A wind field refinement method in mountainous areas based on laser wind radar

Through the fine method of wind farms in mountainous areas based on laser wind measurement radar, combined with geographic information system and WRF model, the problem of insufficient information acquisition of wind farms in mountainous areas is solved, and the refined measurement of high-space resolution wind farms is realized and The accuracy of wind energy resource assessment has been improved.

CN118731980BActive Publication Date: 2024-11-12CHENGDU YUANWANG DETECTION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411232911.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-04
Publication Date
2024-11-12
Estimated Expiration
2044-09-04

AI Technical Summary

Technical Problem

In mountainous areas, it is difficult for the existing technology to obtain wind field information with high spatial and spatial resolution, resulting in insufficient accuracy of wind energy resource assessment, weather forecast model improvement and environmental protection measures, and the boundaries of traditional computational fluid mechanics models. The assignment method fails to effectively consider geographical factor differences, resulting in wind field distortion.

Method used

Using a fine wind farm method in mountainous areas based on laser wind measurement radar, the wind speed and wind direction difference values ​​of PPI and RHI scanning modes are screened by collecting and processing high-spatial resolution data of laser wind measurement radar, combining geographic information system and WRF model. , generate three-dimensional wind fields with high spatiotemporal resolution, and construct refined atmospheric wind fields through random forest models, optimize physical parameterization schemes and CFD boundary wind fields migration.

Benefits of technology

Achieve refined measurement of high-spatial resolution wind farms in mountainous areas, improve the accuracy of wind energy resource assessment and weather forecasts, scientifically handle the impact of geographical factors on wind speed and wind direction, and enhance the reliability and continuity of wind farm information. sex.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118731980B_ABST
    Figure CN118731980B_ABST
Patent Text Reader

Abstract

The present invention relates to a method for refining wind fields in mountainous areas based on laser wind radar, and belongs to the field of electrical digital data processing. The present invention is based on PPI and RHI volume scanning modes, and utilizes the characteristics of the existence of cross-coplanar areas in the two volume scanning modes to reconstruct the refined wind field in the idle area, thereby constructing a three-dimensional wind field with high spatial resolution of the laser wind radar, which provides a basis for the accurate construction of subsequent wind fields with higher spatial resolution. Secondly, when assigning values ​​to CFD boundary wind fields, the present invention is completely different from traditional methods. The influence of different geographical factors on wind speed and wind direction is analyzed, and the influence weights of different geographical factors are determined, thereby providing a more scientific basis for the migration of CFD boundary wind fields. Finally, when WRF generates a medium-resolution wind field, a wind speed and wind direction equilibrium deviation index is designed based on the characteristics of wind speed and wind direction, providing a scientific evaluation index for the optimization selection of physical parameter schemes.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of electrical digital data processing, and in particular to a wind field refinement method in mountainous areas based on a laser wind radar. Background Art

[0002] Due to the complexity of the terrain in mountainous areas, wind speed and direction may change significantly. The accurate acquisition of refined wind field information in the area is conducive to the evaluation of wind energy pre-development, the improvement of weather forecast models, and the formulation of emergency plans in the field of environmental protection. Specifically, through wind field measurement, the distribution and intensity of wind energy resources can be accurately evaluated, helping to determine the best location for the construction of wind farms, thereby improving the efficiency of wind energy utilization. In addition, the terrain in mountainous areas affects the distribution of wind fields, which in turn affects local meteorological conditions. By measuring and analyzing wind field data, weather forecast models can be improved and the accuracy of meteorological forecasts can be improved, especially in the prediction of extreme weather events such as local storms and strong winds. In mountainous areas with complex terrain, wind erosion may lead to soil degradation and vegetation destruction, thereby affecting the ecological environment. Refined wind field measurement is essential for accurately understanding the wind erosion process in mountainous areas. Mastering the accurate characteristics of wind fields can also help formulate environmental protection measures to reduce the negative impact of wind erosion on the ecosystem.

[0003] At present, there are several methods for obtaining wind field information in mountainous areas with complex terrain, such as ground meteorological stations, ground radar remote sensing, UAV measurement, sounding balloons, and weather numerical simulation and modeling. However, in mountainous areas, due to the sparse and uneven distribution of ground meteorological stations, it is difficult to obtain precipitation information with high spatial resolution in this area by relying solely on its wind measurement equipment. Weather radar networking can obtain wind field information with high temporal and spatial resolution, but its detection capability is easily affected by the obstruction of high-altitude mountains around the area, resulting in the loss of low-level spatial wind field information. The wind field information obtained by relying solely on weather numerical simulation and modeling is often insufficient in accuracy and reliability, while other methods such as UAV measurement and sounding balloons cannot obtain refined wind field information in a large area for continuous time. At the same time, the lack of high spatial resolution initial wind field will also affect the accuracy of subsequent weather forecast models. In addition, high spatial resolution wind field information can be generated by combining computational fluid dynamics models. However, in the boundary assignment of computational fluid dynamics models, the traditional method is an assignment method based on geometric height coordinate systems. This method does not take into account factors such as the difference in geographical factors between the assigned location and the location to be assigned, which often causes the wind field to be distorted after the boundary assignment and does not conform to the actual situation of wind field continuity. Therefore, in terms of the detailed characterization of wind fields in mountainous areas, the current fusion method still has a lot of room for improvement. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method for refining wind fields in mountainous areas based on a laser wind radar, thereby solving the shortcomings of the prior art.

[0005] The object of the present invention is achieved by the following technical scheme: a wind field refinement method in mountainous areas based on laser wind radar, the wind field refinement method comprising:

[0006] S1. Input the collected data into the geographic information system to generate geographic factor datasets of high-level, medium-level and low-level spatial resolution grids and NDVI datasets of different spatial resolution grids;

[0007] S2. Based on the collected high-temporal and spatial resolution data of the laser wind radar, the areas of the two scanning modes of PPI and RHI are screened out, and the weight values ​​of the wind speed and wind direction differences are calculated, and finally a three-dimensional laser wind radar wind field with high temporal and spatial resolution is generated;

[0008] S3, input the data collected and generated in step S1 into WRF to generate atmospheric wind field information with medium spatial resolution, form a variety of physical parameterization schemes, and introduce an evaluation index to determine the most suitable WRF physical parameterization scheme for the region;

[0009] S4. Select the geographical factor combination with the highest degree of fit based on the atmospheric wind field information with medium spatial resolution and the geographical factor data set with medium spatial resolution, and determine the weight of each geographical factor under different wind speed and wind direction range conditions;

[0010] S5. Calculate the difference sequence between the geographic factors of the CFD boundary and the geographic factors of the atmospheric wind field with medium spatial resolution, combine the weight of each geographic factor under different wind speed and wind direction interval conditions, generate the mapping relationship from the medium spatial resolution grid to the CFD boundary, and generate the wind speed and wind direction of the CFD boundary wind field;

[0011] S6. Obtain atmospheric wind field information with high spatial resolution based on the generated CFD boundary wind field, build a random forest model, and finally generate a three-dimensional laser wind radar refined atmospheric wind field.

[0012] The steps of S1 include:

[0013] Collect DEM data, NDVI data, NCEP reanalysis data, laser wind radar high temporal and spatial resolution data, meteorological sounding stations and ground meteorological station instantaneous wind speed data. The laser wind radar high temporal and spatial resolution data include radial velocity, azimuth, range library, and beam elevation angle in PPI and RHI scanning modes.

[0014] Input the DEM data into the geographic information system, set the processing range, input and output dataset types, conversion coordinate system, conversion pixel size, mask type and ground curvature operator, and generate geographic factor datasets of high-level, medium-level and low-level spatial resolution grids;

[0015] Based on the target grid with the set spatial resolution, the distance between a grid point in the target grid and the four similar grids in the original grid is calculated, and the four distance values ​​are divided by the sum of the four distances to obtain four result values, which are the four weight values ​​generated for the target interpolation point. Finally, NDVI datasets of high-level, medium-level, and low-level spatial resolution grids are generated.

[0016] The steps of S2 include:

[0017] According to the multiple PPI data of high temporal and spatial resolution of the laser wind radar, based on the geometric relationship between radial velocity, azimuth and beam elevation, the profile of radial velocity distribution with azimuth in each range library is statistically calculated, and then fitted into a sine or cosine curve to find the maximum value Vmax and minimum value Vmin of the curve;

[0018] According to the relationship between the maximum value Vmax and the minimum value Vmin and the beam elevation angle α and the azimuth angle β, the horizontal wind speed Vh and the horizontal wind direction θ in the PPI scanning mode are calculated;

[0019] Based on multiple RHI data with high temporal and spatial resolution from the laser wind radar and the radial velocity V in spherical coordinates RHI The relationship between the beam elevation angle α and the azimuth angle β is used to calculate the horizontal wind speed V of the radial wind speed in spherical coordinates. RHI_H and vertical wind speed V RHI_V , the wind direction at this time is the azimuth angle β;

[0020] Based on the wind field information on the scanning surface under the PPI and RHI scanning modes, the coplanar areas of the two scanning modes are screened out, and then the horizontal wind speed and wind direction data sequences under the PPI and RHI modes in the coplanar areas are extracted respectively, and then the horizontal wind speed and wind direction difference sequences at the same azimuth, the same beam elevation angle and the same distance library are calculated respectively, and the relative error sequences of wind speed and wind direction of the two scanning modes in the coplanar area are formed respectively, and the average relative deviation of the two sequences is calculated respectively as the weight value of the wind speed and wind direction difference, which is represented by W1 and W2;

[0021] A three-dimensional grid with a spatial resolution consistent with the original spatial resolution of the laser wind radar is set. According to the idle area between multiple PPI scans, the grid point R that is closest to a certain three-dimensional grid point M in the idle area and the edge of the coplanar area of ​​the two scanning modes is found. The wind speed V_R and wind direction D_R of point R are extracted, and based on the obtained weight values ​​W1 and W2, a three-dimensional laser wind radar wind field with high temporal and spatial resolution is finally generated.

[0022] The steps of S3 include:

[0023] The obtained low-level spatial resolution grid geographic factor dataset, NDVI data and NCEP reanalysis data are used as the input of WRF, and WRF is set to a three-layer grid nesting mode with different spatial resolutions res1, res2, res3 and time accumulation steps of step1, step2, step3. The spatial resolution of the innermost layer is consistent with the medium-level spatial resolution.

[0024] Select the default physical parameterization scheme, run the WRF model to generate atmospheric wind field information with medium spatial resolution, combine the various radiation, land surface process and boundary schemes provided by WRF to form a variety of physical parameterization schemes with different configurations, run the WRF model in a loop, and compare the output results with the observed profile data of the sounding station;

[0025] The wind speed and wind direction balance deviation index is designed as the evaluation index, and the minimum value of this index is used as the basis for selecting the optimal physical parameterization scheme, so as to determine the WRF physical parameterization scheme that is most suitable for the region.

[0026] The steps of S4 include:

[0027] The atmospheric wind field information with medium spatial resolution is combined with the geographic factor data set of medium spatial resolution grid, and the geographic factors are freely combined into multiple combinations with the number of factors not less than 2. Then, the fit between each factor and multiple combined factors and the atmospheric wind field information is calculated respectively, so as to select the geographic factor combination with the highest fit.

[0028] According to the selected geographical factors and the wind field information of the laser wind radar with high temporal and spatial resolution, the fluctuation range of wind speed and wind direction is divided into N1 and N2 intervals respectively, and then the geographical factors of different wind speed and wind direction areas are statistically analyzed.

[0029] Based on the statistical results, the probability distribution of these geographical factors is calculated and normalized to determine the weight of each geographical factor under different wind speed and wind direction ranges.

[0030] The step of S5 comprises:

[0031] According to the determined geographic factor combination types and the generated geographic factor data of different spatial resolutions, and then according to the demarcated CFD boundary area, the CFD boundary geographic factor data of medium spatial resolution and the geographic factor data of medium spatial resolution are extracted, and then the geographic factors of these two spatial resolutions are spatially matched, and the difference between each geographic factor of the CFD boundary and the corresponding geographic factor of the most adjacent medium spatial resolution is calculated to form n geographic factor difference sequences △a1, △a2,…,△an;

[0032] According to the atmospheric wind field information S_wrf and D_wrf with medium spatial resolution, combined with the weight of each geographical factor under different wind speed and wind direction range conditions, a mapping relationship from the medium spatial resolution grid to the CFD boundary is generated, and the wind speed Speed_c and wind direction direc_c of the CFD boundary wind field are generated.

[0033] The step of S6 comprises:

[0034] According to the generated CFD boundary wind field, set the grid model of the CFD calculation area, and set the high-level spatial resolution in the horizontal and vertical directions to CFD_res1 and CFD_res2 respectively, and run the CFD software to obtain the atmospheric wind field information with high-level spatial resolution;

[0035] The atmospheric wind field information with high spatial resolution is downsampled to a spatial scale consistent with the spatial resolution of the laser wind radar, and used together with the geographic factor data of the same spatial scale as the input data set;

[0036] The three-dimensional wind field data of the laser wind radar is used as the output data set, and the input data set and the output data set are combined into a training set to build a refined atmospheric wind field model based on random forest.

[0037] The high-level spatial resolution atmospheric wind field information output by CFD and the high-level spatial resolution geographic factor data are input into the trained refined atmospheric wind field model to generate a three-dimensional laser wind radar refined atmospheric wind field.

[0038] The mapping relationship from the medium-level spatial resolution grid to the CFD boundary includes:

[0039] Speed_c=S_wrf×(△a1×sw i,1 +△a2×sw i,2 +…+△an×sw i,n ),

[0040] direc_c=D_wrf×(△a1×dw j,1 +△a2×dw j,2 +…+△an×dwj,n );

[0041] Among them, sw i,1 ,sw i,2 ,…,sw i,n Represents the weight factor for different wind speed ranges, dw j,1 ,dw j,2 ,…,dw j,n Represents the weight factors of different wind direction intervals. The values ​​of i and j are 1, 2, ..., N1 and 1, 2, ..., N2 respectively.

[0042] The relationship between the maximum value Vmax and the minimum value Vmin and the beam elevation angle α and the azimuth angle β includes:

[0043] Vmax=Vf×sinα+Vh×cosα, Vmin=Vf×sinα-Vh×cosα, Vh=(Vmax-Vmin) / 2cosα, where Vf is the vertical wind speed, and the azimuth angle β when the minimum value Vmin appears is the wind direction θ.

[0044] The horizontal wind speed V of the radial wind speed under the spherical coordinates is calculated RHI_H and vertical wind speed V RHI_V Includes: V RHI_H =V RHI × cosα, V RHI_V =V RHI ×sinα.

[0045] The invention has the following advantages: a wind field refinement method for mountainous areas based on a laser wind radar, based on PPI and RHI volume scanning modes, utilizing the characteristics of the existence of cross-coplanar areas in the two volume scanning modes, reconstructing the refined wind field in the idle area, thereby constructing a three-dimensional wind field with high spatial resolution of the laser wind radar, which provides a basis for the subsequent accurate construction of a wind field with higher spatial resolution; analyzing the influence of different geographical factors on wind speed and wind direction, determining the influence weights of different geographical factors, thereby providing a more scientific basis for the migration of CFD boundary wind fields; when WRF generates a medium-resolution wind field, a wind speed and wind direction balance deviation index is designed according to the characteristics of wind speed and wind direction, providing a scientific evaluation index for the optimization selection of physical parameter schemes. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the process of the present invention;

[0047] Figure 2 Flowcharts generated for different spatial resolutions of geographic information and NDVI data;

[0048] Figure 3 This is a flow chart for realizing high spatial resolution three-dimensional wind field of laser wind radar;

[0049] Figure 4 This is the flow chart for implementing boundary wind field migration in computational fluid dynamics model. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the present application provided below in conjunction with the drawings is not intended to limit the scope of protection of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application. The present invention is further described below in conjunction with the drawings.

[0051] like Figure 1 As shown, the present invention specifically relates to a method for refining wind fields in mountainous areas based on a laser wind radar, which specifically includes the following contents:

[0052] Step 1: Collect digital elevation data (DEM) and normalized difference vegetation index data (NDVI); collect NCEP reanalysis data, including surface temperature, humidity, pressure, wind speed, wind direction and cloud parameter information; collect high temporal and spatial resolution data from laser wind radar, meteorological sounding stations and instantaneous wind speed data from ground meteorological stations.

[0053] Among them, cloud parameter information data mainly refers to data sets such as cloud base height, cloud percentage, cloud ice water content, cloud liquid water content, and total cloud cover.

[0054] The high temporal and spatial resolution data of the laser wind radar mainly refers to the radial velocity, azimuth, range library, beam elevation angle and other data sets under the plane position scan (PPI) and range height scan (RHI) modes.

[0055] Step 2: If Figure 2As shown in the figure, according to the data collected in step 1, the DEM data is input into the ArcGIS geographic information system, and after setting the processing range, input and output dataset types, conversion coordinate system, conversion pixel size, mask type, and ground curvature operator, geographic factor datasets with different spatial resolutions of 10 km, 1 km, and 10 meters can be generated; based on the target grid with the set spatial resolution, the distance between a certain grid point in the target grid and the four similar grids in the original grid is calculated. The four distance values ​​are divided by the sum of the four distances, and the four result values ​​are the four weight values ​​generated by the target interpolation point, and finally the NDVI datasets with different spatial resolution grids are generated.

[0056] Among them, different spatial resolution grids refer to low-level spatial resolution grids that are consistent with the original spatial resolution of NCEP reanalysis data, medium-level spatial resolution grids that are consistent with the output spatial resolution of weather numerical models, and high-level spatial resolution grids that are consistent with the output spatial resolution of computational fluid dynamics models (CFD).

[0057] The geographic factor dataset refers to altitude, slope, aspect, position, surface relief, and surface roughness.

[0058] Step 3: If Figure 3 As shown, according to the multiple PPI data of high temporal and spatial resolution of the laser wind radar collected in step 1, based on the geometric relationship between radial velocity, azimuth and beam elevation angle, the profile of the radial velocity distribution with azimuth angle in each distance library is statistically calculated, and then fitted into a sine or cosine curve to find the maximum value Vmax and minimum value Vmin of the curve. According to the relationship between the maximum value Vmax and the minimum value Vmin and the beam elevation angle α and the azimuth angle β, the horizontal wind speed Vh and horizontal wind direction θ in the PPI scanning mode can be calculated; according to the multiple RHI data of high temporal and spatial resolution of the laser wind radar, and the radial velocity V in spherical coordinates RHI The relationship between the beam α and the azimuth angle β is used to calculate the horizontal wind speed V of the radial wind speed in spherical coordinates. RHI_H and vertical wind speed V RHI_V , the wind direction at this time is the azimuth angle β;

[0059] Based on the wind field information on the scanning surface under the PPI and RHI scanning modes, the coplanar areas of the two scanning modes are screened out, and then the horizontal wind speed and wind direction data sequences under the PPI and RHI modes in the coplanar areas are extracted respectively. Then, the horizontal wind speed and wind direction difference sequences at the same azimuth, the same beam elevation angle and the same distance library are calculated respectively, and the relative error sequences of wind speed and wind direction of the two scanning modes in the coplanar area are formed respectively. The average relative deviation of the two sequences is calculated respectively as the weight value of the wind speed and wind direction difference, which is represented by W1 and W2.

[0060] A three-dimensional grid with a spatial resolution consistent with the original spatial resolution of the laser wind radar is set, and then according to the gap area between multiple PPI scans, the grid point R that is closest to a certain three-dimensional grid point M in the gap area and the edge of the coplanar area of ​​the two scanning modes is found; then the wind speed V_R and wind direction value D_R of point R are extracted, and based on the obtained weight values ​​W1 and W2, a three-dimensional laser wind radar wind field with a high spatiotemporal resolution of 30 meters is finally generated (V_R_N=V_R×W1, D_R_N=D_R×W2).

[0061] Among them, the relationship between the maximum value Vmax and the minimum value Vmin and the beam elevation angle α and the azimuth angle β refers to Vmax=Vf×sinα+Vh×cosα, Vmin=Vf×sinα-Vh×cosα, Vh=(Vmax-Vmin) / 2cosα, where Vf and Vh are the wind components in the vertical and horizontal directions, respectively, and the wind direction when the minimum value appears is the azimuth angle β.

[0062] Horizontal wind speed V RHI_H and vertical wind speed V RHI_V The calculation method refers to V RHI_H =V RHI × cosα, V RHI_V =V RHI ×sinα.

[0063] Step 4: The geographic factor dataset, NDVI data and NCEP reanalysis data of the 10 km spatial resolution grid obtained in step 2 are used as the input of the mesoscale weather forecast model (WRF). WRF is set to a three-layer grid nested mode with different spatial resolutions and time accumulation steps of 10 km, 3 km, 1 km and 90 seconds, 30 seconds and 10 seconds, respectively. The innermost spatial resolution is 1 km. Select the default physical parameterization scheme of the model and run the WRF mode to generate atmospheric wind field information with a spatial resolution of 1 km. Combine the various radiation, land surface process and boundary schemes provided by WRF to form a variety of physical parameterization schemes with different configurations. Run the WRF mode in a loop and compare the output results with the observed profile data of the sounding station. Design the wind speed and wind direction balance deviation index as an evaluation index, and use the minimum value of this index as the basis for selecting the optimal physical parameterization scheme, so as to determine the most suitable WRF physical parameterization scheme for the region.

[0064] The wind speed and wind direction balance deviation index refers to the absolute value of the product of the minimum root mean square error of wind speed and the minimum annular variance of wind direction.

[0065] Step 5: Based on the atmospheric wind field information with a spatial resolution of 1 km obtained in step 4, combined with the geographic factor data set with a spatial resolution of 1 km obtained in step 2, the geographic factors are freely combined into multiple combinations with the number of factors not less than 2, and then the degree of fit between each factor and multiple combined factors and the atmospheric wind field information is calculated respectively, so as to select the geographic factor combination with the highest degree of fit.

[0066] Among them, the geographical factor combination with the highest fit, assuming the number of factors is n, is represented by a1, a2,…, an.

[0067] Step 6: Based on the combination of geographical factors selected in step 5, combined with the high temporal and spatial resolution laser wind radar wind field information obtained in step 3, the wind speed and wind direction fluctuation ranges are divided into 10 and 12 intervals respectively, and then the geographical factors in different wind speed and wind direction areas are statistically analyzed. The probability distribution of these geographical factors is calculated based on the statistical results, and normalized to determine the weight of each geographical factor under different wind speed and wind direction intervals. The weight factors of different wind speed intervals are expressed by sw i,1 ,sw i,2 ,…,sw i,n To express, the weight factor of different wind direction intervals is dw j,1 ,dw j,2 ,…,dw j,n to represent it, where the values ​​of i and j are 1, 2, …, 10 and 1, 2, …, 12 respectively.

[0068] Step 7: Figure 4 As shown, according to the geographical factor combination types determined in step 5 and the geographical factor data of different spatial resolutions generated in step 2, based on the demarcated CFD boundary area, the CFD boundary geographical factor data of 10 m spatial resolution and the geographical factor data of 1 km spatial resolution are extracted, and then the geographical factors of these two levels of spatial resolution are spatially matched, and the difference between each geographical factor of the CFD boundary and the geographical factor corresponding to the most adjacent 1 km spatial resolution is calculated to form n geographical factor difference sequences △a1, △a2,…,△an.

[0069] Based on the atmospheric wind field information S_wrf and D_wrf with a spatial resolution of 1 km obtained in step 4, combined with the weight of each geographical factor under different wind speed and wind direction interval conditions obtained in step 6, a mapping relationship from the 1 km spatial resolution grid to the 10 m spatial resolution CFD boundary is generated, and the wind speed Speed_c and wind direction direc_c of the CFD boundary wind field are generated.

[0070] The mapping relationship from the 1 km spatial resolution grid to the 10 m spatial resolution CFD boundary is:

[0071] Speed_c=S_wrf×(△a1×sw i,1 +△a2×sw i,2 +…+△an×sw i,n ),

[0072] direc_c=D_wrf×(△a1×dw j,1 +△a2×dw j,2 +…+△an×dw j,n ).

[0073] Step 8: Based on the CFD boundary wind field generated in step (7), set up the network model of the CFD calculation area, and set the high spatial resolution in the horizontal and vertical directions to 10 meters and 10 meters respectively. Run the CFD software to obtain the atmospheric wind field information with a spatial resolution of 10 meters.

[0074] Step 9: Downsample the atmospheric wind field information with a spatial resolution of 10 meters obtained in step 8 to a spatial scale consistent with the spatial resolution of 30 meters of the laser wind radar, and use it together with the geographical factor data with a spatial resolution of 30 meters as the input data set. Use the three-dimensional wind field data of the laser wind radar obtained in step 3 as the output data set. The input data set and the output data set together form a training set to construct a refined atmospheric wind field model based on random forest.

[0075] Step 10: According to step 8 and step 2, the atmospheric wind field information with a spatial resolution of 10 meters output by CFD and the geographic factor data with a spatial resolution of 10 meters are input into the refined atmospheric wind field model trained in step 9 to generate a three-dimensional laser wind radar atmospheric wind field with a spatial resolution of 10 meters.

[0076] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used for various other combinations, modifications and improvements, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art do not deviate from the spirit and scope of the present invention, and should be within the scope of protection of the claims attached to the present invention.

Claims

1. A method for refining wind fields in mountainous areas based on laser wind radar, characterized in that: The wind field refinement method comprises: S1. Input the collected data into the geographic information system to generate geographic factor datasets of high-level, medium-level and low-level spatial resolution grids and NDVI datasets of different spatial resolution grids; S2. Based on the collected high-temporal and spatial resolution data of the laser wind radar, the areas of the PPI and RHI scanning modes are screened out, and the weight values ​​of the wind speed and wind direction differences are calculated, and finally a three-dimensional laser wind radar wind field with high temporal and spatial resolution is generated; S3, input the data collected and generated in step S1 into WRF to generate atmospheric wind field information with medium spatial resolution, form a variety of physical parameterization schemes, and introduce an evaluation index to determine the most suitable WRF physical parameterization scheme for the region; S4. Select the geographical factor combination with the highest degree of fit based on the atmospheric wind field information with medium spatial resolution and the geographical factor data set with medium spatial resolution, and determine the weight of each geographical factor under different wind speed and wind direction range conditions; S5. Calculate the difference sequence between the geographic factors of the CFD boundary and the geographic factors of the atmospheric wind field with medium spatial resolution, combine the weight of each geographic factor under different wind speed and wind direction interval conditions, generate the mapping relationship from the medium spatial resolution grid to the CFD boundary, and generate the wind speed and wind direction of the CFD boundary wind field; S6. Obtain high spatial resolution atmospheric wind field information based on the generated CFD boundary wind field, build a random forest model, and finally generate a three-dimensional laser wind radar refined atmospheric wind field; The steps of S3 include: The obtained low-level spatial resolution grid geographic factor dataset, NDVI data and NCEP reanalysis data are used as the input of WRF, and WRF is set to a three-layer grid nesting mode with different spatial resolutions res1, res2, res3 and time accumulation steps of step1, step2, step3. The spatial resolution of the innermost layer is consistent with the medium-level spatial resolution. Select the default physical parameterization scheme, run the WRF model to generate atmospheric wind field information with medium spatial resolution, combine the various radiation, land surface process and boundary schemes provided by WRF to form a variety of physical parameterization schemes with different configurations, run the WRF model in a loop, and compare the output results with the observed profile data of the sounding station; The wind speed and wind direction balance deviation index is designed as an evaluation index, and the minimum value of this index is used as the basis for selecting the optimal physical parameterization scheme, so as to determine the most suitable WRF physical parameterization scheme for the region; The step of S5 comprises: According to the determined geographic factor combination types and the generated geographic factor data of different spatial resolutions, and then according to the demarcated CFD boundary area, the CFD boundary geographic factor data of medium spatial resolution and the geographic factor data of medium spatial resolution are extracted, and then the geographic factors of these two spatial resolutions are spatially matched, and the difference between each geographic factor of the CFD boundary and the corresponding geographic factor of the most adjacent medium spatial resolution is calculated to form n geographic factor difference sequences △a1, △a2,…,△an; According to the atmospheric wind field information S_wrf and D_wrf of the medium spatial resolution, combined with the weight of each geographical factor under different wind speed and wind direction interval conditions, the mapping relationship from the medium spatial resolution grid to the CFD boundary is generated, and the wind speed Speed_c and wind direction direc_c of the CFD boundary wind field are generated; The mapping relationship from the medium-level spatial resolution grid to the CFD boundary includes: Speed_c=S_wrf×(△a1×sw i,1 +△a2×sw i,2 +…+△an×sw i,n ), direc_c=D_wrf×(△a1×dw j,1 +△a2×dw j,2 +…+△an×dw j,n ); Among them, sw i,1 ,sw i,2 ,…,sw i,n Represents the weight factor for different wind speed ranges, dw j,1 ,dw j,2 ,…,dw j,n Represents the weight factors of different wind direction intervals. The values ​​of i and j are 1, 2, ..., N1 and 1, 2, ..., N2 respectively.

2. The method for refining wind field in mountainous areas based on laser wind radar according to claim 1 is characterized in that: The steps of S1 include: Collect DEM data, NDVI data, NCEP reanalysis data, laser wind radar high temporal and spatial resolution data, meteorological sounding stations and ground meteorological station instantaneous wind speed data. The laser wind radar high temporal and spatial resolution data include radial velocity, azimuth, range library, and beam elevation angle in PPI and RHI scanning modes. Input the DEM data into the geographic information system, set the processing range, input and output dataset types, conversion coordinate system, conversion pixel size, mask type and ground curvature operator, and generate geographic factor datasets of high-level, medium-level and low-level spatial resolution grids; Based on the target grid with the set spatial resolution, the distance between a grid point in the target grid and the four similar grids in the original grid is calculated, and the four distance values ​​are divided by the sum of the four distances to obtain four result values, which are the four weight values ​​generated for the target interpolation point. Finally, NDVI datasets of high-level, medium-level, and low-level spatial resolution grids are generated.

3. The method for refining wind field in mountainous areas based on laser wind radar according to claim 2 is characterized in that: The steps of S2 include: According to the multiple PPI data of high temporal and spatial resolution of the laser wind radar, based on the geometric relationship between radial velocity, azimuth and beam elevation, the profile of radial velocity distribution with azimuth in each range library is statistically calculated, and then fitted into a sine or cosine curve to find the maximum value Vmax and minimum value Vmin of the curve; According to the relationship between the maximum value Vmax and the minimum value Vmin and the beam elevation angle α and the azimuth angle β, the horizontal wind speed Vh and the horizontal wind direction θ in the PPI scanning mode are calculated; Based on multiple RHI data with high temporal and spatial resolution from the laser wind radar and the radial velocity V in spherical coordinates RHI The relationship between the beam elevation angle α and the azimuth angle β is used to calculate the horizontal wind speed V of the radial wind speed in spherical coordinates. RHI_H and vertical wind speed V RHI_V , the wind direction at this time is the azimuth angle β; Based on the wind field information on the scanning surface under the PPI and RHI scanning modes, the coplanar areas of the two scanning modes are screened out, and then the horizontal wind speed and wind direction data sequences under the PPI and RHI modes in the coplanar areas are extracted respectively, and then the horizontal wind speed and wind direction difference sequences at the same azimuth, the same beam elevation angle and the same distance library are calculated respectively, and the relative error sequences of wind speed and wind direction of the two scanning modes in the coplanar area are formed respectively, and the average relative deviation of the two sequences is calculated respectively as the weight value of the wind speed and wind direction difference, which is represented by W1 and W2; A three-dimensional grid with a spatial resolution consistent with the original spatial resolution of the laser wind radar is set. According to the idle area between multiple PPI scans, the grid point R that is closest to a certain three-dimensional grid point M in the idle area and the edge of the coplanar area of ​​the two scanning modes is found. The wind speed V_R and wind direction D_R of point R are extracted, and based on the obtained weight values ​​W1 and W2, a three-dimensional laser wind radar wind field with high temporal and spatial resolution is finally generated.

4. The method for refining wind field in mountainous areas based on laser wind radar according to claim 1, characterized in that: The steps of S4 include: The atmospheric wind field information with medium spatial resolution is combined with the geographic factor data set of medium spatial resolution grid, and the geographic factors are freely combined into multiple combinations with the number of factors not less than 2. Then, the fit between each factor and multiple combined factors and the atmospheric wind field information is calculated respectively, so as to select the geographic factor combination with the highest fit. According to the selected geographical factors and the wind field information of the laser wind radar with high temporal and spatial resolution, the fluctuation range of wind speed and wind direction is divided into N1 and N2 intervals respectively, and then the geographical factors of different wind speed and wind direction areas are statistically analyzed. Based on the statistical results, the probability distribution of these geographical factors is calculated and normalized to determine the weight of each geographical factor under different wind speed and wind direction ranges.

5. The method for refining wind field in mountainous areas based on laser wind radar according to claim 1, characterized in that: The step of S6 comprises: According to the generated CFD boundary wind field, set the grid model of the CFD calculation area, and set the high-level spatial resolution in the horizontal and vertical directions to CFD_res1 and CFD_res2 respectively, and run the CFD software to obtain the atmospheric wind field information with high-level spatial resolution; The atmospheric wind field information with high spatial resolution is downsampled to a spatial scale consistent with the spatial resolution of the laser wind radar, and used together with the geographic factor data of the same spatial scale as the input data set; The three-dimensional wind field data of the laser wind radar is used as the output data set, and the input data set and the output data set are combined into a training set to build a refined atmospheric wind field model based on random forest. The high-level spatial resolution atmospheric wind field information output by CFD and the high-level spatial resolution geographic factor data are input into the trained refined atmospheric wind field model to generate a three-dimensional laser wind radar refined atmospheric wind field.

6. The method for refining wind field in mountainous areas based on laser wind radar according to claim 3 is characterized in that: The relationship between the maximum value Vmax and the minimum value Vmin and the beam elevation angle α and the azimuth angle β includes: Vmax=Vf×sinα+Vh×cosα, Vmin=Vf×sinα-Vh×cosα, Vh=(Vmax-Vmin) / 2cosα, where Vf is the vertical wind speed, and the azimuth angle β when the minimum value Vmin appears is the wind direction θ.

7. The method for refining wind field in mountainous areas based on laser wind radar according to claim 3 is characterized by: The horizontal wind speed V of the radial wind speed under the spherical coordinates is calculated RHI_H and vertical wind speed V RHI_V Includes: V RHI_H =V RHI × cosα, V RHI_V =V RHI ×sinα.