Urban low-altitude micro-scale wind field prediction method and device and electronic equipment

Abstract

The invention relates to the technical field of low-altitude wind field prediction. The invention provides an urban low-altitude micro-scale wind field prediction method and device and electronic equipment. Comprising the steps of obtaining low-altitude wind field data corresponding to a target area; dividing the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the data volume of the first low-altitude wind field data is smaller than that of the second low-altitude wind field data; acquiring wind field simulation data of the target area, and inputting the low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data; and inputting the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data. According to the method, rapid prediction and high-precision prediction of the micro-scale wind field of the target area are realized, so that key data support is provided for route planning and safe flight of a low-altitude aircraft.

Description

Technical Field

[0001] The present invention relates to the technical field of low-altitude wind field prediction, and in particular to a method, device and electronic equipment for predicting urban low-altitude micro-scale wind fields. Background Art

[0002] With the rapid development of the low-altitude economy, low-altitude aircraft (especially manned aircraft) are playing an increasingly important role in urbanization, and their safe operation has become a focus of public attention. Urban low-altitude airspace, the primary space for the development of unmanned aerial vehicles and the low-altitude economy, is most significantly impacted by the meteorological factor known as wind, which is known to be highly variable. Therefore, accurately predicting wind speed and direction is crucial for ensuring low-altitude air routes.

[0003] As a key factor influencing aircraft route planning and flight safety, the complexity and variability of urban low-altitude wind fields pose a severe challenge to the stable operation of aircraft. Compared to high-altitude wind environments, urban low-altitude wind fields are influenced by multiple factors, including urban structure, building distribution, heat emissions, and land use types. They are also affected by the combined effects of atmospheric circulation and various local circulations, resulting in extremely complex wind field characteristics. Specifically, they are characterized by variable and uneven incoming flow directions, widespread local wind phenomena, strong turbulence in the wind field, and significant influence from the coupling of multiple physical field factors.

[0004] Currently, research on the observation and forecasting of complex urban wind fields is primarily focused on the mesoscale range, and its spatial resolution is far from sufficient to resolve the high-precision wind field information required by low-altitude aircraft. While the mesoscale meteorological model WRF plays an important role in weather forecasting, its spatial resolution is typically on the order of kilometers, which cannot meet the flight safety forecast requirements of low-altitude aircraft at scales of hundreds or even tens of meters. Furthermore, while the products currently available on the market that can test low-altitude complex wind fields, such as weather radars, anemometers, and six-component small weather stations, can accurately obtain wind field information, the spatial resolution of the data obtained is poor (due to a limited number of deployments) and the temporal resolution is also low, making it difficult to achieve high-precision analysis of low-altitude complex wind fields.

[0005] Therefore, in order to address the problem of insufficient temporal and spatial accuracy in low-altitude wind field prediction, a method for predicting urban low-altitude micro-scale wind fields is developed, which has important practical significance and application value for ensuring the route planning and flight safety of low-altitude aircraft. Summary of the Invention

[0006] The present invention provides a method, device and electronic equipment for predicting urban low-altitude micro-scale wind fields, which are used to solve the defects of insufficient time and space accuracy in low-altitude wind field prediction in the existing technology, thereby ensuring more accurate route planning of low-altitude aircraft.

[0007] The present invention provides a method for predicting urban low-altitude micro-scale wind fields, comprising:

[0008] Obtain low-altitude wind field data corresponding to the target area;

[0009] Dividing the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the amount of the first low-altitude wind field data is less than the amount of the second low-altitude wind field data;

[0010] Acquire wind field simulation data of the target area; the wind field simulation data is data obtained by simulating the target area using a nested coupling mode of a mesoscale model WRF and CFD;

[0011] Inputting the first low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data;

[0012] Inputting the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data;

[0013] A wind field prediction result for the target area is determined according to the second wind field prediction data.

[0014] According to the urban low-altitude micro-scale wind field prediction method provided by the present invention, obtaining low-altitude wind field data corresponding to the target area includes:

[0015] The wind field data of the target area is obtained using a weather radar, anemometer or a six-component small weather station; the wind field data includes: wind speed and wind direction.

[0016] According to the urban low-altitude micro-scale wind field prediction method provided by the present invention, obtaining the wind field simulation data of the target area includes:

[0017] Obtaining a three-dimensional entity model of a building complex corresponding to the target area;

[0018] Using the mesoscale model WRF to simulate the low-altitude wind field data of the target area to obtain an initial wind field simulation result;

[0019] The initial wind field simulation result is used as the boundary condition of CFD simulation, and the three-dimensional solid model of the building complex is subjected to secondary simulation based on the boundary condition, thereby obtaining wind field simulation data of the target area.

[0020] According to the urban low-altitude micro-scale wind field prediction method provided by the present invention, obtaining the three-dimensional entity model of the building complex corresponding to the target area includes:

[0021] Obtain GIS building vector data and terrain elevation DEM data of the target area; the vector data includes the planar location, outline, floor area, and height of the building;

[0022] A three-dimensional entity model of the building complex is generated based on the GIS building vector data and the terrain elevation DEM data.

[0023] According to the urban low-altitude micro-scale wind field prediction method provided by the present invention, the first wind field prediction model and the second wind field prediction model are constructed using an XGBoost model, a CNN model, a LSTM model, a Transformer model, a LightGBM model, or a CatBoost model.

[0024] According to the urban low-altitude micro-scale wind field prediction method provided by the present invention, before dividing the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, the method further includes:

[0025] The acquired low-altitude wind field data corresponding to the target area are preprocessed, and the preprocessing includes data cleaning to remove outliers and noise in the data; and data normalization to unify wind field data of different magnitudes into the same numerical range.

[0026] The present invention also provides a device for predicting urban low-altitude micro-scale wind fields, comprising:

[0027] An acquisition unit, used to acquire low-altitude wind field data corresponding to a target area;

[0028] a dividing unit, configured to divide the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the amount of the first low-altitude wind field data is smaller than the amount of the second low-altitude wind field data;

[0029] The acquisition unit is further configured to acquire wind field simulation data of the target area; the wind field simulation data is data obtained by simulating the target area using a nested coupling mode of a mesoscale model WRF and CFD;

[0030] a prediction unit, configured to input the first low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data;

[0031] The prediction unit is further configured to input the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data;

[0032] A determination unit is configured to determine a wind field prediction result for the target area based on the second wind field prediction data.

[0033] The urban low-altitude micro-scale wind field prediction method, device and electronic equipment provided by the present invention achieve rapid prediction of the micro-scale wind field in the target area by using the data obtained by simulating the target area through the nested coupling mode of the mesoscale model WRF and CFD, and the first low-altitude wind field data corresponding to the target area as the input of the first wind field prediction model; and then use the first wind field prediction data and the second low-altitude wind field data as the input of the second wind field prediction model to achieve high-precision prediction of the micro-scale wind field in the target area, thereby providing key data support for the route planning and safe flight of low-altitude aircraft. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is one of the flow charts of the urban low-altitude micro-scale wind field prediction method provided by the present invention;

[0035] Figure 2 Flowchart 2 of the urban low-altitude micro-scale wind field prediction method provided by the present invention;

[0036] Figure 3 This is a schematic diagram of the process of modeling and simulating a micro-scale wind farm in an urban area according to an embodiment of the present invention;

[0037] Figure 4 This is a schematic diagram of the structure of the urban low-altitude micro-scale wind field prediction device provided by the present invention;

[0038] Figure 5 This is a schematic structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0039] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0040] Figure 1 This is one of the flow charts of the urban low-altitude micro-scale wind field prediction method provided by the present invention, Figure 2 The present invention provides a flow chart of the urban low-altitude micro-scale wind field prediction method. Figure 1 As shown, the method includes:

[0041] Step 101: Obtain low-altitude wind field data corresponding to the target area.

[0042] Specifically, the target area of ​​the present invention is the area where urban low-altitude wind farms are located. Wind field data for these areas is based on existing sparse meteorological observation data. Specifically, core wind field parameters such as wind speed and direction in the target area are directly acquired using existing meteorological observation equipment (such as weather radar, anemometers, and six-component small weather stations). The data acquired by these devices has low temporal and spatial resolution and is limited by factors such as the number of low-altitude airspace deployments and operational costs, resulting in insufficient resolution for complex low-altitude wind fields. However, the data is highly reliable and can reflect the true characteristics of the local wind field.

[0043] Step 102: Divide the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the data volume of the first low-altitude wind field data is smaller than the data volume of the second low-altitude wind field data.

[0044] Specifically, this application divides the data into "a small amount of sparse data (first low-altitude wind field data)" and "a larger amount of data (second low-altitude wind field data)", and uses these two parts of data as input to the first wind field prediction model and the second wind field prediction model respectively, so that the first low-altitude wind field data can be used to preliminarily integrate simulation information to cover the overall situation, thereby improving the prediction efficiency of the first wind field prediction model; at the same time, the second low-altitude wind field data is used to fine-tune the prediction accuracy of the model.

[0045] Step 103: Acquire wind field simulation data of the target area; the wind field simulation data is data obtained by simulating the target area using a nested coupling mode of a mesoscale model WRF and CFD.

[0046] Specifically, current numerical simulations struggle to accurately determine weather boundary conditions, resulting in significant deviations between simulation results and actual data. Furthermore, simulations lack effective integration with meteorological test data. Numerical simulation models used for weather forecasting are computationally intensive when used to scale to the space required by low-altitude aircraft, making timeliness difficult to guarantee. This application utilizes the mesoscale model WRF to provide a kilometer-scale circulation background (also accounting for the influence of physical processes such as water vapor, long- and short-wave radiation, cumulus clouds, and the underlying surface). Computational Fluid Dynamics (CFD) offers the advantages of high spatiotemporal resolution and the ability to create highly accurate models of realistic buildings and terrain. By combining the advantages of both, this application achieves more accurate and refined wind field information. Specifically, WRF provides realistic boundary conditions for CFD, which in turn outputs high-precision wind field data in a core area (e.g., 2 km x 2 km) to construct a high-resolution wind field database covering the target area.

[0047] In addition, the data obtained by simulating the target area using the nested coupling mode of the mesoscale model WRF and CFD can cover areas where sensors are not deployed, thereby filling the spatial gaps in the observation data; it also realizes wind field details at the meter scale (such as flow around buildings and the influence of terrain undulations), thereby improving the ability to predict complex wind field characteristics. The accurate low-altitude micro-scale wind field prediction results can directly serve the route planning of low-altitude aircraft, helping aircraft to avoid dangerous wind environments (such as strong turbulence, sudden gusts, etc.), reduce the risk of flight accidents, and ensure the operation safety of manned aircraft.

[0048] Step 104: Input the first low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data.

[0049] Specifically, the first wind farm prediction model can be constructed using a machine learning method such as an XGBoost model, a CNN model, an LSTM model, a Transformer model, a LightGBM model, or a CatBoost model. In the embodiment of the present invention, the first wind farm prediction model constructed by the XGBoost model is used as an example for description.

[0050] The present invention utilizes the XGBoost model as the first wind field prediction model to fuse the first low-altitude wind field data with the wind field simulation data, thereby achieving an effective fusion of the authenticity of the observation data and the high resolution of the simulation data, so that the observation data calibrates the deviation of the simulation data, and the simulation data supplements the spatial coverage of the observation data, thereby preliminarily reconstructing the global wind field characteristics of the target area.

[0051] Step 105: Input the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data.

[0052] Specifically, the second wind field prediction model of the present application is based on the first wind field prediction data and inputs a large amount of second low-altitude wind field data, so that the second wind field prediction model can further correct the error of the wind field data, thereby improving the prediction accuracy of the wind field in the target area.

[0053] Step 106: Determine a wind field prediction result for the target area based on the second wind field prediction data.

[0054] Specifically, the second wind field prediction data includes parameters such as wind speed and wind direction at any location in the target area, and can be directly used to reflect the spatiotemporal distribution characteristics of the wind field, such as turbulence intensity and wind profile changes.

[0055] The urban low-altitude micro-scale wind field prediction method provided by the present invention realizes rapid prediction of the micro-scale wind field in the target area by using the data obtained by simulating the target area through the nested coupling mode of the mesoscale model WRF and CFD, and the low-altitude wind field data corresponding to the target area as the input of the first wind field prediction model; and then uses the first wind field prediction data and the second low-altitude wind field data as the input of the second wind field prediction model to realize high-precision prediction of the micro-scale wind field in the target area.

[0056] Furthermore, the present invention obtains wind field simulation data of the target area, including: obtaining a three-dimensional solid model of the building complex corresponding to the target area; using the mesoscale model WRF to simulate the low-altitude wind field data of the target area to obtain an initial simulation result; using the initial simulation result as the boundary condition of the CFD simulation, and performing a secondary simulation on the three-dimensional solid model of the building complex based on the boundary condition, thereby obtaining the wind field simulation data of the target area.

[0057] Specifically, first, a three-dimensional solid model of the building complex corresponding to the target area is collected and constructed. This process requires the use of professional surveying and mapping technology, geographic information system (GIS) data, and architectural design materials to ensure that the model can highly realistically present the actual form, relative position, and spatial distribution of all buildings in the target area, laying the foundation for subsequent accurate simulation of the interaction between the wind field and the buildings. In this embodiment of the present invention, the three-dimensional solid model of the building complex is obtained through the following methods:

[0058] First, obtain the GIS building vector data and terrain elevation DEM data of the target area; the vector data contains the plane position, outline, floor area, and height of the building; and generate a three-dimensional solid model of the building complex based on the GIS building vector data and terrain elevation DEM data.

[0059] Secondly, the mesoscale model WRF was used to simulate low-altitude wind field data in the target area, generating low-resolution initial simulation results. The mesoscale model WRF is a numerical simulation tool widely used in atmospheric science. It considers various physical processes and dynamic mechanisms of atmospheric motion. By setting and calculating relevant meteorological parameters, it generates macroscopic simulation results of the low-altitude wind field in the target area. These results reflect the trends and characteristics of wind field variations at a larger spatial scale.

[0060] The initial wind field simulation results are then used as boundary conditions for computational fluid dynamics (CFD) simulations. CFD simulations focus on studying the detailed flow of fluids (in this embodiment, air) within complex local geometric structures (i.e., buildings). Inputting the initial wind field simulation results into the CFD model as boundary conditions acts as an external atmospheric constraint for the CFD simulation, enabling it to accurately simulate the flow of air around the building complex while taking into account the macroscopic meteorological context.

[0061] Finally, a CFD simulation is performed on the three-dimensional solid model of the building complex based on the set boundary conditions. By numerically solving the fluid dynamics equations, the physical processes of air flow, turbulence generation, and diffusion over and around the building surfaces are simulated. This provides detailed wind field distribution information for the target area, which constitutes the wind field simulation data for the target area.

[0062] The following combination Figure 3 , the solution provided by the present invention is further elaborated:

[0063] First, this application obtains Geographic Information System (GIS) building vector data and terrain elevation DEM (Digital Elevation Model) data for the target area. The vector data includes the planar location, outline, floor area, and height of the buildings, while the DEM data reflects terrain elevation information. Using the Grasshopper (GH) plug-in in Rhino software as a development tool, a first GH program is created to batch-generate 3D solid models of urban building complexes based on the GIS building vector data and terrain elevation DEM data. A second GH program is simultaneously created to construct a terrain surface model based on the DEM data, automatically calibrating the relative positions of the 3D solid model and the terrain surface model. Based on the ground roughness classification criteria in the building structure load code, a third GH program is created to automatically obtain the ground roughness classification of the target area under different wind directions, forming a 3D solid model of the building complex that includes building complex, terrain, and ground roughness information. Then, using the obtained 3D solid model of the building complex, a wind field simulation is performed on the target area using a nested coupling mode of the mesoscale meteorological model WRF (Weather Research and Forecasting Model) and computational fluid dynamics (CFD), as follows:

[0064] WRF is used to simulate mesoscale circulation information within a kilometer-scale range in the target area, covering physical processes such as water vapor, long- and short-wave radiation, cumulus clouds, and the underlying surface, and its output results are used as boundary conditions for CFD simulations. Based on the boundary conditions, a high-resolution (meter-level) numerical simulation is performed on the three-dimensional solid model of the building complex in the core area of ​​the target area (e.g., 2km×2km scale). Combined with high-precision terrain elevation data, microscale wind field data including wind speed, wind direction, and turbulence characteristics are output. The WRF simulation results loaded with different assimilation schemes are compared and analyzed, and the boundary condition settings are optimized. Through the above-mentioned coupled simulation, a high-resolution wind field database for the target area is constructed, and this high-resolution wind field database is used as the wind field simulation data for the target area.

[0065] By pre-building a 3D solid model of the building complex based on GIS and DEM data, this application accurately reproduces the target area's microscopic features, such as building distribution and terrain undulations. This provides realistic geometric boundary conditions for wind field simulation, addressing the simulation bias caused by traditional mesoscale models that ignore local building details. Furthermore, the mesoscale model WRF provides a macroscopic circulation background (including physical processes such as water vapor and radiation) over kilometer-scale areas, providing reliable boundary conditions for CFD. CFD excels at high-resolution (meter-level) microscale simulations, enabling detailed characterization of local wind field characteristics such as flow around buildings and terrain effects. This nested coupling of the two ensures global simulation rationality while achieving meter-level accuracy in core areas (e.g., 2 km × 2 km), overcoming the resolution limitations of a single model.

[0066] Since urban low-altitude wind fields are significantly affected by factors such as building distribution and topography, this application accurately incorporates these local factors through a three-dimensional solid model of the building complex, and combines the WRF-CFD nested coupling mode to simulate the superposition effect of "atmospheric circulation + local circulation". This is more in line with the actual characteristics of urban low-altitude wind fields, such as "variable incoming flow, strong turbulence, and multi-physical field coupling", and solves the problem that traditional mesoscale models (such as WRF used alone) cannot capture the details of microscale wind fields.

[0067] The present invention achieves the unity of high resolution, high authenticity and high applicability of wind field simulation data through high-precision modeling + mode coupling, thereby providing core data support for the accurate prediction of urban low-altitude micro-scale wind fields.

[0068] The first wind field prediction model provided by the present invention is constructed using a machine learning model, which is trained through a loss function.

[0069] Specifically, the first wind farm prediction model constructed using the XGBoost model is used as an example. Observational data for urban low-altitude wind farms is often sparse due to limited sensor deployment (e.g., covering only take-off and landing points or key flight routes). The XGBoost model, through its ensemble learning framework (iterative optimization of multiple decision trees), effectively processes high-dimensional sparse data. It discovers underlying patterns from the fusion of a small amount of observational data with wind farm simulation data, avoids overfitting caused by data sparsity, and improves the ability to generalize wind farm characteristics in unobserved areas. Furthermore, XGBoost is more robust to missing values ​​and outliers, making it more adaptable to the real-world scenarios of complex urban wind farms. Furthermore, the XGBoost model uses regularization terms to control model complexity and employs parallel computing to optimize node selection, resulting in fast training and high computational efficiency. It can quickly fuse sparse observational data with WRF-CFD nested coupled simulation data to produce the first wind farm prediction data. This feature addresses the computationally intensive and time-consuming challenges of traditional CFD simulations, avoids the excessive reliance on computing power of complex deep learning models, and meets the timeliness requirements of wind farm predictions for low-altitude aircraft. The core goal of wind field prediction is to accurately output continuous physical quantities such as wind speed and direction. A typical loss function, represented by the mean squared error (E), effectively measures the deviation between the predicted value and the true value, assigning greater weight to larger errors (such as sudden changes in wind speed caused by turbulence). Using E as a training metric drives the XGBoost model to focus on optimizing the prediction accuracy of key wind field features (such as peak wind speed and wind direction mutation points), thereby preventing the cumulative error from affecting the safety assessment of low-altitude aircraft.

[0070] Furthermore, before dividing the low-altitude wind field data into the first low-altitude wind field data and the second low-altitude wind field data, it also includes: preprocessing the acquired low-altitude wind field data corresponding to the target area, the preprocessing including data cleaning, removing outliers and noise in the data; and data normalization, unifying wind field data of different magnitudes into the same numerical range.

[0071] Specifically, urban low-altitude wind farms are subject to sensor failures, electromagnetic interference, and transient fluctuations caused by extreme weather events. Observational data is prone to outliers (e.g., sudden increases in wind speed to physically unreasonable ranges) or noise (e.g., high-frequency random fluctuations). By removing these interferences through cleaning, we can prevent abnormal data from misleading model training, ensuring that the wind farm data (wind speed and direction) fed into the model is more realistic, thereby improving the model's prediction accuracy.

[0072] The urban low-altitude micro-scale wind field prediction device provided by the present invention is described below. The urban low-altitude micro-scale wind field prediction device described below and the urban low-altitude micro-scale wind field prediction method described above can be referenced to each other.

[0073] Figure 4 The schematic diagram of the structure of the urban low-altitude micro-scale wind field prediction device provided by the present invention includes:

[0074] An acquisition unit 401 is used to acquire low-altitude wind field data corresponding to a target area;

[0075] A division unit 402 is configured to divide the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the amount of the first low-altitude wind field data is smaller than the amount of the second low-altitude wind field data;

[0076] The acquisition unit 401 is further configured to acquire wind field simulation data of the target area; the wind field simulation data is data obtained by simulating the target area using a nested coupling mode of a mesoscale model WRF and CFD;

[0077] The prediction unit 403 is configured to input the first low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data;

[0078] The prediction unit 403 is further configured to input the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data;

[0079] The determining unit 404 is configured to determine a wind field prediction result for the target area according to the second wind field prediction data.

[0080] Figure 5 An example of a physical structure diagram of an electronic device is shown below. Figure 5 As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communications interface 520, and the memory 530 communicate with each other via the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the urban low-altitude micro-scale wind field prediction method, which includes: obtaining low-altitude wind field data corresponding to the target area;

[0081] Dividing the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the amount of the first low-altitude wind field data is less than the amount of the second low-altitude wind field data;

[0082] Acquire wind field simulation data of the target area; the wind field simulation data is data obtained by simulating the target area using a nested coupling mode of a mesoscale model WRF and CFD;

[0083] Inputting the first low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data;

[0084] Inputting the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data;

[0085] A wind field prediction result for the target area is determined according to the second wind field prediction data.

[0086] In addition, the logic instructions in the aforementioned memory 530 can be implemented in the form of a software functional unit and, when sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0087] On the other hand, the present invention further provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can perform the urban low-altitude micro-scale wind field prediction method provided by the above methods, which includes:

[0088] Obtain low-altitude wind field data corresponding to the target area;

[0089] Dividing the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the amount of the first low-altitude wind field data is less than the amount of the second low-altitude wind field data;

[0090] Acquire wind field simulation data of the target area; the wind field simulation data is data obtained by simulating the target area using a nested coupling mode of a mesoscale model WRF and CFD;

[0091] Inputting the first low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data;

[0092] Inputting the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data;

[0093] A wind field prediction result for the target area is determined according to the second wind field prediction data.

[0094] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for predicting urban low-altitude micro-scale wind fields provided by the above methods is implemented, the method comprising: obtaining low-altitude wind field data corresponding to a target area;

[0095] Dividing the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the amount of the first low-altitude wind field data is less than the amount of the second low-altitude wind field data;

[0096] Acquire wind field simulation data of the target area; the wind field simulation data is data obtained by simulating the target area using a nested coupling mode of a mesoscale model WRF and CFD;

[0097] Inputting the first low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data;

[0098] Inputting the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data;

[0099] A wind field prediction result for the target area is determined according to the second wind field prediction data.

[0100] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0101] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for predicting urban low-altitude micro-scale wind fields, characterized in that: include: Obtain low-altitude wind field data corresponding to the target area; Dividing the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the amount of the first low-altitude wind field data is less than the amount of the second low-altitude wind field data; Acquiring wind field simulation data of the target area; The wind field simulation data is data obtained by simulating the target area using a nested coupling mode of the mesoscale model WRF and CFD; Inputting the first low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data; Inputting the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data; A wind field prediction result for the target area is determined according to the second wind field prediction data.

2. The urban low-altitude micro-scale wind field prediction method according to claim 1 is characterized in that: The obtaining of low-altitude wind field data corresponding to the target area includes: The wind field data of the target area is obtained using a weather radar, anemometer or a six-component small weather station; the wind field data includes: wind speed and wind direction.

3. The urban low-altitude micro-scale wind field prediction method according to claim 1 or 2, characterized in that: The obtaining of wind field simulation data of the target area includes: Obtaining a three-dimensional entity model of a building complex corresponding to the target area; Using the mesoscale model WRF to simulate the low-altitude wind field data of the target area to obtain an initial wind field simulation result; The initial wind field simulation result is used as the boundary condition of CFD simulation, and the three-dimensional solid model of the building complex is subjected to secondary simulation based on the boundary condition, thereby obtaining wind field simulation data of the target area.

4. The urban low-altitude micro-scale wind field prediction method according to claim 3 is characterized in that: The acquiring of the three-dimensional entity model of the building complex corresponding to the target area comprises: Obtain GIS building vector data and terrain elevation DEM data of the target area; the vector data includes the planar location, outline, floor area, and height of the building; A three-dimensional entity model of the building complex is generated based on the GIS building vector data and the terrain elevation DEM data.

5. The urban low-altitude micro-scale wind field prediction method according to claim 1 or 2, characterized in that: The first wind field prediction model and the second wind field prediction model are constructed using an XGBoost model, a CNN model, a LSTM model, a Transformer model, a LightGBM model, or a CatBoost model.

6. The urban low-altitude micro-scale wind field prediction method according to claim 1 is characterized in that: Before dividing the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, the method further includes: The acquired low-altitude wind field data corresponding to the target area are preprocessed, and the preprocessing includes data cleaning to remove outliers and noise in the data; and data normalization to unify wind field data of different magnitudes into the same numerical range.

7. A device for predicting urban low-altitude micro-scale wind fields, characterized in that: include: An acquisition unit, used to acquire low-altitude wind field data corresponding to a target area; a dividing unit, configured to divide the low-altitude wind field data into first low-altitude wind field data and second low-altitude wind field data, wherein the amount of the first low-altitude wind field data is smaller than the amount of the second low-altitude wind field data; The acquisition unit is further configured to acquire wind field simulation data of the target area; The wind field simulation data is data obtained by simulating the target area using a nested coupling mode of the mesoscale model WRF and CFD; a prediction unit, configured to input the first low-altitude wind field data and the wind field simulation data into a pre-trained first wind field prediction model to obtain first wind field prediction data; The prediction unit is further configured to input the first wind field prediction data and the second low-altitude wind field data into a pre-trained second wind field prediction model to obtain second wind field prediction data; A determination unit is configured to determine a wind field prediction result for the target area based on the second wind field prediction data.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the urban low-altitude micro-scale wind field prediction method as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the urban low-altitude micro-scale wind field prediction method according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the urban low-altitude micro-scale wind field prediction method according to any one of claims 1 to 6 is implemented.

Images