Fine prediction method for dense power transmission channel wind field based on WRF mode tuning

Abstract

The invention discloses a dense power transmission channel wind field fine prediction method based on WRF mode tuning, and the method comprises the following steps: 1, obtaining the land data of a dense power transmission channel region to replace the default data of a WRF mode, and constructing the underlying surface topographic data of the dense power transmission channel region in the WRF mode; step 2, acquiring strong convection gale process instance data of a dense power transmission channel region in historical time, inputting the data into a WRF mode, and simulating the strong convection gale process instance by the WRF mode by using a plurality of physical process parameterization schemes to obtain an optimized WRF mode; and step 3, acquiring real-time wind field forecast data of the dense power transmission channel area, and inputting the real-time wind field forecast data into the optimized WRF mode to obtain a fine forecast result of the wind field of the dense power transmission channel area. Decision support is provided for power transmission channel gale forecasting and early warning, and the problem that the spatial-temporal scale is insufficient in actual operation of a power grid is effectively solved.

Description

Technical Field

[0001] The present invention relates to the field of wind field prediction methods for power transmission line areas, and in particular to a method for fine-grained wind field prediction of dense power transmission channels based on WRF model tuning. Background Art

[0002] Numerical forecasting technology in the meteorological field has been widely used in meteorological disaster warning for power grids, but the forecast grids are generally 5km×5km, 3km×3km, etc., and the model forecast time interval is 3 hours. For meteorological disasters caused by short-term severe convective weather such as local gales, the time and space scales are difficult to meet the actual operation risk assessment needs of the power grid. In addition, the transmission towers in the power transmission channel of the power grid are often located in special terrain and complex underlying surfaces. The existing numerical meteorological forecasts do not fully consider the impact of the underlying surface terrain factors, and the towers in these places are more vulnerable to natural disasters.

[0003] In response to the problem that the existing numerical meteorological forecasts for transmission channels do not fully consider the impact of the underlying terrain factors in the dense transmission channel area, Zhang Ligang et al. (2023) used CFD (computational fluid dynamics, hereinafter referred to as CFD) numerical simulation technology to simulate the wind field of transmission lines under typical offshore V-shaped valley terrain at different wind direction angles, and compared the differences in wind field calculation results under three scales of terrain with radius of 3km, 5km and 8km based on the Reynolds time-averaged model simulation method. However, this simulation mainly stays in the theoretical analysis and has not been accurately implemented in the actual application of dense transmission channel areas. Summary of the invention

[0004] The present invention provides a method for fine prediction of wind fields in dense transmission channels based on WRF model tuning, so as to solve the problem that the numerical meteorological forecast for transmission channels in the prior art does not fully consider the influence of the underlying topographic factors of the dense transmission channel area, resulting in low wind field forecast accuracy.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] The method for fine-grained wind farm prediction of dense transmission channels based on WRF model tuning includes the following steps:

[0007] Step 1: Obtain land cover data of dense transmission channel areas to replace the default data of the WRF model, and set the WRF model, thereby constructing and generating callable underlying surface terrain data of dense transmission channel areas in the WRF model;

[0008] Step 2, obtaining individual case data of strong convective wind processes in the dense transmission channel area in historical time as input data of the WRF model, and selecting a variety of physical process parameterization schemes in the WRF model, and then after the WRF model loads the underlying surface terrain data of the dense transmission channel area, the selected multiple physical process parameterization schemes are used to simulate the individual cases of strong convective wind processes based on the individual case data of strong convective wind processes in historical time, and the optimal physical process parameterization scheme is determined according to the simulation results, thereby obtaining a WRF model that loads the underlying surface data of the dense transmission channel area and is set as the optimal physical process parameterization scheme, as the optimized WRF model;

[0009] Step 3: Acquire real-time wind farm forecast data in the dense transmission channel area, input the real-time wind farm forecast data into the optimized WRF model obtained in step 2, and obtain a detailed prediction result of the wind field in the dense transmission channel area from the optimized WRF model.

[0010] Furthermore, the land cover data of the dense transmission channel area obtained in step 1 is the MCD12Q1 data of the dense transmission channel area collected by MODIS.

[0011] Furthermore, in step 1, the MCD12Q1 data is spliced, abnormal data is removed, projected, and resampled before replacing the default data of the WRF mode.

[0012] Further, in step 1, after the default data of the WRF mode is replaced, the settings of the WRF mode include:

[0013] Change the data type of WRF mode to classified data;

[0014] Set category_min=1, category_max=21category_max=21;

[0015] In the WRF mode, the four categories of water, lake, ice and snow, and city are set according to the classification values ​​in IGBP, and the WRF default land types, namely forest tundra, mixed tundra, bare tundra, and lake, are used for categories 18-21 in the WRF mode;

[0016] Added landmask_water, interp_option and rel_path to the GEOGRID.TBL file in WRF mode;

[0017] Therefore, the underlying surface topography data of the dense transmission channel area is constructed and generated in the WRF model.

[0018] Furthermore, in step 2, the physical process parameterization schemes selected in the WRF model include five types of boundary layer parameterization schemes, namely YSU, ACM2, QNSE, MYJ, and Boulac, and five types of cloud microphysics schemes, namely Lin, WSM5, WSM6, Thompson, and NSSL 2-moment.

[0019] Furthermore, in step 2, the simulation results of various physical process parameterization schemes are evaluated and the RMSE scores are calculated, and the optimal physical process parameterization scheme is selected based on the RMSE scores.

[0020] Furthermore, in step 3, the fine prediction result data of the wind field is interpolated to the observation site location through the proximity interpolation method, and the root mean square error score between the ground wind speed at the observation site and the model predicted wind speed is calculated, thereby performing real-time verification on the accuracy of the fine prediction results of the wind field obtained by the optimized WRF model.

[0021] The present invention depicts the complex terrain of the dense transmission channel area through the WRF model, and simulates the individual cases of strong convective wind processes in historical events in the dense transmission channel area through the WRF model, and selects the most effective physical process parameterization scheme, thereby fully considering the underlying surface terrain characteristics of the dense transmission channel area, and can accurately predict the wind field conditions in the dense transmission channel area, provide decision-making support for the power system to deal with meteorological disasters, and improve the accuracy of wind field forecasts for dense transmission channels at the kilometer scale.

[0022] The present invention refers to the characteristics of dense transmission channels, constructs the underlying surface terrain data of dense transmission channel areas in the WRF model, selects the optimal physical process parameterization scheme, and uses the optimized WRF model to make refined gale forecasts for transmission channels. It not only combines meteorology and electricity and makes full use of meteorological data, but also improves the spatiotemporal precision and element accuracy according to the characteristics of transmission channels, provides decision support for gale forecast and warning for transmission channels, and effectively solves the problem of insufficient spatiotemporal scale in the actual operation of power grids. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flowchart of a method according to an embodiment of the present invention.

[0024] Figure 2 It is a flow chart of modification and setting of the WRF mode in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0026] This embodiment discloses a method for fine-grained wind prediction of dense transmission channels based on WRF model tuning. According to the characteristics of the transmission channels, by studying the fine-grained characterization method of complex underlying surface features in high-resolution regional numerical weather models, the parameterization scheme of the atmospheric boundary layer and cloud microphysical processes is optimized and improved, and the accuracy of kilometer-scale numerical forecasts is improved.

[0027] WRF stands for Weather Research and Forecasting Model, which is a unified mesoscale weather forecast model developed by the National Center for Environmental Prediction (NCEP), the National Center for Atmospheric Research (NCAR) and other American scientific research institutions. The WRF model is portable, easy to maintain, expandable, efficient and convenient, and is widely used in numerical weather forecasting, data assimilation, parameterized physics research, downscaling climate simulation and other research. WRF is a regional climate model. Using the GFS global forecast data released by the United States as the input of the WRF model, more detailed forecast data can be obtained, and kilometer-level and minute-level forecasts can be achieved in the forecast area.

[0028] This embodiment is based on the WRF model, combined with the complex terrain characteristics of dense transmission channels, and optimizes the model terrain and optimizes the parameterization scheme to achieve a refined forecast of the wind field in the transmission channel area. Figure 1 As shown, this embodiment includes the following steps:

[0029] Step 1: Construct and generate callable underlying surface terrain data of dense transmission channel areas in the WRF model.

[0030] Since the surface friction of different underlying surfaces is different, the dynamic surface roughness can be calculated by the average wind speed and friction speed of different underlying surfaces, as shown in formulas (1) and (2):

[0031]

[0032] In formula (1) and (2): ψ m (ζ) is the correction function of the layer stability; Z 0m is the dynamic surface roughness; z is the observation height; d is the zero plane displacement, which is taken as 0.03m; U is the average wind speed; u * is the friction velocity; k = 0.35 is the Karman constant; x = (1-16ζ) 1 / 4 ;ζ=(zd) / L is the atmospheric stability parameter, L is the length.

[0033] In addition, the energy exchange process based on surface heat and radiation balance is an important way of interaction between the earth and the atmosphere, and different underlying surfaces have different corresponding reflectivities.

[0034] Therefore, taking the above reasons into comprehensive consideration, this embodiment combines the terrain characteristics and urban characteristics of the dense transmission channel area in the WRF model, and locally optimizes the underlying surface data by integrating the description algorithm of the optimized static / dynamic physical characterization quantities to correctly reflect the terrain characteristics of the underlying surface area of ​​the dense transmission channel.

[0035] Specific as Figure 2 As shown, in this embodiment, the land cover data of the dense power transmission channel area is first obtained. The land cover data of the dense power transmission channel area obtained is the MCD12Q1 data of the dense power transmission channel area collected by the Moderate Resolution Imaging Spectroradiometer MODIS developed by NASA.

[0036] In this embodiment, the MCD12Q1 data of the dense transmission channel area is spliced, abnormal data is removed, projected, and resampled before being used to replace the default data of the WRF mode. The projection uses Lambert projection, the unit is degree, which needs to be converted to 500m, and the band uses LC_Type1. The resampling method is Bilinear. The projection type is Geographic, the pixel resolution is 0.004491576, and the projection parameter selects WGS84.

[0037] The processed MCD12Q1 data of the dense transmission channel area is in hdf format. Use convert_geotiff to convert the MCD12Q1 data in hdf format into geogrid binary format data, and then use the format-converted MCD12Q1 data to replace the default data in the WRF model.

[0038] Next, the WRF mode is set up, and the data type of the WRF mode is modified to classified data. In the WRF mode, category_min=1, category_max=21 are set, and the four categories of water bodies, lakes, ice and snow, and cities are set according to the classification values ​​in the IGBP in the WRF mode. The 18th to 21st categories use the default land types of the WRF mode, namely forest tundra, mixed tundra, bare tundra, and lake.

[0039] Then, the format-converted MCD12Q1 data is processed in the set WRF mode to obtain the terrain data file and Index file covering the dense transmission channel area, thereby completing the construction of the underlying surface terrain data of the dense transmission channel area.

[0040] Finally, modify the GEOGRID.TBL file of WPS in the WRF mode, add landmask_water, interp_option and rel_path to the GEOGRID.TBL file, and call the underlying surface terrain data of the dense transmission channel area for simulation in the WRF mode.

[0041] Step 2: First, obtain the FNL data of individual cases of severe convective wind processes in the dense transmission channel area in historical time as the input data of the WRF model after constructing the underlying surface topography data of the dense transmission channel area and setting it in step 1.

[0042] FNL stands for Final Operational Global Analysis, which is a global reanalysis data provided by the National Center for Environmental Prediction (NECP) / National Center for Atmospheric Research (NCAR) of the United States, with a spatial resolution of 0.25°×0.25° and a temporal resolution of 6h. In this embodiment, FNL data of individual cases of severe convective gale processes in the dense transmission channel area in historical time are used as input data of the WRF model. After the initial field data is downscaled, the FNL reanalysis data is used to provide the initial value and lateral boundary conditions for the Guangxi wind power gale forecast model through bilinear interpolation.

[0043] Next, select a variety of physical process parameterization schemes in the WRF model. Due to the limited resolution of the numerical model, the model cannot recognize sub-grid processes such as cumulus convection, cloud microphysical processes, long- and short-wave radiation, and boundary layer processes. For these sub-grid processes, physical process parameterization schemes need to be used to characterize them. There are many optional physical process parameterization schemes in the WRF model, including cloud microphysical processes, long-wave radiation schemes, land surface process schemes, short-wave radiation schemes, etc. Each physical process has multiple schemes to choose from in the WRF model. Selecting different scheme configurations according to the characteristics of different research areas is a necessary choice to improve the simulation effect.

[0044] Since the use of high-resolution grids in the entire WRF model requires a large amount of computing data and cannot meet the timeliness requirements, a double-layer nesting method is used in the WRF model of this embodiment to achieve high resolution in the target area, and the parent area uses coarse resolution to provide boundary conditions. In the WRF model of this embodiment, in order to ensure that the maximum forecast effect is achieved with effective computing resources, the FNL analysis field is downscaled using a double nesting form of 15km:1km, interpolated into a 1km grid (sub-grid) in the target area, and the space outside the target area is interpolated into a 15km grid (parent grid). At the same time, a boundary buffer is set in the model so that the weather scale information in the parent grid can be effectively and accurately transmitted to the innermost refined grid. The physical process parameterization schemes selected and tested include boundary layer schemes YSU, ACM2, QNSE, MYJ, Boulac, and cloud microphysics schemes Lin, WSM5, WSM6, Thompson, and NSSL 2-moment, which are closely related to severe convective weather. In the test, other parameterization schemes were set by default to RRTM longwave radiation scheme; Goddard shortwave radiation scheme; Noah land-surface model; Kain-Fritsch cumulus parameterization scheme was used in large model areas, and cumulus parameterization scheme was not used in small model areas.

[0045] In the WRF mode of this embodiment, the boundary layer parameterization scheme and cloud microphysics scheme in the physical process parameterization scheme are used to simulate the individual cases of severe convective wind processes based on the FNL data of the individual cases of severe convective wind processes in historical time. Among them, the boundary layer parameterization scheme uses 5 types of schemes, namely YSU, ACM2, QNSE, MYJ, and Boulac, and the cloud microphysics scheme uses 5 types of schemes, namely Lin, WSM5, WSM6, Thompson, and NSSL 2-moment, thereby simulating the simulation results of ten physical process parameterization schemes in total.

[0046] In this embodiment, the simulation results of ten physical process parameterization schemes are evaluated and the root mean square error (RMSE) score is calculated, and the optimal physical process parameterization scheme is selected according to the RMSE score.

[0047] Finally, the optimal physical process parameterization scheme is set in the WRF model that loads the underlying surface data of the dense transmission channel area, thereby obtaining the optimized WRF model.

[0048] Step 3: Acquire real-time wind farm forecast data in the dense transmission channel area, input the real-time wind farm forecast data into the optimized WRF model obtained in step 2, and obtain a detailed prediction result of the wind field in the dense transmission channel area from the optimized WRF model.

[0049] In this embodiment, the real-time wind farm forecast data obtained in the dense transmission channel area is the real-time GFS wind farm forecast data. GFS stands for Global Forecast System, which is the global forecast system of the National Center for Environmental Prediction of the United States; GFS data can be used as input data for the WRF model, with a spatial resolution of 0.25°×0.25° and a temporal resolution of 3h.

[0050] In this embodiment, the optimized WRF model is used to make a fine forecast of the wind field at future times based on the real-time GFS wind farm forecast data in the dense transmission channel area. The fine forecast result has a time resolution of 6 minutes, a spatial resolution of 1 km, and a forecast time limit of 72 hours.

[0051] Furthermore, in this embodiment, Figure 1 The final test step is to interpolate the fine forecast result data with a spatial resolution of 1KM to the observation site location through the nearest interpolation method, and calculate the root mean square error (RMSE) score between the ground wind speed at the observation site and the model predicted wind speed, thereby performing real-time testing on the accuracy of the fine wind field forecast results obtained by the optimized WRF model.

[0052] The preferred embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. The embodiments described in the present invention are merely descriptions of the preferred embodiments of the present invention, and do not limit the concept and scope of the present invention. The various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction, and such combinations should also be regarded as the contents disclosed in the present disclosure as long as they do not violate the concept of the present invention. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations.

[0053] The present invention is not limited to the specific details of the above-mentioned embodiments. Within the technical concept of the present invention and without departing from the design concept of the present invention, various modifications and improvements made to the technical solution of the present invention by technical personnel in this field should fall within the protection scope of the present invention. The technical contents for which protection is sought in the present invention have been fully recorded in the claims.

Claims

1. A fine prediction method for wind farms in dense transmission channels based on WRF model tuning, characterized in that: The following steps are involved: Step 1: Obtain land cover data of dense transmission channel areas to replace the default data of the WRF model, and set the WRF model, thereby constructing and generating callable underlying surface terrain data of dense transmission channel areas in the WRF model; Step 2, obtaining individual case data of strong convective wind processes in the dense transmission channel area in historical time as input data of the WRF model, and selecting a variety of physical process parameterization schemes in the WRF model, and then after the WRF model loads the underlying surface terrain data of the dense transmission channel area, the selected multiple physical process parameterization schemes are used to simulate the individual cases of strong convective wind processes based on the individual case data of strong convective wind processes in historical time, and the optimal physical process parameterization scheme is determined according to the simulation results, thereby obtaining a WRF model that loads the underlying surface data of the dense transmission channel area and is set as the optimal physical process parameterization scheme, as the optimized WRF model; Step 3: Acquire real-time wind farm forecast data in the dense transmission channel area, input the real-time wind farm forecast data into the optimized WRF model obtained in step 2, and obtain a detailed prediction result of the wind field in the dense transmission channel area from the optimized WRF model.

2. The method for fine prediction of wind farm in dense transmission channels based on WRF model tuning according to claim 1 is characterized in that: The land cover data of the dense transmission channel area obtained in step 1 is the MCD12Q1 data of the dense transmission channel area collected by MODIS.

3. The method for fine prediction of wind farm in dense transmission channels based on WRF model tuning according to claim 2 is characterized in that: In step 1, the MCD12Q1 data is spliced, abnormal data is removed, projected, and resampled before replacing the default data of the WRF mode.

4. The method for fine prediction of wind farm in dense transmission channels based on WRF model tuning according to claim 1 is characterized in that: In step 1, after the default data of the WRF mode is replaced, the settings of the WRF mode include: Change the data type of WRF mode to classified data; Set category_min=1, category_max=21 category_max=21; In the WRF mode, the four categories of water, lake, ice and snow, and city are set according to the classification values ​​in IGBP, and the WRF default land types, namely forest tundra, mixed tundra, bare tundra, and lake, are used for categories 18-21 in the WRF mode; Added landmask_water, interp_option and rel_path to the GEOGRID.TBL file in WRF mode; Therefore, the underlying surface topography data of the dense transmission channel area is constructed and generated in the WRF model.

5. The method for fine prediction of wind field in dense transmission channels based on WRF model tuning according to claim 1 is characterized in that: In step 2, the physical process parameterization schemes selected in the WRF model include five types of boundary layer parameterization schemes, namely YSU, ACM2, QNSE, MYJ, and Boulac, and five types of cloud microphysics schemes, namely Lin, WSM5, WSM6, Thompson, and NSSL 2-moment.

6. The method for fine prediction of wind farm in dense transmission channels based on WRF model tuning according to claim 1 is characterized in that: In step 2, the simulation results of various physical process parameterization schemes are evaluated and the RMSE scores are calculated, and the optimal physical process parameterization scheme is selected based on the RMSE scores.

7. The method for fine prediction of wind farm in dense transmission channels based on WRF model tuning according to claim 1 is characterized in that: In step 3, the fine prediction result data of the wind field is interpolated to the observation site location through the nearest interpolation method, and the root mean square error score between the ground wind speed at the observation site and the model predicted wind speed is calculated, thereby verifying the accuracy of the fine prediction results of the wind field obtained by the optimized WRF model in real time.

Images