Multi-scale cloud system dynamic evolution simulation modeling method and system based on digital twinning

Abstract

The invention provides a multi-scale cloud system dynamic evolution simulation modeling method and system based on digital twinning, and relates to the technical field of digital twinning, and the method comprises the steps: obtaining cloud system multi-dimensional data through a multi-source cooperative sensing network, extracting cloud system primitive features through a recursive segmentation algorithm after data registration and quality evaluation, and dividing cloud cluster regions; then, constructing a deep digital twin mapping network of cloud system primitive features and a real cloud system, and carrying out feature verification and parameter optimization; and finally, carrying out self-adaptive multi-grid discretization on the cloud system, establishing cross-regional coupling constraint containing a conservation law, representing a dynamic evolution process of the cloud system as a partial differential equation set, carrying out integrated solution with the deep digital twinborn mapping network, and dynamically correcting an evolution prediction result by using real-time digital twinborn data.

Description

Technical Field

[0001] The present invention relates to digital twin technology, and in particular to a multi-scale cloud system dynamic evolution simulation modeling method and system based on digital twin. Background Art

[0002] Existing cloud evolution prediction technologies mainly rely on a single sensor to obtain cloud feature information, which is difficult to fully reflect the complex structure and dynamic change characteristics of the cloud system. In the process of multi-source data fusion, there are problems such as insufficient spatiotemporal registration accuracy and lack of data quality assessment mechanism, which affect the accuracy of cloud feature extraction.

[0003] At present, the cloud boundary division method lacks a clear physical basis, and it is difficult to accurately distinguish the differential characteristics of the core area and the transition area. At the same time, the traditional feature extraction algorithm cannot effectively adapt to the dynamic changes in the morphology and physical characteristics of the cloud system, resulting in inaccurate description of the cloud system evolution characteristics. In terms of numerical simulation, the existing methods use a fixed grid division strategy, which makes it difficult to specifically characterize the evolution characteristics of different regions, and the coupling effect between regions is not sufficiently considered, making it difficult to balance computational efficiency and accuracy.

[0004] In addition, existing cloud evolution prediction methods generally lack a feedback mechanism for real-time observation data, and there is a large deviation between the prediction results and the actual situation. Physical constraints are not fully considered in the prediction correction process, making it difficult to ensure the physical rationality of the correction results. To solve the above problems, it is urgent to establish a cloud evolution prediction method that can comprehensively utilize multi-source data, accurately extract cloud characteristics, efficiently perform numerical simulations, and realize dynamic corrections. Summary of the invention

[0005] The embodiments of the present invention provide a multi-scale cloud system dynamic evolution simulation modeling method and system based on digital twins, which can solve the problems in the prior art.

[0006] According to a first aspect of the embodiments of the present invention, Provides a multi-scale cloud system dynamic evolution simulation modeling method based on digital twins, including: A multi-source collaborative sensing network is constructed using multispectral cameras, lidar and millimeter-wave radar to obtain cloud near-ground data, and satellite remote sensors are used to obtain cloud top data and ground meteorological stations to obtain meteorological element data. Phase consistency analysis is performed on the acquired multi-source data, and a spatial transformation matrix between the sensor coordinate systems is established to obtain the registered data. The registered data is then evaluated for quality based on the signal-to-noise ratio, spatial resolution and temporal consistency, and valid data that meets the evaluation threshold is screened out. Based on the valid data, a recursive segmentation algorithm is used to extract primitive features of the cloud system, the extreme points of the reflectivity gradient and the temperature gradient are calculated to divide the core area of ​​the cloud cluster, and the parameter attenuation curve around the core area is calculated to divide the transition area; based on the regional division results, the contour curvature, grayscale distribution, texture features and multi-scale edge features of the cloud system are extracted, and a deep digital twin mapping network of the primitive features of the cloud system and the real cloud system is constructed, and the morphological feature parameters output by the deep digital twin mapping network are verified with the meteorological measured data, and the parameters of the deep digital twin mapping network are dynamically optimized based on the verification results; An adaptive multi-grid method is used to perform fine discretization on the core area. The grid scale is automatically adjusted with the turbulence intensity and thermal parameter gradient of the cloud cluster. A non-uniform grid that considers the evolution of vortices within the cloud system is established for the transition area. Based on the grid division results, the dynamic evolution of the cloud system is simulated and modeled, and cross-regional coupling constraints including mass conservation, momentum conservation, and energy conservation are established. The dynamic evolution process of the cloud system is characterized as a set of partial differential equations including turbulent diffusion, thermodynamic parameters, and regional coupling. The set of partial differential equations is integrated with the deep digital twin mapping network, and the regional decomposition method is used to solve the overall evolution characteristics of the cloud system. The evolution prediction results are dynamically corrected using real-time digital twin data.

[0007] In an optional embodiment, Based on the effective data, the steps of extracting cloud primitive features by using a recursive segmentation algorithm, calculating the extreme points of reflectivity gradient and temperature gradient to divide the core area of ​​the cloud cluster, and calculating the parameter attenuation curve around the core area to divide the transition area include: A recursive segmentation algorithm is used to extract cloud primitive features. The reflectivity mean and reflectivity standard deviation, temperature mean and temperature standard deviation are calculated for each sub-region. When the reflectivity standard deviation or temperature standard deviation of a sub-region is greater than a preset segmentation threshold, the region is further segmented until the reflectivity standard deviation and temperature standard deviation are less than the preset segmentation threshold or the minimum segmentation scale is reached. The feature vectors of the primitive region are recorded to form a primitive feature set. Based on the primitive feature set, the reflectivity gradient and temperature gradient between adjacent primitives are calculated, the dynamic weight coefficients of the reflectivity and temperature are calculated according to the variance of the reflectivity gradient and the temperature gradient, and the reflectivity gradient, the temperature gradient and their cross terms are weightedly fused to obtain a comprehensive gradient field; The integrated gradient field is decomposed into multiple scales using a Gaussian function, and the local maximum point is extracted from the gradient field at each scale. The local maximum point is established in correspondence between different scales to form an extreme point chain; the duration of the extreme point chain in the scale space is used as a stability index, and the extreme point with the highest stability index is selected as the candidate point of the core area; the candidate points are subjected to spatial connectivity analysis, and the set of candidate points whose Euclidean space distance is less than the set minimum distance threshold is defined as the core area; Calculate the horizontal wind field gradient of the periphery of the core area, calculate the local turbulence intensity according to the horizontal wind field gradient, and combine the local turbulence intensity with reflectivity and temperature to obtain a corrected parameter value; divide the radial distance from the core area to the periphery into multiple intervals, fit the parameter attenuation curve with a double exponential function in each interval according to the corrected parameter value, and impose derivative continuity constraints at the interval connection; calculate the second-order derivative and the third-order derivative of the parameter attenuation curve, take the weighted sum of the absolute value of the second-order derivative and the integral of the third-order derivative as the boundary identification objective function, and solve the position that minimizes the boundary identification objective function as the transition area boundary.

[0008] In an optional embodiment, The steps of constructing a deep digital twin mapping network of the cloud primitive features and the real cloud system, verifying the morphological feature parameters output by the deep digital twin mapping network with the meteorological measured data, and dynamically optimizing the parameters of the deep digital twin mapping network based on the verification results include: The contour curvature, grayscale distribution, texture features and multi-scale edge features of the cloud system are input into a deep digital twin mapping network for feature processing, wherein the deep digital twin mapping network includes a feature encoder, a hierarchical dynamic memory module for storing cloud morphological patterns, and a multi-task decoder; the feature encoder encodes the features based on an adaptive sampling strategy of cloud morphological complexity to obtain encoded features, and the cloud morphological complexity is calculated by a weighted combination of contour curvature variance and texture entropy; The hierarchical dynamic memory module updates the memory unit through the attention mechanism and the gated update method that introduces the physical constraints of temperature gradient and pressure field, and dynamically fuses the memory state and the encoding features to obtain the fused features; based on the fused features, a multi-granularity spatiotemporal attention mechanism is used to process the cloud morphology dependency, and the spatial dependency features are captured by adaptive receptive field adjustment based on the cloud center distance, and a recursive attention unit is introduced to model the temporal dependency, so as to obtain a feature representation with spatiotemporal dependency; the feature representation is input into the multi-task decoder for morphological parameter prediction and reconstruction, and a joint loss function including task loss, physical constraint loss, memory update loss and temporal smoothing loss is constructed to optimize the prediction results, and the weight coefficient of the joint loss function is dynamically adjusted according to the prediction error; The prediction results output by the multi-task decoder are accurately evaluated based on the shape similarity of the Hausdorff distance, and are physically verified through area-perimeter ratio constraints and neighborhood consistency constraints, wherein the area-perimeter ratio constraint limits the value range of the shape coefficient, and the neighborhood consistency constraint ensures that the changes in morphological parameters at adjacent moments conform to the laws of meteorological development, and obtains verification performance indicators; based on the verification performance indicators, a dynamic learning strategy is adopted to optimize the deep digital twin mapping network, wherein the learning rate is dynamically adjusted based on a periodic attenuation function, and the sample weight coefficient is calculated based on the prediction error, and the sample weight coefficient is calculated by the exponential function of the prediction error and the temperature parameter, and the sample weight coefficient is multiplied by the task loss, physical constraint loss, memory update loss and timing smoothing loss respectively to update the network parameters.

[0009] In an optional embodiment, The adaptive multi-grid method is used to finely discretize the core area. The grid scale is automatically adjusted with the cloud turbulence intensity and thermal parameter gradient. The steps of establishing a non-uniform grid for the transition area considering the evolution of the vortex inside the cloud system include: The turbulent kinetic energy spectrum and vorticity field distribution in the core area are calculated based on the meteorological radar reflectivity and Doppler velocity field data. The turbulence intensity distribution at different scales is obtained by analyzing the turbulent kinetic energy spectrum. The vortex structure in the core area is identified by the vortex threshold method, and the vortex center position and vortex intensity are extracted. The temperature field and pressure field of the cloud area are constructed by using meteorological sounding data, the temperature gradient and the pressure gradient are calculated, and the absolute value of the temperature gradient and the absolute value of the pressure gradient are weightedly combined to obtain the thermal parameter gradient intensity index; the turbulence intensity distribution of different scales and the thermal parameter gradient intensity index are input into an exponential function, and a grid scale mapping function is obtained by weighted combination; A quadtree structure is used for multiple grid division, and the grid unit is subdivided according to the ratio of the output value of the grid scale mapping function to the initial grid scale, and the grid unit that meets the subdivision threshold condition is divided into four subgrids, and the difference between adjacent grid levels is limited; A vortex intensity evolution relationship is constructed based on the vortex center position and the vortex intensity, and a grid density attenuation coefficient of a transition region is calculated by a hyperbolic tangent function according to the vortex center distance, wherein the grid density attenuation coefficient increases with increasing distance from the vortex center; The mesh shape quality, orthogonality index and feature capture capability are weightedly combined to obtain a mesh quality evaluation index. When the mesh quality evaluation index is lower than a preset threshold, the mesh cells to be optimized are determined, and the position information of the mesh cells to be optimized is input into the mesh scale mapping function to recalculate the local mesh scale. The mesh topology is updated based on the recalculated mesh scale and the physical field is interpolated and reconstructed.

[0010] In an optional embodiment, Based on the grid division results, the dynamic evolution of the cloud system is simulated and modeled, and cross-region coupling constraints including mass conservation, momentum conservation and energy conservation are established. The dynamic evolution process of the cloud system is characterized as a set of partial differential equations including turbulent diffusion, thermodynamic parameters and regional coupling. The partial differential equations are integrated with the deep digital twin mapping network, and the regional decomposition method is used to solve the overall evolution characteristics of the cloud system. The steps of dynamically correcting the evolution prediction results using real-time digital twin data include: Establish cross-region coupling constraints of mass conservation, momentum conservation and energy conservation at the interface between the core region and the transition region. The mass conservation constraint is based on the flux continuity conditions of the density field and the velocity field, the momentum conservation constraint is based on the pressure and viscous stress tensors, and the energy conservation constraint is based on the specific internal energy, thermal conductivity and temperature field. A partial differential equation system for cloud system evolution including turbulent diffusion terms, thermodynamic parameter terms and regional coupling terms is constructed, wherein the turbulent diffusion terms are characterized by turbulent kinetic energy, turbulent diffusion coefficient and dissipation rate, the thermodynamic parameter terms are characterized by temperature field, pressure field and density field, the regional coupling terms are determined by interface flux balance, and the partial differential equation system relates the cross-region coupling constraint to the turbulent diffusion terms, thermodynamic parameter terms and regional coupling terms; Inputting the partial differential equation group into the deep digital twin mapping network, establishing a mapping loss function including a physical constraint loss term and a conservation constraint loss term, and optimizing the parameters of the deep digital twin mapping network by minimizing the mapping loss function; The output results of the deep digital twin mapping network are solved by adopting a regional decomposition method, and the physical parameters of the transition region are fixed based on the distribution characteristics of the turbulent diffusion term and the thermodynamic parameter term to solve the core region, and the updated physical parameters of the core region are used as boundary conditions to solve the physical parameters of the transition region, and the physical parameters of the core region and the transition region are iterated repeatedly until the physical parameters of the core region and the transition region reach the preset convergence conditions, and the physical parameters include velocity field, density field, turbulent kinetic energy, turbulent diffusion coefficient, dissipation rate, temperature field, pressure field, specific internal energy, and thermal conductivity; The cloud system near-ground data and the cloud system top data are received as real-time digital twin data, a Kalman gain matrix is ​​constructed based on the prediction error covariance of the physical parameters of the core area and the transition area and the observation error covariance of the real-time digital twin data, error analysis is performed between the real-time digital twin data and the solution result of the regional decomposition method, and the Kalman gain matrix is ​​used to dynamically correct the cloud system evolution prediction result.

[0011] In an optional embodiment, Inputting the partial differential equation group into the deep digital twin mapping network, establishing a mapping loss function including a physical constraint loss term and a conservation constraint loss term, and optimizing the parameters of the deep digital twin mapping network by minimizing the mapping loss function comprises: The physical constraint loss term includes turbulent dynamics loss and thermodynamic coupling loss. The turbulent dynamics loss is constructed based on the spatiotemporal derivatives of turbulent kinetic energy and dissipation rate and the turbulence generation term. The weight coefficient of the turbulent dynamics loss is adaptively adjusted by the local turbulence intensity and the turbulence Reynolds number. The thermodynamic coupling loss is constructed based on the convection-diffusion equation of the temperature field and the pressure field. The weight coefficient of the thermodynamic coupling loss is dynamically determined by the nonlinear function of the temperature gradient and the pressure gradient. The conservation constraint loss term is constructed based on the cross-region coupling constraint, wherein the conservation constraint loss term includes a mass conservation constraint, a momentum conservation constraint, and an energy conservation constraint, and a stress tensor correction term based on vorticity correlation is introduced; Constructing a feature-driven combined loss function, wherein the combined loss function performs a weighted combination of the physical constraint loss term and the conservation constraint loss term, wherein the weight coefficient of the combined loss function is dynamically adjusted by the cloud system morphology complexity, temperature gradient, and pressure gradient, wherein the cloud system morphology complexity is calculated by a weighted combination of profile curvature variance and texture entropy; A multi-scale residual correction mechanism is introduced, and the multi-scale residual correction mechanism includes local residuals and global residuals. The local residuals are calculated based on turbulence characteristic quantities, and the global residuals are calculated based on thermodynamic quantities. The weight coefficients of the local residuals and the global residuals are dynamically determined by the ratio of the turbulence characteristic scale to the cloud characteristic scale. The parameters of the deep digital twin mapping network are optimized and updated by minimizing the combined loss function and combining the multi-scale residual correction results, wherein the learning rate is adaptively adjusted based on the gradient of physical quantities and the prediction error, and the physical quantities include velocity field, vorticity field, pressure field, density field, temperature field, turbulent kinetic energy and dissipation rate. At the same time, a periodic attenuation mechanism is introduced. At the end of each training cycle, the learning rate is periodically adjusted according to the degree of satisfaction of the physical constraint loss term and the conservation constraint loss term.

[0012] In an optional embodiment, The steps of receiving the cloud system near-ground data and the cloud system top data as real-time digital twin data, constructing a Kalman gain matrix based on the prediction error covariance of the physical parameters of the core area and the transition area and the observation error covariance of the real-time digital twin data, performing error analysis on the real-time digital twin data and the solution result of the regional decomposition method, and dynamically correcting the cloud system evolution prediction result using the Kalman gain matrix include: Acquire cloud near-ground data including reflectivity, radial velocity and spectral width, fuse multi-site radar data using three-dimensional variational assimilation method, construct three-dimensional wind field, temperature field and water vapor field distribution in the near-ground layer, and determine the ground-based observation error covariance matrix based on historical observation statistics; obtain cloud top data including cloud top height, cloud top temperature and cloud top radiation flux, extract cloud characteristic parameters using infrared-visible light dual-channel inversion algorithm, obtain cloud motion information through multi-time sequence registration, and determine the satellite observation error covariance matrix through channel noise analysis; Based on the difference prediction errors between the prediction results of the physical parameters in the core area and the transition area and the corresponding true values, a prediction error covariance matrix is ​​constructed, and a local correlation constraint based on the spatial distance of the physical parameters is introduced to correct the prediction error covariance matrix, wherein the local correlation constraint is determined by the correlation length scale; A Kalman gain matrix is ​​constructed based on the prediction error covariance matrix, the ground-based observation error covariance matrix and the satellite observation error covariance matrix, and an adaptive adjustment mechanism based on cloud morphology complexity is introduced to correct the Kalman gain matrix, wherein the cloud morphology complexity is calculated by a weighted combination of profile curvature variance and texture entropy; The cloud system near-ground data and the cloud system top data are mapped to the observation space, the observation residual vector with the predicted value is calculated, and the predicted state is updated based on the observation residual vector and the modified Kalman gain matrix; the turbulence characteristic constraint, the mass conservation constraint and the energy conservation constraint are introduced into the updated predicted state, and the turbulence characteristic constraint is constructed based on the non-negativity of turbulent kinetic energy and dissipation rate; the constrained predicted state is used as the initial condition for prediction at the next moment, forming a closed-loop iterative process of dynamic correction.

[0013] According to a second aspect of the embodiments of the present invention, Provide a multi-scale cloud system dynamic evolution simulation modeling system based on digital twins, including: The first unit is used to construct a multi-source collaborative sensing network using multispectral cameras, lidar and millimeter wave radar to obtain cloud system near-ground data, and to obtain cloud system top data and meteorological element data from ground meteorological stations in combination with satellite remote sensors; to perform phase consistency analysis on the acquired multi-source data, to establish a spatial transformation matrix between the sensor coordinate systems to obtain the registered data, to perform quality assessment on the registered data based on the signal-to-noise ratio, spatial resolution and temporal consistency, and to screen out valid data that meets the assessment threshold; The second unit is used to extract the primitive features of the cloud system based on the valid data by using a recursive segmentation algorithm, calculate the extreme points of the reflectivity gradient and the temperature gradient to divide the core area of ​​the cloud cluster, and calculate the parameter attenuation curve around the core area to divide the transition area; based on the regional division result, extract the contour curvature, grayscale distribution, texture features and multi-scale edge features of the cloud system, construct a deep digital twin mapping network of the primitive features of the cloud system and the real cloud system, perform feature verification on the morphological feature parameters output by the deep digital twin mapping network and the meteorological measured data, and dynamically optimize the parameters of the deep digital twin mapping network based on the verification results; The third unit is used to use an adaptive multi-grid method to perform fine discretization on the core area. The grid scale is automatically adjusted with the turbulence intensity and thermal parameter gradient of the cloud cluster, and a non-uniform grid is established for the transition area considering the evolution of vortices within the cloud system. Based on the grid division results, the dynamic evolution of the cloud system is simulated and modeled, and cross-regional coupling constraints including mass conservation, momentum conservation and energy conservation are established. The dynamic evolution process of the cloud system is characterized as a set of partial differential equations including turbulent diffusion, thermodynamic parameters and regional coupling. The set of partial differential equations is integrated with the deep digital twin mapping network, and the regional decomposition method is used to solve the overall evolution characteristics of the cloud system. The evolution prediction results are dynamically corrected using real-time digital twin data.

[0014] According to a third aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0015] A fourth aspect of the embodiments of the present invention is: A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0016] The present invention uses a multi-source collaborative sensing network to obtain comprehensive cloud data, and extracts cloud primitive features through a deep digital twin mapping network, achieving an effective fusion of cloud morphological features and meteorological measured data, thereby improving the accuracy of cloud simulation modeling. The refined grid division strategy and cross-region coupling constraints further enhance the simulation capability of the dynamic evolution of cloud systems.

[0017] The present invention characterizes the dynamic evolution process of the cloud system as a set of partial differential equations, and solves them by combining a deep digital twin mapping network and a regional decomposition method, thereby realizing the prediction of the overall evolution characteristics of the cloud system. The prediction results are dynamically corrected using real-time digital twin data, further improving the accuracy and reliability of the prediction.

[0018] The present invention uses an adaptive multi-grid method to finely discretize the core area, automatically adjusts the grid scale according to the cloud turbulence intensity and thermal parameter gradient, and avoids unnecessary waste of computing resources. The application of the regional decomposition method also improves the efficiency of solving the overall evolution characteristics of the cloud system. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flowchart of a multi-scale cloud system dynamic evolution simulation modeling method based on digital twins according to an embodiment of the present invention; Figure 2 It is a structural schematic diagram of a multi-scale cloud system dynamic evolution simulation modeling system based on digital twins according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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.

[0021] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0022] Figure 1 Schematic diagram of the process of a multi-scale cloud system dynamic evolution simulation modeling method based on digital twins according to an embodiment of the present invention. Figure 1 As shown, the method includes: S1. Use multi-spectral cameras, lidar and millimeter-wave radar to build a multi-source collaborative perception network to obtain cloud near-ground data, and combine satellite remote sensors to obtain cloud top data and ground meteorological stations to obtain meteorological element data; perform phase consistency analysis on the acquired multi-source data, establish a spatial transformation matrix between sensor coordinate systems to obtain registered data, and evaluate the quality of the registered data based on signal-to-noise ratio, spatial resolution, and temporal consistency, and screen out valid data that meets the evaluation threshold; S2. Based on the valid data, a recursive segmentation algorithm is used to extract the primitive features of the cloud system, the extreme points of the reflectivity gradient and the temperature gradient are calculated to divide the core area of ​​the cloud cluster, and the parameter attenuation curve around the core area is calculated to divide the transition area; based on the regional division results, the contour curvature, grayscale distribution, texture features and multi-scale edge features of the cloud system are extracted, and a deep digital twin mapping network of the primitive features of the cloud system and the real cloud system is constructed, and the morphological feature parameters output by the deep digital twin mapping network are verified with the meteorological measured data, and the parameters of the deep digital twin mapping network are dynamically optimized based on the verification results; S3. An adaptive multi-grid method is used to perform fine discretization on the core area. The grid scale is automatically adjusted with the turbulence intensity and thermal parameter gradient of the cloud cluster. A non-uniform grid is established for the transition area considering the evolution of vortices within the cloud system. Based on the grid division results, the dynamic evolution of the cloud system is simulated and modeled, and cross-regional coupling constraints including mass conservation, momentum conservation, and energy conservation are established. The dynamic evolution process of the cloud system is characterized as a set of partial differential equations including turbulent diffusion, thermodynamic parameters, and regional coupling. The set of partial differential equations is integrated with the deep digital twin mapping network, and the regional decomposition method is used to solve the overall evolution characteristics of the cloud system. The evolution prediction results are dynamically corrected using real-time digital twin data.

[0023] In an optional embodiment, Based on the effective data, the steps of extracting cloud primitive features by using a recursive segmentation algorithm, calculating the extreme points of reflectivity gradient and temperature gradient to divide the core area of ​​the cloud cluster, and calculating the parameter attenuation curve around the core area to divide the transition area include: A recursive segmentation algorithm is used to extract cloud primitive features. The reflectivity mean and reflectivity standard deviation, temperature mean and temperature standard deviation are calculated for each sub-region. When the reflectivity standard deviation or temperature standard deviation of a sub-region is greater than a preset segmentation threshold, the region is further segmented until the reflectivity standard deviation and temperature standard deviation are less than the preset segmentation threshold or the minimum segmentation scale is reached. The feature vectors of the primitive region are recorded to form a primitive feature set. Based on the primitive feature set, the reflectivity gradient and temperature gradient between adjacent primitives are calculated, the dynamic weight coefficients of the reflectivity and temperature are calculated according to the variance of the reflectivity gradient and the temperature gradient, and the reflectivity gradient, the temperature gradient and their cross terms are weightedly fused to obtain a comprehensive gradient field; The integrated gradient field is decomposed into multiple scales using a Gaussian function, and the local maximum point is extracted from the gradient field at each scale. The local maximum point is established in correspondence between different scales to form an extreme point chain; the duration of the extreme point chain in the scale space is used as a stability index, and the extreme point with the highest stability index is selected as the candidate point of the core area; the candidate points are subjected to spatial connectivity analysis, and the set of candidate points whose Euclidean space distance is less than the set minimum distance threshold is defined as the core area; Calculate the horizontal wind field gradient of the periphery of the core area, calculate the local turbulence intensity according to the horizontal wind field gradient, and combine the local turbulence intensity with reflectivity and temperature to obtain a corrected parameter value; divide the radial distance from the core area to the periphery into multiple intervals, fit the parameter attenuation curve with a double exponential function in each interval according to the corrected parameter value, and impose derivative continuity constraints at the interval connection; calculate the second-order derivative and the third-order derivative of the parameter attenuation curve, take the weighted sum of the absolute value of the second-order derivative and the integral of the third-order derivative as the boundary identification objective function, and solve the position that minimizes the boundary identification objective function as the transition area boundary.

[0024] Exemplarily, first, perform cloud primitive feature extraction. Input the cloud data to be processed, such as satellite cloud image data. Set the segmentation thresholds of reflectivity and temperature, as well as the minimum segmentation scale of the region. Recursively segment the cloud data: calculate the reflectivity mean, reflectivity standard deviation, temperature mean and temperature standard deviation of each sub-region. If the reflectivity standard deviation or temperature standard deviation of the sub-region is greater than the preset segmentation threshold, and the scale of the sub-region is greater than the minimum segmentation scale, continue to segment the sub-region. Otherwise, stop segmenting, and record the feature vector of the sub-region (such as reflectivity mean, reflectivity standard deviation, temperature mean, temperature standard deviation, region area, region location, etc.) into the primitive feature set. For example, for a piece of cloud data, the initial reflectivity standard deviation is 20, the temperature standard deviation is 5, the segmentation thresholds are set to 10 and 3, and the minimum segmentation scale is 1 square kilometer. After recursive segmentation, several sub-regions are finally obtained, whose reflectivity standard deviation and temperature standard deviation are both less than the set threshold or the scale is less than the minimum segmentation scale, and the features of these sub-regions form a primitive feature set.

[0025] Then, calculate the comprehensive gradient field. Traverse all adjacent primitives in the primitive feature set and calculate the reflectivity gradient and temperature gradient between them. For example, the reflectivities of two adjacent primitives are 60 and 70 respectively, and the distance is 2 kilometers, then the reflectivity gradient is (70-60) / 2=5 / kilometer. According to the variance of the reflectivity gradient and temperature gradient of all adjacent primitive pairs, calculate the dynamic weight coefficients of reflectivity and temperature. For example, if the variance of the reflectivity gradient is large, then assign a larger weight coefficient to the reflectivity gradient. Perform weighted fusion on the reflectivity gradient, temperature gradient and their cross terms (the product of the reflectivity gradient and the temperature gradient) to obtain a comprehensive gradient field.

[0026] Next, extract the core area. Use Gaussian function to decompose the integrated gradient field at multiple scales, for example, use Gaussian kernels with different standard deviations to convolve the gradient field. Extract local maximum points for the gradient field at each scale. Establish correspondence between these local maximum points at different scales to form an extreme point chain. For example, if a point is a local maximum point at three scales, then these three points form an extreme point chain. The duration of the extreme point chain in the scale space is used as a stability index. For example, if an extreme point chain exists at five scales, its stability index is 5. Select the extreme point with the highest stability index as the candidate point for the core area. Perform spatial connectivity analysis on the candidate points, and define the set of candidate points whose Euclidean space distance is less than the set minimum distance threshold as the core area. For example, if the distance between two candidate points is 1 km, which is less than the set minimum distance threshold of 2 km, then the two points belong to the same core area.

[0027] Finally, divide the transition area. Calculate the horizontal wind field gradient outside the core area. Calculate the local turbulence intensity based on the horizontal wind field gradient. Combine the local turbulence intensity with reflectivity and temperature to obtain the corrected parameter value. Divide the radial distance from the core area to the periphery into multiple intervals. For example, divide the distance into three intervals: 0-5 km, 5-10 km, and 10-15 km. In each interval, use a double exponential function to fit the parameter attenuation curve according to the corrected parameter value, and impose derivative continuity constraints at the interval connection to ensure the smoothness of the curve. For example, in the 0-5 km interval, fit a double exponential function, and in the 5-10 km interval fit another double exponential function, and ensure that the derivatives of the two functions at 5 km are equal. Calculate the second and third derivatives of the parameter attenuation curve. The weighted sum of the absolute value of the second-order derivative and the integral of the third-order derivative is used as the boundary identification objective function. Solve the position that minimizes the boundary identification objective function as the transition area boundary.

[0028] The present invention uses a recursive segmentation algorithm to more finely characterize the internal structure of the cloud system and extract more accurate primitive features. The multi-scale analysis and the selection of stability indicators can effectively suppress the influence of noise and improve the accuracy and stability of core area identification. The influence of horizontal wind field gradient and local turbulence intensity is taken into account, and a double exponential function is used to fit the parameter attenuation curve, which can more accurately describe the changing trend of cloud system parameters, thereby more accurately dividing the transition area.

[0029] In an optional embodiment, The steps of constructing a deep digital twin mapping network of the cloud primitive features and the real cloud system, verifying the morphological feature parameters output by the deep digital twin mapping network with the meteorological measured data, and dynamically optimizing the parameters of the deep digital twin mapping network based on the verification results include: The contour curvature, grayscale distribution, texture features and multi-scale edge features of the cloud system are input into a deep digital twin mapping network for feature processing, wherein the deep digital twin mapping network includes a feature encoder, a hierarchical dynamic memory module for storing cloud morphological patterns, and a multi-task decoder; the feature encoder encodes the features based on an adaptive sampling strategy of cloud morphological complexity to obtain encoded features, and the cloud morphological complexity is calculated by a weighted combination of contour curvature variance and texture entropy; The hierarchical dynamic memory module updates the memory unit through the attention mechanism and the gated update method that introduces the physical constraints of temperature gradient and pressure field, and dynamically fuses the memory state and the encoding features to obtain the fused features; based on the fused features, a multi-granularity spatiotemporal attention mechanism is used to process the cloud morphology dependency, and the spatial dependency features are captured by adaptive receptive field adjustment based on the cloud center distance, and a recursive attention unit is introduced to model the temporal dependency, so as to obtain a feature representation with spatiotemporal dependency; the feature representation is input into the multi-task decoder for morphological parameter prediction and reconstruction, and a joint loss function including task loss, physical constraint loss, memory update loss and temporal smoothing loss is constructed to optimize the prediction results, and the weight coefficient of the joint loss function is dynamically adjusted according to the prediction error; The prediction results output by the multi-task decoder are accurately evaluated based on the shape similarity of the Hausdorff distance, and are physically verified through area-perimeter ratio constraints and neighborhood consistency constraints, wherein the area-perimeter ratio constraint limits the value range of the shape coefficient, and the neighborhood consistency constraint ensures that the changes in morphological parameters at adjacent moments conform to the laws of meteorological development, and obtains verification performance indicators; based on the verification performance indicators, a dynamic learning strategy is adopted to optimize the deep digital twin mapping network, wherein the learning rate is dynamically adjusted based on a periodic attenuation function, and the sample weight coefficient is calculated based on the prediction error, and the sample weight coefficient is calculated by the exponential function of the prediction error and the temperature parameter, and the sample weight coefficient is multiplied by the task loss, physical constraint loss, memory update loss and timing smoothing loss respectively to update the network parameters.

[0030] Exemplarily, first, cloud feature extraction is performed. The acquired cloud image data is preprocessed, including denoising, contrast enhancement and other operations to improve image quality. Then, various morphological features of the cloud are extracted, including contour curvature, grayscale distribution, texture features and multi-scale edge features. Contour curvature describes the shape complexity of the cloud by calculating the curvature change of the cloud edge. Grayscale distribution reflects the density and thickness information of the cloud by counting the number of pixels with different grayscale values. Texture features use methods such as grayscale co-occurrence matrix to extract texture information of the cloud, such as roughness, smoothness, etc. Multi-scale edge features detect edges of cloud images using convolution kernels of different sizes to capture edge information of different scales. For example, edge detection is performed using convolution kernels of sizes such as 3x3, 5x5, and 7x7. Assume that the grayscale value range of a cloud image is 0-255, its contour curvature range is 0-1, the texture feature value range is 0-10, and the multi-scale edge intensity range is 0-1.

[0031] Next, a deep digital twin mapping network is constructed. The network consists of a feature encoder, a hierarchical dynamic memory module, and a multi-task decoder. The feature encoder receives the extracted cloud features and performs adaptive sampling according to the cloud morphological complexity. The cloud morphological complexity is calculated by a weighted combination of the contour curvature variance and the texture entropy. For example, the contour curvature variance weight is 0.6 and the texture entropy weight is 0.4. The encoder encodes the features into a compact feature vector. The hierarchical dynamic memory module is used to store cloud morphological patterns. The module updates the memory unit through an attention mechanism and a gated update method that introduces physical constraints of temperature gradients and pressure fields. The memory unit stores the morphological features of different types of clouds. For example, the memory unit stores the typical morphological features of different types of clouds, such as cumulus, stratus, and cirrus. The multi-task decoder receives the fused features and performs morphological parameter prediction and reconstruction.

[0032] Then, the cloud morphological parameters are predicted and reconstructed. The encoded features are fused with the output of the memory module to obtain the fused features. A multi-granular spatiotemporal attention mechanism is used to process the cloud morphological dependencies. The adaptive receptive field adjustment based on the distance from the cloud center captures spatial dependency features. For example, the closer the area is to the center, the smaller the receptive field is, and the farther the area is from the center, the larger the receptive field is. A recursive attention unit is introduced to model the temporal dependency. For example, the feature information at the previous moment is used to predict the cloud morphology at the current moment. The obtained feature representation with spatiotemporal dependency is input into the multi-task decoder for morphological parameter prediction and reconstruction. For example, the area, perimeter, height and other parameters of the cloud are predicted, and the three-dimensional morphology of the cloud is reconstructed.

[0033] Next, the loss function is constructed and optimized. A joint loss function including task loss, physical constraint loss, memory update loss and temporal smoothing loss is constructed to optimize the prediction results. Task loss is used to measure the difference between the predicted parameters and the true parameters. Physical constraint loss is used to ensure that the prediction results conform to physical laws, such as area perimeter ratio constraints and neighborhood consistency constraints. Memory update loss is used to constrain the update process of memory units. Temporal smoothing loss is used to ensure the smoothness of the prediction results at adjacent moments. The weight coefficient of the joint loss function is dynamically adjusted according to the prediction error. For example, if the physical constraint loss is large, its weight coefficient is increased.

[0034] Finally, the network parameters are dynamically optimized. The prediction results output by the multi-task decoder are evaluated for accuracy based on the shape similarity of the Hausdorff distance, and are physically verified through the area-perimeter ratio constraint and the neighborhood consistency constraint. The area-perimeter ratio constraint limits the value range of the shape coefficient, for example, the value range of the shape coefficient is 0.5-1.5. The neighborhood consistency constraint ensures that the changes in the morphological parameters at adjacent moments conform to the laws of meteorological development. For example, the change in the area of ​​the cloud system at adjacent moments does not exceed 10%. After obtaining the verification performance indicators, the dynamic learning strategy is used to optimize the deep digital twin mapping network. The learning rate is dynamically adjusted based on the periodic decay function. For example, the initial value of the learning rate is 0.01, and it decays to 0.9 times the original value every 10 cycles. The sample weight coefficient is calculated based on the prediction error. The sample weight coefficient is calculated by the exponential function of the prediction error and the temperature parameter. The network parameters are updated after the sample weight coefficient is multiplied by the task loss, physical constraint loss, memory update loss, and temporal smoothing loss.

[0035] The present invention constructs a deep digital twin mapping network and combines multi-granularity spatiotemporal attention mechanism and physical constraints to more accurately capture the morphological characteristics and evolution laws of cloud systems, thereby improving the prediction accuracy of cloud morphological parameters; the multi-task decoder can reconstruct the three-dimensional morphology of cloud systems, provide more intuitive and comprehensive cloud information, and help to better understand the formation and development process of clouds; the dynamic memory module and dynamic learning strategy can enable the network to adapt to different types of cloud systems and different meteorological conditions, thereby improving the generalization ability and robustness of the network.

[0036] In an optional embodiment, The adaptive multi-grid method is used to finely discretize the core area. The grid scale is automatically adjusted with the cloud turbulence intensity and thermal parameter gradient. The steps of establishing a non-uniform grid for the transition area considering the evolution of the vortex inside the cloud system include: The turbulent kinetic energy spectrum and vorticity field distribution in the core area are calculated based on the meteorological radar reflectivity and Doppler velocity field data. The turbulence intensity distribution at different scales is obtained by analyzing the turbulent kinetic energy spectrum. The vortex structure in the core area is identified by the vortex threshold method, and the vortex center position and vortex intensity are extracted. The temperature field and pressure field of the cloud area are constructed by using meteorological sounding data, the temperature gradient and the pressure gradient are calculated, and the absolute value of the temperature gradient and the absolute value of the pressure gradient are weightedly combined to obtain the thermal parameter gradient intensity index; the turbulence intensity distribution of different scales and the thermal parameter gradient intensity index are input into an exponential function, and a grid scale mapping function is obtained by weighted combination; A quadtree structure is used for multiple grid division, and the grid unit is subdivided according to the ratio of the output value of the grid scale mapping function to the initial grid scale, and the grid unit that meets the subdivision threshold condition is divided into four subgrids, and the difference between adjacent grid levels is limited; A vortex intensity evolution relationship is constructed based on the vortex center position and the vortex intensity, and a grid density attenuation coefficient of a transition region is calculated by a hyperbolic tangent function according to the vortex center distance, wherein the grid density attenuation coefficient increases with increasing distance from the vortex center; The mesh shape quality, orthogonality index and feature capture capability are weightedly combined to obtain a mesh quality evaluation index. When the mesh quality evaluation index is lower than a preset threshold, the mesh cells to be optimized are determined, and the position information of the mesh cells to be optimized is input into the mesh scale mapping function to recalculate the local mesh scale. The mesh topology is updated based on the recalculated mesh scale and the physical field is interpolated and reconstructed.

[0037] For example, first, obtain weather radar reflectivity and Doppler velocity field data. Take a severe convective weather process as an example, use Doppler weather radar to obtain volume scan data with a time resolution of 6 minutes and a spatial resolution of 1km×1°. Perform quality control on the acquired radar-based data to remove noise and clutter interference.

[0038] Then, the turbulence intensity and vortex characteristics of the core area are calculated. The turbulence kinetic energy spectrum is calculated using radar reflectivity data, and the turbulence intensity distribution at different scales is analyzed. Assume that the turbulence kinetic energy spectrum at a certain grid point is E(k), where k is the wave number. By integrating E(k) in different wavenumber bands, turbulent energy at different scales can be obtained. At the same time, the vorticity field distribution is calculated using Doppler velocity field data, and the vortex structure in the core area is identified using the vortex threshold method, and the vortex center position and vortex intensity are extracted. The coordinate position of the vortex center and the maximum vortex value of the vortex area are recorded as the vortex intensity.

[0039] Next, construct the thermal parameter gradient intensity index of the cloud area. Use meteorological sounding data to construct the temperature field and pressure field of the cloud area. Assume that the time resolution of the sounding data is 12 hours and the spatial resolution is 500 meters. By interpolating the sounding data, the temperature field and pressure field consistent with the spatial resolution of the radar data are obtained. Then, calculate the temperature gradient and pressure gradient, and perform weighted combination of the absolute value of the temperature gradient and the absolute value of the pressure gradient to obtain the thermal parameter gradient intensity index. For example, set the temperature gradient weight to 0.7 and the pressure gradient weight to 0.3, and take the weighted sum of the two to obtain the thermal parameter gradient intensity index. Subsequently, construct a grid scale mapping function. Input the turbulence intensity distribution and thermal parameter gradient intensity index of different scales into the exponential function, and obtain the grid scale mapping function through weighted combination.

[0040] Then, adaptive multi-grid division is performed. The quadtree structure is used for multi-grid division, and the grid unit is subdivided according to the ratio of the output value of the grid scale mapping function to the initial grid scale. Assume that the initial grid scale is 1km and the subdivision threshold is set to 2. If the output value of the grid scale mapping function of a grid unit is 0.5km, the grid unit meets the subdivision condition and is divided into four sub-grids, each with a scale of 0.5km. At the same time, the difference between adjacent grid levels is limited to no more than 1 to ensure a smooth transition of the grid.

[0041] Next, a non-uniform grid is constructed in the transition region. The vortex intensity evolution relationship is constructed based on the vortex center position and vortex intensity, and the grid density attenuation coefficient of the transition region is calculated by the hyperbolic tangent function according to the distance from the vortex center. Assuming that the grid density at the vortex center is 1, the grid density gradually decreases as the distance from the vortex center increases.

[0042] Finally, mesh quality optimization is performed. The mesh shape quality, orthogonality index, and feature capture capability are weighted and combined to obtain the mesh quality evaluation index. Assuming that the mesh shape quality weight is 0.4, the orthogonality index weight is 0.3, and the feature capture capability weight is 0.3, the weighted sum of the three is used to obtain the mesh quality evaluation index. When the mesh quality evaluation index is lower than the preset threshold, the mesh cells to be optimized are determined, and the location information of the mesh cells to be optimized is input into the mesh scale mapping function to recalculate the local mesh scale. Based on the recalculated mesh scale, the mesh topology is updated and the physical field is interpolated and reconstructed.

[0043] The present invention can more accurately capture the fine structure and evolution process inside the cloud cluster. Through the adaptive grid adjustment strategy, it realizes the reasonable allocation of computing resources, improves computing efficiency, enhances the forecast accuracy of severe convective weather processes, and provides a more reliable scientific basis for disaster prevention and mitigation.

[0044] In an optional embodiment, Based on the grid division results, the dynamic evolution of the cloud system is simulated and modeled, and cross-region coupling constraints including mass conservation, momentum conservation and energy conservation are established. The dynamic evolution process of the cloud system is characterized as a set of partial differential equations including turbulent diffusion, thermodynamic parameters and regional coupling. The partial differential equations are integrated with the deep digital twin mapping network, and the regional decomposition method is used to solve the overall evolution characteristics of the cloud system. The steps of dynamically correcting the evolution prediction results using real-time digital twin data include: Establish cross-region coupling constraints of mass conservation, momentum conservation and energy conservation at the interface between the core region and the transition region. The mass conservation constraint is based on the flux continuity conditions of the density field and the velocity field, the momentum conservation constraint is based on the pressure and viscous stress tensors, and the energy conservation constraint is based on the specific internal energy, thermal conductivity and temperature field. A partial differential equation system for cloud system evolution including turbulent diffusion terms, thermodynamic parameter terms and regional coupling terms is constructed, wherein the turbulent diffusion terms are characterized by turbulent kinetic energy, turbulent diffusion coefficient and dissipation rate, the thermodynamic parameter terms are characterized by temperature field, pressure field and density field, the regional coupling terms are determined by interface flux balance, and the partial differential equation system relates the cross-region coupling constraint to the turbulent diffusion terms, thermodynamic parameter terms and regional coupling terms; Inputting the partial differential equation group into the deep digital twin mapping network, establishing a mapping loss function including a physical constraint loss term and a conservation constraint loss term, and optimizing the parameters of the deep digital twin mapping network by minimizing the mapping loss function; The output results of the deep digital twin mapping network are solved by adopting a regional decomposition method, and the physical parameters of the transition region are fixed based on the distribution characteristics of the turbulent diffusion term and the thermodynamic parameter term to solve the core region, and the updated physical parameters of the core region are used as boundary conditions to solve the physical parameters of the transition region, and the physical parameters of the core region and the transition region are iterated repeatedly until the physical parameters of the core region and the transition region reach the preset convergence conditions, and the physical parameters include velocity field, density field, turbulent kinetic energy, turbulent diffusion coefficient, dissipation rate, temperature field, pressure field, specific internal energy, and thermal conductivity; The cloud system near-ground data and the cloud system top data are received as real-time digital twin data, a Kalman gain matrix is ​​constructed based on the prediction error covariance of the physical parameters of the core area and the transition area and the observation error covariance of the real-time digital twin data, error analysis is performed between the real-time digital twin data and the solution result of the regional decomposition method, and the Kalman gain matrix is ​​used to dynamically correct the cloud system evolution prediction result.

[0045] Exemplarily, first, perform spatial discretization. Divide the three-dimensional space where the cloud system is located into a series of non-overlapping grid cells. In order to capture the details of the core area (such as the strong convection area), a finer grid division is used for the core area, while a coarser grid is used for the surrounding transition area. For example, the grid resolution of the core area can be set to 1 km in the horizontal direction and 0.5 km in the vertical direction, and the grid resolution of the transition area can be set to 5 km in the horizontal direction and 2 km in the vertical direction.

[0046] Then, the coupling constraints between the core region and the transition region are established. At the interface between the core region and the transition region, the conservation of physical quantities needs to be guaranteed. The mass conservation constraint requires that the mass flowing into the interface is equal to the mass flowing out of the interface, which can be achieved by the flux balance of density and velocity on the interface. The momentum conservation constraint requires that the pressure and viscosity forces on both sides of the interface are balanced. The energy conservation constraint requires that the energy exchange on both sides of the interface remain balanced, which can be achieved by the flux balance of specific internal energy, thermal conductivity and temperature on the interface.

[0047] Next, a set of partial differential equations for cloud system evolution is constructed. This set of equations describes the spatiotemporal evolution of various physical quantities in the cloud system, and mainly includes three parts: turbulent diffusion terms, thermodynamic parameter terms, and regional coupling terms. The turbulent diffusion term describes the influence of turbulent motion in the cloud system on physical quantities, such as the diffusion of momentum, heat, and water vapor. It can be expressed in terms of turbulent kinetic energy, turbulent diffusion coefficient, and dissipation rate. The thermodynamic parameter term describes the influence of thermodynamic parameters such as temperature, pressure, and density on the evolution of the cloud system. The regional coupling term describes the interaction between the core region and the transition region, which is determined by the flux balance condition on the interface. This set of partial differential equations links factors such as conservation constraints, turbulent diffusion, thermodynamic parameters, and regional coupling to form a complete cloud system evolution model.

[0048] Subsequently, a deep digital twin mapping network is constructed. The above set of partial differential equations is input into a deep neural network, which learns the complex nonlinear relationship of cloud system evolution. In order to train the network, a loss function needs to be constructed, which contains two parts: a physical constraint loss term and a conservation constraint loss term. The physical constraint loss term is used to ensure that the output of the network satisfies the laws of physics, such as conservation of mass, momentum, and energy. The conservation constraint loss term is used to ensure that the conservation constraints between the core region and the transition region are satisfied. By minimizing the loss function, the parameters of the network can be optimized so that it can accurately predict the evolution of the cloud system. For example, a neural network containing convolutional layers and recurrent layers can be used, and the mean square error can be used as the loss function.

[0049] Next, the regional decomposition method is used to solve the problem. The output of the deep digital twin mapping network is used as the input of the regional decomposition method. First, according to the distribution characteristics of turbulent diffusion and thermodynamic parameters, the physical parameters of the transition region (such as velocity, temperature, pressure, etc.) are fixed to solve the physical parameters of the core region. Then, the updated physical parameters of the core region are used as boundary conditions to solve the physical parameters of the transition region. This cycle is repeated until the physical parameters of the core region and the transition region meet the preset convergence conditions, such as the difference between the results of two iterations is less than a certain threshold.

[0050] Finally, dynamic corrections are made using real-time digital twin data. Near-ground and top data of the cloud system are obtained in real time, such as satellite observation data, radar data, and ground meteorological station data. The prediction error covariance of the physical parameters in the core area and the transition area, as well as the error covariance of the real-time observation data, are calculated. The Kalman gain matrix is ​​constructed based on these two covariance matrices. The real-time observation data is fused with the solution results of the regional decomposition method using the Kalman gain matrix to dynamically correct the cloud system evolution prediction results and improve the prediction accuracy. For example, assuming that the observation data includes temperature and wind speed, the Kalman filter algorithm can be used to fuse the observation data with the model prediction results to correct the predicted values ​​of temperature and wind speed.

[0051] By combining conservation laws, deep learning and real-time data, the present invention can more accurately capture the complex dynamic evolution process of cloud systems, thereby improving prediction accuracy; by introducing conservation laws into the model, the physical consistency of the model is ensured, avoiding physical unreasonable phenomena that may occur in traditional numerical models; through the regional decomposition method, complex computing tasks are decomposed into multiple subtasks and solved in parallel, thereby improving computing efficiency and making real-time prediction of cloud systems possible.

[0052] In an optional embodiment, Inputting the partial differential equation group into the deep digital twin mapping network, establishing a mapping loss function including a physical constraint loss term and a conservation constraint loss term, and optimizing the parameters of the deep digital twin mapping network by minimizing the mapping loss function comprises: The physical constraint loss term includes turbulent dynamics loss and thermodynamic coupling loss. The turbulent dynamics loss is constructed based on the spatiotemporal derivatives of turbulent kinetic energy and dissipation rate and the turbulence generation term. The weight coefficient of the turbulent dynamics loss is adaptively adjusted by the local turbulence intensity and the turbulence Reynolds number. The thermodynamic coupling loss is constructed based on the convection-diffusion equation of the temperature field and the pressure field. The weight coefficient of the thermodynamic coupling loss is dynamically determined by the nonlinear function of the temperature gradient and the pressure gradient. The conservation constraint loss term is constructed based on the cross-region coupling constraint, wherein the conservation constraint loss term includes a mass conservation constraint, a momentum conservation constraint, and an energy conservation constraint, and a stress tensor correction term based on vorticity correlation is introduced; Constructing a feature-driven combined loss function, wherein the combined loss function performs a weighted combination of the physical constraint loss term and the conservation constraint loss term, wherein the weight coefficient of the combined loss function is dynamically adjusted by the cloud system morphology complexity, temperature gradient, and pressure gradient, wherein the cloud system morphology complexity is calculated by a weighted combination of profile curvature variance and texture entropy; A multi-scale residual correction mechanism is introduced, and the multi-scale residual correction mechanism includes local residuals and global residuals. The local residuals are calculated based on turbulence characteristic quantities, and the global residuals are calculated based on thermodynamic quantities. The weight coefficients of the local residuals and the global residuals are dynamically determined by the ratio of the turbulence characteristic scale to the cloud characteristic scale. The parameters of the deep digital twin mapping network are optimized and updated by minimizing the combined loss function and combining the multi-scale residual correction results, wherein the learning rate is adaptively adjusted based on the gradient of physical quantities and the prediction error, and the physical quantities include velocity field, vorticity field, pressure field, density field, temperature field, turbulent kinetic energy and dissipation rate. At the same time, a periodic attenuation mechanism is introduced. At the end of each training cycle, the learning rate is periodically adjusted according to the degree of satisfaction of the physical constraint loss term and the conservation constraint loss term.

[0053] Exemplarily, first, a deep digital twin mapping network needs to be constructed. The network can adopt a deep learning architecture, such as a convolutional neural network (CNN) or a recurrent neural network (RNN). The input data includes the initial state and boundary conditions of the cloud system evolution, such as the values ​​of physical quantities such as velocity field, temperature field, and pressure field at the initial moment and boundary. The output is the state of the cloud system at a future moment, such as the predicted values ​​of the velocity field, temperature field, and pressure field at a certain moment in the future. The specific structure of the network can be adjusted according to actual conditions, such as the number of convolutional layers, the size of the convolution kernel, the depth of the network, etc.

[0054] Next, a mapping loss function is constructed to train the network. This loss function contains multiple parts, which are used to constrain the network's prediction results to satisfy physical laws and conservation laws.

[0055] The physical constraint loss term is used to constrain the prediction results of the network to conform to the laws of fluid dynamics and thermodynamics. Among them, the turbulence dynamics loss term is used to constrain the spatiotemporal variation of turbulent kinetic energy and dissipation rate. The weight coefficient of turbulence dynamics loss is adaptively adjusted according to the local turbulence intensity and turbulence Reynolds number. For example, the greater the turbulence intensity, the greater the weight coefficient. The thermodynamic coupling loss term is used to constrain the convection and diffusion laws of the temperature field and pressure field. Its weight coefficient is dynamically determined according to the nonlinear function of the temperature gradient and pressure gradient. For example, the greater the temperature gradient, the greater the weight coefficient.

[0056] Conservation constraint losses are used to constrain the predictions of the network to satisfy the conservation laws of mass, momentum, and energy. For example, the mass conservation loss can be constructed by calculating the difference between the predicted change in the density field within the control volume and the mass flux through the surface of the control volume. Momentum and energy conservation losses can be constructed in a similar way. In addition, the accuracy of the conservation constraints can be improved by introducing a stress tensor correction term based on vorticity correlation.

[0057] The physical constraint loss term and the conservation constraint loss term are weighted combined to construct a feature-driven combined loss function. The weight coefficient of the combined loss function is dynamically adjusted according to the cloud morphology complexity, temperature gradient, and pressure gradient. The cloud morphology complexity can be calculated by a weighted combination of the contour curvature variance and texture entropy. For example, the more complex the cloud morphology, the larger the contour curvature variance and texture entropy, and the larger the corresponding weight coefficient.

[0058] In order to further improve the prediction accuracy, a multi-scale residual correction mechanism is introduced. This mechanism includes local residuals and global residuals. Local residuals are calculated based on turbulence characteristic quantities, such as turbulence pulsation velocity. Global residuals are calculated based on thermodynamic quantities, such as the average value of temperature and pressure. The weight coefficients of local residuals and global residuals are dynamically determined according to the ratio of the turbulence characteristic scale to the cloud characteristic scale. For example, the smaller the turbulence characteristic scale is relative to the cloud characteristic scale, the larger the weight coefficient of the local residual is.

[0059] Finally, the network parameters are optimized and updated by minimizing the combined loss function and combining the multi-scale residual correction results. The learning rate is adaptively adjusted based on the physical quantity gradient and the prediction error. For example, the larger the physical quantity gradient, the larger the prediction error, and the larger the learning rate. At the same time, a periodic attenuation mechanism is introduced. At the end of each training cycle, the learning rate is adjusted in stages according to the satisfaction degree of the physical constraint loss term and the conservation constraint loss term. For example, the smaller the value of the loss term, the higher the model prediction accuracy, and the learning rate can be appropriately reduced.

[0060] By combining physical constraints and conservation constraints and introducing a multi-scale residual correction mechanism, the present invention can effectively improve the prediction accuracy of cloud evolution and more accurately capture the evolution process of the cloud system; by adaptively adjusting the weight coefficient and learning rate of the loss function, the network can better adapt to different cloud forms and evolution processes and enhance the generalization ability of the network; through deep learning technology, the computational cost of cloud evolution simulation can be effectively reduced, the computing efficiency can be improved, and rapid prediction of cloud evolution can be achieved.

[0061] In an optional embodiment, The steps of receiving the cloud system near-ground data and the cloud system top data as real-time digital twin data, constructing a Kalman gain matrix based on the prediction error covariance of the physical parameters of the core area and the transition area and the observation error covariance of the real-time digital twin data, performing error analysis on the real-time digital twin data and the solution result of the regional decomposition method, and dynamically correcting the cloud system evolution prediction result using the Kalman gain matrix include: Acquire cloud near-ground data including reflectivity, radial velocity and spectral width, fuse multi-site radar data using three-dimensional variational assimilation method, construct three-dimensional wind field, temperature field and water vapor field distribution in the near-ground layer, and determine the ground-based observation error covariance matrix based on historical observation statistics; obtain cloud top data including cloud top height, cloud top temperature and cloud top radiation flux, extract cloud characteristic parameters using infrared-visible light dual-channel inversion algorithm, obtain cloud motion information through multi-time sequence registration, and determine the satellite observation error covariance matrix through channel noise analysis; Based on the difference prediction errors between the prediction results of the physical parameters in the core area and the transition area and the corresponding true values, a prediction error covariance matrix is ​​constructed, and a local correlation constraint based on the spatial distance of the physical parameters is introduced to correct the prediction error covariance matrix, wherein the local correlation constraint is determined by the correlation length scale; A Kalman gain matrix is ​​constructed based on the prediction error covariance matrix, the ground-based observation error covariance matrix and the satellite observation error covariance matrix, and an adaptive adjustment mechanism based on cloud morphology complexity is introduced to correct the Kalman gain matrix, wherein the cloud morphology complexity is calculated by a weighted combination of profile curvature variance and texture entropy; The cloud system near-ground data and the cloud system top data are mapped to the observation space, the observation residual vector with the predicted value is calculated, and the predicted state is updated based on the observation residual vector and the modified Kalman gain matrix; the turbulence characteristic constraint, the mass conservation constraint and the energy conservation constraint are introduced into the updated predicted state, and the turbulence characteristic constraint is constructed based on the non-negativity of turbulent kinetic energy and dissipation rate; the constrained predicted state is used as the initial condition for prediction at the next moment, forming a closed-loop iterative process of dynamic correction.

[0062] Exemplarily, first, obtain cloud system near-ground data. Use the Doppler weather radar network deployed on the ground to scan and collect radar echo data containing information such as reflectivity, radial velocity and spectral width. Then, use the three-dimensional variational assimilation method to fuse the observation data from multiple radar sites to construct the three-dimensional wind field, temperature field and water vapor field distribution near the ground. At the same time, based on the statistical results of past observation data, determine the error covariance matrix of ground-based radar observations for subsequent data fusion and error correction.

[0063] Next, obtain the cloud top data. Use the infrared and visible light sensors carried by meteorological satellites to obtain data such as cloud top height, cloud top temperature, and cloud top radiation flux. Use the infrared-visible light dual-channel inversion algorithm to extract cloud characteristic parameters, such as cloud optical thickness and effective particle radius. By aligning multi-time satellite image sequences, obtain cloud motion information, such as cloud movement speed and direction. In addition, by analyzing the channel noise of the sensor, determine the error covariance matrix of satellite observations.

[0064] Then, the prediction error covariance matrix is ​​constructed and corrected. The prediction results of the physical parameters in the core area and the transition area obtained based on the regional decomposition method are compared with the corresponding true values, and their differences, i.e., the prediction errors, are calculated. Using these prediction errors, the prediction error covariance matrix is ​​constructed. In order to improve the prediction accuracy, the local correlation constraint based on the spatial distance of the physical parameters is introduced to correct the prediction error covariance matrix. Specifically, the local correlation constraint is defined by the correlation length scale. The closer the distance between the grid points, the higher the correlation of their physical parameters.

[0065] Next, the Kalman gain matrix is ​​constructed and corrected. The Kalman gain matrix is ​​constructed using the previously obtained prediction error covariance matrix, ground-based observation error covariance matrix, and satellite observation error covariance matrix. In order to adapt to the complexity of different cloud morphologies, an adaptive adjustment mechanism based on the complexity of cloud morphology is introduced to correct the Kalman gain matrix. The complexity of cloud morphology is calculated by a weighted combination of contour curvature variance and texture entropy. For example, for clouds with complex contours and rich textures, a higher weight is given, thereby increasing the Kalman gain and paying more attention to the influence of observation data.

[0066] Subsequently, the predicted state is updated. The cloud system near-ground data and cloud system top data are mapped to the observation space, and the observation residual vector with the predicted value is calculated. The predicted state is updated using the calculated observation residual vector and the corrected Kalman gain matrix to obtain a more accurate cloud system state estimate.

[0067] Finally, constraints are introduced to the updated prediction state and iterated. Turbulence characteristic constraints, mass conservation constraints, and energy conservation constraints are introduced to the updated prediction state to ensure the physical consistency of the prediction results. Turbulence characteristic constraints are constructed based on the non-negativity of turbulent kinetic energy and dissipation rate. The constrained prediction state is used as the initial condition for the prediction at the next moment, forming a closed-loop iterative process of dynamic correction, and realizing continuous and dynamic correction of cloud system evolution.

[0068] Assume that at a certain moment, the radar observed a wind speed of 10 m / s at the bottom of the cloud system, and the satellite observed a cloud top height of 5,000 m. The regional decomposition method predicted a wind speed of 8 m / s and a cloud top height of 4,500 m. After error analysis and Kalman gain correction, the final predicted wind speed was corrected to 9.5 m / s and the cloud top height was corrected to 4,800 m, which are closer to the actual observed values.

[0069] The present invention can effectively reduce prediction errors and improve the accuracy of cloud evolution prediction by fusing ground-based radar and satellite observation data and combining the prediction results of the regional decomposition method; introduce an adaptive adjustment mechanism based on the complexity of cloud morphology, which can dynamically adjust the Kalman gain matrix according to the characteristics of different cloud systems, so as to better adapt to different types of cloud systems and improve the robustness of prediction; introduce turbulence characteristic constraints, mass conservation constraints and energy conservation constraints to ensure the physical consistency of the prediction results and make the prediction results more in line with the actual situation.

[0070] Figure 2 Schematic diagram of the structure of a multi-scale cloud system dynamic evolution simulation modeling system based on digital twins according to an embodiment of the present invention. Figure 2 As shown, the system comprises: The first unit is used to construct a multi-source collaborative sensing network using multispectral cameras, lidar and millimeter wave radar to obtain cloud system near-ground data, and to obtain cloud system top data and meteorological element data from ground meteorological stations in combination with satellite remote sensors; to perform phase consistency analysis on the acquired multi-source data, to establish a spatial transformation matrix between the sensor coordinate systems to obtain the registered data, to perform quality assessment on the registered data based on the signal-to-noise ratio, spatial resolution and temporal consistency, and to screen out valid data that meets the assessment threshold; The second unit is used to extract the primitive features of the cloud system based on the valid data by using a recursive segmentation algorithm, calculate the extreme points of the reflectivity gradient and the temperature gradient to divide the core area of ​​the cloud cluster, and calculate the parameter attenuation curve around the core area to divide the transition area; based on the regional division result, extract the contour curvature, grayscale distribution, texture features and multi-scale edge features of the cloud system, construct a deep digital twin mapping network of the primitive features of the cloud system and the real cloud system, perform feature verification on the morphological feature parameters output by the deep digital twin mapping network and the meteorological measured data, and dynamically optimize the parameters of the deep digital twin mapping network based on the verification results; The third unit is used to use an adaptive multi-grid method to perform fine discretization on the core area. The grid scale is automatically adjusted with the turbulence intensity and thermal parameter gradient of the cloud cluster, and a non-uniform grid is established for the transition area considering the evolution of vortices within the cloud system. Based on the grid division results, the dynamic evolution of the cloud system is simulated and modeled, and cross-regional coupling constraints including mass conservation, momentum conservation and energy conservation are established. The dynamic evolution process of the cloud system is characterized as a set of partial differential equations including turbulent diffusion, thermodynamic parameters and regional coupling. The set of partial differential equations is integrated with the deep digital twin mapping network, and the regional decomposition method is used to solve the overall evolution characteristics of the cloud system. The evolution prediction results are dynamically corrected using real-time digital twin data.

[0071] According to a third aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0072] A fourth aspect of the embodiments of the present invention is: A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0073] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-scale cloud system dynamic evolution simulation modeling method based on digital twins, characterized in that: include: A multi-source collaborative sensing network is constructed using multispectral cameras, lidar and millimeter-wave radar to obtain cloud near-ground data, and satellite remote sensors are used to obtain cloud top data and ground meteorological stations to obtain meteorological element data. Phase consistency analysis is performed on the acquired multi-source data, and a spatial transformation matrix between the sensor coordinate systems is established to obtain the registered data. The registered data is then evaluated for quality based on the signal-to-noise ratio, spatial resolution and temporal consistency, and valid data that meets the evaluation threshold is screened out. Based on the valid data, a recursive segmentation algorithm is used to extract primitive features of the cloud system, the extreme points of the reflectivity gradient and the temperature gradient are calculated to divide the core area of ​​the cloud cluster, and the parameter attenuation curve around the core area is calculated to divide the transition area; based on the regional division results, the contour curvature, grayscale distribution, texture features and multi-scale edge features of the cloud system are extracted, and a deep digital twin mapping network of the primitive features of the cloud system and the real cloud system is constructed, and the morphological feature parameters output by the deep digital twin mapping network are verified with the meteorological measured data, and the parameters of the deep digital twin mapping network are dynamically optimized based on the verification results; An adaptive multi-grid method is used to perform fine discretization on the core area. The grid scale is automatically adjusted with the turbulence intensity and thermal parameter gradient of the cloud cluster. A non-uniform grid that considers the evolution of vortices within the cloud system is established for the transition area. Based on the grid division results, the dynamic evolution of the cloud system is simulated and modeled, and cross-regional coupling constraints including mass conservation, momentum conservation, and energy conservation are established. The dynamic evolution process of the cloud system is characterized as a set of partial differential equations including turbulent diffusion, thermodynamic parameters, and regional coupling. The set of partial differential equations is integrated with the deep digital twin mapping network, and the regional decomposition method is used to solve the overall evolution characteristics of the cloud system. The evolution prediction results are dynamically corrected using real-time digital twin data.

2. The method according to claim 1, characterized in that Based on the valid data, a recursive segmentation algorithm is used to extract cloud primitive features. The steps of calculating the extreme points of reflectivity gradient and temperature gradient to divide the core area of ​​the cloud cluster and calculating the parameter attenuation curve around the core area to divide the transition area include: A recursive segmentation algorithm is used to extract cloud primitive features. The reflectivity mean and reflectivity standard deviation, temperature mean and temperature standard deviation are calculated for each sub-region. When the reflectivity standard deviation or temperature standard deviation of a sub-region is greater than a preset segmentation threshold, the region is further segmented until the reflectivity standard deviation and temperature standard deviation are less than the preset segmentation threshold or the minimum segmentation scale is reached. The feature vectors of the primitive region are recorded to form a primitive feature set. Based on the primitive feature set, the reflectivity gradient and temperature gradient between adjacent primitives are calculated, the dynamic weight coefficients of the reflectivity and temperature are calculated according to the variance of the reflectivity gradient and the temperature gradient, and the reflectivity gradient, the temperature gradient and their cross terms are weightedly fused to obtain a comprehensive gradient field; The integrated gradient field is decomposed into multiple scales using a Gaussian function, and the local maximum point is extracted from the gradient field at each scale. The local maximum point is established in correspondence between different scales to form an extreme point chain; the duration of the extreme point chain in the scale space is used as a stability index, and the extreme point with the highest stability index is selected as the candidate point of the core area; the candidate points are subjected to spatial connectivity analysis, and the set of candidate points whose Euclidean space distance is less than the set minimum distance threshold is defined as the core area; Calculate the horizontal wind field gradient of the periphery of the core area, calculate the local turbulence intensity according to the horizontal wind field gradient, and combine the local turbulence intensity with reflectivity and temperature to obtain a corrected parameter value; divide the radial distance from the core area to the periphery into multiple intervals, fit the parameter attenuation curve with a double exponential function in each interval according to the corrected parameter value, and impose derivative continuity constraints at the interval connection; calculate the second-order derivative and the third-order derivative of the parameter attenuation curve, take the weighted sum of the absolute value of the second-order derivative and the integral of the third-order derivative as the boundary identification objective function, and solve the position that minimizes the boundary identification objective function as the transition area boundary.

3. The method according to claim 1, characterized in that The steps of constructing a deep digital twin mapping network of the cloud primitive features and the real cloud system, verifying the morphological feature parameters output by the deep digital twin mapping network with the meteorological measured data, and dynamically optimizing the parameters of the deep digital twin mapping network based on the verification results include: The contour curvature, grayscale distribution, texture features and multi-scale edge features of the cloud system are input into a deep digital twin mapping network for feature processing, wherein the deep digital twin mapping network includes a feature encoder, a hierarchical dynamic memory module for storing cloud morphological patterns, and a multi-task decoder; the feature encoder encodes the features based on an adaptive sampling strategy of cloud morphological complexity to obtain encoded features, and the cloud morphological complexity is calculated by a weighted combination of contour curvature variance and texture entropy; The hierarchical dynamic memory module updates the memory unit through the attention mechanism and the gated update method that introduces the physical constraints of temperature gradient and pressure field, and dynamically fuses the memory state and the encoding features to obtain the fused features; based on the fused features, a multi-granularity spatiotemporal attention mechanism is used to process the cloud morphology dependency, and the spatial dependency features are captured by adaptive receptive field adjustment based on the cloud center distance, and a recursive attention unit is introduced to model the temporal dependency, so as to obtain a feature representation with spatiotemporal dependency; the feature representation is input into the multi-task decoder for morphological parameter prediction and reconstruction, and a joint loss function including task loss, physical constraint loss, memory update loss and temporal smoothing loss is constructed to optimize the prediction results, and the weight coefficient of the joint loss function is dynamically adjusted according to the prediction error; The prediction results output by the multi-task decoder are accurately evaluated based on the shape similarity of the Hausdorff distance, and are physically verified through area-perimeter ratio constraints and neighborhood consistency constraints, wherein the area-perimeter ratio constraint limits the value range of the shape coefficient, and the neighborhood consistency constraint ensures that the changes in morphological parameters at adjacent moments conform to the laws of meteorological development, and obtains verification performance indicators; based on the verification performance indicators, a dynamic learning strategy is adopted to optimize the deep digital twin mapping network, wherein the learning rate is dynamically adjusted based on a periodic attenuation function, and the sample weight coefficient is calculated based on the prediction error, and the sample weight coefficient is calculated by the exponential function of the prediction error and the temperature parameter, and the sample weight coefficient is multiplied by the task loss, physical constraint loss, memory update loss and timing smoothing loss respectively to update the network parameters.

4. The method according to claim 1, characterized in that: The adaptive multi-grid method is used to finely discretize the core area. The grid scale is automatically adjusted with the cloud turbulence intensity and thermal parameter gradient. The steps of establishing a non-uniform grid for the transition area considering the evolution of the vortex inside the cloud system include: The turbulent kinetic energy spectrum and vorticity field distribution in the core area are calculated based on the meteorological radar reflectivity and Doppler velocity field data. The turbulence intensity distribution at different scales is obtained by analyzing the turbulent kinetic energy spectrum. The vortex structure in the core area is identified by the vortex threshold method, and the vortex center position and vortex intensity are extracted. The temperature field and pressure field of the cloud area are constructed by using meteorological sounding data, the temperature gradient and the pressure gradient are calculated, and the absolute value of the temperature gradient and the absolute value of the pressure gradient are weightedly combined to obtain the thermal parameter gradient intensity index; the turbulence intensity distribution of different scales and the thermal parameter gradient intensity index are input into an exponential function, and a grid scale mapping function is obtained by weighted combination; A quadtree structure is used for multiple grid division, and the grid unit is subdivided according to the ratio of the output value of the grid scale mapping function to the initial grid scale, and the grid unit that meets the subdivision threshold condition is divided into four subgrids, and the difference between adjacent grid levels is limited; A vortex intensity evolution relationship is constructed based on the vortex center position and the vortex intensity, and a grid density attenuation coefficient of a transition region is calculated by a hyperbolic tangent function according to the vortex center distance, wherein the grid density attenuation coefficient increases with increasing distance from the vortex center; The mesh shape quality, orthogonality index and feature capture capability are weightedly combined to obtain a mesh quality evaluation index. When the mesh quality evaluation index is lower than a preset threshold, the mesh cells to be optimized are determined, and the position information of the mesh cells to be optimized is input into the mesh scale mapping function to recalculate the local mesh scale. The mesh topology is updated based on the recalculated mesh scale and the physical field is interpolated and reconstructed.

5. The method according to claim 1, characterized in that Based on the grid division results, the dynamic evolution of the cloud system is simulated and modeled, and cross-region coupling constraints including mass conservation, momentum conservation and energy conservation are established. The dynamic evolution process of the cloud system is characterized as a set of partial differential equations including turbulent diffusion, thermodynamic parameters and regional coupling. The partial differential equations are integrated with the deep digital twin mapping network, and the regional decomposition method is used to solve the overall evolution characteristics of the cloud system. The steps of dynamically correcting the evolution prediction results using real-time digital twin data include: Establish cross-region coupling constraints of mass conservation, momentum conservation and energy conservation at the interface between the core region and the transition region. The mass conservation constraint is based on the flux continuity conditions of the density field and the velocity field, the momentum conservation constraint is based on the pressure and viscous stress tensors, and the energy conservation constraint is based on the specific internal energy, thermal conductivity and temperature field. A partial differential equation system for cloud system evolution including turbulent diffusion terms, thermodynamic parameter terms and regional coupling terms is constructed, wherein the turbulent diffusion terms are characterized by turbulent kinetic energy, turbulent diffusion coefficient and dissipation rate, the thermodynamic parameter terms are characterized by temperature field, pressure field and density field, the regional coupling terms are determined by interface flux balance, and the partial differential equation system relates the cross-region coupling constraint to the turbulent diffusion terms, thermodynamic parameter terms and regional coupling terms; Inputting the partial differential equation group into the deep digital twin mapping network, establishing a mapping loss function including a physical constraint loss term and a conservation constraint loss term, and optimizing the parameters of the deep digital twin mapping network by minimizing the mapping loss function; The output results of the deep digital twin mapping network are solved by adopting a regional decomposition method, and the physical parameters of the transition region are fixed based on the distribution characteristics of the turbulent diffusion term and the thermodynamic parameter term to solve the core region, and the updated physical parameters of the core region are used as boundary conditions to solve the physical parameters of the transition region, and the physical parameters of the core region and the transition region are iterated repeatedly until the physical parameters of the core region and the transition region reach the preset convergence conditions, and the physical parameters include velocity field, density field, turbulent kinetic energy, turbulent diffusion coefficient, dissipation rate, temperature field, pressure field, specific internal energy, and thermal conductivity; The cloud system near-ground data and the cloud system top data are received as real-time digital twin data, a Kalman gain matrix is ​​constructed based on the prediction error covariance of the physical parameters of the core area and the transition area and the observation error covariance of the real-time digital twin data, error analysis is performed between the real-time digital twin data and the solution result of the regional decomposition method, and the Kalman gain matrix is ​​used to dynamically correct the cloud system evolution prediction result.

6. The method according to claim 5, characterized in that Inputting the partial differential equation group into the deep digital twin mapping network, establishing a mapping loss function including a physical constraint loss term and a conservation constraint loss term, and optimizing the parameters of the deep digital twin mapping network by minimizing the mapping loss function comprises: The physical constraint loss term includes turbulent dynamics loss and thermodynamic coupling loss. The turbulent dynamics loss is constructed based on the spatiotemporal derivatives of turbulent kinetic energy and dissipation rate and the turbulence generation term. The weight coefficient of the turbulent dynamics loss is adaptively adjusted by the local turbulence intensity and the turbulence Reynolds number. The thermodynamic coupling loss is constructed based on the convection-diffusion equation of the temperature field and the pressure field. The weight coefficient of the thermodynamic coupling loss is dynamically determined by the nonlinear function of the temperature gradient and the pressure gradient. The conservation constraint loss term is constructed based on the cross-region coupling constraint, wherein the conservation constraint loss term includes a mass conservation constraint, a momentum conservation constraint, and an energy conservation constraint, and a stress tensor correction term based on vorticity correlation is introduced; Constructing a feature-driven combined loss function, wherein the combined loss function performs a weighted combination of the physical constraint loss term and the conservation constraint loss term, wherein the weight coefficient of the combined loss function is dynamically adjusted by the cloud system morphology complexity, temperature gradient, and pressure gradient, wherein the cloud system morphology complexity is calculated by a weighted combination of profile curvature variance and texture entropy; A multi-scale residual correction mechanism is introduced, and the multi-scale residual correction mechanism includes local residuals and global residuals. The local residuals are calculated based on turbulence characteristic quantities, and the global residuals are calculated based on thermodynamic quantities. The weight coefficients of the local residuals and the global residuals are dynamically determined by the ratio of the turbulence characteristic scale to the cloud characteristic scale. The parameters of the deep digital twin mapping network are optimized and updated by minimizing the combined loss function and combining the multi-scale residual correction results, wherein the learning rate is adaptively adjusted based on the gradient of physical quantities and the prediction error, and the physical quantities include velocity field, vorticity field, pressure field, density field, temperature field, turbulent kinetic energy and dissipation rate. At the same time, a periodic attenuation mechanism is introduced. At the end of each training cycle, the learning rate is periodically adjusted according to the degree of satisfaction of the physical constraint loss term and the conservation constraint loss term.

7. The method according to claim 5, characterized in that The steps of receiving the cloud system near-ground data and the cloud system top data as real-time digital twin data, constructing a Kalman gain matrix based on the prediction error covariance of the physical parameters of the core area and the transition area and the observation error covariance of the real-time digital twin data, performing error analysis on the real-time digital twin data and the solution result of the regional decomposition method, and dynamically correcting the cloud system evolution prediction result using the Kalman gain matrix include: Acquire cloud near-ground data including reflectivity, radial velocity and spectral width, fuse multi-site radar data using three-dimensional variational assimilation method, construct three-dimensional wind field, temperature field and water vapor field distribution in the near-ground layer, and determine the ground-based observation error covariance matrix based on historical observation statistics; obtain cloud top data including cloud top height, cloud top temperature and cloud top radiation flux, extract cloud characteristic parameters using infrared-visible light dual-channel inversion algorithm, obtain cloud motion information through multi-time sequence registration, and determine the satellite observation error covariance matrix through channel noise analysis; Based on the difference prediction errors between the prediction results of the physical parameters in the core area and the transition area and the corresponding true values, a prediction error covariance matrix is ​​constructed, and a local correlation constraint based on the spatial distance of the physical parameters is introduced to correct the prediction error covariance matrix, wherein the local correlation constraint is determined by the correlation length scale; A Kalman gain matrix is ​​constructed based on the prediction error covariance matrix, the ground-based observation error covariance matrix and the satellite observation error covariance matrix, and an adaptive adjustment mechanism based on cloud morphology complexity is introduced to correct the Kalman gain matrix, wherein the cloud morphology complexity is calculated by a weighted combination of profile curvature variance and texture entropy; The cloud system near-ground data and the cloud system top data are mapped to the observation space, the observation residual vector with the predicted value is calculated, and the predicted state is updated based on the observation residual vector and the modified Kalman gain matrix; the turbulence characteristic constraint, the mass conservation constraint and the energy conservation constraint are introduced into the updated predicted state, and the turbulence characteristic constraint is constructed based on the non-negativity of turbulent kinetic energy and dissipation rate; the constrained predicted state is used as the initial condition for prediction at the next moment, forming a closed-loop iterative process of dynamic correction.

8. A multi-scale cloud system dynamic evolution simulation modeling system based on digital twins, used to implement the method described in any one of claims 1 to 7, characterized in that: include: The first unit is used to construct a multi-source collaborative sensing network using multispectral cameras, lidar and millimeter wave radar to obtain cloud system near-ground data, and to obtain cloud system top data and meteorological element data from ground meteorological stations in combination with satellite remote sensors; to perform phase consistency analysis on the acquired multi-source data, to establish a spatial transformation matrix between the sensor coordinate systems to obtain the registered data, to perform quality assessment on the registered data based on the signal-to-noise ratio, spatial resolution and temporal consistency, and to screen out valid data that meets the assessment threshold; The second unit is used to extract the primitive features of the cloud system based on the valid data by using a recursive segmentation algorithm, calculate the extreme points of the reflectivity gradient and the temperature gradient to divide the core area of ​​the cloud cluster, and calculate the parameter attenuation curve around the core area to divide the transition area; based on the regional division result, extract the contour curvature, grayscale distribution, texture features and multi-scale edge features of the cloud system, construct a deep digital twin mapping network of the primitive features of the cloud system and the real cloud system, perform feature verification on the morphological feature parameters output by the deep digital twin mapping network and the meteorological measured data, and dynamically optimize the parameters of the deep digital twin mapping network based on the verification results; The third unit is used to use an adaptive multi-grid method to perform fine discretization on the core area. The grid scale is automatically adjusted with the turbulence intensity and thermal parameter gradient of the cloud cluster, and a non-uniform grid is established for the transition area considering the evolution of vortices within the cloud system. Based on the grid division results, the dynamic evolution of the cloud system is simulated and modeled, and cross-regional coupling constraints including mass conservation, momentum conservation and energy conservation are established. The dynamic evolution process of the cloud system is characterized as a set of partial differential equations including turbulent diffusion, thermodynamic parameters and regional coupling. The set of partial differential equations is integrated with the deep digital twin mapping network, and the regional decomposition method is used to solve the overall evolution characteristics of the cloud system. The evolution prediction results are dynamically corrected using real-time digital twin data.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method described in any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images