Low-altitude atmosphere three-dimensional wind field inversion method based on vertical observation of wind profile radar

Abstract

The invention discloses a low-altitude atmosphere three-dimensional wind field inversion method based on wind profile radar vertical observation, relates to the technical field of three-dimensional wind field inversion, and aims to solve the problem of poor inversion effect caused by inaccurate model parameters in the wind field inversion process. The difference between simulation data and actual observation data is evaluated from multiple angles, a comprehensive optimization basis is provided, parameter calculation optimization is performed by adopting a particle swarm optimization method, the method has the characteristics of high convergence speed and high global search capability, is suitable for optimization of complex problems, and is used for confirming model requirements according to a wind field inversion rule. The model is selected by combining the evaluation result of the model library, it is ensured that the selected model has high pertinence and applicability, the prediction precision and reliability of the model are improved, model initialization is conducted through wind field fusion data, data information of multiple sources is fully fused, and the comprehensiveness and accuracy of the data are improved.

Description

Technical Field

[0001] The invention relates to the technical field of three-dimensional wind field inversion, and in particular to a method for inverting a low-altitude atmospheric three-dimensional wind field based on vertical observation of a wind profiler radar. Background Art

[0002] Three-dimensional wind field inversion refers to the use of specific techniques and methods to infer the wind field distribution in the three-dimensional space of the atmosphere from limited observation data.

[0003] The Chinese patent document with the announcement number CN118091666B discloses a deep learning wind profile radar full-beam three-dimensional wind field inversion method and system, which mainly decodes and inverts the acquired wind profile radar raw data to obtain the three-dimensional wind field initial information, and then performs time consistency averaging processing; matches and standardizes the historical wind field observation data and the wind profile radar raw data to form observation grids, and introduces terrain data as features to construct a training data set; constructs a deep learning-based three-dimensional wind field inversion model based on the improved deep neural network CU-Net Pro, and optimizes and trains the model using the training data set; standardizes and time-matches the three-dimensional inverted wind data of the wind profile radar to be corrected, and then inputs it into the model to invert the three-dimensional inverted wind correction value. Although the above patent solves the problem of three-dimensional inversion, there are still the following problems in actual operation: 1. Failure to obtain accurate data through more and more complete acquisition ports results in inaccurate acquisition of wind farm raw data.

[0004] 2. The collected wind field data is not further processed, and no targeted model is established for the wind field data, resulting in poor inversion results when performing wind field inversion.

[0005] 3. The wind field inversion model was not simulated more effectively, and no targeted simulation optimization and adjustment were performed based on the simulation results, resulting in poor inversion results. Summary of the invention

[0006] The purpose of the present invention is to provide a low-altitude atmospheric three-dimensional wind field inversion method based on vertical observation of a wind profiler radar. The root mean square error, correlation coefficient and absolute error are covered through difference analysis, the difference between simulated data and actual observation data is evaluated from multiple angles, and a comprehensive optimization basis is provided. The particle swarm optimization method is used for parameter calculation optimization, which has the characteristics of fast convergence speed and strong global search capability, and is suitable for the optimization of complex problems. The model requirements are confirmed according to the wind field inversion rules, and the model is selected in combination with the evaluation results of the model library, which ensures that the selected model has high pertinence and applicability, and helps to improve the prediction accuracy and reliability of the model. The wind field fusion data is used to initialize the model, and the data information from multiple sources is fully integrated, which improves the comprehensiveness and accuracy of the data, and can solve the problems in the prior art.

[0007] To achieve the above object, the present invention provides the following technical solutions: A three-dimensional wind field inversion method for the low-altitude atmosphere based on vertical observation of wind profiler radar, wherein the three-dimensional wind field inversion method is to construct a dynamic model according to the acquired data, simulate the wind field with the constructed dynamic model, generate a three-dimensional wind field by using an inversion algorithm based on the simulation results and the actual observation data, and then visualize the generated three-dimensional wind field; The three-dimensional wind field inversion method comprises: collecting and quality controlling data provided by wind profile radar and ground meteorological station; Data from different sources are fused and data preprocessed after data fusion is completed; Select a kinetic model from the model library, initialize the selected kinetic model using the data after data preprocessing, and set initial state parameters; The initialized dynamic model is used to simulate the wind field, and the difference between the wind field simulation data and the actual observation data is analyzed. According to the difference analysis results, the initialized dynamic model is optimized and adjusted using the optimization algorithm; The optimized and adjusted dynamic model is used to generate a three-dimensional wind field, and the generated three-dimensional risk is inverted using visualization technology to display the results.

[0008] Preferably, data provided by wind profiler radar and ground meteorological stations are collected and quality controlled, including: The vertical wind speed profile data from the wind profile radar is collected, and the vertical wind speed profile data includes vertical wind speed and horizontal wind speed data of the time series; Collect data from ground weather stations, including basic meteorological element data such as temperature, pressure, humidity, wind speed and wind direction; Conduct quality control on the data provided by wind profiler radar and ground meteorological station respectively; The quality control of wind profile radar data includes: echo quality check, depolarization ratio check, vertical wind field consistency check and horizontal wind field comparison; The quality control of ground weather station data includes: internal consistency check, spatiotemporal continuity check, threshold check and statistical quality control; At the same time, outliers in the data provided by wind profiler radars and ground meteorological stations are corrected or eliminated, and missing data are interpolated; After quality control is completed, actual observation data is obtained.

[0009] Preferably, data from different sources are fused, and after the data fusion is completed, data preprocessing is performed, including: Fuse the wind profile radar data and ground meteorological station data that have completed quality control in the actual observation data; Before data fusion, the wind profile radar data and ground meteorological station data that have completed quality control in the actual observation data are quality assessed. The quality assessment uses the isolation forest algorithm to identify outliers and noise points in wind speed, wind direction and other data, and marks or removes them. The reliability of ground meteorological station data is assessed using the Bayesian network model, and the rationality of the data is judged based on the correlation between different meteorological elements. Finally, the data with quality assessment within the high quality standard is selected for data fusion. The data fusion process includes: matching wind profiler radar data with ground meteorological station data, where data matching includes time matching and space matching; time matching is to match wind profiler radar data with ground meteorological station data in time according to timestamps; space matching is to match wind profiler radar data with ground meteorological station data in space according to geographic locations.

[0010] Preferably, data from different sources are fused, and after the data fusion is completed, data preprocessing is performed, which also includes: Perform data interpolation on the wind profile radar data and ground meteorological station data after data matching, and the data interpolation includes vertical interpolation and horizontal interpolation; Among them, vertical interpolation is to use linear interpolation or bilinear interpolation on the wind speed profile data of the wind profile radar to interpolate the data to the observation height of the ground meteorological station and supplement the data of different altitude layers; horizontal interpolation is to expand the observation data of the ground meteorological station to the area covered by the radar through the Kriging interpolation method; The wind profile radar data and ground meteorological station data after data interpolation are weighted by using the weighted average method; Before weighted processing, a weighted strategy is formulated, which includes time weight, space weight and data quality weight; Time weight is to assign a higher weight to near-real-time observation data during the fusion process of wind profile radar data and ground meteorological station data after data interpolation. If the timestamp of the data is very close to the current time, the data can be given a higher weight; Spatial weight is calculated by calculating the spatial distance between radar data and weather station data, and higher weight is given to points with closer distance; The data quality weight is dynamically adjusted based on the quality assessment of each data source; The weighted wind profile radar data and the data from the ground meteorological station data are combined to obtain fused data of the wind profile radar data and the ground meteorological station data; Perform data preprocessing on the fused actual observation data; After data preprocessing, wind field fusion data is obtained.

[0011] Preferably, a kinetic model is selected from a model library, the selected kinetic model is initialized using the data after data preprocessing, and initial state parameters are set, including: First, the models in the model library are evaluated, including physical properties, mathematical expressions, scope of application and performance records; Confirm the model requirements according to the wind field inversion rules and the evaluation results of the models in the model library, and select the corresponding dynamic model; Initialize the selected dynamic model using wind field fusion data; The initialization process includes: confirming the initialization indicators in the wind field fusion data, and the initialization standards include wind speed, wind direction, temperature, pressure and humidity data; According to the initialization index in the wind field fusion data, the initialization state parameters of the selected dynamic model are set; Among them, the initialization state parameters are set to use the dynamic model and its corresponding initialization method, and the initialization indicators in the wind field fusion data are input through a function or class method, and after the input, the initial state configuration of the selected dynamic model is started; After the initialization state parameters are set, the wind field dynamics model is obtained.

[0012] Preferably, the initialized dynamic model is subjected to wind field simulation, and the wind field simulation data and the actual observation data are subjected to difference analysis. According to the difference analysis result, the initialized dynamic model is optimized and adjusted by using an optimization algorithm, including: The wind field dynamics model is used to simulate the wind field. Before the wind field simulation is performed, the temporal and spatial scope and resolution of the simulation are first confirmed. At the same time, the physical parameters and boundary conditions of the wind field dynamics model are configured. Finally, the simulation time step and total simulation time are set. After the above steps are completed, the wind field data in the wind field dynamics model is confirmed, and the wind field data includes the spatially distributed wind speed, wind direction, temperature and pressure; A wind field model is established according to the wind field dynamics model, and simulation conditions of the wind field model are set according to the wind field data, and finally a target wind field model is obtained; Divide the target area in the target wind field model into a number of grids using a structured grid method; The wind field parameters of each grid are calculated through time deduction. The steps of calculating the wind field parameters include: Input the parameters of the wind field data into the target wind field model and set the initial state in time and space; For each time step, the discrete equation is used for time integration to calculate the state of the target wind field model at the next time step; In each time step, iterative updates are performed in sequence until the set total time is reached; At each time step, the calculated wind field data is saved in the corresponding data set, and the data set is marked as simulated wind field data; The simulated wind field data and the actual observation data are analyzed for differences. The difference analysis includes root mean square error, correlation coefficient and absolute error. The root mean square error is used to quantify the average deviation between the simulated wind field data and the actual observation data; the correlation coefficient is used to evaluate the similarity between the simulated wind field data and the actual observation data; the absolute error is used to calculate the absolute difference between the simulated wind field data and the actual observation data at each grid point or time step; According to the difference analysis results, the data in the simulated wind farm data that has the greatest impact on the target wind farm model is confirmed; According to the confirmed impact data, the particle swarm optimization method is used to optimize the parameter calculation; Adjust the parameters of the target wind farm model according to the optimized parameters, re-simulate the wind farm after the parameters are adjusted, and iterate the new simulation parameters again until the simulation accuracy of the target wind farm model is within a preset range; After the optimization and adjustment of the target wind field model is completed, the vertical wind field inversion model is obtained.

[0013] Preferably, the resolution of the simulation is confirmed, including: Retrieve the available memory space corresponding to the wind farm simulation system; Extract the rated memory capacity of the wind farm simulation system; Determine the available memory percentage value corresponding to the wind farm simulation according to the ratio between the available memory space corresponding to the wind farm simulation system and the rated memory capacity; Extracting the change rate of wind field data in the vertical direction, wherein the wind field data includes spatially distributed wind speed, wind direction, temperature and pressure; and the change rate of wind field data in the vertical direction includes the average change rate of wind speed, the average change rate of wind direction angle, the average change rate of temperature and the average change rate of pressure; Extracting the frequency of change of the wind field data in each unit time; The resolution of the wind field simulation in the vertical and horizontal directions is obtained according to the change frequency of the wind field data in each unit time, the change rate of the wind field data in the vertical direction and the percentage value of the available memory corresponding to the wind field simulation.

[0014] Preferably, obtaining the resolution of the wind field simulation in the vertical and horizontal directions according to the frequency of change of the wind field data in each unit time, the rate of change of the wind field data in the vertical direction, and the percentage value of available memory corresponding to the wind field simulation includes: Extracting the maximum value of the change rate among the change rates of the wind field data in the vertical direction; wherein the change rate of the wind field data in the vertical direction includes the average change rate of wind speed, the average change rate of wind direction angle, the average change rate of temperature and the average change rate of pressure; Normalizing the maximum value of the rate of change of the wind field data in the vertical direction and the value of the available memory ratio corresponding to the wind field simulation to obtain the normalized maximum value of the rate of change and the value of the available memory ratio; Compare the normalized maximum value of the rate of change and the normalized value of the available memory percentage; When the maximum value of the normalized change rate does not exceed the value of the normalized available memory ratio, the simulation resolution is confirmed according to the grid size corresponding to the preset initial vertical resolution and initial horizontal resolution; When the maximum value of the change rate after normalization exceeds the value of the available memory ratio after normalization, the grid sizes corresponding to the initial vertical resolution and the initial horizontal resolution are adjusted using the change frequency of the wind field data in each unit time, the change rate of the wind field data in the vertical direction, and the value of the available memory ratio corresponding to the wind field simulation to obtain the grid sizes corresponding to the adjusted vertical resolution and horizontal resolution; The grid sizes corresponding to the adjusted vertical resolution and horizontal resolution are obtained by the following formula: ; Among them, L c and L s Indicates the grid size corresponding to the adjusted vertical and horizontal resolutions; L c0 and L s0represents the grid size corresponding to the initial vertical resolution and the initial horizontal resolution; P represents the percentage of available memory corresponding to the wind field simulation; B 01 , B 02 , B 03 and B 04 They represent the average change rate of wind speed, wind direction angle, temperature and pressure respectively; n represents the number of unit time experienced by wind field data collection; f 01i 、f 02i 、f 03i and f 04i represents the frequency of change of wind speed, wind direction, temperature and pressure corresponding to the i-th unit time; f 01c 、f 02c 、f 03c and f 04c Indicates the preset reference value of the frequency of change of wind speed, wind direction, temperature and pressure; The simulation resolution was confirmed by adjusting the grid size corresponding to the vertical and horizontal resolutions.

[0015] Preferably, the target area in the target wind field model is divided into a plurality of grids using a structured grid method, including: Extracting a target area in a target wind field model; Comparing the area value of the target area in the target wind field model with a preset area value reference value to obtain a ratio between the area value of the target area and the preset area value reference value; Comparing a ratio between an area value of the target area and a preset area value reference value with a preset ratio threshold; When the ratio between the area value of the target area and the preset area value reference value does not exceed the preset ratio threshold, the target area is structurally divided using the initial target area grid size to obtain a plurality of grids; When the ratio between the area value of the target area and the preset area value reference value exceeds a preset ratio threshold, the size change rate between the grid size of the adjusted vertical resolution and the grid size of the initial vertical resolution is retrieved as a first size change rate value; Retrieving a size change rate between a grid size of the adjusted horizontal resolution and a grid size of the initial horizontal resolution as a second size change rate value; The initial target area grid size is adjusted by using the ratio between the area value of the target area and the preset area value reference value in combination with the first size change rate value and the second size change rate value to obtain an adjusted target area grid size; The adjusted target area grid size is obtained by the following formula: ; Where D represents the adjusted target area grid size; D0 represents the initial target area grid size; P d The ratio between the area value of the target area and the preset area value reference value; P c Indicates the value of the first dimensional change rate; P s Indicates the value of the second dimension change rate; The target area is structurally divided according to the adjusted target area grid size to obtain a plurality of grids.

[0016] Preferably, the optimized and adjusted dynamic model is used to generate a three-dimensional wind field, and the generated three-dimensional risk is inverted and displayed using visualization technology, including: The vertical wind field inversion model is used to generate a three-dimensional wind field using a three-dimensional model generator, wherein the wind field parameter data in the vertical wind field inversion model is input into the three-dimensional model generator, and after the three-dimensional model generator is started, a three-dimensional wind field model generated by the vertical wind field inversion model is obtained; Visualize the three-dimensional wind field model. The visualization includes: using streamline diagrams to show the flow direction and intensity of the wind field, using vector field visualization to show the wind speed and direction of each grid point in three-dimensional space, and visualizing the spatial distribution of temperature and humidity; The three-dimensional wind field model after visualization processing is visualized and outputted, and three-dimensional wind field visualization data of the vertical wind field inversion model is obtained.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention provides a method for inverting the low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar. Data interpolation can supplement and refine the data, making the data more precise and accurate in both vertical and horizontal directions. Horizontal interpolation can expand the data coverage of ground meteorological stations so that they can cover the area monitored by radar, thereby expanding the spatial coverage of the data.

[0018] 2. The low-altitude atmospheric three-dimensional wind field inversion method based on vertical observation of wind profile radar provided by the present invention confirms the model requirements according to the wind field inversion rules, and selects the model in combination with the evaluation results of the model library, ensuring that the selected model has a high degree of pertinence and applicability. This helps to improve the prediction accuracy and reliability of the model, and uses wind field fusion data for model initialization, fully integrating data information from multiple sources, and improving the comprehensiveness and accuracy of the data.

[0019] 3. The present invention provides a method for inverting the three-dimensional wind field of the low-altitude atmosphere based on vertical observation of wind profiler radar. The difference analysis covers the root mean square error, correlation coefficient and absolute error, evaluates the difference between the simulated data and the actual observation data from multiple angles, provides a comprehensive optimization basis, and uses particle swarm optimization to perform parameter calculation optimization. This is an emerging evolutionary computing algorithm with the characteristics of fast convergence speed and strong global search capability, and is suitable for the optimization of complex problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the three-dimensional wind field inversion process of vertical observation by the wind profiler radar of the present invention; Figure 2 It is a schematic diagram of the three-dimensional wind field inversion steps of vertical observation by wind profiler radar of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] In order to solve the problem in the prior art that when acquiring the actual data of the wind farm, more and more perfect acquisition ports are not used to accurately acquire the data, resulting in inaccurate acquisition of the original data of the wind farm, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions: A three-dimensional wind field inversion method for the low-altitude atmosphere based on vertical observation of wind profiler radar, wherein the three-dimensional wind field inversion method is to construct a dynamic model according to the acquired data, simulate the wind field with the constructed dynamic model, generate a three-dimensional wind field by using an inversion algorithm based on the simulation results and the actual observation data, and then visualize the generated three-dimensional wind field; The three-dimensional wind field inversion method comprises: collecting and quality controlling data provided by wind profile radar and ground meteorological station; Data from different sources are fused and data preprocessed after data fusion is completed; Select a kinetic model from the model library, initialize the selected kinetic model using the data after data preprocessing, and set initial state parameters; The initialized dynamic model is used to simulate the wind field, and the difference between the wind field simulation data and the actual observation data is analyzed. According to the difference analysis results, the initialized dynamic model is optimized and adjusted using the optimization algorithm; The optimized and adjusted dynamic model is used to generate a three-dimensional wind field, and the generated three-dimensional risk is inverted using visualization technology to display the results.

[0023] Collect and quality control data from wind profiler radars and ground-based meteorological stations, including: The vertical wind speed profile data from the wind profile radar is collected, and the vertical wind speed profile data includes vertical wind speed and horizontal wind speed data of the time series; Collect data from ground weather stations, including basic meteorological element data such as temperature, pressure, humidity, wind speed and wind direction; Conduct quality control on the data provided by wind profiler radar and ground meteorological station respectively; The quality control of wind profile radar data includes: echo quality check, depolarization ratio check, vertical wind field consistency check and horizontal wind field comparison; The quality control of ground weather station data includes: internal consistency check, spatiotemporal continuity check, threshold check and statistical quality control; At the same time, outliers in the data provided by wind profiler radars and ground meteorological stations are corrected or eliminated, and missing data are interpolated; After quality control is completed, actual observation data is obtained.

[0024] Specifically, by implementing strict quality control measures, such as echo quality check, depolarization ratio check, vertical wind field consistency check and horizontal wind field comparison (for wind profiler radar data), as well as internal consistency check, spatiotemporal continuity check, threshold check and statistical quality control (for ground meteorological station data), the accuracy of the data can be significantly improved. It covers all key data collected from wind profiler radars and ground meteorological stations, including vertical wind speed profile data (vertical wind speed and horizontal wind speed) and basic meteorological element data (temperature, pressure, humidity, wind speed and wind direction), ensuring the comprehensiveness and completeness of the data, and clarifying the processing strategy for outliers, that is, correction or elimination, which helps to eliminate errors or unreasonable values ​​in the data and further improve the reliability of the data. Missing data is interpolated to ensure the continuity and completeness of the data, which is crucial for subsequent data analysis and application. Through systematic data collection and quality control processes, the level of automation in data processing can be improved, manual intervention can be reduced, and work efficiency can be improved.

[0025] In order to solve the problem in the prior art that the collected wind field data is not further processed and the wind field data is not modeled, which leads to poor inversion effect when performing wind field inversion, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions: Data from different sources are fused and data preprocessed after data fusion is completed, including: Fuse the wind profile radar data and ground meteorological station data that have completed quality control in the actual observation data; Before data fusion, the wind profile radar data and ground meteorological station data that have completed quality control in the actual observation data are quality assessed. The quality assessment uses the isolation forest algorithm to identify outliers and noise points in wind speed, wind direction and other data, and marks or removes them. The reliability of ground meteorological station data is assessed using the Bayesian network model, and the rationality of the data is judged based on the correlation between different meteorological elements. Finally, the data with quality assessment within the high quality standard is selected for data fusion. The data fusion process includes: matching wind profiler radar data with ground meteorological station data, where data matching includes time matching and space matching; time matching is to match wind profiler radar data with ground meteorological station data in time according to timestamps; space matching is to match wind profiler radar data with ground meteorological station data in space according to geographic locations.

[0026] Perform data interpolation on the wind profile radar data and ground meteorological station data after data matching, and the data interpolation includes vertical interpolation and horizontal interpolation; Among them, vertical interpolation is to use linear interpolation or bilinear interpolation on the wind speed profile data of the wind profile radar to interpolate the data to the observation height of the ground meteorological station and supplement the data of different altitude layers; horizontal interpolation is to expand the observation data of the ground meteorological station to the area covered by the radar through the Kriging interpolation method; The wind profile radar data and ground meteorological station data after data interpolation are weighted by using the weighted average method; Before weighted processing, a weighted strategy is formulated, which includes time weight, space weight and data quality weight; Time weight is to assign a higher weight to near-real-time observation data during the fusion process of wind profile radar data and ground meteorological station data after data interpolation. If the timestamp of the data is very close to the current time, the data can be given a higher weight; Spatial weight is calculated by calculating the spatial distance between radar data and weather station data, and higher weight is given to points with closer distance; The data quality weight is dynamically adjusted based on the quality assessment of each data source; The weighted wind profile radar data and the data from the ground meteorological station data are combined to obtain fused data of the wind profile radar data and the ground meteorological station data; Perform data preprocessing on the fused actual observation data; After data preprocessing, wind field fusion data is obtained.

[0027] Specifically, through data fusion, data from different sources (wind profiler radar and ground meteorological station) can be integrated to enhance data integrity. This integration can fill in the data gaps or missing data that may exist in a single data source and improve the overall availability of data. Data matching (including time matching and space matching) ensures that the fused data is consistent in time and space, which helps to reduce errors caused by inconsistent data sources. Data interpolation (vertical interpolation and horizontal interpolation) can supplement and refine data, making the data more detailed and accurate in both vertical and horizontal directions. Horizontal interpolation (such as Kriging interpolation) can expand the data coverage of ground meteorological stations so that they can cover the area monitored by radar, thereby expanding the spatial coverage of the data. Time matching ensures the continuity of data in time, which is helpful for time series analysis or prediction. A variety of data fusion and preprocessing techniques (such as linear interpolation, bilinear interpolation, Kriging interpolation and weighted average method) are used. These methods can be selected and adjusted according to the characteristics and needs of the data, thereby improving the flexibility and adaptability of data fusion. By integrating and optimizing data from different sources, the scheme can provide more comprehensive and accurate meteorological information, which is of great significance to fields such as meteorological forecasting, climate research and environmental monitoring. Among them, quality assessment can significantly improve the accuracy of data by identifying and processing outliers and noise points. This helps to reduce data errors caused by factors such as equipment failure, data transmission errors or environmental interference. The reliability of ground meteorological station data is evaluated by using the Bayesian network model, which can comprehensively consider the correlation between multiple meteorological elements, so as to more accurately judge the rationality of the data. This evaluation method helps to screen out high-quality data and provide a reliable basis for subsequent data fusion. By evaluating the quality of the data, it can be ensured that only high-quality data is used for fusion. This helps to reduce the error accumulation in the data fusion process and improve the accuracy and reliability of the fused data. The formulation of a weighting strategy can ensure that in the data fusion process, data from different sources and of different qualities can be reasonably weighted according to their importance. This helps to make the fusion results more in line with the actual situation and improve the scientific nature of data fusion. The formulation of the weighting strategy can be adjusted and optimized according to actual needs. For example, in terms of time weight, near-real-time data can be given a higher weight according to the real-time requirements of the data; in terms of spatial weight, it can be reasonably allocated according to the spatial distribution and data density of radars and weather stations. This flexibility helps to adapt to different application scenarios and data characteristics. By formulating a clear weighting strategy, the data fusion process can be simplified and unnecessary calculation and processing steps can be reduced. This helps to improve the efficiency of data fusion and reduce computing and time costs.

[0028] Select a kinetic model from the model library, initialize the selected kinetic model using the data after data preprocessing, and set the initial state parameters, including: First, the models in the model library are evaluated, including physical properties, mathematical expressions, scope of application and performance records; Confirm the model requirements according to the wind field inversion rules and the evaluation results of the models in the model library, and select the corresponding dynamic model; Initialize the selected dynamic model using wind field fusion data; The initialization process includes: confirming the initialization indicators in the wind field fusion data, and the initialization standards include wind speed, wind direction, temperature, pressure and humidity data; According to the initialization index in the wind field fusion data, the initialization state parameters of the selected dynamic model are set; Among them, the initialization state parameters are set to use the dynamic model and its corresponding initialization method, and the initialization indicators in the wind field fusion data are input through a function or class method, and after the input, the initial state configuration of the selected dynamic model is started; After the initialization state parameters are set, the wind field dynamics model is obtained.

[0029] Specifically, by first evaluating the models in the model library, including physical characteristics, mathematical expressions, scope of application and performance records, the scientificity and accuracy of the selected model are ensured. This helps to select the most suitable dynamic model for the current wind farm characteristics, confirm the model requirements according to the wind farm inversion rules, and select the model in combination with the evaluation results of the model library, ensuring that the selected model is highly targeted and applicable. This helps to improve the prediction accuracy and reliability of the model. The model is initialized using wind farm fusion data, which fully integrates data information from multiple sources and improves the comprehensiveness and accuracy of the data. This helps the model to more accurately reflect the actual situation of the wind farm. The initialization indicators and standards are clarified in the initialization process, including parameter data such as wind speed, wind direction, temperature, pressure and humidity, ensuring the standardization and normalization of initialization. This helps to improve the efficiency and accuracy of model initialization. The initialization indicators in the wind farm fusion data are input into the dynamic model through function or class methods, realizing flexible configuration and initialization of the model. This helps to adapt to the changes in different wind farm characteristics and needs, and facilitates the subsequent expansion and optimization of the model. After the initialization state parameters are set, the wind farm dynamic model can be obtained for subsequent wind farm analysis and prediction. The entire process is efficient and practical, helping to quickly respond to wind farm management needs.

[0030] In order to solve the problem that the wind field inversion model is not simulated more effectively in the existing technology, and the simulation optimization and adjustment are not carried out according to the simulation results, which leads to poor inversion effect, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions: The initialized dynamic model is used to simulate the wind field, and the difference analysis is performed between the wind field simulation data and the actual observation data. According to the difference analysis results, the optimization algorithm is used to optimize and adjust the initialized dynamic model, including: The wind field dynamics model is used to simulate the wind field. Before the wind field simulation is performed, the temporal and spatial scope and resolution of the simulation are first confirmed. At the same time, the physical parameters and boundary conditions of the wind field dynamics model are configured. Finally, the simulation time step and total simulation time are set. After the above steps are completed, the wind field data in the wind field dynamics model is confirmed, and the wind field data includes the spatially distributed wind speed, wind direction, temperature and pressure; A wind field model is established according to the wind field dynamics model, and simulation conditions of the wind field model are set according to the wind field data, and finally a target wind field model is obtained; Divide the target area in the target wind field model into a number of grids using a structured grid method; The wind field parameters of each grid are calculated through time deduction. The steps of calculating the wind field parameters include: Input the parameters of the wind field data into the target wind field model and set the initial state in time and space; For each time step, the discrete equation is used for time integration to calculate the state of the target wind field model at the next time step; In each time step, iterative updates are performed in sequence until the set total time is reached; At each time step, the calculated wind field data is saved in the corresponding data set, and the data set is labeled as simulated wind field data.

[0031] The simulated wind field data and the actual observation data are analyzed for differences. The difference analysis includes root mean square error, correlation coefficient and absolute error. The root mean square error is used to quantify the average deviation between the simulated wind field data and the actual observation data; the correlation coefficient is used to evaluate the similarity between the simulated wind field data and the actual observation data; the absolute error is used to calculate the absolute difference between the simulated wind field data and the actual observation data at each grid point or time step; According to the difference analysis results, the data in the simulated wind farm data that has the greatest impact on the target wind farm model is confirmed; According to the confirmed impact data, the particle swarm optimization method is used to optimize the parameter calculation; Adjust the parameters of the target wind farm model according to the optimized parameters, re-simulate the wind farm after the parameters are adjusted, and iterate the new simulation parameters again until the simulation accuracy of the target wind farm model is within a preset range; After the optimization and adjustment of the target wind field model is completed, the vertical wind field inversion model is obtained.

[0032] Specifically, the efficiency and accuracy of the simulation process are ensured through refined initialization settings, including confirming the temporal and spatial scope and resolution of the simulation, configuring physical parameters and boundary conditions, setting the time step and the total simulation time. The structured grid method is used to divide the target area, which improves the standardization and computational efficiency of grid generation, while ensuring the accuracy of numerical operations. The simulated wind field data includes spatially distributed wind speed, wind direction, temperature and pressure, which fully reflects the characteristics of the wind field. The difference analysis covers the root mean square error, correlation coefficient and absolute error, evaluates the difference between the simulated data and the actual observed data from multiple angles, and provides a comprehensive optimization basis. The particle swarm optimization method is used for parameter calculation optimization. This is an emerging evolutionary computing algorithm with the characteristics of fast convergence speed and strong global search ability, which is suitable for the optimization of complex problems. According to the results of the difference analysis, the data with the greatest impact are optimized and adjusted to ensure the pertinence and effectiveness of the optimization process. By adjusting the parameters iteratively for multiple times and re-simulating the wind field, the real wind field situation is continuously approached, which improves the simulation accuracy of the model. The preset range of simulation accuracy ensures the controllability of the optimization process and the reliability of the results.

[0033] Specifically, the resolution of the simulation is confirmed, including: Retrieve the available memory space corresponding to the wind farm simulation system; Extract the rated memory capacity of the wind farm simulation system; Determine the available memory percentage value corresponding to the wind farm simulation according to the ratio between the available memory space corresponding to the wind farm simulation system and the rated memory capacity; Extracting the change rate of wind field data in the vertical direction, wherein the wind field data includes spatially distributed wind speed, wind direction, temperature and pressure; and the change rate of wind field data in the vertical direction includes the average change rate of wind speed, the average change rate of wind direction angle, the average change rate of temperature and the average change rate of pressure; Extracting the frequency of change of the wind field data in each unit time; The resolution of the wind field simulation in the vertical and horizontal directions is obtained according to the change frequency of the wind field data in each unit time, the change rate of the wind field data in the vertical direction and the percentage value of the available memory corresponding to the wind field simulation.

[0034] The technical effect of the above technical solution is: by calling the available memory space and rated memory capacity of the wind farm simulation system, the available memory ratio is calculated. This enables the system to clearly understand its own memory resource status and reasonably plan the wind farm simulation task according to the availability of memory. For example, when the available memory ratio is high, the complexity of the simulation can be appropriately increased; when the memory is tight, an optimization strategy is adopted or the simulation scale is reduced to avoid simulation interruption or system performance degradation due to insufficient memory, which effectively improves the utilization efficiency of memory resources. Extracting multiple change rates of wind farm data in the vertical direction (average change rate of wind speed, average change rate of wind direction angle, average change rate of temperature and average change rate of pressure), as well as the change frequency per unit time, can effectively improve the comprehensiveness and accuracy of the dynamic characteristics of the wind field. At the same time, the vertical direction change rate reflects the characteristic difference of the wind field in the height dimension, and the change frequency reflects the fluctuation of wind field data over time. The above technical solution provides rich and accurate input for wind field simulation, making the simulation results closer to the actual wind field situation and improving the accuracy and reliability of the simulation.

[0035] On the other hand, the vertical and horizontal resolutions of wind farm simulation are obtained by integrating the frequency of wind farm data changes per unit time, the rate of change in the vertical direction, and the percentage of available memory. The simulation resolution can be adaptively adjusted according to the dynamic characteristics of wind farm data and the system memory resource status, improving the mobility of simulation resolution adjustment and its matching and adjustment sensitivity with the dynamic characteristics of wind farm data, thereby balancing simulation accuracy and system performance under different conditions, and improving the overall performance and adaptability of the wind farm simulation system.

[0036] Specifically, the resolution of the wind field simulation in the vertical and horizontal directions is obtained according to the change frequency of the wind field data in each unit time, the change rate of the wind field data in the vertical direction, and the percentage value of the available memory corresponding to the wind field simulation, including: Extracting the maximum value of the change rate among the change rates of the wind field data in the vertical direction; wherein the change rate of the wind field data in the vertical direction includes the average change rate of wind speed, the average change rate of wind direction angle, the average change rate of temperature and the average change rate of pressure; Normalizing the maximum value of the rate of change of the wind field data in the vertical direction and the value of the available memory ratio corresponding to the wind field simulation to obtain the normalized maximum value of the rate of change and the value of the available memory ratio; Compare the normalized maximum value of the rate of change and the normalized value of the available memory percentage; When the maximum value of the normalized change rate does not exceed the value of the normalized available memory ratio, the simulation resolution is confirmed according to the grid size corresponding to the preset initial vertical resolution and initial horizontal resolution; When the maximum value of the change rate after normalization exceeds the value of the available memory ratio after normalization, the grid sizes corresponding to the initial vertical resolution and the initial horizontal resolution are adjusted using the change frequency of the wind field data in each unit time, the change rate of the wind field data in the vertical direction, and the value of the available memory ratio corresponding to the wind field simulation to obtain the grid sizes corresponding to the adjusted vertical resolution and horizontal resolution; The grid sizes corresponding to the adjusted vertical resolution and horizontal resolution are obtained by the following formula: ; Among them, L c and L s Indicates the grid size corresponding to the adjusted vertical and horizontal resolutions; L c0 and L s0 represents the grid size corresponding to the initial vertical resolution and the initial horizontal resolution; P represents the percentage of available memory corresponding to the wind field simulation; B 01 , B 02 , B 03 and B 04 They represent the average change rate of wind speed, wind direction angle, temperature and pressure respectively; n represents the number of unit time experienced by wind field data collection; f 01i 、f 02i 、f 03i and f 04i represents the frequency of change of wind speed, wind direction, temperature and pressure corresponding to the i-th unit time; f 01c 、f 02c 、f 03c and f 04c Indicates the preset reference value of the frequency of change of wind speed, wind direction, temperature and pressure; The simulation resolution was confirmed by adjusting the grid size corresponding to the vertical and horizontal resolutions.

[0037] The technical effect of the above technical solution is: by extracting the maximum value of the change rate of wind farm data in the vertical direction, the most variable key factor of the wind farm in the vertical direction can be grasped, providing a representative key indicator for subsequent analysis and decision-making, and more accurately grasping the dynamic characteristics of the wind farm. At the same time, the maximum value of the change rate is compared with the value of the available memory ratio, and different resolution confirmation strategies are adopted according to the comparison results, so that the system can make more targeted and flexible decisions based on the actual situation of wind farm changes and memory resources, avoiding a one-size-fits-all approach, and improving the system's adaptability to different wind farm conditions and memory conditions.

[0038] On the other hand, the maximum value of the rate of change and the percentage of available memory are normalized so that the two can be compared and analyzed on the same scale, and the simulation resolution can be determined more reasonably according to the memory resources. When the maximum value of the rate of change does not exceed the percentage of available memory, the preset initial resolution is used to ensure that when resources are sufficient and the wind field changes are relatively stable, the simulation can run in a relatively stable and efficient manner, making full use of system resources. When the maximum value of the rate of change exceeds the percentage of available memory, it means that the wind field changes more drastically and more computing resources are required to ensure the accuracy of the simulation. At this time, the grid size corresponding to the initial resolution is adjusted using multi-dimensional data. Under the premise of ensuring the simulation accuracy, it can avoid excessive consumption of memory resources due to excessive resolution, prevent the system from freezing or crashing, and achieve a balance between simulation accuracy and system performance. At the same time, according to the change frequency of wind field data, the vertical change rate and the percentage of available memory, the grid size corresponding to the resolution is adjusted, so that the simulation resolution can be adjusted in real time with the dynamic changes of the wind field and the changes of memory resources. It can not only improve the accuracy, efficiency and accuracy of capturing the details and changes of the wind field, but also improve the matching between the simulation and the memory percentage.

[0039] At the same time, the grid size corresponding to the adjusted vertical resolution and horizontal resolution is calculated through the formula, and the simulation resolution is confirmed according to the adjusted size, which provides an accurate quantitative basis for the adjustment of the simulation resolution and ensures that the appropriate resolution can be obtained under various wind field conditions and memory resources. This ensures the quality and reliability of wind field simulation and makes the simulation results closer to the actual wind field conditions.

[0040] Specifically, the target area in the target wind field model is divided into several grids using a structured grid method, including: Extracting a target area in a target wind field model; Comparing the area value of the target area in the target wind field model with a preset area value reference value to obtain a ratio between the area value of the target area and the preset area value reference value; Comparing a ratio between an area value of the target area and a preset area value reference value with a preset ratio threshold; When the ratio between the area value of the target area and the preset area value reference value does not exceed the preset ratio threshold, the target area is structurally divided using the initial target area grid size to obtain a plurality of grids; When the ratio between the area value of the target area and the preset area value reference value exceeds a preset ratio threshold, the size change rate between the grid size of the adjusted vertical resolution and the grid size of the initial vertical resolution is retrieved as a first size change rate value; Retrieving a size change rate between a grid size of the adjusted horizontal resolution and a grid size of the initial horizontal resolution as a second size change rate value; The initial target area grid size is adjusted by using the ratio between the area value of the target area and the preset area value reference value in combination with the first size change rate value and the second size change rate value to obtain an adjusted target area grid size; The adjusted target area grid size is obtained by the following formula: ; Where D represents the adjusted target area grid size; D0 represents the initial target area grid size; P d The ratio between the area value of the target area and the preset area value reference value; P c Indicates the value of the first dimensional change rate; P s Indicates the value of the second dimension change rate; The target area is structurally divided according to the adjusted target area grid size to obtain a plurality of grids.

[0041] The technical effect of the above technical solution is: by extracting the target area in the target wind field model and comparing its area value with the preset area value reference value, the size characteristics of the target area can be accurately grasped relative to the preset standard, providing a basic basis for the subsequent grid division strategy, so that the division process can make reasonable decisions for target areas of different sizes. Comparing the above ratio with the preset ratio threshold, and taking different processing methods according to the comparison results, the system can flexibly select the grid division strategy according to the actual area of ​​the target area, which can effectively improve the accuracy of grid size determination.

[0042] At the same time, when the ratio exceeds the threshold, the change rate between the adjusted vertical resolution and horizontal resolution grid size and the initial grid size is retrieved as the first size change rate and the second size change rate value, taking into account the impact of the resolution change in the vertical and horizontal directions of the wind field simulation on the grid size, making the grid size adjustment more comprehensive and accurate. By using the ratio of the target area area value to the preset area value reference value, combined with the two size change rate values, the initial target area grid size is adjusted through a specific formula, providing an accurate quantitative method for determining the grid size. This precise calculation based on multiple factors can accurately adjust the grid size according to the specific conditions of the target area and the resolution change requirements, ensuring that each grid can cover the target area more reasonably and improving the accuracy of grid division.

[0043] On the other hand, for areas whose areas are not much different from the preset reference values, the initial target area grid size is directly used for structural division, which avoids unnecessary complex calculations, improves the efficiency of grid division, and can quickly complete the division of a large number of conventional areas, saving computing resources and time costs. For areas with larger areas or areas that differ greatly from the preset values, by introducing the size change rate to adjust the grid size, a finer grid division can be achieved in these complex areas, thereby more accurately capturing the detailed information of the target area in the wind farm simulation and improving the accuracy of the simulation. This approach of adopting different strategies according to different situations achieves a balance between computing efficiency and simulation accuracy, which not only ensures the overall computing speed, but also provides high-precision simulation results when needed.

[0044] The optimized and adjusted dynamic model is used to generate a three-dimensional wind field, and the generated three-dimensional risk is inverted using visualization technology to display the results, including: The vertical wind field inversion model is used to generate a three-dimensional wind field using a three-dimensional model generator, wherein the wind field parameter data in the vertical wind field inversion model is input into the three-dimensional model generator, and after the three-dimensional model generator is started, a three-dimensional wind field model generated by the vertical wind field inversion model is obtained; Visualize the three-dimensional wind field model. The visualization includes: using streamline diagrams to show the flow direction and intensity of the wind field, using vector field visualization to show the wind speed and direction of each grid point in three-dimensional space, and visualizing the spatial distribution of temperature and humidity; The three-dimensional wind field model after visualization processing is visualized and outputted, and three-dimensional wind field visualization data of the vertical wind field inversion model is obtained.

[0045] Specifically, through the three-dimensional model generator and visualization processing technology, complex wind field data can be presented in an intuitive three-dimensional form. This intuitiveness enables researchers and managers to more easily understand and analyze the flow direction and intensity of the wind field, the distribution of wind speed and wind direction, and the spatial distribution of temperature and humidity. The wind field parameter data of the vertical wind field inversion model is used as input to ensure the accuracy of the three-dimensional wind field model. This accuracy is crucial for scientific research, meteorological forecasting, environmental assessment and other fields. The application of three-dimensional model generator and visualization processing technology can efficiently process and analyze a large amount of wind field data. This improves the speed and efficiency of data processing, allowing researchers to obtain the required analysis results more quickly, and supports real-time updating and visualization output of data, which means that researchers can obtain the latest wind field data at any time and analyze and display it in real time. This is particularly important for occasions that require real-time monitoring and early warning. It not only supports traditional wind field display methods such as streamline diagrams and vector field visualization, but also displays additional information such as the spatial distribution of temperature and humidity. This versatility enables the solution to meet the needs of different fields and research purposes. Through intuitive 3D wind farm visualization data, researchers and managers can better understand the characteristics and changing trends of wind farms and make more informed decisions. This is of great significance to fields such as weather forecasting, wind energy development, and environmental protection. It can monitor and warn of potential risks in wind farms in real time, such as extreme weather events and equipment failures. This helps to take measures in advance to reduce the losses and impacts caused by risks.

[0046] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0047] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A method for inverting low-altitude atmospheric three-dimensional wind fields based on vertical observation of wind profiler radar, characterized in that: The three-dimensional wind field inversion method is to construct a dynamic model based on the acquired data, simulate the wind field with the constructed dynamic model, generate a three-dimensional wind field using the simulation results and the actual observation data using an inversion algorithm, and then visualize the generated three-dimensional wind field; The three-dimensional wind field inversion method comprises: collecting and quality controlling data provided by wind profile radar and ground meteorological station; Data from different sources are fused and data preprocessed after data fusion is completed; Select a kinetic model from the model library, initialize the selected kinetic model using the data after data preprocessing, and set initial state parameters; The initialized dynamic model is used to simulate the wind field, and the difference between the wind field simulation data and the actual observation data is analyzed. According to the difference analysis results, the initialized dynamic model is optimized and adjusted using the optimization algorithm; The optimized and adjusted dynamic model is used to generate a three-dimensional wind field, and the generated three-dimensional risk is inverted using visualization technology to display the results.

2. The method for inverting low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar according to claim 1 is characterized in that: Collect and quality control data from wind profiler radars and ground-based meteorological stations, including: The vertical wind speed profile data from the wind profile radar is collected, and the vertical wind speed profile data includes vertical wind speed and horizontal wind speed data of the time series; Collect data from ground weather stations, including basic meteorological element data such as temperature, pressure, humidity, wind speed and wind direction; Conduct quality control on the data provided by wind profiler radar and ground meteorological station respectively; The quality control of wind profile radar data includes: echo quality check, depolarization ratio check, vertical wind field consistency check and horizontal wind field comparison; The quality control of ground weather station data includes: internal consistency check, spatiotemporal continuity check, threshold check and statistical quality control; At the same time, outliers in the data provided by wind profiler radars and ground meteorological stations are corrected or eliminated, and missing data are interpolated; After quality control is completed, actual observation data is obtained.

3. The method for inverting low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar according to claim 2 is characterized in that: Data from different sources are fused and data preprocessed after data fusion is completed, including: Fuse the wind profile radar data and ground meteorological station data that have completed quality control in the actual observation data; Before data fusion, the wind profile radar data and ground meteorological station data that have completed quality control in the actual observation data are quality assessed. The quality assessment uses the isolation forest algorithm to identify outliers and noise points in the wind speed and wind direction data, and marks or removes them. The reliability of the ground meteorological station data is assessed using the Bayesian network model, and the rationality of the data is judged based on the correlation between different meteorological elements. Finally, the data with quality assessment within the high quality standard is selected for data fusion. The data fusion process includes: matching wind profiler radar data with ground meteorological station data, where data matching includes time matching and space matching; time matching is to match wind profiler radar data with ground meteorological station data in time according to timestamps; space matching is to match wind profiler radar data with ground meteorological station data in space according to geographic locations.

4. The method for inverting low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar according to claim 3 is characterized in that: Data from different sources are fused and data preprocessed after data fusion, including: Perform data interpolation on the wind profile radar data and ground meteorological station data after data matching, and the data interpolation includes vertical interpolation and horizontal interpolation; Among them, vertical interpolation is to use linear interpolation or bilinear interpolation on the wind speed profile data of the wind profile radar to interpolate the data to the observation height of the ground meteorological station and supplement the data of different altitude layers; horizontal interpolation is to expand the observation data of the ground meteorological station to the area covered by the radar through the Kriging interpolation method; The wind profile radar data and ground meteorological station data after data interpolation are weighted by using the weighted average method; Before weighted processing, a weighted strategy is formulated, which includes time weight, space weight and data quality weight; Time weight is the weight assigned to near-real-time observation data during the fusion process of wind profile radar data and ground meteorological station data after data interpolation. If the timestamp of the data is very close to the current time, the data will be given a higher weight; Spatial weights are calculated by calculating the spatial distance between radar data and weather station data, and weights are assigned based on points that are close; The data quality weight is dynamically adjusted based on the quality assessment of each data source; The weighted wind profile radar data and the data from the ground meteorological station data are combined to obtain fused data of the wind profile radar data and the ground meteorological station data; Perform data preprocessing on the fused actual observation data; After data preprocessing, wind field fusion data is obtained.

5. The method for inverting low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar according to claim 4 is characterized in that: Select a kinetic model from the model library, initialize the selected kinetic model using the data after data preprocessing, and set the initial state parameters, including: First, the models in the model library are evaluated, including physical properties, mathematical expressions, scope of application and performance records; Confirm the model requirements according to the wind field inversion rules and the evaluation results of the models in the model library, and select the corresponding dynamic model; Initialize the selected dynamic model using wind field fusion data; The initialization process includes: confirming the initialization indicators in the wind field fusion data, and the initialization standards include wind speed, wind direction, temperature, pressure and humidity data; According to the initialization index in the wind field fusion data, the initialization state parameters of the selected dynamic model are set; Among them, the initialization state parameters are set to use the dynamic model and its corresponding initialization method, and the initialization indicators in the wind field fusion data are input through a function or class method, and after the input, the initial state configuration of the selected dynamic model is started; After the initialization state parameters are set, the wind field dynamics model is obtained.

6. The method for inverting low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar according to claim 5, characterized in that: The initialized dynamic model is used to simulate the wind field, and the difference analysis is performed between the wind field simulation data and the actual observation data. According to the difference analysis results, the optimization algorithm is used to optimize and adjust the initialized dynamic model, including: The wind field dynamics model is used to simulate the wind field. Before the wind field simulation is performed, the temporal and spatial scope and resolution of the simulation are first confirmed. At the same time, the physical parameters and boundary conditions of the wind field dynamics model are configured. Finally, the simulation time step and total simulation time are set. After the steps are completed, the wind field data in the wind field dynamics model is confirmed, and the wind field data includes the spatially distributed wind speed, wind direction, temperature and pressure; A wind field model is established according to the wind field dynamics model, and simulation conditions of the wind field model are set according to the wind field data, and finally a target wind field model is obtained; Divide the target area in the target wind field model into a number of grids using a structured grid method; The wind field parameters of each grid are calculated through time deduction. The steps of calculating the wind field parameters include: Input the parameters of the wind field data into the target wind field model and set the initial state in time and space; For each time step, the discrete equation is used for time integration to calculate the state of the target wind field model at the next time step; In each time step, iterative updates are performed in sequence until the set total time is reached; At each time step, the calculated wind field data is saved in the corresponding data set, and the data set is marked as simulated wind field data; The simulated wind field data and the actual observation data are analyzed for differences. The difference analysis includes root mean square error, correlation coefficient and absolute error. The root mean square error is used to quantify the average deviation between the simulated wind field data and the actual observation data; the correlation coefficient is used to evaluate the similarity between the simulated wind field data and the actual observation data; the absolute error is used to calculate the absolute difference between the simulated wind field data and the actual observation data at each grid point or time step; According to the difference analysis results, the data in the simulated wind farm data that has the greatest impact on the target wind farm model is confirmed; According to the confirmed impact data, the particle swarm optimization method is used to optimize the parameter calculation; Adjust the parameters of the target wind farm model according to the optimized parameters, re-simulate the wind farm after the parameters are adjusted, and iterate the new simulation parameters again until the simulation accuracy of the target wind farm model is within a preset range; After the optimization and adjustment of the target wind field model is completed, the vertical wind field inversion model is obtained.

7. The method for inverting low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar according to claim 6, characterized in that: Verify the resolution of the simulation, including: Retrieve the available memory space corresponding to the wind farm simulation system; Extract the rated memory capacity of the wind farm simulation system; Determine the available memory percentage value corresponding to the wind farm simulation according to the ratio between the available memory space corresponding to the wind farm simulation system and the rated memory capacity; Extracting the change rate of wind field data in the vertical direction, wherein the wind field data includes spatially distributed wind speed, wind direction, temperature and pressure; and the change rate of wind field data in the vertical direction includes the average change rate of wind speed, the average change rate of wind direction angle, the average change rate of temperature and the average change rate of pressure; Extracting the frequency of change of the wind field data in each unit time; The resolution of the wind field simulation in the vertical and horizontal directions is obtained according to the change frequency of the wind field data in each unit time, the change rate of the wind field data in the vertical direction and the percentage value of the available memory corresponding to the wind field simulation.

8. The method for inverting low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar according to claim 7, characterized in that: The resolution of the wind field simulation in the vertical and horizontal directions is obtained according to the frequency of change of the wind field data in each unit time, the rate of change of the wind field data in the vertical direction, and the percentage of available memory corresponding to the wind field simulation, including: Extracting the maximum value of the change rate among the change rates of the wind field data in the vertical direction; wherein the change rate of the wind field data in the vertical direction includes the average change rate of wind speed, the average change rate of wind direction angle, the average change rate of temperature and the average change rate of pressure; Normalizing the maximum value of the rate of change of the wind field data in the vertical direction and the value of the available memory ratio corresponding to the wind field simulation to obtain the normalized maximum value of the rate of change and the value of the available memory ratio; Compare the normalized maximum value of the rate of change and the normalized value of the available memory percentage; When the maximum value of the normalized change rate does not exceed the value of the normalized available memory ratio, the simulation resolution is confirmed according to the grid size corresponding to the preset initial vertical resolution and initial horizontal resolution; When the maximum value of the change rate after normalization exceeds the value of the available memory ratio after normalization, the grid sizes corresponding to the initial vertical resolution and the initial horizontal resolution are adjusted using the change frequency of the wind field data in each unit time, the change rate of the wind field data in the vertical direction, and the value of the available memory ratio corresponding to the wind field simulation to obtain the grid sizes corresponding to the adjusted vertical resolution and horizontal resolution; The simulation resolution was confirmed by adjusting the grid size corresponding to the vertical and horizontal resolutions.

9. The method for inverting low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar according to claim 8, characterized in that: The target area in the target wind field model is divided into several grids using the structured grid method, including: Extracting a target area in a target wind field model; Comparing the area value of the target area in the target wind field model with a preset area value reference value to obtain a ratio between the area value of the target area and the preset area value reference value; Comparing a ratio between an area value of the target area and a preset area value reference value with a preset ratio threshold; When the ratio between the area value of the target area and the preset area value reference value does not exceed the preset ratio threshold, the target area is structurally divided using the initial target area grid size to obtain a plurality of grids; When the ratio between the area value of the target area and the preset area value reference value exceeds a preset ratio threshold, the size change rate between the grid size of the adjusted vertical resolution and the grid size of the initial vertical resolution is retrieved as a first size change rate value; Retrieving a size change rate between a grid size of the adjusted horizontal resolution and a grid size of the initial horizontal resolution as a second size change rate value; The initial target area grid size is adjusted by using the ratio between the area value of the target area and the preset area value reference value in combination with the first size change rate value and the second size change rate value to obtain an adjusted target area grid size; The target area is structurally divided according to the adjusted target area grid size to obtain a plurality of grids.

10. The method for inverting low-altitude atmospheric three-dimensional wind field based on vertical observation of wind profiler radar according to claim 9, characterized in that: The optimized and adjusted dynamic model is used to generate a three-dimensional wind field, and the generated three-dimensional risk is inverted using visualization technology to display the results, including: The vertical wind field inversion model is used to generate a three-dimensional wind field using a three-dimensional model generator, wherein the wind field parameter data in the vertical wind field inversion model is input into the three-dimensional model generator, and after the three-dimensional model generator is started, a three-dimensional wind field model generated by the vertical wind field inversion model is obtained; Visualize the three-dimensional wind field model. The visualization includes: using streamline diagrams to show the flow direction and intensity of the wind field, using vector field visualization to show the wind speed and direction of each grid point in three-dimensional space, and visualizing the spatial distribution of temperature and humidity; The three-dimensional wind field model after visualization processing is visualized and outputted, and three-dimensional wind field visualization data of the vertical wind field inversion model is obtained.

Images