Meteorological reanalysis data set generation method and device, equipment and medium

Abstract

The invention relates to the technical field of meteorological analysis, and discloses a meteorological reanalysis data set generation method, device and equipment and a medium, and the method comprises the steps: obtaining target topographic data and multi-source meteorological data; the target topographic data is topographic data with specified resolution in the target area; determining a resolution grid corresponding to the target area according to the target topographic data and the energy site information in the target area; performing data assimilation according to the multi-source meteorological data at the target moment and the background field at the target moment so as to generate an atmospheric analysis field at the target moment on the resolution grid; and generating a meteorological reanalysis data set of the time sequence based on the atmospheric analysis field at each moment. According to the scheme, the health degree of the generator can be accurately evaluated based on the advanced data of the generator and the real-time data of the generator, so that the operation reliability of the generator is improved.

Description

Technical Field

[0001] The present application relates to the field of meteorological analysis technology, and in particular to a method, apparatus, device and medium for generating a meteorological reanalysis data set. Background Art

[0002] Meteorological reanalysis datasets utilize numerical weather prediction models and data assimilation techniques to integrate various meteorological observational data (such as surface observations, high-altitude soundings, satellites, and radar) from the past period into atmospheric circulation models. This generates a long-term, consistent, high-temporal and spatial resolution global or regional three-dimensional atmospheric state dataset. Conventional meteorological reanalysis datasets typically use a fixed, equally spaced horizontal grid resolution (for example, the commonly used ERA5 and NCEP / NCAR reanalysis datasets have resolutions between 0.25° and 2.5°) and provide time series data for standard atmospheric variables (such as temperature, pressure, humidity, wind speed, wind direction, and radiation). This fixed-resolution reanalysis data performs well in studies of climate mean states or large-scale weather systems.

[0003] However, the accuracy of existing meteorological reanalysis datasets is usually low, which makes it difficult to meet the needs of refined operation of new energy and optimized scheduling of water, wind and solar energy multi-energy complementarity. Summary of the Invention

[0004] In view of this, the present application provides a method, device, equipment and medium for generating a meteorological reanalysis dataset, which improves the accuracy of generating a meteorological reanalysis dataset. The technical solution is as follows.

[0005] In a first aspect, a method for generating a meteorological reanalysis dataset is provided, the method comprising:

[0006] Acquire target terrain data and multi-source meteorological data; the target terrain data is terrain data with a specified resolution within the target area; the multi-source meteorological data includes several types of meteorological observation data;

[0007] Determine a resolution grid corresponding to the target area based on target terrain data and energy site information within the target area; each grid position in the resolution grid has its own corresponding resolution;

[0008] Performing data assimilation based on the multi-source meteorological data at the target moment and the background field at the target moment to generate an atmospheric analysis field at the target moment on the resolution grid; the background field at the target moment is the atmospheric analysis field at the moment before the target moment;

[0009] Based on the atmospheric analysis fields at each moment, a time series meteorological reanalysis dataset is generated.

[0010] In an optional embodiment, determining the resolution grid corresponding to the target area based on the target terrain data and the energy site information within the target area includes:

[0011] Obtaining terrain characteristics of each location within the target area based on the target terrain data; the terrain characteristics include slope, aspect, curvature, terrain relief, terrain shielding, landmark roughness, and potential local circulation;

[0012] The resolution of the grid at each location in the target area is determined according to the energy site information at each location in the target area and the terrain characteristics of the each location.

[0013] In an optional embodiment, determining the resolution of the grid at each location in the target area based on the energy site information at each location in the target area and the terrain features of each location includes:

[0014] According to the distance information between each location in the target area and the energy site, query the rule base to obtain the first resolution;

[0015] querying the rule base according to the terrain features of each location in the target area to obtain a second resolution;

[0016] The higher one of the first resolution and the second resolution is used as the resolution of the grid at each position of the target area.

[0017] In an optional implementation, if the target position of the target area complies with several levels of rules in the rule base, several layers of grids with different levels of resolution are nested on the target position.

[0018] In an optional embodiment, performing data assimilation based on the multi-source meteorological data at the target time and the background field at the target time to generate the atmospheric analysis field at the target time on the resolution grid includes:

[0019] According to the geographical location corresponding to the multi-source meteorological data at the target time, the multi-source meteorological data at the target time is matched to the finest grid at the geographical location;

[0020] Based on the background field on the multi-layer grid at the target time and the multi-source meteorological data on each layer of grid, an atmospheric analysis field on the multi-layer grid is generated.

[0021] In an optional embodiment, the method further includes:

[0022] According to the atmospheric analysis field on the multi-layer grid at the target moment, a forecast of the first duration is run on the multi-layer grid to generate a background field at the next moment of the target moment.

[0023] In an optional embodiment, generating a time series meteorological reanalysis dataset based on the atmospheric analysis field at each moment includes:

[0024] If the forecast running time reaches the second duration, the atmospheric analysis fields on each grid generated every first duration are output as the meteorological reanalysis dataset.

[0025] In a second aspect, a device for generating a meteorological reanalysis dataset is provided, the device comprising:

[0026] A data input module is used to obtain target terrain data and multi-source meteorological data; the target terrain data is terrain data with a specified resolution within the target area; the multi-source meteorological data includes several types of meteorological observation data;

[0027] A resolution grid determination module is used to determine a resolution grid corresponding to a target area based on target terrain data and energy site information within the target area; each grid position in the resolution grid has its own corresponding resolution;

[0028] A meteorological reanalysis generation module is used to perform data assimilation at a resolution grid based on multi-source meteorological data at a target time and a background field at the target time, so as to generate an atmospheric analysis field at the target time on the resolution grid; the background field at the target time is the atmospheric analysis field at the previous moment before the target time;

[0029] The data output module is used to generate a time series meteorological reanalysis dataset based on the atmospheric analysis field at each moment.

[0030] In a fourth aspect, an electronic device is provided, comprising a processor and a storage medium, wherein the storage medium stores program instructions executable by the processor, and the processor executes the program instructions to perform the above-mentioned meteorological reanalysis dataset generation method.

[0031] In a fifth aspect, a computer-readable storage medium is provided, wherein the storage medium stores at least one instruction, and the at least one instruction is loaded by a processor to execute the above-mentioned meteorological reanalysis dataset generation method.

[0032] The technical solution provided by this application may have the following beneficial effects:

[0033] In this application, the target terrain data and multi-source meteorological data are first obtained, and then the resolution grid corresponding to the target area is determined based on the target terrain data and the energy site information in the target area, where different grids can have different resolutions. Then, data assimilation is performed based on the multi-source meteorological data at the target moment and the background field at the target moment to generate the atmospheric analysis field at the target moment on the resolution grid. At this time, the atmospheric analysis fields of different resolutions can be generated based on the grid areas of different resolutions on the resolution grid. Finally, a time series meteorological reanalysis data set is generated based on the atmospheric analysis fields at each moment. The above scheme can generate high-precision meteorological reanalysis data where needed according to the needs of energy sites and terrain, thereby improving the accuracy of meteorological reanalysis data set generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0035] Figure 1 The present invention is a flowchart of a method for generating a meteorological reanalysis dataset according to an exemplary embodiment.

[0036] Figure 2 The present invention is a flowchart of a method for generating a meteorological reanalysis dataset according to an exemplary embodiment.

[0037] Figure 3 This is a system logic diagram of a meteorological reanalysis involved in an embodiment of the present application.

[0038] Figure 4 Schematic diagram of the structure of a meteorological reanalysis data set generation device provided in an embodiment of the present application.

[0039] Figure 5 It is a structural diagram of an electronic device provided by an optional embodiment of the present invention. DETAILED DESCRIPTION

[0040] Meteorological reanalysis datasets utilize numerical weather prediction models and data assimilation techniques to integrate various meteorological observational data (such as surface observations, high-altitude soundings, satellite data, and radar) over a period of time into atmospheric circulation models. This produces a set of long-term, consistent, high-temporal and spatial resolution global or regional three-dimensional atmospheric state datasets. These datasets play a vital role in global climate research, climate change monitoring, numerical model validation, and as driving data for various applied models (such as hydrological, agricultural, environmental, and energy models).

[0041] Conventional meteorological reanalysis datasets typically use a fixed, equally spaced horizontal grid resolution (for example, the commonly used ERA5 and NCEP / NCAR reanalysis datasets have resolutions between 0.25° and 2.5°) and provide time series data for standard atmospheric variables (such as temperature, pressure, humidity, wind speed, wind direction, and radiation). This fixed-resolution reanalysis data performs well in studying the climate mean state or large-scale weather systems.

[0042] However, with the transformation of the global energy structure, the proportion of renewable energy generation, represented by wind and solar energy, is increasing. The power output of wind power and photovoltaic power generation is highly sensitive to meteorological conditions (especially wind speed and solar radiation), and has significant volatility and intermittency. In order to ensure the stable operation of the power grid and improve the absorption capacity of renewable energy, accurate wind power and photovoltaic power forecasting is crucial. In recent years, the integrated system of water, wind and solar energy has become an important direction for improving energy utilization efficiency and grid stability. The optimal scheduling and control of this complex system has put forward higher requirements on the accuracy and reliability of the input meteorological forecast and reanalysis data.

[0043] At the same time, many regions around the world rich in wind and solar energy resources, especially those with abundant hydropower resources, are often located in mountainous and river valley regions with complex topography. Complex topography has a significant impact on local meteorological conditions. For example, mountains, river valleys, slope changes, and differences in surface roughness can give rise to complex micro- and local-scale meteorological phenomena such as valley winds, climbing winds, downslope winds, canyon channeling, local turbulence, topographically induced uplift or subsidence, and spatial variations in solar radiation due to topographic shielding. These phenomena occur on spatial scales ranging from tens of meters to kilometers, and their magnitude and complexity far exceed those found in flat areas.

[0044] Existing conventional meteorological reanalysis datasets use a fixed and usually coarse resolution, and their grid scale (e.g., tens to hundreds of kilometers) is much larger than the scale of many key local meteorological phenomena caused by complex terrain. This means that terrain changes and microclimate characteristics within a single grid point cannot be effectively captured and reflected, resulting in the inability of reanalysis data to accurately describe the true atmospheric state in these areas, especially the spatial distribution and vertical shear of wind speed, and solar radiation changes in complex terrain areas. There are significant deviations. Wind power and photovoltaic power prediction models driven by these low-precision meteorological reanalysis data often have large prediction errors in complex mountainous and valley areas, making it difficult to meet the needs of refined operation of new energy and optimized scheduling of water, wind, and solar energy multi-energy complementarity.

[0045] In order to obtain high-precision meteorological data in complex terrain areas, one method is to perform numerical weather forecast simulations at higher resolutions. Although this can capture fine-scale features to a certain extent, the computing resources required for long-term, large-scale, high-resolution (such as 100-meter-level) numerical simulations are extremely large, the data storage volume is huge, and the cost of generating reanalysis data sets (which require retrospective simulation and assimilation of historical periods) is extremely high, making it impractical. Another method is to increase ground meteorological observation stations or use remote sensing (such as satellites, radars) and telemetry (such as wind towers, Lidar, Sodar) data. However, the distribution of observation points is often sparse and uneven, especially in remote areas with complex terrain, where observation data are even more limited. Relying solely on observational data cannot provide a spatially continuous, consistent, and complete historical meteorological record.

[0046] At present, although there are methods that combine data assimilation technology to improve the accuracy of numerical simulations, the existing meteorological reanalysis generation process is usually based on assimilation and model integration based on a preset fixed-resolution grid. These methods do not directly link high-resolution terrain feature analysis with the needs of specific applications (wind and solar power prediction), and use this as a basis to intelligently determine the optimal spatial resolution distribution of meteorological reanalysis data, thereby generating a customized dataset with non-uniform resolution that can more effectively capture the local meteorological characteristics of complex terrain. Therefore, there is an urgent need for a new method and system that can overcome the limitations of existing technologies in processing meteorological data in complex terrain areas, and generate a set of meteorological reanalysis datasets specifically for complex mountainous and valley terrains, with high accuracy and high reliability, and suitable for wind and solar renewable energy power prediction and multi-energy complementary scheduling by intelligently and adaptively adjusting the spatial resolution and effectively fusing multi-source observation data.

[0047] In order to solve the above problems, an embodiment of the present application provides a method for generating a meteorological reanalysis dataset. Figure 1 This is a flow chart of a method for generating a meteorological reanalysis dataset according to an exemplary embodiment. The method is applied to an electronic device and includes:

[0048] Step 101 : Acquire target terrain data and multi-source meteorological data; the target terrain data is terrain data with a specified resolution within a target area; the multi-source meteorological data includes several types of meteorological observation data.

[0049] In this embodiment of the present application, the target terrain data is a digital elevation model. In this embodiment of the present application, the resolution of the target terrain data should be sufficiently fine so that the details of complex terrain can be captured based on the target terrain data. For example, a DEM with a resolution of 30 meters, 90 meters, or even higher can be used. In this embodiment of the present application, the target terrain data may also include land use / land cover (LULC) data (such as MODIS LULC, Corine Land Cover, etc.) to determine key parameters that affect near-surface meteorology, such as surface roughness.

[0050] Multi-source meteorological data refers to various types of meteorological observation data in the region and its surrounding areas, such as temperature, humidity, air pressure, wind speed and direction, precipitation and other data observed by conventional ground meteorological stations; temperature, humidity, air pressure, wind profile and other data obtained by high-altitude detection; aviation meteorological data, satellite meteorological data and radar meteorological data, etc.

[0051] Step 102 : Determine a resolution grid corresponding to the target area based on the target terrain data and the energy site information within the target area; each grid at each position in the resolution grid has its own corresponding resolution.

[0052] In an embodiment of the present application, after obtaining the target terrain data and multi-source meteorological data, the high-resolution target terrain data and multi-source meteorological data can be used, combined with the energy site information in the target area, to analyze the terrain characteristics of each region that affect local meteorology and energy output, and quantify its demand or importance for resolution.

[0053] Optionally, in the embodiments of the present application, terrain features that may affect local weather and energy output may include slope, aspect, curvature, terrain relief, terrain shielding analysis, and the like. For example, curvature may include planar curvature and cross-sectional curvature, which can reflect the unevenness of the terrain and is related to airflow divergence and convergence or confluence and accumulation of water; ground relief is the standard deviation or maximum and minimum elevation difference calculated within a certain window, which is used to quantify the ruggedness of the terrain; and terrain shielding analysis calculates the terrain shielding that each location may receive from direct solar radiation throughout the day, which in turn affects photovoltaic power generation.

[0054] Energy site information is the specific location information of energy sites such as wind farms and photovoltaic farms. Different resolutions need to be set for locations at different distances from the energy sites. For example, locations closer to wind farms obviously require more precise meteorological identification requirements, so the grid at that location needs to be set to a higher resolution.

[0055] Step 103 , performing data assimilation based on the multi-source meteorological data at the target time and the background field at the target time, to generate the atmospheric analysis field at the target time on the resolution grid; the background field at the target time is the atmospheric analysis field at the previous moment before the target time.

[0056] In the embodiment of the present application, in order to achieve meteorological reanalysis of the target area, a standard reanalysis cycle process can be performed based on multi-source meteorological data. Specifically, for each reanalysis time t, the analysis field at time t-1 is used as the initial field, and an atmospheric numerical model is run to perform a short-term forecast (for example, using the WRF model) until time t, and then the forecast result is used as the background field at time t;

[0057] At this point, the background field at time t and the multi-source meteorological data are assimilated to generate the analysis field at time t on the resolution grid. The above process is iterated until the reanalysis data for the required time period are generated.

[0058] Step 104: Generate a time series meteorological reanalysis dataset based on the atmospheric analysis field at each moment.

[0059] After obtaining the atmospheric analysis field at each moment, a time series meteorological reanalysis dataset can be generated according to the moment.

[0060] In summary, in this application, the target terrain data and multi-source meteorological data are first obtained, and then the resolution grid corresponding to the target area is determined based on the target terrain data and the energy site information in the target area, where different grids can have different resolutions. Then, data assimilation is performed based on the multi-source meteorological data at the target moment and the background field at the target moment to generate the atmospheric analysis field at the target moment on the resolution grid. At this time, the atmospheric analysis fields of different resolutions can be generated based on the grid areas of different resolutions on the resolution grid. Finally, a time series meteorological reanalysis data set is generated based on the atmospheric analysis fields at each moment. The above scheme can generate high-precision meteorological reanalysis data where needed according to the needs of energy sites and terrain, thereby improving the accuracy of meteorological reanalysis data set generation.

[0061] Figure 2This is a flow chart of a method for generating a meteorological reanalysis dataset according to an exemplary embodiment. The method is applied to an electronic device, which is provided with a meteorological reanalysis dataset generation system, such as Figure 3 Figure 1 shows a system logic diagram for meteorological reanalysis. This system receives data input from various sources, including global / regional background field reanalysis data, high-resolution terrain data, and multi-source meteorological observation data. The terrain feature and energy site analysis module processes the high-resolution terrain data and new energy site information. The intelligent resolution grid determination module constructs a non-uniform intelligent resolution grid based on the terrain analysis results and the distribution of energy sites. The multi-source observation data integration and preprocessing module is responsible for collecting, quality controlling, and formatting various meteorological observation data. These data are input into the data assimilation module based on the intelligent resolution grid and combined with the background field of the atmospheric numerical model to generate the analysis field. The meteorological reanalysis generation module runs the atmospheric numerical model on the intelligent resolution grid and periodically performs the data assimilation process. Ultimately, the data organization and output module forms a meteorological reanalysis dataset with terrain intelligent resolution.

[0062] The following combination Figure 2 and Figure 3 The method for generating a meteorological reanalysis dataset in an embodiment of the present application is described, and the method includes:

[0063] Step 201 : Acquire target terrain data and multi-source meteorological data; the target terrain data is terrain data with a specified resolution within a target area; the multi-source meteorological data includes several types of meteorological observation data.

[0064] for Figure 3 For the data input module in the , this module is responsible for collecting and managing various types of raw data used to generate reanalysis datasets. The raw data includes the following types.

[0065] Global / regional background field reanalysis data: These provide large-scale atmospheric initial fields and lateral boundary conditions (typically with a resolution of tens to hundreds of kilometers). For example, ECMWF's ERA5, NCEP / NCAR's FNL analysis data, or other applicable global or regional reanalysis products can be used. These data provide constraints for subsequent high-resolution simulations and assimilation based on local regions.

[0066] High-resolution terrain data: This is the foundation for achieving "terrain-smart resolution." This requires a digital elevation model (DEM) with a resolution fine enough to capture the details of complex terrain, such as 30-meter, 90-meter, or even higher-resolution DEMs (such as SRTM, ASTER GDEM, and data provided by provincial / national surveying and mapping agencies). Land use / land cover (LULC) data (such as MODIS LULC and Corine Land Cover) are also required to determine key parameters affecting near-surface meteorology, such as surface roughness.

[0067] Multi-source meteorological observation data: collect various types of meteorological observation data in and around the target area. These data are the key to the data assimilation process and are used to correct the model background field. Including: conventional ground meteorological station observations: temperature, humidity, air pressure, wind speed and direction, precipitation, etc.; high-altitude sounding data: temperature, humidity, pressure, wind profile; aviation meteorological reports; satellite data: infrared and visible light channel radiance data, atmospheric motion vectors (AMVs), surface temperature, cloud products, radiation products, etc.; radar data: weather radar reflectivity, radial velocity data. In addition, the embodiment of the present application can also realize encrypted observations in and around complex terrain areas, especially measured data from wind farms and photovoltaic power stations, including: wind speed and direction, temperature, and humidity observations at multiple layers of wind towers; wind speed and direction observations in wind turbine cabins (correction and quality control are required); irradiance meter observations of photovoltaic power stations (total horizontal irradiation, direct radiation, scattered radiation), ambient temperature, component temperature, etc. These local high-density observations are crucial to improving the reanalysis accuracy of complex terrain areas.

[0068] Step 202: Obtain terrain characteristics of each location within the target area based on the target terrain data; the terrain characteristics include slope, aspect, curvature, terrain relief, terrain shielding, landmark roughness, and potential local circulation.

[0069] like Figure 3 As shown, the system also includes a terrain feature and energy site analysis module. This module uses high-resolution DEM and LULC data, combined with new energy site information, to analyze and determine the terrain features that affect local weather and energy output, and quantify their resolution requirements or importance. In this embodiment of the application, a terrain feature analysis algorithm can be used to perform grid calculations on the high-resolution DEM to generate a series of terrain features.

[0070] For example, the Horn algorithm and Zevenbergen & Thorne algorithm can be used to calculate the slope and aspect; the plane curvature and profile curvature can be calculated to reflect the changes in the convexity of the terrain; the standard deviation of the elevation or the maximum and minimum elevation difference within a certain window can be calculated to obtain the terrain undulation; the terrain shielding that each grid point may receive from direct solar radiation during the day can be calculated; based on the LULC data, different land cover types can be mapped to corresponding surface roughness values ​​by looking up tables or using more complex models; by analyzing the valley direction, dominant slope aspect, etc., a preliminary judgment can be made on potential areas of channel effects such as valley winds and watershed winds.

[0071] Step 203 : determining the resolution of the grid at each location in the target area according to the energy site information at each location in the target area and the terrain features of the location.

[0072] Optionally, based on the distance information between each location in the target area and the energy site, the rule base is queried to obtain a first resolution; based on the terrain characteristics of each location in the target area, the rule base is queried to obtain a second resolution; and the higher of the first resolution and the second resolution is used as the resolution of the grid for each location in the target area.

[0073] In this embodiment of the present application, the terrain feature and energy site analysis module is also used to implement energy site information processing. Specifically, it can receive the location and attribute information of wind farms (wind turbine coordinates, hub height, rated power, blade diameter, etc.) and photovoltaic power plants (power plant boundaries, installed capacity, main orientation, inclination angle, etc.), and convert this information into spatial point, line, or surface feature layers.

[0074] The Terrain Feature and Energy Site Analysis module also performs spatial overlay analysis on the terrain feature layer and energy site location information. For example, it calculates the average slope, TRI, and distance to the nearest wind turbine / photovoltaic panel within each area. Based on these quantitative characteristics, a preliminary assessment of the area's meteorological resolution requirements is made. For example, locations with high TRI, proximity to wind farms, and confined valleys require high wind speed resolution; locations with steep slopes, facing specific directions, and susceptible to shadowing require high solar radiation resolution.

[0075] like Figure 3 As shown, the system also includes an intelligent resolution grid determination module, which is responsible for constructing a non-uniform (intelligent resolution) calculation / data grid system covering the target area based on the terrain feature analysis results, energy site locations, and preset rules or models.

[0076] In the embodiments of the present application, the module is also responsible for rule-based zoning and grid specification. Specifically, in the embodiments of the present application, a series of rules are defined, for example: "all areas within a radius of R1 meters around a wind turbine or a photovoltaic power station site, the target resolution is set to the highest level (e.g. 50-200 meters)"; "areas with terrain ruggedness (TRI) greater than T1, the target resolution is at least R2 level (e.g. 200-500 meters)"; "areas with slope greater than S1 and located in a specific valley direction, the target resolution is R3 level (e.g. 500-1000 meters)"; "flat areas far from energy facilities, the target resolution can be set to a lower level (e.g. 1-3 kilometers)".

[0077] In this module, when implementing the setting of resolution, the sensitivity of meteorological variables can be further considered, for example, the resolution requirement of the wind field is higher than that of the temperature field. By spatially superimposing the areas defined by these rules, the resolution conflict in the overlapping area is solved (usually the highest resolution requirement is taken). Near the boundary of the determined area with different resolutions, a transition zone is considered to be set or a smooth transition strategy is adopted to avoid numerical problems caused by resolution jumps.

[0078] In the embodiments of the present application, if the target position of the target area meets several rules of different levels in the rule library, then the target position is nested with several layers of grids with different levels of resolution.

[0079] In this module, a spatial index structure such as quadtree (2D) or octree (3D) is used, starting from the entire simulation area, according to the pre-set subdivision criteria (such as: whether the terrain ruggedness within the node exceeds the threshold? Whether it contains energy stations? Whether there are multiple high importance terrain features in the node?), the spatial area is recursively divided into smaller sub-areas. The subdivision process continues until the sub-area meets certain stopping conditions (such as reaching the pre-set minimum resolution level, or the terrain and station distribution within the sub-area is uniform enough). Each leaf node (or the final grid cell) is assigned a specific resolution level.

[0080] Through the above process, a computational grid with different size grid cells is finally generated. For the implementation of nested grid models such as WRF, this module outputs a series of nested grid definitions (including center point, grid point number, resolution, nesting ratio, etc.), where the finest nesting layer covers the areas that require high resolution, and the layout of these areas is based on intelligent resolution decision based on terrain.

[0081] Through the above-mentioned intelligent resolution grid determination module, a spatial grid structure with non-uniform resolution for the operation and data storage of atmospheric numerical models is defined. This can be a configuration file of nested grids, a data structure describing non-uniform grid coordinates and connection relationships, or a resolution distribution map layer.

[0082] Step 204 : Match the multi-source meteorological data at the target time to the finest grid at the geographical location corresponding to the multi-source meteorological data at the target time.

[0083] In the embodiment of this application, Figure 3 As shown, the system also includes a multi-source observation data integration and preprocessing module, which is responsible for collecting the original meteorological observation data from the data input module and performing necessary processing before data assimilation.

[0084] This module can also realize data collection and decoding, that is, obtaining real-time and historical observation data from various data sources (such as FTP servers, API interfaces, databases), and performing format decoding (such as converting BUFR, GRIB, ASCII and other formats into an internal unified format).

[0085] This module can also implement quality control, perform strict quality checks on observation data, and remove or mark erroneous data. Quality control methods include:

[0086] Range checking: Checks whether the observed value is within a reasonable physical range.

[0087] Step size check: Checks whether the changes in the observed values ​​of consecutive time steps exceed reasonable limits.

[0088] Spatial consistency check: Check whether the value of a certain observation point is consistent with the surrounding observation points. You can refer to the background field for inspection.

[0089] Vertical consistency check: Check the vertical consistency between layers of profile observations (such as sounding, Lidar, Sodar).

[0090] Blacklist / Whitelist check: Filter sites based on historical performance.

[0091] This module also implements error assignment, which is to determine the estimated observation error variance for each valid observation based on the observation type, instrument accuracy, observation environment, and QC results. This is an important input to the data assimilation process (usually forming the observation error covariance matrix R).

[0092] This module can also implement data assimilation formatting, that is, converting the observation data after quality control and error assignment into the specific format required by the selected data assimilation system (such as the OBS format of WRF-DA, the Thredds format of the DART system, etc.).

[0093] Optionally, this module can also implement data screening and thinning. Specifically, based on the density of the intelligent resolution grid and the capabilities of the data assimilation system, it can thin out observations that are too dense in certain areas (especially low-resolution areas) or select only observations that are representative of the current grid scale. For high-resolution grid areas, dense observations should be retained and utilized as much as possible.

[0094] Step 205 : Generate an atmospheric analysis field on the multi-layer grid based on the background field on the multi-layer grid at the target time and the multi-source meteorological data on each layer of grid.

[0095] Optionally, based on the atmospheric analysis field on the multi-layer grid at the target moment, a forecast of the first duration is run on the multi-layer grid to generate a background field at the next moment of the target moment.

[0096] In the embodiment of the present application, a meteorological reanalysis generation module and a data assimilation module based on intelligent resolution grid are also included, such as Figure 3 As shown in the figure, the two interact iteratively to generate the atmospheric analysis field for the specified time period.

[0097] Specifically, for the data assimilation module based on the smart resolution grid, this module is the key to achieving analysis accuracy. It is responsible for effectively integrating the preprocessed multi-source observation data into the background field of the atmospheric numerical model and generating the analysis field on the smart resolution grid.

[0098] In the embodiment of the present application, the data assimilation system can adopt a mature data assimilation system and configure it to adapt to a non-uniform or nested grid structure. For example, the data assimilation system that can be adopted includes a variational assimilation system and an ensemble Kalman filter system.

[0099] In this embodiment of the present application, the data assimilation module based on the intelligent resolution grid also needs to implement the mapping of the background field, obtaining the meteorological variable values ​​on the grid from the short-term forecast output (background field) of the atmospheric numerical model (i.e., step 204). If the model runs on a nested grid, the background field naturally corresponds to the nested grid structure. If the model itself supports non-uniform grids, the background field is directly located on the non-uniform grid.

[0100] In an embodiment of the present application, the data assimilation module based on the smart resolution grid also has an observation operator: for each observation, an observation operator is required to map the state variables on the model grid to the observation type and position. For example, the ground temperature observation corresponds to the bottom temperature of the model, the wind tower observation corresponds to the wind field at the corresponding height and horizontal position of the model, and the satellite radiance requires a radiation transfer model to convert the model temperature and humidity profile into radiance. For smart resolution grids, the observation operator needs to be able to handle the mapping (such as interpolation) between the observation point and its nearest model grid point, especially in areas with large resolution changes. For observations near the encrypted area of ​​the smart resolution grid, since the grid is closer to the observation point, the discretization error of the observation operator is smaller and the assimilation effect is better.

[0101] In the embodiment of the present application, the data assimilation module based on the intelligent resolution grid also needs to calculate the error covariance: the background error covariance matrix B reflects the uncertainty of the model forecast and the spatial correlation of the error. In variational assimilation, B is usually obtained through static climatological methods or ensemble methods. The observation error covariance matrix R reflects the observation error and is provided by the 5.5 module. On the intelligent resolution grid, the spatial correlation scale of the background error may need to take into account the influence of terrain and become non-uniform.

[0102] In this embodiment, the data assimilation module based on the smart resolution grid executes a selected assimilation algorithm (e.g., minimizing an objective function or calculating an ensemble mean and perturbations) to calculate an analysis increment, which is a correction to the background field. The analysis increment is then added to the background field to generate an analysis field (Analysis) at the current moment, which is the best estimate of the atmospheric state. This analysis field is located on the smart resolution grid. Finally, the output is the assimilated and corrected moment-by-moment atmospheric state analysis field located on the smart resolution grid.

[0103] As for the meteorological reanalysis generation module, this module uses the output of the data assimilation module based on the intelligent resolution grid to drive the atmospheric numerical model to perform cyclic assimilation and generate a time series meteorological reanalysis data set.

[0104] The first step in this module is to select an atmospheric numerical model. This involves choosing a non-hydrostatic model suitable for mesoscale to microscale simulations and capable of supporting intelligent resolution grid structures (e.g., nested or non-uniform grids). For example, the Weather Research and Forecasting (WRF) model is a common choice. It supports multiple nestings, allowing you to configure nested domains of varying resolutions to cover regions with varying intelligent resolution requirements. Other possible models include MPAS (Model for Prediction Across Scales).

[0105] Then a reanalysis cycle is performed: a standard reanalysis cycle process can be performed based on the multi-source meteorological data. Specifically, for each reanalysis time t, the analysis field at time t-1 is used as the initial field, and a short-term forecast is performed (for example, by using the WRF model) to predict to time t, and then the prediction result is used as the background field at time t; at this time, the background field at time t and the multi-source meteorological data are assimilated to generate the analysis field at time t on the resolution grid. The above process is iteratively performed until the reanalysis data of the required time period is generated.

[0106] In the model, the side boundary conditions of the simulation area can also be set, for example, obtained from higher-level reanalysis data (such as global reanalysis) and interpolated to the model grid points.

[0107] For the meteorological reanalysis generation module, it can output key meteorological variables at each reanalysis time on the intelligent resolution grid, including but not limited to: U / V / W wind speed components (multi-layer height, especially near the hub height), potential temperature, humidity, air pressure, surface temperature, surface air pressure, surface heat flux, latent heat flux, friction velocity, surface roughness, downward shortwave / longwave radiation flux (direct, scattered, diffuse components), upward shortwave / longwave radiation flux, cloud amount, cloud type, precipitation, etc. These variables should be selected to have a direct impact on the wind power, photovoltaic, and multi-energy complementary scheduling model.

[0108] Step 206, based on the atmospheric analysis field at each time, a time series of meteorological reanalysis data sets are generated.

[0109] In the embodiments of the present application, if the running time of the prediction reaches the second length of time, the atmospheric analysis field generated every first length of time on each grid is output to the meteorological reanalysis data set.

[0110] The system also includes a data organization and output module, which is responsible for organizing, storing, and providing a convenient access interface for the generated intelligent resolution meteorological reanalysis data. Considering the non-uniform resolution characteristics of the data, a suitable data structure needs to be selected.

[0111] Optionally, if a nested grid model is used, the output data of each layer of nested grid can be directly saved. The data set can be organized into multiple file sets, each file set corresponding to a nested level. When used, the data of the finest nested layer is selected according to the position.

[0112] Optionally, scientific data formats such as NetCDF or HDF5 are used. Groups can be defined in the file to represent different regions or resolution levels. Variable data can be stored in the corresponding groups, and the spatial position is described using index or non-uniform coordinate axes.

[0113] Alternatively, a tile-based multi-resolution architecture can be used: the entire area is divided into regular tiles, each of which stores data at different resolutions based on its internal intelligent resolution requirements. Tiles can overlap to facilitate interpolation.

[0114] Data storage formats are usually NetCDF-4 or HDF5, which support complex data structures, compression, and metadata.

[0115] This module also defines a data access interface, providing an application programming interface (API) or data reading library, enabling downstream wind and solar power forecasting models and other applications to easily query and extract data by time and spatial region (point, rectangle, polygon). The interface should automatically identify and provide the corresponding intelligent resolution data based on the query location. It should also support streaming data access.

[0116] The following example illustrates the above solution using a practical application scenario:

[0117] Target area: A typical mountain valley wind power cluster area located along the lower reaches of a certain river (referred to as "Zone A of the Lower Reaches of a Certain River Valley Wind Power Cluster"). The terrain in this area is extremely complex, with the elevation rapidly rising from approximately 800 meters at the bottom of the valley to over 2,000 meters on the surrounding ridges. It has a typical dry and hot valley climate and significant valley wind circulation. Wind farms are distributed along the valley and on the slopes of some tributaries, with wind turbines installed at an altitude of approximately 1,000 to 1,800 meters. Large hydropower stations and a small number of photovoltaic power stations are also planned or already built in this area, and there is an urgent need for the coordinated scheduling of water, wind, and solar energy.

[0118] Select a representative complete historical period, for example, from 00:00 (UTC) on January 1, 2020 to 18:00 (UTC) on December 31, 2022, a total of 3 years of time series.

[0119] The European Centre for Medium-Range Weather Forecasts (ECMWF) ERA5 reanalysis data are used as external driving fields. ERA5 provides global coverage with a horizontal resolution of approximately 0.25° x 0.25° and a temporal resolution of 1 hour. We will use ERA5 hourly data as the lateral boundary conditions (LBCs) and initial fields (used to start the reanalysis cycle) for the outermost grid in the regional model simulation.

[0120] The digital elevation model (DEM) used SRTM 30-meter resolution DEM data. GlobeLand30 30-meter resolution land cover (LULC) data was used. Detailed information on all installed wind turbines within Area A of the wind farm cluster was obtained from the wind farm owner. This information included the precise geographic coordinates (latitude and longitude) of each turbine; the turbine hub height; and the turbine model (which could be used to infer blade diameter, rated power, and standard power curves).

[0121] Collect all available encrypted meteorological observation data in and around the target area, especially the measured data from the wind farm itself:

[0122] Conventional weather stations: Collect surface (SYNOP) and upper-air (TEMP) observation data from all national and provincial weather stations within the target area and a 200-kilometer radius.

[0123] Satellite data: Collect infrared and visible channel radiance data from conventional geostationary satellites (such as Himawari-8 or FY series) covering the area, as well as atmospheric motion vectors (AMVs) and cloud classification products extracted from them.

[0124] Weather Radar: Radial velocity and reflectivity data are collected from national weather radars (such as CINRAD) covering the area (if supported by the assimilation system).

[0125] Wind farm intensified observation:

[0126] Wind tower data: This tool collects 10-minute or hourly data from three wind towers installed within the wind farm area. These towers are installed at various representative terrain locations, at standard heights of 10, 50, 80, and 100 meters (and possibly higher), to measure wind speed, direction, temperature, humidity, and air pressure.

[0127] SCADA data: Collect 10-minute or second-by-second data recorded by the Supervisory Control and Data Acquisition (SCADA) system for all wind turbines within the wind farm. Specifically, data includes nacelle wind speed, nacelle wind direction, ambient temperature, power output, and turbine operating status. This nacelle wind speed must be converted to an estimated equivalent freestream wind speed at hub height using a pre-established transfer function or pattern correction method, and quality control and error estimation must be performed.

[0128] Telemetry data (if available): If telemetry equipment such as Sodar or Lidar is installed at or near the wind farm, collect the wind profile data it provides.

[0129] Photovoltaic power station data (if assimilating wind-solar related variables): If there is a photovoltaic power station nearby, collect observation data such as total horizontal irradiation, direct irradiation, diffuse irradiation, and module temperature.

[0130] Using SRTM 30-meter DEM data and tools from GIS libraries such as GDAL or ArcPy, the following high-resolution terrain feature raster layers are calculated:

[0131] Slope: The Horn algorithm is used to calculate the slope value at each 30-meter grid point.

[0132] Aspect: A standard 3x3 neighborhood algorithm is used to calculate the aspect value for each 30-meter grid point.

[0133] Topographic relief (TRI): The standard deviation of elevation is calculated within a sliding window with a radius of 1 km (approximately 33x33 DEM grid points) to obtain a grid representing the ruggedness of the area.

[0134] Surface roughness length (Z0): Based on GlobeLand30 data, different land cover types (such as forest, grassland, built-up area, and water bodies) are mapped to surface roughness length values ​​at 30-meter grid points according to typical values ​​or locally corrected values.

[0135] Wind turbine spatial distribution analysis: Import all wind turbine coordinate points into the GIS environment to generate a wind turbine location point feature layer. Resolution requirement assessment basis: Based on the terrain characteristics and wind turbine distribution, establish resolution requirement assessment rules:

[0136] The location of the wind turbine is the core high-resolution area;

[0137] In areas with steep slopes and high relief, local meteorological changes are drastic and require high resolution;

[0138] Areas where wind speeds accelerate, decelerate, and turn due to terrain changes (such as canyon tunnels, ridge tops, and sudden slope changes) require high resolution.

[0139] Areas with large variations in surface roughness affect near-surface wind speeds and require higher resolution;

[0140] Consider the impact of terrain within a certain range upstream and downstream of the wind direction.

[0141] Then, a multi-nested grid method based on regular partitioning is used to construct the intelligent resolution grid structure. The entire reanalysis area is set as a large range, and different levels of high-resolution nested regions are defined within it. The locations and ranges of these nested regions are intelligently determined based on the analysis results of the intelligent resolution grid determination module.

[0142] The specific grid points are defined as follows:

[0143] Domain 01 (d01): The outermost region, covering the entire lower Jinsha River valley and the surrounding area, with a horizontal resolution of 5.4 km. It is used to receive ERA5 LBCs and pass large-scale information downward;

[0144] Domain 02 (d02): The first nesting level covers the main area of ​​the lower Jinsha River valley and the outer extent of the wind farm cluster. The resolution is set to 1.8 km (nesting ratio 3:1). This area is based on Rule A: the coverage of areas with a terrain relief index (TRI) greater than 30 or a slope greater than 10 degrees.

[0145] Domain 03 (d03): The second nesting level focuses on the core area of ​​wind turbine cluster A and its surrounding windward and leeward slopes directly affected by terrain. The resolution is set to 600 meters (nesting ratio 3:1). This area is based on Rule B: the coverage range is determined by the area within the horizontal distance of any wind turbine less than 10 kilometers, the terrain relief index (TRI) is greater than 80, or the slope is greater than 20 degrees.

[0146] Domain 04 (d04): The third nesting level provides detailed coverage of wind turbine-dense areas, typical valley entrances, major ridgelines, and areas with strong local wind characteristics identified through terrain analysis. The resolution is set to 200 meters (nesting ratio 3:1). This domain is based on Rule C: coverage is determined by the area within 3 kilometers of any wind turbine, or the terrain relief index (TRI) is greater than 150, or the slope is greater than 25 degrees.

[0147] Grid generation tool: Use the geogrid program in WPS (WRF Preprocessing System) that comes with the WRF model to generate geographic input files that meet these resolution and range requirements based on the nested domain parameters defined above and high-resolution terrain data.

[0148] Generates the namelist.wps file required for the WRF model. This file contains configuration information such as the center point, number of grid points, resolution, and nesting ratio of the four nested domains d01, d02, d03, and d04. This configuration information is the result of intelligent resolution decisions. It also generates geographic input files such as geo_em.d0*.nc.

[0149] Then, all meteorological observation data are integrated and preprocessed using the Multi-Source Observation Data Integration and Preprocessing module. A unified data processing script (e.g., Python script) is used to convert data from different sources and formats into a unified format (e.g., OBS format) readable by the WRF Data Assimilation System (WRF-DA). In particular, when processing SCADA data, a transfer function is required to convert nacelle wind speed into hub-height wind speed estimates. Furthermore, quality control and error assignment are performed on all observation data, ultimately generating a unified format observation file readable by the WRF-DA system.

[0150] In this example, the data assimilation system uses the WRF-DA system, which is compatible with WRF. WRF-DA supports data assimilation of nested grids within the WRF model it is running. In this example, the three-dimensional variational (3D-Var) assimilation method is used. 3D-Var solves the analysis field by minimizing a cost function that weighs the differences between the analysis field and the background field, as well as the differences between the analysis field and the observations, and takes into account the background error covariance (B) and the observation error covariance (R).

[0151] The assimilation process is as follows: a 6-hour assimilation window is used, coinciding with the reanalysis cycle step size. This means that at each analysis time (e.g., 00Z, 06Z, 12Z, and 18Z), all qualified observations within the window (e.g., 3 hours before and 3 hours after the analysis time) are assimilated. The background field at each analysis time is the 6-hour forecast generated by the WRF model from the previous analysis time, located on four nested grids: d01, d02, d03, and d04. The input consists of preprocessed and quality-controlled observation data files. The WRF-DA system automatically matches the observations to the finest nested grid covering that location based on their geographic location. For example, wind mast and SCADA data within a wind farm are primarily assimilated onto the d04 grid (200-meter resolution) or the d03 grid (600-meter resolution). The observation error covariance R is provided by module 6.9.5. The background error covariance B can be obtained using the static climatological B matrix provided by WRF-DA, or samples can be obtained through batch mode to calculate the stream-dependent B.

[0152] The main variables assimilated include: horizontal wind speed components (U, V), temperature (T), humidity (Q), surface pressure (PSFC), or vertical integrated quality in the unconstrained model. The WRF-DA system calculates the analysis increments on the grid points d01, d02, d03, and d04 to generate the analysis field for each grid. The assimilation process in finer grids (d04, d03) can more effectively utilize the intensified observations in the area covered by these grids, propagate the observation information to a smaller spatial scale, and correct the background field deviation of the model in these complex terrain areas. Finally, the three-dimensional analysis field (including model variables) is output on the four nested grids d01, d02, d03, and d04 at each analysis time (such as 00Z, 06Z, 12Z, and 18Z).

[0153] For the meteorological reanalysis generation module, the atmospheric numerical model used is the WRF model, version V4.3 for example. It is configured as a non-hydrostatic, compressible model. In addition, a physical parameterization scheme suitable for mesoscale to microscale simulations and performing well in complex terrain and at different resolutions can be selected.

[0154] Optional physical parameterization schemes include the MYNN scheme, which better handles turbulence over complex terrain; the Noah-MP scheme, which considers interactions between vegetation, soil, and snow cover; the longwave and shortwave schemes of the Rapid Transmitted Model Global (RRTMG) model, for example, which considers radiative transport of atmospheric gases, clouds, and aerosols; and the Kain-Fritsch (new Eta) scheme. Cumulus parameterization is typically not used on fine grids such as d03 and d04, as the model explicitly resolves convection.

[0155] Perform a 6-hour cycle assimilation reanalysis based on a preset three-year time period. For each cycle:

[0156] Using the analysis field at the previous analysis time (e.g., 00Z) as the initial field, the WRF model runs a 6-hour forecast on the nested grids d01-d04 to generate the background field at the current time (e.g., 06Z). The lateral boundary conditions are updated every 6 hours from ERA5; the background field at the current time and the observation data within the assimilation window are input into WRF-DA.

[0157] WRF-DA generates the analysis field for the current moment and outputs it to the d01-d04 grid. This analysis field is then used as the initial field for the next cycle (forecasting from 06Z to 12Z). This process is repeated for a three-year cycle (the second time period mentioned above). Finally, a three-dimensional atmospheric state analysis file is output every six hours on the four nested grids d01-d04.

[0158] The Data Organization and Output module then organizes the three years of WRF wrfout files into separate folders by year, month, day, and time. The file naming convention includes date and time information. Because WRF wrfout files inherently contain data in a nested grid structure, this organization naturally supports multi-resolution data.

[0159] The final "Terrain-Intelligent Resolution Meteorological Reanalysis Dataset" is a collection of all wrfout files. Each wrfout file contains three-dimensional variable data for grids d01, d02, d03, and d04. The d04 grid covers the complex terrain at the core of the wind farm and has the highest resolution (200 meters). d03 covers the surrounding complex terrain and wind farm area and has the next highest resolution (600 meters), and so on.

[0160] Save all standard variables output by the WRF model, especially ensuring that variables important to wind power prediction are saved, such as: U and V wind speed components at multiple levels (such as ground, 10m, 50m, 80m, 100m, 120m, etc.); boundary layer height (PBLH); surface temperature, surface pressure, surface heat flux, latent heat flux; various radiation fluxes (downward shortwave / longwave, upward shortwave / longwave); terrain height; surface roughness, etc.

[0161] Alternatively, a simple data access script or library can be developed (for example, using Python and libraries such as netCDF4, xarray, and PyNIO). Users enter the geographic coordinates (latitude and longitude), altitude, and time of the query, and the tool automatically locates the file corresponding to the time; determines which nested grid level the query point falls within; and uses spatial interpolation methods (such as bilinear interpolation) to extract the meteorological variable value at that point at that altitude and time from the data of the finest overlay grid. This tool hides the details of the underlying multi-resolution data structure, making it transparent to the user and providing an "intelligent resolution" access experience.

[0162] In addition, the module can also generate a detailed metadata file describing the dataset generation method, model and assimilation configuration, observation data type used, grid definition, variable list, unit, coordinate system, reanalysis period, data quality information, etc.

[0163] By using the intelligent resolution meteorological reanalysis dataset generated by this solution, it is expected that the wind power forecast accuracy for this complex mountain valley wind power cluster area will be significantly higher than that using conventional reanalysis data. Especially in periods and areas with strong topographic influences, the error rate is expected to be reduced by 5% or more, the forecast bias is reduced, and rapid power fluctuations are better captured. In addition, the multi-energy complementary scheduling support capability will be evaluated: This dataset will be input into the scheduling optimization model of the hydro-wind-solar multi-energy complementary system to evaluate its improvement over the use of conventional reanalysis data in terms of increasing wind and solar power absorption rate, optimizing the combination of hydro-wind-solar generators, reducing backup costs, and improving system stability.

[0164] In summary, in this application, the target terrain data and multi-source meteorological data are first obtained, and then the resolution grid corresponding to the target area is determined based on the target terrain data and the energy site information in the target area, where different grids can have different resolutions. Then, data assimilation is performed based on the multi-source meteorological data at the target moment and the background field at the target moment to generate the atmospheric analysis field at the target moment on the resolution grid. At this time, the atmospheric analysis fields of different resolutions can be generated based on the grid areas of different resolutions on the resolution grid. Finally, a time series meteorological reanalysis data set is generated based on the atmospheric analysis fields at each moment. The above scheme can generate high-precision meteorological reanalysis data where needed according to the needs of energy sites and terrain, thereby improving the accuracy of meteorological reanalysis data set generation.

[0165] The present application also provides a device for generating a meteorological reanalysis dataset in an embodiment. This device is used to implement the above-mentioned embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0166] The present invention provides a device for generating a meteorological reanalysis data set. Figure 4 : is a schematic diagram of the structure of a meteorological reanalysis dataset generation device provided in an embodiment of the present application, the device comprising:

[0167] The data input module 401 is used to obtain target terrain data and multi-source meteorological data; the target terrain data is terrain data with a specified resolution within the target area; the multi-source meteorological data includes several types of meteorological observation data;

[0168] The resolution grid determination module 402 is used to determine the resolution grid corresponding to the target area based on the target terrain data and the energy site information within the target area; the grid at each position in the resolution grid has its own corresponding resolution;

[0169] The meteorological reanalysis generation module 403 is configured to perform data assimilation at a resolution grid based on the multi-source meteorological data at the target time and the background field at the target time, so as to generate an atmospheric analysis field at the target time on the resolution grid; the background field at the target time is the atmospheric analysis field at the previous moment before the target time;

[0170] The data output module 404 is used to generate a time series meteorological reanalysis data set based on the atmospheric analysis field at each moment.

[0171] In summary, in this application, the target terrain data and multi-source meteorological data are first obtained, and then the resolution grid corresponding to the target area is determined based on the target terrain data and the energy site information in the target area, where different grids can have different resolutions. Then, data assimilation is performed based on the multi-source meteorological data at the target moment and the background field at the target moment to generate the atmospheric analysis field at the target moment on the resolution grid. At this time, the atmospheric analysis fields of different resolutions can be generated based on the grid areas of different resolutions on the resolution grid. Finally, a time series meteorological reanalysis data set is generated based on the atmospheric analysis fields at each moment. The above scheme can generate high-precision meteorological reanalysis data where needed according to the needs of energy sites and terrain, thereby improving the accuracy of meteorological reanalysis data set generation.

[0172] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0173] The above-mentioned apparatus is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above-mentioned functions.

[0174] See also Figure 5 , Figure 5 FIG. 1 is a schematic diagram of the structure of an electronic device provided by an optional embodiment of the present invention. The electronic device may be a computer device, such as Figure 5 As shown, the electronic device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed in the electronic device, including instructions stored in or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface).

[0175] The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0176] The memory 20 stores instructions that can be executed by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0177] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and applications required for at least one function; the data storage area may store data created based on the use of an electronic device for displaying a mini-program landing page, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk, or a solid-state drive; the memory 20 may also include a combination of the above-mentioned types of memory.

[0178] The electronic device further includes a communication interface 30 for the electronic device to communicate with other devices or a communication network.

[0179] The embodiment of the present invention also provides a computer-readable storage medium. The above-mentioned method according to the embodiment of the present invention can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0180] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for generating a meteorological reanalysis dataset, characterized in that: The method comprises: Acquire target terrain data and multi-source meteorological data; the target terrain data is terrain data with a specified resolution within the target area; the multi-source meteorological data includes several types of meteorological observation data; Determine a resolution grid corresponding to the target area based on target terrain data and energy site information within the target area; each grid position in the resolution grid has its own corresponding resolution; Performing data assimilation based on the multi-source meteorological data at the target moment and the background field at the target moment to generate an atmospheric analysis field at the target moment on the resolution grid; the background field at the target moment is the atmospheric analysis field at the moment before the target moment; Based on the atmospheric analysis fields at each moment, a time series meteorological reanalysis dataset is generated.

2. The method according to claim 1, characterized in that Determining the resolution grid corresponding to the target area based on the target terrain data and the energy site information within the target area includes: Obtaining terrain characteristics of each location within the target area based on the target terrain data; the terrain characteristics include slope, aspect, curvature, terrain relief, terrain shielding, landmark roughness, and potential local circulation; The resolution of the grid at each location in the target area is determined based on the energy site information at each location in the target area and the terrain features of the each location.

3. The method according to claim 2, characterized in that Determining the resolution of the grid at each location in the target area based on the energy site information at each location in the target area and the terrain features of each location includes: According to the distance information between each location in the target area and the energy site, query the rule base to obtain the first resolution; querying the rule base according to the terrain features of each location in the target area to obtain a second resolution; The higher of the first resolution and the second resolution is used as the resolution of the grid at each position of the target area.

4. The method according to claim 3, characterized in that If the target position of the target area complies with several levels of rules in the rule base, several layers of grids with different levels of resolution are nested on the target position.

5. The method according to claim 4, characterized in that The performing data assimilation based on the multi-source meteorological data at the target time and the background field at the target time to generate the atmospheric analysis field at the target time on the resolution grid includes: According to the geographical location corresponding to the multi-source meteorological data at the target time, the multi-source meteorological data at the target time is matched to the finest grid at the geographical location; Based on the background field on the multi-layer grid at the target time and the multi-source meteorological data on each layer of grid, an atmospheric analysis field on the multi-layer grid is generated.

6. The method according to claim 5, characterized in that The method further comprises: According to the atmospheric analysis field on the multi-layer grid at the target moment, a forecast of the first duration is run on the multi-layer grid to generate a background field at the next moment of the target moment.

7. The method according to claim 6, characterized in that The meteorological reanalysis dataset is generated based on the atmospheric analysis field at each moment, including: If the forecast running time reaches the second duration, the atmospheric analysis fields on each grid generated every first duration are output as the meteorological reanalysis dataset.

8. A meteorological reanalysis data set generation device, characterized in that: The device comprises: A data input module is used to obtain target terrain data and multi-source meteorological data; the target terrain data is terrain data with a specified resolution within the target area; the multi-source meteorological data includes several types of meteorological observation data; A resolution grid determination module is used to determine a resolution grid corresponding to a target area based on target terrain data and energy site information within the target area; each grid position in the resolution grid has its own corresponding resolution; A meteorological reanalysis generation module is used to perform data assimilation at a resolution grid based on multi-source meteorological data at a target time and a background field at the target time, so as to generate an atmospheric analysis field at the target time on the resolution grid; the background field at the target time is the atmospheric analysis field at the previous moment before the target time; The data output module is used to generate a time series meteorological reanalysis dataset based on the atmospheric analysis field at each moment.

9. An electronic device, characterized in that: The electronic device includes a processor and a storage medium, wherein the storage medium stores program instructions executable by the processor, and the processor executes the program instructions to perform the meteorological reanalysis dataset generation method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The storage medium stores at least one instruction, and the at least one instruction is loaded by a processor to execute the meteorological reanalysis dataset generation method according to any one of claims 1 to 7.

Images