A high-frequency sampled WRF-CFD coupled wind field simulation method

Abstract

The application discloses a high-frequency-sampling WRF-CFD coupling wind field simulation method, relates to the technical field of wind field coupling simulation, and comprises the following steps: determining a region to be analyzed and one-way coupling data requirements; determining a plurality of virtual site layout positions, and establishing a spatial corresponding relationship between the virtual sites and CFD driving positions; configuring a time series site output module tslist, outputting atmospheric state variable profile data at the virtual site positions at a WRF mode time step interval; constructing a CFD model; converting the atmospheric state variable profile data into CFD original driving data; performing data processing on the CFD original driving data and inputting the CFD model to carry out microscale wind field simulation; and the like.

Description

Technical Field

[0001] This invention relates to the field of wind field coupling simulation technology, and in particular to a high-frequency sampling WRF-CFD coupled wind field simulation method. Background Technology

[0002] In the field of wind field simulation, numerical models at different scales each have their own applicable scope. Mesoscale meteorological models (such as the WRF model) are widely used in wind resource simulation and assessment and wind speed forecasting, but their typical horizontal resolution is at the kilometer level, which is insufficient to distinguish the near-surface wind field characteristics under the influence of microscale structures such as complex local terrain and urban building clusters. Microscale CFD (Computational Fluid Dynamics) models can finely analyze local wind field structures at meter-level resolution, but their boundary conditions often rely on empirical formulas, making it difficult to reflect the wind field structure of the real atmospheric boundary layer. Coupled with the WRF (Weather Research and Forecasting Model) to provide reasonable boundary conditions for CFD models to conduct downscaled wind field simulations has become one of the mainstream technical approaches in this field.

[0003] Existing WRF-CFD coupling methods generally follow a similar technical approach in the data transfer stage: extracting the wind field and related variable fields of the region to be analyzed from historical WRF output files (wrfout files), and then transforming them into driving conditions for the CFD model through data processing. Typically, the preset save interval for wrfout files ranges from several minutes to tens of minutes. However, WRF models internally predict atmospheric state evolution at integration step intervals (usually on the order of seconds), and minute-level output save intervals are far from sufficient to capture rapid changes in the wind field at the WRF integration step scale. Simultaneously, the spatial scale difference between mesoscale WRF grids and microscale CFD grids can reach two to three orders of magnitude.

[0004] In highly nonlinear wind fields under complex terrain conditions, mapping the aforementioned driving data, whose spatiotemporal resolution is limited, to the CFD computational domain after data processing not only makes it difficult to guarantee the accuracy of the processing results, but also lacks constraints on the physical consistency of coupled variables in the time dimension. Directly improving temporal resolution by shortening the save interval of wrfout files would lead to a high-frequency save task for large-scale multivariable fields, resulting in a significant data storage and read / write burden. Therefore, it is urgent to develop a more efficient coupling technique between mesoscale numerical weather prediction models and CFD models to improve the time-varying fidelity of transferring mesoscale meteorological field information to microscale models. Summary of the Invention

[0005] The purpose of this invention is to provide a high-frequency sampling WRF-CFD coupled wind field simulation method, which solves the problem that the existing WRF-CFD coupling method mainly relies on the wrfout file for instantaneous data transmission based on a preset save interval during the unidirectional coupling process. This results in the CFD driving data construction being limited by the preset save interval of the wrfout file, making it difficult to fully utilize the temporal evolution information of the wind field and related atmospheric variable fields during the WRF model integration process. Increasing the save frequency of the wrfout file will increase the data storage and read / write burden.

[0006] To achieve the above objectives, this invention provides a high-frequency sampling WRF-CFD coupled wind field simulation method, comprising the following steps: Step 100: Determine the unidirectional coupling data requirements based on the region to be analyzed. The unidirectional coupling data requirements include the driving method of the CFD model and the WRF output data required to construct the CFD driving data. Step 200: Based on the unidirectional coupling data requirements, determine the deployment locations of multiple virtual sites and establish the spatial correspondence between the virtual sites and the CFD drive locations. Step 300: Configure the WRF mode time series station output module tslist according to the virtual station deployment location, and carry out WRF regional meteorological simulation, outputting atmospheric state variable profile data at the virtual station location according to the WRF mode time step interval; Step 400: Based on the driving method of the CFD model, the location of the virtual stations, and the atmospheric state variable profile data, determine the spatial range of the CFD computational domain, the dominant inflow direction, the inflow boundary, the outflow boundary, and other boundary conditions, thereby constructing the CFD model of the region to be analyzed. Step 500: Extract the target variable data required for constructing CFD driving data from the atmospheric state variable profile data through the set program. The target variable data corresponding to different target variables are converted from the WRF model coordinate system to the CFD model coordinate system as CFD raw driving data through a preset method. Step 600: Process the raw CFD driving data to form CFD driving data used to drive the CFD model calculation. Step 700: Input the CFD driving data into the CFD model to drive the CFD model to perform microscale wind field simulation of the area to be analyzed.

[0007] Furthermore, the driving methods of the CFD model include boundary condition driving method, internal source term perturbation driving method, or assimilation constraint driving method.

[0008] Furthermore, in step 100, the WRF computation domain includes an inner nested domain and an outer nested domain, and the innermost nested domain of the WRF needs to cover the corresponding geographical range of the CFD flow field inlet boundary.

[0009] Furthermore, in step 200, determining the deployment locations of multiple virtual sites based on the unidirectional coupling data requirements includes: When the CFD model is driven by boundary conditions, virtual sites are deployed along the geographical locations corresponding to the target boundary of the CFD computation domain. When the CFD model is driven by internal source term perturbation or assimilation constraint, virtual sites are deployed in the geographical locations corresponding to the target area within the CFD computation domain.

[0010] Further, in step 200, establishing the spatial correspondence between the virtual site and the CFD driving location includes: Establish the spatial correspondence between virtual sites and CFD drive locations using any of the following methods: First, select horizontal control positions on the target boundary or internal area of ​​the CFD computation domain to apply driving data, determine the geographic coordinates corresponding to each horizontal control position according to the transformation relationship between the CFD model coordinate system and the real geographic coordinate system, and set up virtual sites accordingly. Secondly, select WRF horizontal grid points in the WRF computation domain and deploy virtual stations at the selected horizontal grid points. Construct a local Cartesian coordinate system for the CFD computation domain based on the selected horizontal grid points and their spatial arrangement. Use the corresponding coordinates of the selected horizontal grid points in the local Cartesian coordinate system as the CFD driving position.

[0011] Furthermore, the method of establishing the spatial correspondence between the virtual site and the CFD drive position also includes: checking the deviation between the set virtual site position and the virtual site position actually read and used by the WRF. When the horizontal distance corresponding to the deviation is not greater than the preset horizontal error threshold, the spatial correspondence is considered to meet the requirements; otherwise, the deployment position of the virtual site needs to be re-determined.

[0012] Furthermore, when the virtual site deployment location is represented by latitude and longitude coordinates, the storage accuracy of the virtual site coordinates is improved by modifying the WRF source code, so that the spatial correspondence between the set virtual site location and the virtual site location actually read and used by WRF meets the preset requirements.

[0013] Further, in step 300, configuring the WRF mode time series site output module tslist according to the virtual site deployment location includes: writing the name, identifier and location information of each virtual site into the tslist file, and completing the corresponding configuration in the namelist.input file. The corresponding configuration includes setting the upper limit of the number of virtual sites, the size of the time series output buffer, the maximum number of vertical layers of the output profile data, and enabling decrossing wind vector output.

[0014] Further, in step 400, the spatial range of the CFD computational domain, the dominant inflow direction, the inflow boundary, the outflow boundary, and other boundary conditions are determined based on the CFD model's driving method, the location of the virtual stations, and atmospheric state variable profile data, thereby constructing a CFD model of the region to be analyzed, including the following steps: Step 401: The spatial range of the CFD computing domain can be determined based on the driving method of the CFD model and the deployment location of the virtual sites. Step 402: Extract atmospheric state variable profile data at the corresponding locations of the four side boundaries of the CFD computational domain. Calculate the unit width flow time series in the U and V directions within the WRF model coordinate system during the analysis period, where U represents west to east and V represents south to north. Based on the absolute value, positive and negative directions, and temporal variation characteristics of the unit width flow in the U and V directions, determine the dominant inflow direction of the CFD computational domain, and accordingly determine the inflow boundary, outflow boundary, and other boundary condition settings of the CFD computational domain.

[0015] Furthermore, the target variable is any one of wind vector, temperature, pressure, water vapor, turbulence-related variables, or other atmospheric state variables that can be used to construct CFD driving data.

[0016] Therefore, the high-frequency sampling WRF-CFD coupled wind field simulation method described above has the following beneficial effects: 1. The tslist module, which is conventionally used only to extract time series simulation results from a small number of actual meteorological stations, wind towers, or observation points, is extended to a high-frequency sampling method in WRF-CFD coupling. Multiple virtual stations are deployed in the geographical locations corresponding to the area to be analyzed according to the CFD driving requirements. Compared with existing methods, it can directly obtain atmospheric state variable profile data corresponding to the CFD driving location during WRF operation, enabling mesoscale meteorological field information to be transmitted to the CFD model in a more targeted manner, thus expanding the tslist module as a data sampling method in WRF-CFD coupling. 2. By using the tslist module to output atmospheric state variable profile data at virtual site locations according to the WRF mode time step interval, the construction of CFD-driven data no longer relies on shortening the save interval of the wrfout file. Furthermore, since the tslist module only outputs profile data at the site location rather than a full-field snapshot, it can obtain complete temporal evolution information of the WRF mode integration process at the corresponding location while significantly reducing the data storage and read / write burden caused by the high-frequency output of the global variable field in the wrfout file. 3. The virtual site deployment, tslist high-frequency profile output, profile data mapping, and driving data processing are linked together into a complete meso-microscale coupling technology link. The tslist module outputs atmospheric state variable profile data at WRF model time step intervals, so that the WRF model integration process information at the corresponding location in the area to be analyzed can enter the CFD driving data construction process in the form of high-frequency time series profiles. This makes the construction of CFD driving data no longer limited to the discrete instantaneous field provided by the preset saving time of the wrfout file, but can utilize richer mesoscale time series evolution information, thereby significantly improving the time-varying fidelity of the mesoscale meteorological field transferred to the CFD model.

[0017] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0018] Figure 1 This is a flowchart of a high-frequency sampling WRF-CFD coupled wind field simulation method according to the present invention. Detailed Implementation

[0019] The following detailed description of embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely illustrates selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0020] Please see Figure 1 A high-frequency sampling WRF-CFD coupled wind field simulation method includes the following steps: Step 100: Determine the unidirectional coupling data requirements based on the region to be analyzed, and determine the WRF computation domain and the CFD computation domain; First, the region to be analyzed for wind field simulation is determined. This region is the actual geographical area of ​​interest in this method, such as areas used for wind resource assessment, wind disaster risk assessment, engineering wind environment analysis, or low-altitude operational wind environment assessment. The region to be analyzed differs from the WRF or CFD computational domains used in this method; rather, it is the higher-level analysis object served by both. The WRF computational domain is used to simulate mesoscale meteorological processes in and around the region to be analyzed, while the CFD computational domain is used to conduct more refined wind field simulations within or within the region to be analyzed.

[0021] Based on the region to be analyzed, the unidirectional coupling data requirements are further determined, including the driving method of the CFD model and the WRF output data required to construct the CFD driving data. The driving methods of the CFD model include boundary condition driving, internal source term perturbation driving, or assimilation constraint driving.

[0022] It should be noted that the driving method of CFD models and the WRF output data required to construct CFD driving data generally use boundary condition driving methods (such as lateral boundary driving) coupled with wind vector data. These methods can also be adjusted according to the user's technical capabilities and available computing resources. Using unconventional methods other than boundary condition driving methods has technical barriers and unique advantages. For example, the homogenization constraint driving method combined with periodic boundaries has the advantages of not needing to fix the inflow boundary and being able to adapt to wind direction changes, making it suitable for long-term simulations over multiple days, but it has higher technical requirements. In addition, data coupling can couple more variables according to computing resources, but this will inevitably lead to greater computing power requirements.

[0023] In this embodiment, the boundary condition driven method is used as an example for explanation, that is, the WRF output data is processed into time-varying boundary conditions on the target boundary of the CFD computation domain. For the internal source term perturbation driven method or the assimilation constraint driven method, the data acquisition, coordinate mapping and data processing logic are similar to the boundary condition driven method. The difference is that the driving data application position is changed from the target boundary of the CFD computation domain to the internal region of the target of the CFD computation domain.

[0024] After determining the region to be analyzed and the unidirectional coupling data requirements, the WRF computation domain and CFD computation domain are further determined as follows: When determining the WRF computational domain, it is determined layer by layer from the innermost high-resolution nested domain outwards. The outer nested domains are set using conventional methods, while only the innermost nested domain is determined based on coupling requirements. First, the CFD computational domain is initially designed based on experience, considering the region to be analyzed and available computing resources, determining the approximate flow length and spanwise width. When the wind direction during the analysis period is known, the innermost nested domain of the WRF can be determined based on the corresponding geographical range of the CFD flow field inlet boundary. In this case, the innermost nested domain of the WRF is set to completely cover this geographical range, and each side boundary extends at least 20 grid points outwards to prevent boundary effects. When the wind direction during the analysis period is unknown, the innermost nested domain of the WRF is determined using the following method: Considering all possible wind directions, the formula for calculating the distance from the center of the region to be analyzed in the CFD computational domain to the farthest point on the flow field inlet boundary (i.e., the inlet corner) is as follows: ; in, The flow direction distance from the center of the region to be analyzed in the CFD computational domain to the inlet boundary; The span of the CFD computational domain; This is the distance from the center of the region to be analyzed in the CFD computational domain to the farthest point on the flow field inlet boundary.

[0025] To ensure that the innermost nested domain of the WRF can cover the corresponding geographical area of ​​the CFD flow field inlet boundary regardless of wind direction, the geographical location of the center of the region to be analyzed is considered. Center of the circle Create a circular region with a radius of [radius value]. To prevent negative impacts from boundary effects, extend the circular region outwards by at least 20 additional physical distances corresponding to the innermost nested domain grid points of the WRF. That is, the horizontal geographic range that the innermost nested domain of WRF must cover is... With center and radius as A circular area.

[0026] The innermost nested domain in WRF is set to the geographic location of the center and the center of the region to be analyzed. For overlapping rectangles, their horizontal boundaries should extend from the center outwards to the east, west, south, and north at least [length missing]. distance.

[0027] Step 200: Determine the location and spatial correspondence of the virtual sites; Based on the unidirectional coupling data requirements determined in step 100, the locations of multiple virtual stations corresponding to the region to be analyzed are determined. These virtual stations are high-frequency sampling points deployed according to the meso-microscale coupling requirements, and are not actual observation stations. When using a boundary condition-driven approach, virtual stations can be deployed along the geographical locations corresponding to the target boundary of the CFD computational domain; when using an internal source term perturbation-driven approach or an assimilation constraint-driven approach, virtual stations can be deployed at geographical locations corresponding to the internal regions of the target within the CFD computational domain. The number and spacing of each virtual station are determined according to the requirements, mainly depending on the CFD computational domain, the driving method of the CFD model, and the required spatial resolution.

[0028] Specifically, the spatial correspondence between virtual sites and CFD drive locations can be established using one of the following two methods: One approach, adopted in this embodiment, involves first determining the target boundary or internal region of the CFD computation domain, and then selecting several horizontal control positions on the target boundary or internal region for applying driving data. Each horizontal control position corresponds to a vertical driving profile in the CFD model. Subsequently, based on the transformation relationship between the CFD model coordinate system and the real geographic coordinate system, the latitude and longitude coordinates corresponding to each horizontal control position are determined, and virtual stations are set accordingly. For this approach, it is necessary to verify the deviation between the set virtual station positions and the virtual station positions actually read and used by WRF. When the horizontal distance corresponding to this deviation is not greater than a preset horizontal error threshold, the virtual station placement position can be considered to meet the spatial correspondence requirement with the CFD driving position; otherwise, the placement position of the virtual station needs to be redefined. The preset horizontal error threshold can be determined based on the virtual station placement spacing, the CFD model background grid scale, and the coupling simulation accuracy requirements; in this embodiment, the preset horizontal error threshold is taken as no greater than half of the CFD model background grid scale. The second method involves first selecting one or more WRF horizontal grid points in the WRF computation domain and then deploying virtual stations at the selected horizontal grid points. Subsequently, based on the selected horizontal grid points and their spatial arrangement, a local Cartesian coordinate system for the CFD computation domain is constructed, and the corresponding coordinates of the selected horizontal grid points in this local Cartesian coordinate system are used as the horizontal control positions of the target boundary or the internal region of the target in the CFD computation domain, thereby establishing a spatial correspondence between the virtual stations and the CFD horizontal control positions.

[0029] The location of virtual stations can be represented by latitude and longitude coordinates or WRF grid index. When representing the location of virtual stations by latitude and longitude coordinates, the WRF source code can be further modified to improve the storage accuracy of the virtual station coordinates, so that the spatial correspondence between the set virtual station location and the virtual station location actually read and used by WRF meets the preset requirements. In this embodiment, by modifying the wrf_tsin.F file, the storage accuracy of the virtual station coordinates is changed from the default... Upgraded to .

[0030] Step 300: Set up a virtual site and perform WRF regional weather simulation; Specifically, after determining the deployment locations of the virtual sites in step 200, the time series site output module tslist in WRF mode is configured. Specifically, the name, identifier, and location information of each virtual site are written into the tslist file, and the corresponding configurations are completed in the namelist.input file. These configurations include setting the upper limit for the number of virtual sites, the size of the time series output buffer, the maximum number of vertical layers for output profile data, and enabling decrossing wind vector output.

[0031] Subsequently, regional meteorological simulations using WRF are conducted. During the integration process, the WRF model outputs atmospheric state variable profile data at the virtual station locations at WRF model time step intervals, forming complete time-series data. Atmospheric state variables include wind vectors, potential temperature, geopotential height, water vapor mixing ratio, and air pressure, which are output by default from the tslist module. Other variables (such as turbulent kinetic energy) can also be included by modifying the WRF source code to extend the output.

[0032] Existing WRF-CFD coupling methods typically extract instantaneous variable fields at preset save times from WRF historical output files (wrfout) and then convert them into driving conditions for the CFD model. Since the wrfout save interval is usually several minutes to tens of minutes, these methods struggle to fully utilize the temporal evolution information of wind fields and related atmospheric variable fields during WRF model integration. If the wrfout save interval is shortened to improve temporal resolution, high-frequency saving of a large range of multivariable fields within the WRF computational domain is required, resulting in a significant data storage and read / write burden. In contrast, this invention utilizes the tslist module to output atmospheric state variable profile data at virtual site locations at WRF model time step intervals, eliminating the need to shorten the wrfout save interval as a prerequisite for constructing CFD driving data. Furthermore, since the tslist module only outputs profile data at the site location rather than a full-field snapshot, it can obtain complete temporal evolution information of the WRF model integration process at the corresponding location while reducing the data storage and read / write burden caused by the high-frequency output of global variable fields from wrfout.

[0033] Step 400: Construct a CFD model of the region to be analyzed; Based on the requirements for refined wind field simulation of the area to be analyzed, a CFD model is constructed for the area or a local area thereof. During the model construction process, the spatial extent of the CFD computational domain is determined according to the driving mode of the CFD model determined in step 100 and the location of the virtual stations determined in step 200. On this basis, combined with the atmospheric state variable profile data obtained in step 300, the wind speed and direction characteristics within the CFD computational domain are preliminarily analyzed in the WRF model coordinate system, and the dominant incoming flow direction, inflow boundary, outflow boundary, and other boundary condition settings of the CFD computational domain are determined.

[0034] In this embodiment, the wind speed and direction characteristic analysis employs a unit width flow analysis method in two horizontally orthogonal directions. Specifically, atmospheric state variable profile data are extracted from the corresponding positions of the four side boundaries of the CFD computational domain. The unit width flow time series in the U and V directions are calculated respectively within the WRF model coordinate system during the analysis period, where U represents west to east and V represents south to north. Subsequently, based on the absolute value, positive and negative directions, and temporal variation characteristics of the unit width flow in the U and V directions, the main incoming flow direction of the CFD computational domain during the analysis period is determined, and the inflow boundary, outflow boundary, and other boundary conditions of the CFD computational domain are determined accordingly.

[0035] Step 500: Map the atmospheric state variable profile data to the raw CFD driving data; Based on the unidirectional coupling data requirements determined in step 100, target variable data corresponding to the target variables required for constructing CFD driving data is extracted from the atmospheric state variable profile data obtained in step 300. The extraction of target variable data is implemented using a written Python post-processing program. In this embodiment, the target variable data is wind vector profile data, including wind speed components in three orthogonal directions, namely U, V, and W. Specifically, the program reads the .UU, .VV, and .WW files output by tslist according to the virtual site number and target time, and retrieves the data records corresponding to the target time in the files, extracting the U, V, and W wind speed components at each model vertical layer at that time. Subsequently, the wind speed components in the three directions at the same virtual site and the same target time are combined in strata order to obtain the wind vector profile data at the virtual site. Thus, the target variable data required for constructing CFD driving data is formed.

[0036] Subsequently, based on the spatial correspondence established in step 200 and the CFD computation domain determined in step 400, the target variable data is converted from the WRF model coordinate system to the raw CFD driving data in the CFD model coordinate system. For the vertical coordinates, since WRF uses terrain following... coordinates or hybrid - For pressure vertical coordinates, the actual altitude corresponding to the same model layer changes with horizontal position and time. Therefore, the wrrf-python library is used to read the height auxiliary information required for vertical coordinate transformation from the wrrfout file at the corresponding or nearby time, and the WRF model layer height at the virtual site is converted to ground altitude to correspond to the vertical coordinates of the CFD model. For horizontal coordinates, based on the name, identifier, and / or location information of the virtual site, and according to the spatial correspondence established in step 200, the coordinate values ​​of the virtual site in the horizontal coordinate system of the CFD model are determined by querying the correspondence table. Subsequently, the wind vector profile data of the virtual site output by tslist is organized into the above CFD model coordinate system to form the raw CFD driving data in the CFD model coordinate system.

[0037] It should be noted that, in addition to the wind vector in this embodiment, the target variable may also include temperature, pressure, water vapor, turbulence-related variables, or other atmospheric state variables that can be used to construct CFD driving data, depending on the specific calculation requirements.

[0038] Step 600: Process the raw CFD driving data to form CFD driving data; Based on the computational settings, driving requirements, and wind field construction requirements of the CFD model, the raw CFD driving data in the CFD model coordinate system is processed to form CFD driving data used to drive the CFD model calculations. The purpose of this step is to adapt the raw CFD driving data to the input requirements of the CFD model in terms of temporal organization, spatial distribution, and data format, and to perform averaging, smoothing, disturbance overlay, or other wind field construction processing according to specific simulation needs.

[0039] In this embodiment, a boundary condition-driven approach is adopted. The raw CFD driving data obtained in step 500 is organized into the input file format required by the timeVaryingMappedFixedValue boundary condition in OpenFOAM using a written Python program. This includes the coordinates of the boundary data points in the CFD model coordinate system and the corresponding wind vector profile data at multiple time points. The spatial mapping method is specified as planar interpolation in the boundary condition settings.

[0040] In other implementations, the raw CFD driving data can be processed into different forms of CFD driving data according to application requirements. For example, it can be averaged, smoothed, or otherwise processed to form fixed driving data for representative moments or representative time periods; it can also be processed into momentum source terms, thermal source terms, or other source term forms applied to a specified region of the CFD computation domain according to the requirements of the internal source term perturbation driving method; it can also be processed into a reference field, target profile, or constraint term used to constrain the CFD simulation results according to the requirements of the assimilation constraint driving method.

[0041] It should be noted that, depending on the application scenario, data processing is not limited to the above-mentioned formatting, spatial mapping settings, and interpolation mapping. Under different CFD models, different driving methods, and different wind field simulation requirements, data processing can also include data denoising, smoothing filtering, regional averaging, outlier removal, turbulence disturbance superposition, spatial surface fitting, numerical correction based on preset mapping relationships, and processing using machine learning models, or a combination of the above methods.

[0042] Step 700: Conduct microscale CFD wind field simulation based on CFD-driven data.

[0043] The CFD driving data generated in step 600 is input into the CFD model to drive the CFD model to perform microscale wind field simulation of the region to be analyzed. In this embodiment, the CFD driving data is used as the data source for the time-varying boundary conditions of the CFD model. During the integration process, the CFD solver calls the timeVaryingMappedFixedValue boundary condition, reads wind vector profile data at multiple input times, and applies the CFD driving data to the target boundary of the CFD computational domain according to the current calculation time and the preset spatial mapping method, so that the wind vector profile data on the target boundary of the CFD computational domain is updated over time.

[0044] Since the time resolution of the virtual site output is equal to the time step of the WRF mode, which is much higher than the saving frequency of regular wrfout files (several minutes to tens of minutes), it can provide high time resolution driving timing, better capture the rapid changes in mesoscale wind fields, and effectively reduce the errors introduced by time interpolation.

[0045] In other implementations, the CFD model can perform calculations based on the specific form of the CFD driving data formed in step 600. When the CFD driving data is in the form of time-invariant driving data, it can be used for wind field simulation under fixed driving conditions; when the CFD driving data is in the form of internal source terms, it can be used as source terms in the CFD governing equations; when the CFD driving data is in the form of assimilation constraints, it can be used to constrain the CFD simulation results to approximate the mesoscale background process given by the WRF model or the reference field constructed from it. After the CFD calculation is completed, the microscale wind field simulation results of the area to be analyzed, the wind impact results of the target object, and the application evaluation results obtained after further post-processing can be output according to the application requirements.

[0046] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A high-frequency sampling WRF-CFD coupled wind field simulation method, characterized in that, Includes the following steps: Step 100: Determine the unidirectional coupling data requirements based on the region to be analyzed, and determine the WRF computation domain and CFD computation domain; the unidirectional coupling data requirements include the driving method of the CFD model and the WRF output data required to construct the CFD driving data. Step 200: Based on the unidirectional coupling data requirements, determine the deployment locations of multiple virtual sites and establish the spatial correspondence between the virtual sites and the CFD drive locations, including: Establish the spatial correspondence between virtual sites and CFD drive locations using any of the following methods: First, select horizontal control positions on the target boundary or internal area of ​​the CFD computation domain to apply driving data, determine the geographic coordinates corresponding to each horizontal control position according to the transformation relationship between the CFD model coordinate system and the real geographic coordinate system, and set up virtual sites accordingly. Secondly, select WRF horizontal grid points in the WRF computation domain and place virtual stations at the selected horizontal grid points. Construct a local Cartesian coordinate system of the CFD computation domain based on the selected horizontal grid points and their spatial arrangement. Use the corresponding coordinates of the selected horizontal grid points in the local Cartesian coordinate system as the CFD driving position. Step 300: Configure the WRF model time series station output module tslist according to the virtual station deployment location, and conduct WRF regional meteorological simulation. Output atmospheric state variable profile data at the virtual station location according to the WRF model time step interval, including: Write the name, identifier, and location information of each virtual site into the tslist file, and complete the corresponding configuration in the namelist.input file. The corresponding configuration includes setting the upper limit of the number of virtual sites, the size of the time series output buffer, the maximum number of vertical layers of the output profile data, and enabling decrossing wind vector output. Step 400: Based on the CFD model's driving method, the location of virtual stations, and atmospheric state variable profile data, determine the spatial extent of the CFD computational domain, the dominant inflow direction, inflow boundary, outflow boundary, and other boundary condition settings, thereby constructing a CFD model of the region to be analyzed, including: Step 401: The spatial range of the CFD computing domain can be determined based on the driving method of the CFD model and the deployment location of the virtual sites. Step 402: Extract atmospheric state variable profile data at the corresponding positions of the four side boundaries of the CFD computational domain. Calculate the unit width flow time series in the U and V directions within the WRF model coordinate system during the analysis period, where U represents west to east and V represents south to north. Based on the absolute value, positive and negative directions, and time variation characteristics of the unit width flow in the U and V directions, determine the dominant inflow direction of the CFD computational domain, and accordingly determine the inflow boundary, outflow boundary, and other boundary condition settings of the CFD computational domain. Step 500: Extract the target variable data required for constructing CFD driving data from the atmospheric state variable profile data through the set program. The target variable data corresponding to different target variables are converted from the WRF model coordinate system to the CFD model coordinate system as CFD raw driving data through a preset method. Step 600: Process the raw CFD driving data to form CFD driving data used to drive the CFD model calculation. Step 700: Input the CFD driving data into the CFD model to drive the CFD model to perform microscale wind field simulation of the area to be analyzed.

2. The high-frequency sampling WRF-CFD coupled wind field simulation method according to claim 1, characterized in that, The driving methods of the CFD model include boundary condition driving method, internal source term perturbation driving method, or assimilation constraint driving method.

3. The high-frequency sampling WRF-CFD coupled wind field simulation method according to claim 2, characterized in that, In step 100, the WRF computation domain includes an inner nested domain and an outer nested domain. The innermost nested domain of the WRF needs to cover the corresponding geographical range of the CFD flow field inlet boundary.

4. The high-frequency sampling WRF-CFD coupled wind field simulation method according to claim 3, characterized in that, In step 200, determining the deployment locations of multiple virtual sites based on the unidirectional coupling data requirements includes: When the CFD model is driven by boundary conditions, virtual sites are deployed along the geographical locations corresponding to the target boundary of the CFD computation domain. When the CFD model is driven by internal source term perturbation or assimilation constraint, virtual sites are deployed in the geographical locations corresponding to the target area within the CFD computation domain.

5. The high-frequency sampling WRF-CFD coupled wind field simulation method according to claim 4, characterized in that, The method of establishing the spatial correspondence between the virtual site and the CFD drive position also includes: checking the deviation between the set virtual site position and the virtual site position actually read and used by the WRF. When the horizontal distance corresponding to the deviation is not greater than the preset horizontal error threshold, the spatial correspondence is considered to meet the requirements; otherwise, the deployment position of the virtual site needs to be re-determined.

6. The high-frequency sampling WRF-CFD coupled wind field simulation method according to claim 5, characterized in that, When the virtual site deployment location is represented by latitude and longitude coordinates, the storage accuracy of the virtual site coordinates is improved by modifying the WRF source code, so that the spatial correspondence between the set virtual site location and the virtual site location actually read and used by WRF meets the preset requirements.

7. The high-frequency sampling WRF-CFD coupled wind field simulation method according to claim 6, characterized in that, The target variable is any one of the following: wind vector, temperature, pressure, water vapor, turbulence-related variables, or other atmospheric state variables that can be used to construct CFD driving data.

Images