dynamic wind pressure time history data came from the center monitoring point on the surface of the component under typhoon conditions, with a sampling frequency of 1000 Hz and a data length of 10,000 points. The Fast Fourier Transform was performed using the Cooley-Tukey algorithm, and the Hanning window was selected as the window function to reduce spectral leakage. The power spectral density results showed a significant peak at a frequency of 3.2 Hz, the amplitude of which was 15 times the average amplitude of other frequency components. The dominant frequency component was identified in the power spectral density. The dominant frequency is 3.2 Hz, and it is compared with the first-order natural frequency of the photovoltaic array to be designed. The value was obtained from previous modal analysis and is 2.5 Hz. The frequency ratio was calculated. Based on the frequency ratio and damping ratio, the damping ratio... The value is set to 0.02, and the amplification factor of the wind-induced vibration response is calculated using random vibration theory. The calculation uses the following formula:

[0046] in: It is the amplification factor of the wind vibration response. It's the frequency ratio. It is the damping ratio, which is calculated by substituting the values. In some embodiments, the damping ratio is selected based on the characteristics of the support material. The damping ratio is determined based on measured data of similar structures, and the difference between the damping ratio values ​​from the two sources is within 20%, corresponding to a difference of less than 10% in the calculated wind vibration response amplification factor. The simulation results are iterated over for all wind directions, including 0 degrees, 45 degrees, 90 degrees, 135 degrees, 180 degrees, 225 degrees, 270 degrees, and 315 degrees (eight directions in total). The peak positive pressure and peak negative pressure corresponding to each wind direction are identified. For example, at a wind direction angle of 0 degrees, the peak positive pressure on the upper surface of the first row of components in the array is +1250 Pa, and the peak negative pressure on the lower surface is -950 Pa. In specific implementations, the peak positive pressure and peak negative pressure are obtained by iterating through all time-history data of each monitoring point to find the maximum and minimum values. The peak positive and peak negative pressures corresponding to all wind direction angles are mapped onto the surface of the three-dimensional geometric solid model of the photovoltaic array to draw an extreme wind pressure distribution map. In the mapping process, the pressure extreme value at each monitoring point is used as node data, and a continuous pressure cloud map of the entire component surface is generated through a surface interpolation algorithm. The extreme wind pressure distribution map is visualized in the form of a color cloud map, and the pressure value range is marked with a legend.

[0047] In practice, the extreme wind pressure distribution map is read, and the wind pressure value on the map is multiplied by the wind-induced vibration response amplification factor to obtain the design wind load value. For example, for the area marked as +1250 Pa in the map, the design wind load value is calculated as follows: The design wind load value is used as the input load and applied to the structural finite element model of the photovoltaic array. The structural finite element model is a hybrid model of beam and shell elements. The support is simulated using beam elements, and the modules are simulated using shell elements. The load is applied to the shell elements of the modules in the form of surface pressure, and the load direction is perpendicular to the module surface. In the structural design verification module, the ratio of the material's yield strength to allowable stress is set. The material is Q235 steel, with a yield strength of 235 MPa and an allowable stress ratio of 0.8, i.e., an allowable stress of 188 MPa. The axial force and bending moment of the members are automatically calculated. The calculation is based on linear static analysis. It can be understood that the axial force and bending moment are the main basis for the selection of the member cross-section. Based on the calculation results, an iterative search is performed in the member specification library, which contains various cross-sectional parameters of steel, such as the cross-sectional area and moment of inertia of square tubes, round tubes, and H-beams. The smallest cross-sectional specification that meets the strength and stiffness requirements is selected as the member cross-section selection suggestion. For example, for a diagonal brace with a calculated maximum axial pressure of 50 kN and a maximum bending moment of 2 kN·m, the iterative search first checks the smallest cross-section, a 20x20x2 mm square tube, which does not meet the strength requirements. Then, larger specifications are tried in turn, and finally a 40x40x3 mm square tube is selected as the member cross-section selection suggestion because the cross-sectional strength and stiffness of this specification meet the requirements and the mass is the smallest. Meanwhile, based on the distribution of wind load, the arrangement density of the support beams is optimized, and the connection spacing configuration parameters are generated. The wind load distribution shows that the wind load in the edge area of ​​the array is 30% higher than that in the center area. Therefore, the connection spacing of the beams in the edge area is increased from 1.5 meters to 1.0 meters, while the spacing in the center area is kept at 1.5 meters. The connection spacing configuration parameters are given in the form of a list, which includes the beam number and the corresponding connection spacing value.

[0048] Based on the recommended member cross-section selection and connector spacing configuration parameters, a structural reinforcement model of the photovoltaic array's 3D geometric solid model was created, followed by a modal analysis and static equilibrium verification. In the 3D geometric solid model of the photovoltaic array, the original member entities were replaced with entities having the specifications specified in the member cross-section selection recommendations. For example, all diagonal bracing entities with a cross-section of 30x30x2.5 mm were replaced with diagonal bracing entities with a cross-section of 40x40x3 mm. According to the adjusted connector spacing configuration parameters, the positions of bolts or welded joints on the support beams were rearranged. For example, on beam number B-01, the bolt joint position was adjusted from [0, 1.5, 3.0, 4.5] meters to [0, 1.0, 2.0, 3.0, 4.0, 4.5] meters. A free vibration modal analysis was performed on the reinforced structural model to extract its first few natural frequencies and mode shapes. The Lanczos algorithm was used for modal analysis, extracting the first six modes. The first natural frequency is 2.8 Hz, and the mode shape is the translational motion of the entire array along the wind direction. Standard static loads, including a combination of 1.2 times the dead load and 1.4 times the wind load, were applied to the reinforced structural model. Static equilibrium was then performed to check for any abnormal displacements or stress concentrations. The results showed a maximum displacement of 25 mm, located at the top of the array, and a maximum stress of 175 MPa, located at the root of the beam near the densely connected joints. Neither exceeded the allowable limits, and no abnormal displacements or stress concentrations were found. If no problems were found in the modal analysis and static equilibrium verification, the current reinforced modeling state was confirmed as the final design scheme. Output reinforced modeled and verified photovoltaic array design drawings and wind load simulation analysis report. The photovoltaic array design drawings include updated structural layout drawings, member details and node details. The wind load simulation analysis report includes extreme wind pressure distribution map, wind vibration response amplification factor, member internal force calculation results and selection basis.

[0049] See Figure 5This is an extreme wind pressure distribution map, visually presenting the magnitude and distribution of wind pressure at different locations on the array under wind conditions. It is one of the core outputs of refined wind load simulation. The high-pressure core region forms a strong positive pressure core in the central area of ​​the array, with peak wind pressure approaching 1000 Pa. The positive pressure zone diffuses outward in an elliptical shape, with wind pressure gradually decreasing from the center to the periphery, reflecting the stagnation effect of airflow impacting the windward surface of the array. This area is critical for photovoltaic modules to withstand the maximum positive load and requires significant structural reinforcement. The low-pressure edge region forms symmetrical negative pressure zones on the left and right edges of the array, with the lowest wind pressure around -800 Pa, exhibiting a suction effect. The negative pressure zone also shows a gradient distribution, with stronger suction closer to the array edge, a typical result of airflow around the array and vortex shedding effects. Weak negative pressure zones also exist at the upper and lower edges, with lower intensity than those on the sides. The transition zone is a transitional zone between the positive pressure core and the negative pressure edge, with wind pressure approaching 0 Pa, representing the boundary area where the load direction changes. The existence of the transition zone means that different positions of the array are simultaneously subjected to both push and pull loads, which places higher demands on the fatigue and stability of the structure.

[0050] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A refined simulation design method for wind load on photovoltaic arrays, characterized in that, The method includes: Topographic elevation data and surface roughness data were acquired, and a three-dimensional topographic mesh model was constructed in conjunction with the layout drawings of the photovoltaic array. Boolean operations were performed on the three-dimensional geometric solid model to cut it, and the incoming wind speed profile data and turbulence intensity profile data of the area where the photovoltaic array is located were extracted. The incoming wind speed profile data and turbulence intensity profile data are input into the wind load parameter calibration module to generate a set of aerodynamic interference factors and wind pressure coefficient correction coefficients applicable to the site conditions of the photovoltaic array. Based on the aerodynamic interference factor and wind pressure coefficient correction coefficient, the three-dimensional geometric solid model of the photovoltaic array is reconstructed into a mesh to generate a high-resolution computational mesh containing details of component gaps and support structure. Based on the high-resolution computing grid, boundary conditions for two extreme wind fields, namely typhoon and non-typhoon conditions, were loaded, and transient wind load numerical simulations were run to collect dynamic wind pressure time history data on the surface of the photovoltaic array. Statistical analysis was performed on the dynamic wind pressure time history data to extract the extreme wind pressure distribution map and wind vibration response amplification factor of the photovoltaic array under different wind direction angles; The extreme wind pressure distribution map and wind vibration response amplification factor are imported into the structural design verification module to automatically generate recommendations for the selection of rod cross sections and configuration parameters for the spacing of connectors in the photovoltaic array support system.

2. The refined simulation design method for wind load based on photovoltaic arrays as described in claim 1, characterized in that, Topographic elevation and surface roughness data were acquired, and a three-dimensional terrain mesh model was constructed using the photovoltaic array layout drawings. Boolean operations were performed on the three-dimensional geometric model to cut it, and incoming wind speed profiles and turbulence intensity profiles for the photovoltaic array area were extracted, including: Obtain the topographic elevation data and surface roughness data of the site where the photovoltaic array is located, and construct a three-dimensional terrain mesh model in combination with the layout drawings of the photovoltaic array; Import the three-dimensional geometric solid model of the photovoltaic module into the three-dimensional terrain mesh model, and perform Boolean operation to cut the three-dimensional geometric solid model so that its boundary fits the undulating surface of the three-dimensional terrain mesh model; The steady-state wind field initialization simulation was performed on the fitted three-dimensional terrain mesh model using computational fluid dynamics software, and the incoming wind speed profile data and turbulence intensity profile data of the area where the photovoltaic array is located were extracted. The process of acquiring terrain elevation and surface roughness data of the site where the photovoltaic array is located, and constructing a three-dimensional terrain mesh model in conjunction with the layout drawings of the photovoltaic array, includes: Digital elevation model data within the site selection area of ​​the photovoltaic array is retrieved from the geographic information system database and then interpolated and converted into regular terrain raster data. The vegetation cover type and building distribution around the photovoltaic array were surveyed and recorded on-site, and the surface roughness length parameter was quantified and generated. Read the layout drawings of the photovoltaic array and identify the row spacing, column spacing, and tilt and azimuth parameters of each module; The terrain raster data, the surface roughness length parameter, and the array arrangement parameters are imported into 3D modeling software to generate a base surface with terrain undulation features. The array layout is projected onto the base surface according to the array arrangement parameters to initially construct the three-dimensional terrain mesh model.

3. The refined simulation design method for wind load based on photovoltaic arrays as described in claim 2, characterized in that, Importing the 3D geometric solid model of the photovoltaic module into the 3D terrain mesh model, and performing Boolean operations to cut the 3D geometric solid model, including: Construct a cuboid solid model with the same dimensions as the actual photovoltaic module, and set a small tilt angle on the long side of the cuboid solid model to simulate the installation slope; The cuboid solid model is copied in batches to the designated location of the three-dimensional terrain mesh model according to the array arrangement parameters; Traverse the interface between each cuboid solid model and the terrain mesh below, and calculate the cutting contour line at the interface. Using the cutting contour line, perform Boolean difference operation on the corresponding cuboid solid model to remove the model part embedded in the terrain and retain the exposed part that fits the terrain. All entity models processed by Boolean operations are merged into a single photovoltaic array assembly model.

4. The refined simulation design method for wind load based on photovoltaic arrays as described in claim 3, characterized in that, The steady-state wind field initialization simulation was performed on the fitted 3D terrain mesh model using computational fluid dynamics software. Incoming wind speed profiles and turbulence intensity profiles for the photovoltaic array area were extracted, including: A virtual wind tunnel inlet boundary is set on the windward side of the three-dimensional terrain mesh model, and a uniform incoming flow velocity is given. Set symmetrical or sliding wall boundaries on the top and sides of the model, and set non-slip wall boundaries on the leeward side and the ground. Select a turbulence model suitable for high Reynolds number flows and set the key parameters of the turbulence model, including the type of wall function and the analytical requirements of the near-wall mesh; Run steady-state calculations until the flow field residuals converge and the velocity changes at the monitoring points tend to stabilize; Several characteristic measurement points are selected within the height range of the photovoltaic array, and the average wind speed and turbulent kinetic energy data at each measurement point are recorded. The turbulent kinetic energy data is then converted into the turbulence intensity profile data.

5. The refined simulation design method for wind load based on photovoltaic arrays as described in claim 4, characterized in that, The incoming wind speed profile data and turbulence intensity profile data are input into the wind load parameter calibration module to generate a set of aerodynamic interference factors and wind pressure coefficient correction coefficients suitable for the site conditions of the photovoltaic array, including: The wind tunnel test database inside the wind load parameter calibration module is retrieved. The wind tunnel test database contains the aerodynamic interference factor reference value and wind pressure coefficient reference value of the photovoltaic array under different arrangement methods. The incoming wind speed profile data and the turbulence intensity profile data are used as environmental variables. Interpolation queries are performed in the wind tunnel test database to find the benchmark value corresponding to the closest environmental conditions. Based on the measured surface roughness length parameters of the current site and the actual spacing parameters of the photovoltaic array, the queried benchmark values ​​are linearly corrected. The corrected baseline values ​​are combined to form the aerodynamic interference factor and the wind pressure coefficient correction factor.

6. The refined simulation design method for wind load based on photovoltaic arrays as described in claim 5, characterized in that, Based on the aforementioned aerodynamic interference factor and wind pressure coefficient correction coefficient, the three-dimensional geometric solid model of the photovoltaic array is reconstructed into a mesh, generating a high-resolution computational mesh that includes details of module gaps and support structures, including: The aerodynamic interference factor is read, and the mesh density level at the component gap is determined according to the magnitude of the aerodynamic interference factor. The larger the interference factor, the higher the density level. Read the wind pressure coefficient correction factor, and adjust the mesh refinement at the edges and corners of the component according to the magnitude of the wind pressure coefficient correction factor. The larger the correction factor, the higher the refinement. Perform curvature analysis on all surfaces in the photovoltaic array assembly model and automatically insert additional mesh nodes in areas with drastic curvature changes. A hybrid meshing strategy is adopted, using structured quadrilateral meshes for the planar areas of the components and unstructured tetrahedral meshes for the connection between the support and the components; Check the quality parameters of the mesh, remove mesh cells with excessive distortion, and complete the generation of the high-resolution computational mesh.

7. The refined simulation design method for wind load based on photovoltaic arrays as described in claim 6, characterized in that, Based on the high-resolution computational grid, boundary conditions for two extreme wind fields—typhoon and non-typhoon—were loaded. Transient wind load numerical simulations were then run to collect dynamic wind pressure time-history data on the photovoltaic array surface, including: For typhoon conditions, an average wind speed profile with exponential decay characteristics and a high-intensity turbulence pulse signal are loaded onto the virtual wind tunnel inlet boundary. For non-typhoon conditions, an average wind speed profile with logarithmic law characteristics and a low-intensity turbulence pulse signal are loaded onto the virtual wind tunnel inlet boundary. Large eddy simulation mode was activated under both operating conditions to capture transient vortex shedding phenomena in the flow field. Several virtual pressure monitoring points are arranged on the surface of the photovoltaic array, and the pressure value change of each monitoring point within the simulation time step is recorded; The pressure data from each monitoring point are organized according to the time series to form the dynamic wind pressure time history data.

8. The refined simulation design method for wind load based on photovoltaic arrays as described in claim 7, characterized in that, Statistical analysis was performed on the dynamic wind pressure time history data to extract the extreme wind pressure distribution map and wind vibration response amplification factor of the photovoltaic array under different wind direction angles, including: The power spectral density of the wind pressure signal is obtained by performing a fast Fourier transform on the dynamic wind pressure time history data. The dominant frequency component is identified in the power spectral density, and the dominant frequency is compared with the first-order natural frequency of the photovoltaic array to be designed to calculate the frequency ratio. The amplification factor of the wind vibration response is calculated using random vibration theory based on the frequency ratio and damping ratio. By iterating through the simulation results for all wind direction angles, the peak positive pressure and peak negative pressure corresponding to each wind direction angle are found. The peak positive pressure and peak negative pressure corresponding to all wind direction angles are mapped onto the surface of the three-dimensional geometric solid model of the photovoltaic array, and the extreme wind pressure distribution map is drawn.

9. The refined simulation design method for wind load based on photovoltaic arrays as described in claim 8, characterized in that, The extreme wind pressure distribution map and wind vibration response amplification factor are imported into the structural design verification module to automatically generate recommendations for the selection of member cross-sections and configuration parameters for the spacing of connectors in the photovoltaic array support system, including: Read the extreme wind pressure distribution map, multiply the wind pressure value on the map by the wind vibration response amplification factor, and obtain the design wind load value; The design wind load value is used as the input load and applied to the structural finite element model of the photovoltaic array. In the structural design verification module, the ratio of the material's yield strength to its allowable stress is set, and the axial force and bending moment of the members are automatically calculated. Based on the calculation results, an iterative search is performed in the member specification library to select the minimum cross-sectional specification that meets the strength and stiffness requirements as the member cross-section selection suggestion; Simultaneously, based on the distribution of wind load, the arrangement density of the support beams is optimized, and the spacing configuration parameters of the connectors are generated.

10. The refined simulation design method for wind load based on photovoltaic arrays as described in claim 9, characterized in that, Also includes: Based on the recommended selection of the rod cross-section and the configuration parameters of the connector spacing, the three-dimensional geometric solid model of the photovoltaic array is structurally reinforced and modeled, and a modal analysis and static equilibrium verification are performed. Output reinforced modeling and verification photovoltaic array design drawings and wind load simulation analysis report; Based on the recommended selection of the rod cross-section and the configuration parameters of the connector spacing, a structural reinforcement model of the three-dimensional geometric solid model of the photovoltaic array is performed, followed by a modal analysis and static equilibrium verification, including: In the three-dimensional geometric solid model of the photovoltaic array, the original rod solids are replaced with solids having the specifications specified in the rod cross-section selection recommendations; Based on the adjusted connector spacing configuration parameters, rearrange the positions of bolts or welded nodes on the support beam; A free vibration modal analysis was performed on the reinforced structural model to extract its first few natural frequencies and mode shapes. Apply standard static loads to the reinforced structural model, perform static equilibrium solutions, and check for any abnormal displacements or stress concentrations. If no problems are found in modal analysis and static equilibrium verification, the current reinforcement modeling state will be confirmed as the final design scheme.

Images