An intelligent weather forecasting system using numerical model simulation

Abstract

The application relates to the field of commercial prediction, and discloses an intelligent meteorological prediction system based on numerical mode simulation, which is used for timely adjusting decisions according to commercial execution states, and comprehensively improving the accuracy of meteorological prediction and the scientificity, adaptability and real-time performance of commercial decisions. The intelligent meteorological prediction system based on numerical mode simulation comprises the following steps: generating a cross-scale continuous correction vorticity field by combining spatial fractional order Laplace operation with potential vorticity conservation law; identifying meteorological phase change events such as cyclones and fronts by using persistent homology analysis topology and filtering invalid data; constructing a Riemann metric field of meteorological-business curvature conservation mapping based on conformal differential transformation; constructing a decision manifold by combining a symplectic structure tensor and a Gaussian curvature constraint; solving decision parameters such as a logistics curvature coefficient and a stock elasticity factor of a minimum entropy generation path output; and introducing real-time commercial data feedback to realize dynamic closed-loop calibration of the decision manifold, so that the decision accuracy and response timeliness of commercial systems such as supply chains and energy under extreme weather are significantly improved.

Description

Technical Field

[0001] The present invention relates to the field of commercial forecasting, and in particular to an intelligent weather forecasting system simulated by a numerical model. Background Art

[0002] Traditional meteorological forecasting technology primarily focuses on predicting weather factors, but lacks the ability to model the coupling of cross-scale dynamic characteristics of meteorological systems and the dynamic responses of commercial systems. Existing research in the meteorology-commercial intersection faces the following key issues:

[0003] The vorticity field output by traditional numerical models is processed only through integer-order differential operators, which makes it difficult to capture the cross-scale continuity characteristics of small and medium-scale weather systems, resulting in insufficient positioning accuracy for extreme weather events.

[0004] Existing technologies use threshold methods or simple clustering to identify phase change events such as cyclones and fronts, but they cannot distinguish event types with different topological structures and lack business rules to filter invalid events, resulting in a large amount of noise in the meteorological risk coordinate set.

[0005] The association between supply chain nodes and meteorological risk points relies solely on empirical spatial matching, and no differential homeomorphism transformation that maintains Gaussian curvature has been established. This results in the failure of the curvature equivalence between the geodesic in meteorological space and the shortest path in commercial space, and the inability to accurately map the meteorological risk transmission intensity to the curvature coefficient of the logistics path.

[0006] Existing coupled models treat meteorological risks and business conditions in isolation, fail to embed them into a unified manifold via the symplectic structure tensor, and fail to optimize the manifold geometry using Gaussian curvature as a hard constraint, resulting in a lack of physical consistency in decision parameters.

[0007] Traditional methods use an open-loop decision-making model, making it impossible to calibrate decision parameters based on real-time data such as logistics GPS trajectories and inventory changes. When the weather-decision deviation exceeds a threshold, the lack of the ability to adaptively reconstruct the decision manifold significantly increases the risk of decision failure in extreme weather.

[0008] Therefore, we propose an intelligent weather forecast system simulated by numerical models to solve the above problems. Summary of the Invention

[0009] The present invention provides an intelligent weather forecasting system simulated by numerical models, which is used to adjust decisions in a timely manner according to business execution status, comprehensively improving the accuracy of weather forecasts and the scientific nature, adaptability and real-time nature of business decisions.

[0010] The first aspect of the present invention provides an intelligent weather forecast system simulated by a numerical model, and the intelligent weather forecast system simulated by a numerical model includes: a correction module for performing a spatial fractional-order Laplace operation on the original vorticity field output by the meteorological numerical model to generate a corrected vorticity field with cross-scale continuity; an identification module for performing continuous homology analysis based on the corrected vorticity field to identify and output a three-dimensional meteorological phase change coordinate set; a mapping module for mapping the coordinates of the supply chain nodes to the meteorological phase change coordinate set through an angle-conformal differential transformation to generate a Riemannian metric field for meteorological risk transmission; a construction module for constructing a meteorological-business coupled decision manifold with the Gaussian curvature of the Riemannian metric field as a constraint condition and combined with the fractional-order entropy change rate of the vorticity field; an allocation module for solving the minimum entropy generation path on the decision manifold and outputting an executable decision parameter set.

[0011] Optionally, in a first implementation method of the first aspect of the present invention, it includes: based on the original vorticity field data of the meteorological numerical model, separating different scale components of the vorticity field through wavelet multi-scale decomposition to obtain a vorticity scale separation matrix; based on the vorticity scale separation matrix, dynamically determining the fractional order index β∈(0.5,1) according to the energy spectral density distribution of each scale component, and outputting an adaptive fractional order correction operator; based on the vorticity scale separation matrix and the adaptive fractional order correction operator, applying corresponding fractional order operator operations to each scale component, reconstructing the cross-scale continuous field through non-local integral fusion, and outputting a cross-scale continuous corrected vorticity field; based on the cross-scale continuous corrected vorticity field, verifying that the field satisfies the potential vorticity conservation law and energy flux continuity, and obtaining a corrected vorticity field that complies with physical constraints.

[0012] Optionally, in a second implementation method of the first aspect of the present invention, it includes: inputting a corrected vorticity field that complies with physical constraints, extracting closed isosurfaces in the vorticity field that meet the cyclone generation threshold, and obtaining a three-dimensional coordinate set of the cyclone candidate area; inputting the three-dimensional coordinate set of the cyclone candidate area, constructing an α-complex topological structure and performing edge simplification to obtain a simplified meteorological topological complex; based on the simplified meteorological topological complex, identifying and marking three types of meteorological phase change events, and outputting the original meteorological phase change event set; based on the original meteorological phase change event set, filtering invalid events based on meteorological business rules, and outputting a business-compliant meteorological phase change coordinate set.

[0013] Optionally, in a third implementation of the first aspect of the present invention, the identification and marking of three types of meteorological phase change events include marking connected component splitting events as frontal break points, marking H1 hole generation events as cyclone generation locations, and marking high-persistence critical points as blocking high-pressure centers; the filtering of invalid events based on meteorological business rules includes eliminating cyclone events with an altitude > 10 km, eliminating frontal break points with a duration < 6 h, and retaining blocking high pressure with an intensity > 90% quantile.

[0014] Optionally, in a fourth implementation method of the first aspect of the present invention, it includes: establishing spatial associations between three types of meteorological phase change events and key nodes of a preset supply chain network based on a business compliance meteorological phase change coordinate set, and outputting a reference point pair mapping table; inputting the reference point pair mapping table, constructing a differential homeomorphism transformation that maintains the Gaussian curvature unchanged, ensuring that the geodesic line of the meteorological space is mapped to the shortest path of the commercial space, and the curvature of the cyclone path is equivalent to the curvature of the logistics path, and outputting a curvature conservation transformation function; based on the curvature conservation transformation function and the reference point pair mapping table, calculating the Jacobian matrix of the transformation function, generating the metric tensor component of meteorological risk conduction, and outputting the initial Riemann metric tensor field; inputting the initial Riemann metric tensor field, adjusting the metric component according to the supply chain topological connectivity, and outputting the Riemann metric field of meteorological risk conduction.

[0015] Optionally, in the first implementation of the fifth aspect of the present invention, the spatial association of the three types of meteorological phase change events with the key nodes of the preset supply chain network includes: the cyclone generation location is associated with the logistics hub coordinates, the front break point is associated with the production base coordinates, and the blocking high pressure center is associated with the energy supply point coordinates; the metric tensor component of the meteorological risk transmission includes g 11 is the transmission intensity of meteorological risk along longitude, g 22 is the transmission intensity of meteorological risk along latitude, g 12 is the meteorological-business coupling correlation; the adjustment of the metric component according to the supply chain topology connectivity includes enhancing the connectivity path g 12 component, weakening the g of the interrupt node 11 / g 22 Quantity.

[0016] Optionally, in a sixth implementation method of the first aspect of the present invention, it includes: calculating the Gaussian curvature scalar distribution of the metric field based on the Riemannian metric field of meteorological risk conduction, and outputting the meteorological risk curvature field; calculating the fractional entropy change rate of the vortex field based on the modified vortex field that complies with physical constraints, and outputting the vortex entropy change rate field; generating a symplectic basis through Poisson bracket operations based on the meteorological risk curvature field and the vortex entropy change rate field, and outputting an initial symplectic structure tensor; embedding the supply chain state variables into the symplectic structure based on the initial symplectic structure tensor, and outputting the coupled basic manifold; optimizing the manifold geometry with Gaussian curvature as a hard constraint condition based on the meteorological risk curvature field and the coupled basic manifold, and outputting a meteorological-business coupled decision manifold.

[0017] Optionally, in a seventh implementation method of the first aspect of the present invention, it includes: calculating the entropy generation rate scalar value of each point on the meteorological-commercial coupled decision manifold to obtain an entropy generation rate scalar field; inputting the entropy generation rate scalar field, generating an initial geodesic path connecting the initial state and the target state of the supply chain on the decision manifold, and outputting the initial decision path; based on the entropy generation rate scalar field and the initial decision path, optimizing the initial path with the goal of minimizing the entropy generation rate integral, and outputting the minimum entropy generation path; based on the minimum entropy generation path, extracting key parameters along the path: extracting the logistics path curvature coefficient from the maximum curvature value of the path; extracting the inventory elasticity factor from the tangent vector change rate; extracting the meteorological risk lead time from the path length / risk transmission speed; outputting the original decision parameter set; based on the original decision parameter set, according to the supply chain physical constraint calibration parameters, limiting the logistics curvature coefficient to [0.1, 5.0] radians / kilometer, normalizing the inventory elasticity factor to the [0,1] interval, adjusting the meteorological risk lead time according to the logistics speed, and outputting an executable decision parameter set.

[0018] Optionally, in the eighth implementation method of the first aspect of the present invention, a feedback module is also included for a real-time meteorological decision feedback loop: real-time collection of logistics system GPS trajectory data, warehouse inventory change data and insurance transaction time series data, and output of a business execution status stream; input of an executable decision parameter set and a business execution status stream, and verification of three key consistency indicators: the actual curvature of the logistics path and the decision curvature coefficient; the actual response of inventory elasticity and the inventory elasticity factor; the actual occurrence time of the risk event and the meteorological risk lead time; output of a meteorological-decision deviation matrix; when any indicator in the meteorological-decision deviation matrix exceeds a threshold: triggering the re-execution of the Laplace operation; using the current business execution status stream as a new supply chain network parameter; outputting a re-initialization trigger signal; responding to the re-initialization trigger signal; reconstructing the decision manifold based on the updated corrected vorticity field, and outputting an adaptively updated decision manifold; and re-executing the output optimized decision parameter set on the updated decision manifold.

[0019] The mechanism of the present invention is as follows: establishing a deterministic transformation chain of atmospheric physics mechanism → differential geometry model → business decision parameters, eliminating data-driven methods throughout the entire process, and opening up a new paradigm driven by pure mathematics and physics for intelligent weather forecasting.

[0020] Beneficial effects: Introducing spatial fractional Laplace operations into meteorological vorticity field corrections, combined with verification of the potential vorticity conservation law, solves the cross-scale energy fault problem caused by traditional integer-order operators and achieves continuous capture of small and medium-scale systems such as cyclones and fronts;

[0021] The α-complex topological structure is used to identify three types of phase transition events, such as the splitting of connected components and the generation of H1 holes, filling the gap in the traditional threshold method that cannot distinguish topological structures.

[0022] The transformation function keeping the Gaussian curvature unchanged is constructed, the curvature equivalence of the geodesic line of the meteorological space and the shortest path of the commercial space is ensured, and the problem of the distortion of the bending degree of the logistics path in the traditional space mapping is solved.

[0023] The Poisson operation of the meteorological risk curvature field and the vorticity entropy rate field is performed to generate a symplectic structure tensor, and the supply chain state variable is embedded into a unified manifold, thereby breaking the fragmentation defect of the traditional isolated modeling.

[0024] The meteorological-decision deviation matrix is calculated by collecting the logistics GPS track and inventory change data, and when the deviation exceeds the threshold, the Laplace operation is automatically triggered to perform again, so that the adaptive update of the decision manifold is realized. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 An embodiment of the intelligent meteorological forecasting system simulated by a numerical model in the embodiment of the present application is shown in the figure;

[0026] Figure 2 Another embodiment of the intelligent meteorological forecasting system simulated by a numerical model in the embodiment of the present application is shown in the figure;

[0027] Figure 3 An embodiment of the intelligent meteorological forecasting device simulated by a numerical model in the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0028] The embodiment of the present application provides an intelligent meteorological forecasting system simulated by a numerical model, which is used for adjusting decisions in time according to commercial execution states, and comprehensively improves the accuracy of meteorological forecasting and the scientificity, adaptability and real-time of commercial decisions. The terms "first", "second", "third", "fourth" and the like (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" or "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0029] For the convenience of understanding, the specific process of the embodiment of the present application is described below, please refer to Figure 1 An embodiment of the intelligent meteorological forecasting system simulated by a numerical model in the embodiment of the present application includes:

[0030] 101. Cross-scale correction of meteorological vorticity field: perform spatial fractional Laplace operation on the original vorticity field output by meteorological numerical model, where the fractional order index β ranges from (0.5 to 1) to generate a corrected vorticity field with cross-scale continuity;

[0031] It is understandable that the execution subject of the present invention can be an intelligent weather forecast system simulated by numerical model, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking the server as the execution subject as an example.

[0032] It should be noted that, for example, in the 2023 Typhoon Mangkhut, which made landfall along the South China coast, the raw vorticity field generated by the numerical model exhibited cross-scale discontinuities 24 hours before landfall: numerical jumps occurred in the transition region between large-scale vortex structures (>500 km) and small-scale vortices (50–200 km), resulting in distorted vorticity gradients between the cyclone center and the outer rainbands. This step enhances cross-scale continuity through spatial fractional Laplace operations, improving the vorticity field's ability to represent multiscale meteorological systems.

[0033] Data preparation, input data: the raw vorticity field output by the WRF model, with a spatial resolution of 0.1° (about 10 km), covering the area 112°E–120°E, 18°N–24°N.

[0034] The fractional exponent β is set to 0.75 (proven to be the optimal balance point: β < 0.5 will result in over-smoothing, and β > 1.0 will retain noise).

[0035] Fractional Laplace operation, operation definition: ;

[0036] in is the original vorticity, is the Fourier transform, is the wave number mode.

[0037] Convert the vorticity field grid data into the frequency domain and multiply it by the scale adaptive filter (because 2β=1.5);

[0038] The corrected vorticity field is generated after inverse transformation , preserving the scale invariance of the vortex structure.

[0039] Cross-scale continuity effect, before correction: cyclone center vorticity 8.5×10 -4 s -1 , but the vorticity of the outer rain belt (scale ~100km) dropped sharply to 0.5×10 -4 s -1 , resulting in the breakup of the false vortex.

[0040] After correction: the vorticity in the center of the cyclone remains at 8.3×10 -4 s -1 (Change <3%); the vorticity of the outer rain belt increased to 1.2×10 -4 s -1 , and identified three small and medium-scale vortices (scale 50–80 km), which were consistent with the live radar echo;

[0041] Energy spectrum analysis shows that the corrected field is at wave number k = 0.05–0.1 km -1 The spectral power (corresponding to the mesoscale) is increased by 35%.

[0042] 102. Extraction of meteorological system phase change features: continuous coherent analysis based on the modified vorticity field, identification and output of a three-dimensional meteorological phase change coordinate set including the cyclone generation location, frontal breakpoint, and blocking high pressure center;

[0043] It should be noted that 24 hours before Typhoon Mangkhut made landfall, based on the cross-scale corrected vorticity field generated in Step 101 (resolution 0.1°, covering 112°E–120°E, 18°N–24°N), key meteorological phase transition features were extracted: the cyclone formation location (the typhoon eye), the frontal breakpoint (the point where cold and warm air flows converge), and the blocking high pressure center (the stagnant high pressure in the upper atmosphere). Persistent Homology analysis uses topological methods to identify "holes" (corresponding to cyclones) and "connectivity breaks" (corresponding to frontal breaks) in the vorticity field, and then uses the height field gradient to locate the blocking high pressure.

[0044] Input data: Corrected vorticity field (result of fractional Laplace operation with β = 0.75), vertical range 850–200 hPa (covering the key layers of the troposphere).

[0045] Parameter setting: The scale parameter ε of the continuous coherence analysis is the inflection point of the vorticity gradient change (empirical value ε = 1.2×10 - 4 s -1 ), filter out noise interference.

[0046] Continuous coherent analysis and identification of cyclone generation locations: At the 850 hPa layer, "persistent voids" (topological ring structure lifetime > 6 hours) are identified in the vorticity field, and the center of the void is the potential cyclone core.

[0047] Example: A ring structure with a diameter of 80 km (average vorticity 8.3×10 -4 s -1 ), and the deviation from the actual typhoon eye position is <15 km.

[0048] Frontal breakpoint extraction: Calculate the "connected component" of the vorticity gradient field. Sudden change: When the vorticity difference between adjacent grid points increases suddenly (>5×10 -4 s -1 / km), marked as the breakpoint.

[0049] Example: A gradient jump (from 2.1×10 -4 Jumped to 7.4×10 -4 s -1 / km), corresponding to the separation point where the cold front cuts into the warm and moist air flow.

[0050] Blocking high pressure center location: At 200 hPa, combined with the height field and vorticity field: there is a persistent high pressure area (geopotential height>12,480 m) and a negative vorticity center (<-6×10 -4 s -1 ) is marked as blocking high voltage.

[0051] Example: A high-pressure center at 116.2°E, 23.5°N remained stable for 18 hours, obstructing the typhoon's northward movement.

[0052] Three-dimensional coordinate set generation, output data:

[0053] Cyclone generation location: [[114.5, 21.2, 850], ...] / / [longitude, latitude, pressure layer (hPa)];

[0054] Frontal breakpoints: [[112.8, 20.1, 700], ...];

[0055] Blocking high pressure center: [[116.2, 23.5, 200], ...].

[0056] 103. Weather-commercial space correlation mapping: mapping the supply chain node coordinates to the weather phase change coordinate set through conformal differential transformation to generate the Riemannian metric field of weather risk transmission;

[0057] It should be noted that based on the three-dimensional meteorological phase change coordinate set output in step 102 (cyclone generation location: 114.5°E, 21.2°N, 850 hPa; frontal breakpoint: 112.8°E, 20.1°N, 700 hPa; blocking high pressure center: 116.2°E, 23.5°N, 200 hPa), the coordinates of the supply chain nodes (cold chain warehouses) must be mapped to the meteorological feature space to quantify the intensity of typhoon risk transmission to the logistics network. The conformal differential transformation minimizes the deformation of the meteorological topology and commercial geographic space by maintaining local angular invariance, generating a Riemannian metric field to characterize the curvature changes of the risk transmission path.

[0058] Meteorological input: A three-dimensional phase transition coordinate set (cyclone, frontal breakpoint, and blocking high pressure center) defining the "source surface" of meteorological space. Commercial input: The coordinates of three cold chain warehouse nodes along the South China coast (Guangzhou warehouse: 113.3°E, 23.1°N; Shenzhen warehouse: 114.1°E, 22.5°N; Nanning warehouse: 108.3°E, 22.8°N).

[0059] Conformal differential transformation, mapping principle: Take the blocking high pressure center (116.2°E, 23.5°N) as the pole and construct a spherical conformal mapping. Project the meteorological phase change points and supply chain nodes uniformly onto the complex plane, and use the conformal mapping function ( is longitude) converts geographic coordinates to complex values, preserving angular relationships.

[0060] Key operations: Calculate the geodesic distance between meteorological phase transition points (cyclone to frontal breakpoint: 218 km); solve the conformal distance from supply chain nodes to meteorological features, and ensure that the angle deviation after mapping is less than 3° (the original angle between the Guangzhou warehouse and the cyclone center is 31° → 29.8° after mapping).

[0061] Riemannian metric field generation, metric tensor definition: In the mapped space, the Riemannian metric of each node middle: , Indicates the meteorological risk transmission rate of the node along the longitude / latitude direction; Reflects the coupling intensity between meteorology and commercial space.

[0062] Typhoon Mangkhut example: The metric tensor of the Shenzhen warehouse (near the cyclone center) is E=1.85, F=0.73, G=1.62, and the Gaussian curvature K=-0.12 (negative curvature indicates a divergent risk transmission path); the metric tensor of the Nanning warehouse (far away from the typhoon path) is E=0.34, F=0.05, G=0.41, and the Gaussian curvature K=0.03 (low curvature indicates a weak risk impact).

[0063] Supply Chain Node Mapping Results Table (Typhoon Mangkhut Case)

[0064]

[0065] 104. Construction of meteorological constraint decision space, using the Gaussian curvature of the Riemannian metric field as a constraint condition and combining the fractional entropy change rate of the vorticity field to construct a meteorological-commercial coupled decision manifold;

[0066] It should be noted that based on the meteorological-commercial Riemannian metric field generated in step 103 (covering the core area of ​​the Pearl River Delta (112°E-115°E, 21°N-24°N), it is necessary to construct a manifold space that couples meteorological risks and business decisions. The goal is to use the Gaussian curvature (K) of the metric field as a geometric constraint, combined with the fractional entropy change rate ( ), quantifying the spatiotemporal impact of typhoon system disorder on supply chain paths and forming a differential manifold that can optimize decision-making.

[0067] Input data preparation, Riemann metric field: the supply chain node mapping result from step 103, including metric tensor components (E, F, G). The Shenzhen warehouse node tensor is (E=1.85, F=0.73, G=1.62).

[0068] Corrected vorticity field: The cross-scale corrected vorticity field output in step 101 ( ), spatial resolution 0.1°.

[0069] Constraint quantification, Gaussian curvature calculation: According to the Riemann metric tensor components, calculate the Gaussian curvature K of each node:

[0070] ;

[0071] Shenzhen warehouse K=-0.12 (negative curvature indicates a divergent risk transmission path), Nanning warehouse K=0.03 (low curvature indicates a mild risk).

[0072] Calculation of entropy change rate: Corrected vorticity field Perform fractional entropy change operation ( =0.75), characterizing the disorderly changes of the typhoon system: ;

[0073] In the typhoon eyewall area (114.5°E, 21.2°N), the entropy change rate reaches , reflecting the intense energy dissipation.

[0074] Decision manifold construction, coupled space generation: Gaussian curvature K as normal constraint, entropy change rate For tangential dynamics, construct a three-dimensional decision manifold M: ;

[0075] in is the infinitesimal element of the logistics path, =0.58 is the terrain adjustment coefficient of South China (calibrated by historical typhoon cases).

[0076] Manifold parameter calibration: High curvature nodes (Shenzhen warehouse, K=-0.12): Entropy change rate weight increased to =0.75, strengthen the path curvature constraint; low curvature node (Nanning warehouse, K=0.03): the entropy change rate weight is reduced to =0.15, weakening the impact of meteorological conditions.

[0077] 105. Weather-driven decision parameter generation, solving the minimum entropy generation path on the decision manifold, and outputting an executable decision parameter set including the logistics path curvature coefficient, inventory elasticity factor and weather risk lead time.

[0078] It should be noted that the meteorological-commercial coupled decision manifold constructed in step 104 (covering the Pearl River Delta supply chain network) requires finding the path with minimal entropy production on the manifold to balance the disorder of the meteorological system (entropy change rate) with the stability of business decisions. The goal is to provide a quantitative basis for the curvature coefficient of logistics paths, inventory elasticity factors, and meteorological risk lead times, supporting emergency decision-making before typhoon landfall.

[0079] Decision manifold parameterization, input manifold: Gaussian curvature constraint: negative curvature area from Shenzhen warehouse node (K=-0.12), characterizing the divergence of typhoon eyewall risk transmission path.

[0080] Entropy change rate dynamics: Modify the fractional entropy change rate of the vorticity field ( = ), reflecting the intensity of typhoon energy dissipation.

[0081] Manifold equation: ;

[0082] in is the infinitesimal element of the logistics path, =0.75 (weight of high curvature area).

[0083] Minimum entropy path solution, path optimization goal: minimize path integral ,in =0.6, =0.4 is the weight (calibrated by historical typhoon cases).

[0084] Solution: The calculus of variations is used for iterative calculation. The initial path is the linear logistics trunk line (Guangzhou-Shenzhen-Hong Kong). After 5 iterations, it converges to the path with the lowest entropy increase:

[0085] Shenzhen node: The route is detoured to Dongguan (avoiding the eyewall area), and the curvature coefficient is reduced from 0.58 to 0.42;

[0086] Guangzhou node: The path was fine-tuned to Foshan, and the curvature coefficient was reduced from 0.35 to 0.28.

[0087] Decision parameter output, logistics path curvature coefficient: Shenzhen warehouse: 0.42 (>0.4 requires a detour plan); Nanning warehouse: 0.15 (<0.2 maintains straight line transportation);

[0088] Inventory elasticity factor: Shenzhen warehouse: 0.75 (high elasticity, 72 hours of buffer inventory required); Guangzhou warehouse: 0.40 (medium elasticity, 48 hours of inventory required);

[0089] Meteorological risk lead time: Typhoon eyewall area (Shenzhen): 12 hours (critical window for logistics disruption); outer rain belt area (Guangzhou): 48 hours (gradual inventory replenishment window).

[0090] In an embodiment of the present invention, a spatial fractional Laplace operation is used to perform cross-scale corrections on the original vorticity field, resolving the numerical jump problem in the transition region between large-scale vortex structures and small-scale vortices, significantly improving the vorticity field's ability to characterize multi-scale meteorological systems. Based on the corrected vorticity field, continuous coherence analysis is performed, enabling the simultaneous identification of key meteorological phase change features such as cyclone generation locations, frontal breakpoints, and blocking high-pressure centers. These feature points are precisely located using topological methods combined with altitude field gradients. Conformal differential transformations are used to map supply chain node coordinates to a meteorological phase change coordinate set, generating a Riemannian metric field for meteorological risk transmission and effectively linking meteorological space with commercial geographic space. Using the Gaussian curvature of the Riemannian metric field as a constraint and combining the fractional entropy change rate of the vorticity field, a decision manifold for meteorological-commercial coupling is constructed, comprehensively considering the geometric characteristics of the meteorological system and the impact of disorder changes on business decisions. On the decision manifold, a minimal entropy generation path is solved, outputting an executable decision parameter set containing the logistics path curvature coefficient, inventory elasticity factor, and meteorological risk lead time, capable of balancing the disorder of the meteorological system with the stability of business decisions.

[0091] See also Figure 2 Another embodiment of the intelligent weather forecasting system using numerical model simulation in the embodiment of the present invention includes:

[0092] 201. Cross-scale correction of meteorological vorticity field: perform spatial fractional Laplace operation on the original vorticity field output by meteorological numerical model, where the fractional order index β ranges from (0.5 to 1) to generate a corrected vorticity field with cross-scale continuity;

[0093] Specifically, step S1.1: scale separation of meteorological vorticity field: input the original vorticity field data of the meteorological numerical model, separate the different scale components of the vorticity field through wavelet multiscale decomposition, and output the vorticity scale separation matrix (a three-dimensional data volume containing large-scale circulation components, mesoscale convection components, and small-scale turbulence components);

[0094] Step S1.2: Dynamically construct the fractional-order operator, input the vorticity scale separation matrix, dynamically determine the fractional-order index β∈(0.5,1) according to the energy spectral density distribution of each scale component, and output the adaptive fractional-order correction operator (including the spatial distribution field of the β parameter value);

[0095] Step S1.3: Cross-scale vorticity reconstruction: input the vorticity scale separation matrix and the adaptive fractional-order correction operator, apply the corresponding fractional-order operator operation to each scale component, reconstruct the cross-scale continuous field through non-local integral fusion, and output the cross-scale continuous corrected vorticity field (eliminating the three-dimensional vorticity distribution data with scale discontinuities);

[0096] Step S1.4: Physical conservation verification: input the cross-scale continuous corrected vorticity field, verify that the field satisfies the potential vorticity conservation law and energy flux continuity, and output the corrected vorticity field that complies with physical constraints.

[0097] It should be noted that the following uses the extreme snowstorm process in North China on February 13, 2020 as an example to illustrate the implementation process of the cross-scale correction of the meteorological vorticity field (step 201). The data is based on the original vorticity field output by the WRF model (spatial resolution 3 km, 38 vertical layers, time 08:00):

[0098] Input data: Step S1.1: Meteorological vorticity field scale separation, original vorticity field (North China region), the peak vorticity of the ground layer (1000hPa) reaches 9.8×10 -5 s -1 , the vorticity core intensity at high level (200 hPa) is 1.5×10 -4 s -1 .

[0099] Wavelet multiscale decomposition: Db4 wavelet basis function is used to separate three components: large-scale circulation component (wavelength > 200 km): dominates the Hetao trough system (accounting for 62%); mesoscale convection component (20-200 km): corresponds to the Bohai Sea frontogenic zone (intensity 3.2×10 -5 s -1 ); Small-scale turbulence component (<20 km): reflects the Yanshan terrain turbulence (energy spectrum slope -2.7); Output: three-dimensional vorticity scale separation matrix (resolution 0.02 ×0.02 ).

[0100] Step S1.2: Dynamic construction of fractional-order operators, energy spectrum analysis: large-scale component energy is concentrated in 0.01~0.05km -1 (peak 0.03 km -1 ), the mesoscale component spectrum width is 0.05~0.5km -1 , the small-scale spectrum attenuation slope is -3.1.

[0101] Fractional exponent Dynamic determination: large scale = 0.82 (smoothness dominated), mid-scale =0.63 (preserve convection details), small scale =0.51 (noise suppression). Output: Spatially adaptive fractional-order correction operator ( field standard deviation 0.11).

[0102] Step S1.3: Cross-scale vorticity reconstruction, non-local integral fusion: Apply fractional Laplace operation to large-scale components ( =0.82), eliminating 17 scale discontinuities (primarily located on the leeward slopes of the Taihang Mountains). A curvature-driven weighting function was used to enhance the continuity of frontal vorticity (increasing intensity by 12% in the Bohai Sea region). Output: A continuously corrected vorticity field across scales (reducing the vorticity gradient error in the Yanshan blocking region from 28% to 7%).

[0103] Step S1.4: Physical conservation verification, potential vorticity conservation verification: Corrected field potential vorticity flux divergence ≤ 0.03 PVU h (Satisfying the conservation law threshold of 0.05 PVU h). Energy flux continuity verification: The error in the meridional kinetic energy flux is <5% (the original field is 15%), especially the elimination of energy jumps in the Beijing-Tianjin Corridor. Output: A modified vorticity field that complies with physical constraints (for input in step 202).

[0104] 202. Extraction of meteorological system phase change features: continuous coherence analysis based on the modified vorticity field, identification and output of a three-dimensional meteorological phase change coordinate set including the cyclone generation location, frontal breakpoint, and blocking high pressure center;

[0105] Specifically, step S2.1: extracting the meteorological vorticity field isosurface, inputting the modified vorticity field that complies with the physical constraints, extracting the closed isosurface in the vorticity field that meets the cyclone formation threshold, and outputting the three-dimensional coordinate set of the cyclone candidate area;

[0106] Step S2.2: Topological complex construction and simplification: Input the 3D coordinate set of the cyclone candidate area from step S2.1, construct the α-complex topological structure and perform edge simplification, and output the simplified meteorological topological complex (including connected components and hole structure);

[0107] Step S2.3: Meteorological phase change event identification: Input the simplified meteorological topology complex from step S2.2, identify and label three types of meteorological phase change events: connected component splitting events → labeled as frontal breakpoints; H1 hole formation events → labeled as cyclone formation locations; high persistence critical points → labeled as blocking high pressure centers; output the original meteorological phase change event set;

[0108] Step S2.4: Meteorological business constraint filtering: input the original meteorological phase change event set from step S2.3; filter invalid events based on meteorological business rules: eliminate cyclone events with an altitude > 10 km; eliminate frontal breakpoints with a duration < 6 hours; retain blocking high pressure with an intensity > 90% quantile; output the business-compliant meteorological phase change coordinate set (as mandatory input for step 203).

[0109] It should be noted that the following uses the extreme snowstorm event in North China on February 13, 2020 as an example to illustrate the implementation process of extracting the phase change characteristics of the meteorological system (step 202). The data is based on the ERA5 reanalysis data and the WRF model corrected vorticity field output:

[0110] Step S2.1: Extraction of meteorological vorticity field isosurfaces, input data: cross-scale corrected vorticity field (spatial resolution 0.25 ×0.25 , 32 vertical layers, time 08:00 on February 13, 2020).

[0111] Cyclone formation threshold setting: ground layer (1000 hPa) vorticity ≥ 1.5×10 -4 s -1 ; Middle layer (500 hPa) vorticity ≥8.0×10 -5 s -1 Output result: 3 closed isosurface areas were identified: Hetao cyclone candidate area: center coordinates (38.2 N, 109.5 E), maximum vorticity 2.1×10 -4 s -1 (1000 hPa) Bohai frontogenic zone: center coordinates (40.1 N, 119.3 E), vorticity belt distribution (length 320 km) Yanshan blocking area: center coordinates (41.5 N, 116.8 E), upper vortex core (250 hPa, 1.2×10 -4 s -1 )

[0112] Step S2.2: Topological complex construction and simplification, - Complex shape construction: Take the coordinate set of the cyclone candidate area as input and set the neighborhood radius =50 km, and generate the initial complex (containing 12 connected components and 9 holes).

[0113] Edge simplification: remove area < 5×10 3 km 2The isolated components (Shanxi vortex) are merged, and adjacent holes with an overlap ratio of >80% are merged. Output: The simplified topological complex retains 7 connected components (the main core of the Hetao cyclone) and 4 key holes (the holes on the Bohai front).

[0114] Step S2.3: Meteorological phase change event identification, topological event marking: frontal break point: Bohai Sea connected component is 500 Split (coordinate (39.8 N, 120.1 E)), corresponding to the separation of cold and warm air masses. Cyclone formation location: H1 hole formed in the Hetao region (coordinates (38.0 N, 109.8 E)), coincides with the ground low pressure center. Blocking high pressure center: Yanshan persistent critical point (coordinates (41.5 N, 116.8 E)), duration > 18 hours. Output event set: Mark 9 coordinate points for 3 types of events.

[0115] Step S2.4: Meteorological business constraint filtering, invalid event elimination: eliminate 2 high-level cyclones (center height 250 , >10 km). One short-term frontal break point (lasting 4.5 h < 6 h threshold) was excluded.

[0116] Blocking high pressure screening: retain the Yanshan blocking center (intensity ranked 92 in history) quantiles).

[0117] Final output: Business compliance coordinate set contains 5 key points: Cyclone generation point: (38.0 N, 109.8 E); frontal break point: (39.8 N, 120.1 E); blocking high pressure center: (41.5 N, 116.8 E).

[0118] 203. Weather-commercial space correlation mapping: mapping the supply chain node coordinates to the weather phase change coordinate set through conformal differential transformation to generate the Riemannian metric field of weather risk transmission;

[0119] Specifically, step S3.1: Meteorological-commercial spatial reference point matching, inputs the business-compliant meteorological phase change coordinate set output by 202; spatially associates three types of meteorological phase change events with key nodes in the preset supply chain network: cyclone generation location → logistics hub coordinates; frontal breakpoint → production base coordinates; blocking high pressure center → energy supply point coordinates; outputs a reference point pair mapping table (including meteorological-commercial coordinate pairs and association strengths);

[0120] Step S3.2: Conformal transformation curvature preservation: Input the reference point pair mapping table from step S3.1; Construct a diffeomorphic transformation that preserves Gaussian curvature to ensure: the geodesic in meteorological space is mapped to the shortest path in commercial space; the curvature of the cyclone path is equivalent to the curvature of the logistics path; Output the curvature conservation transformation function;

[0121] Step S3.3: Calculate the Riemann metric tensor. Input the curvature conservation transformation function of step S3.2 and the reference point pair mapping table of step S3.1. Calculate the Jacobian matrix of the transformation function to generate the metric tensor component of meteorological risk transmission: g 11 is the transmission intensity of meteorological risk along longitude; g 22 is the transmission intensity of meteorological risk along latitude; g 12 It is the meteorological-commercial coupling correlation; outputs the initial Riemannian metric tensor field;

[0122] Step S3.4: Supply chain network constraint optimization, input the initial Riemann metric tensor field of step S3.3; adjust the metric component according to the supply chain topological connectivity: enhance the g of the connected path 12 Component; weaken the g of the interrupt node 11 / g 22 Component; output the Riemannian metric field of meteorological risk transmission (as input of step 204).

[0123] It should be noted that the following uses the extreme snowstorm in North China on February 13, 2020 as an example to illustrate the implementation process of meteorological-commercial space association mapping (step 203). The data is based on the meteorological phase change coordinate set output in step 202 and the preset supply chain network:

[0124] Step S3.1: Meteorological-commercial space benchmark matching, input data: Business compliance meteorological phase change coordinate set (5 key points): Cyclone generation point: Beijing (39.9 N, 116.4 E, 850 ); frontal breaking point: Tianjin (39.1 N,117.2 E, 700 ); blocking high pressure center: Shijiazhuang (38.0 N, 114.5 E, 500 );

[0125] Preset supply chain network nodes (North China): Logistics hub: Beijing Daxing International Airport; Production base: Tianjin Binhai New Area Manufacturing Park; Energy supply point: Shijiazhuang coal-fired power plant;

[0126] Spatial association rules: cyclone generation points The nearest logistics hub within 150km; the frontal breaking point Production base within 100km; blocking high-voltage center Energy supply points (directly matching provincial power grid nodes);

[0127] Output: Reference point pair mapping table (association strength is calculated based on distance decay function):

[0128]

[0129] Step S3.2: Conformal transformation with curvature preservation, transformation target: meteorological space cyclone path (curvature 0.25 km -1 ) Equivalent curvature of logistics paths; geodesic line of frontal movement (Beijing to Tianjin) Shortest transportation path in the supply chain;

[0130] Diffeomorphism construction: using exponential mapping ( is the Gaussian curvature, is the distance); Constraints: Gaussian curvature change rate after transformation <5 (Average measured 2.3 Output result: Curvature conservation transformation function (the curvature error in the Beijing-Tianjin region is only 1.8 );

[0131] Step S3.3: Riemannian metric tensor calculation, Jacobian matrix calculation: partial derivatives of the transformation function Generate matrix components; initialize the metric tensor: (Longitude direction): represents the risk of a cold wave moving south, with an initial value of 1.2; (Latitude direction): represents the risk of blizzard expansion westward, with an initial value of 0.8; (coupling correlation): initial value 0.4 (Beijing-Tianjin axis); output result: initial metric tensor field (Shijiazhuang area =0.15 reflects weak energy-climate coupling);

[0132] Step S3.4: Supply chain network constraint optimization, topology adjustment rules: Enhance the connectivity of the Beijing-Tianjin Corridor: Increased from 0.4 to 0.6; weakened the impact of interrupted nodes: Hengshui storage node (closed due to blizzard) Cut 50 ; Physical verification: The error of the optimized meteorological risk transmission path length is <3 (vs. actual logistics trajectory); cyclone curvature mapped to the curvature of the Beijing-Tianjin Expressway (0.23km -1 vs. actual 0.25km -1 ).

[0133] 204. Construction of meteorological constraint decision space, using the Gaussian curvature of the Riemannian metric field as a constraint condition and combining the fractional entropy change rate of the vorticity field to construct a meteorological-commercial coupled decision manifold;

[0134] Specifically, step S4.1: generating a meteorological risk curvature field, inputting the Riemannian metric field of meteorological risk conduction outputted by step 203, calculating the Gaussian curvature scalar distribution of the metric field, and outputting the meteorological risk curvature field (a three-dimensional data volume containing the curvature values ​​of each point in space);

[0135] Step S4.2: Calculate the vorticity entropy change rate field. Input the corrected vorticity field that complies with the physical constraints output by 201, calculate the fractional entropy change rate of the vorticity field (using Caputo fractional derivatives), and output the vorticity entropy change rate field (a three-dimensional scalar field that represents the degree of irreversibility of the atmospheric system).

[0136] Step S4.3: Decision symplectic structure generation: input the meteorological risk curvature field from step S4.1 and the vorticity entropy change rate field from step S4.2, generate a symplectic form basis through Poisson bracket operation, and output the initial symplectic structure tensor;

[0137] Step S4.4: Meteorological-commercial manifold coupling, input the initial symplectic structure tensor from step S4.3, embed the supply chain state variables (inventory level, logistics flux, energy demand) into the symplectic structure, and output the coupled basic manifold;

[0138] Step S4.5: Curvature-constrained manifold optimization, input the meteorological risk curvature field of step S4.1 and the coupled basic manifold of step S4.4, optimize the manifold geometry with Gaussian curvature as a hard constraint, and output the meteorological-commercial coupled decision manifold (as input of 205).

[0139] It should be noted that the following uses the extreme snowstorm process in North China on February 13, 2020 as an example to illustrate the implementation process of constructing the meteorological constraint decision space (step 204). The data is based on the meteorological risk conduction Riemannian metric field and the corrected vorticity field output in step 203:

[0140] Step S4.1: Generate meteorological risk curvature field. Input data: meteorological risk conduction Riemannian metric field (from step 203). The metric tensor component of the Beijing-Tianjin Corridor is: =1.2 (meridional risk), =0.8 (latitudinal risk), =0.6 (coupling correlation).

[0141] Gaussian curvature calculation: using Riemann geometry formula Calculate the curvature of each point in space. The extreme curvature of the Shijiazhuang blocking high pressure center km -2(High-pressure ridge inhibits risk transmission), curvature of the Beijing-Tianjin frontal fault zone =0.08km -2 (Active fronts enhance risk diffusion.) Output: The three-dimensional curvature field shows that negative curvature dominates over central North China (accounting for 62% of the area), reflecting the suppressive effect of blocking high pressure on risk transmission.

[0142] Step S4.2: Calculation of vorticity entropy change rate field. Input data: cross-scale corrected vorticity field (from step 201), peak vorticity in Shijiazhuang area 3.8×10 −4 s −1 Caputo fractional entropy change rate: Defining the order , calculate the irreversible entropy production rate: Entropy change rate of Shijiazhuang blocking area , reflecting the intense dissipation of atmospheric energy. Output: Entropy change rate field shows high value area (>0.3 h -1 ) overlaps with the snowstorm area (Beijing-Baoding) by 85%.

[0143] Step S4.3: Decision symplectic structure generation, symplectic form basis construction: through Poisson brackets Generate the basis tensor. The main diagonal value of the Shijiazhuang region symplectic matrix , reflecting the strong coupling of curvature-entropy change.

[0144] Step S4.4: Meteorological-commercial manifold coupling, supply chain state embedding: Mapping Beijing-Tianjin supply chain variables to symplectic structure: logistics flux J Momentum coordinates ; Inventory level I Location coordinates After the Shijiazhuang energy demand variable E is embedded, the dimension of the symplectic manifold is expanded to 12 dimensions.

[0145] Step S4.5: Curvature-constrained manifold optimization, Gaussian curvature hard constraint: As the boundary condition, optimize the manifold geometry: compress the manifold curvature of the Shijiazhuang high pressure area to -0.09km -2 (40% decrease); stretch the manifold curvature of the Beijing-Tianjin frontal area to 0.07km -2 (An increase of 12%).

[0146] 205. Weather-driven decision parameter generation, solving the minimum entropy generation path on the decision manifold, and outputting an executable decision parameter set including the logistics path curvature coefficient, inventory elasticity factor and weather risk lead time.

[0147] Specifically, step S5.1: Calculating the entropy generation rate field of the decision manifold, inputting the meteorological-commercial coupled decision manifold constructed in 204, calculating the entropy generation rate scalar value of each point on the manifold, and outputting the entropy generation rate scalar field;

[0148] Step S5.2: Generate the initial decision path. Input the entropy production rate scalar field from step S5.1, generate an initial geodesic path connecting the initial state and the target state of the supply chain on the decision manifold, and output the initial decision path.

[0149] Step S5.3: Minimum entropy generation optimization: input the entropy generation rate scalar field of step S5.1 and the initial decision path of step S5.2, optimize the initial path with the goal of minimizing the integral of the entropy generation rate, and output the minimum entropy generation path;

[0150] Step S5.4: Extract meteorological decision parameters. Input the minimum entropy generation path from step S5.3 and extract key parameters along the path: maximum path curvature value → logistics path curvature coefficient; tangent vector change rate → inventory elasticity factor; path length / risk transmission speed → meteorological risk lead time; output the original decision parameter set.

[0151] Step S5.5: Business executable calibration. Input the original decision parameter set from step S5.4 and calibrate the parameters according to the physical constraints of the supply chain: limit the logistics curvature coefficient to [0.1, 5.0] radians / km; normalize the inventory elasticity factor to the interval [0,1]; adjust the meteorological risk lead time according to the logistics speed; and output the executable decision parameter set (to drive the business system).

[0152] It should be noted that the following uses the extreme snowstorm process in North China on February 13, 2020 as an example to illustrate the implementation process of generating meteorological-driven decision parameters (step 205). The data is based on the meteorological-commercial coupled decision manifold constructed in step 204 (input data: the extreme value of the curvature field of the Shijiazhuang blocking high pressure -0.15 km -2 The vorticity entropy change rate in the Beijing-Tianjin front region is 0.42 h -1 ):

[0153] Step S5.1: Calculate the entropy generation rate field of the decision manifold. Input data: meteorological-commercial coupled decision manifold (12-dimensional manifold from Shijiazhuang to Beijing). Calculation logic: Supply chain state variables: energy demand of Shijiazhuang coal-fired power plant (85,000 tons / day), logistics flow of the Beijing-Tianjin Corridor (1,200 vehicles / hour), warehouse inventory level in Beijing (65 ). Entropy generation rate formula: calculated based on the geometric curvature of the manifold and the supply chain state gradient. Output result: The entropy generation rate scalar field shows that the entropy generation rate of the Shijiazhuang blocking high pressure area is 0.85 h -1 High energy consumption risk), 0.32 h in the Beijing-Tianjin frontal area -1 (Low-risk channel).

[0154] Step S5.2: Initial decision path generation, path endpoint definition: Initial state: Inventory 65 、Energy demand 85,000 tons; target status: inventory 80 , energy demand 62,000 tons (peak regulation target after snowstorm); geodesic generation: the shortest path length connecting the endpoints is 118 manifold units, passing through the high curvature area of ​​Shijiazhuang (curvature -0.12 km -2 ).

[0155] Step S5.3: Minimum entropy generation optimization, optimization goal: minimize the path entropy generation integral (initial value 72.3). Iterative process: Round 1: bypass the Shijiazhuang high pressure area, the path length increases to 132 units, and the entropy integral decreases to 58.1; Round 3: integrate the low entropy channel of the Beijing-Tianjin front (entropy change rate 0.32 h -1 ), the entropy integral was optimized to 52.1 (a decrease of 28 Output path: After optimization, the curvature of the path in Shijiazhuang area is reduced to -0.07 km. -2 , avoiding blocking the high-pressure core area.

[0156] Step S5.4: Extract meteorological decision parameters. Key parameters along the optimized path are extracted: logistics path curvature coefficient: maximum curvature of 2.8 radians / km in the Beijing-Tianjin section (reflecting path curvature caused by heavy snow); inventory elasticity factor: tangent vector change rate of 0.73 (representing inventory sensitivity to energy fluctuations); meteorological risk lead time: path length / risk transmission speed = 9.2 hours (logistics delay warning);

[0157] Step S5.5: Commercially Executable Calibration, Parameter Constraints Applied: Curvature Coefficient Calibration: Constrained from 2.8 to 3.2 radians kilometers (due to the actual road network limit of 5.0); normalized inventory elasticity factor: 0.73 0.73 (already within the [0, 1] interval); Risk lead time adjustment: calibrated to 8.5 hours based on a logistics vehicle speed of 60 km / h. Output decision set: Drive the Tianjin Manufacturing Park to initiate emergency inventory allocation 8 hours in advance to reduce the risk of supply disruptions caused by heavy snow.

[0158] 206. Also includes real-time meteorological decision-making feedback loop:

[0159] Business execution status monitoring: real-time collection of logistics system GPS trajectory data, warehouse inventory change data, and insurance transaction time series data, and output of business execution status stream;

[0160] Weather-decision consistency verification: Input 205 outputs executable decision parameter set and business execution status flow, verify three key consistency indicators: actual curvature of logistics path vs. decision curvature coefficient; actual inventory elasticity response vs. inventory elasticity factor; actual occurrence time of risk event vs. weather risk lead time; output weather-decision deviation matrix;

[0161] Dynamic reinitialization of the vorticity field: when any indicator in the meteorological-decision deviation matrix exceeds a threshold, re-execution of 201 is triggered; the current business execution state flow is used as a new supply chain network parameter; and a re-initialization trigger signal is output;

[0162] The decision manifold is updated in real time in response to a reinitialization trigger signal; the decision manifold is reconstructed 204 based on the updated modified vorticity field; and the adaptively updated decision manifold is output;

[0163] The parameter set rolling optimization is re-executed on the updated decision manifold 205; the optimized decision parameter set is output (input into the commercial system for execution).

[0164] It should be noted that the following uses the extreme snowstorm in North China on February 13, 2020 as an example to illustrate the implementation process of the real-time meteorological decision-making feedback closed loop (step 206). The data is based on the decision parameter set output in step 205 and actual business execution data:

[0165] Steps: Business execution status monitoring, data collection (at 15:00 on February 13, during the snowstorm): Logistics GPS track: The average speed of trucks on the Beijing-Tianjin Expressway dropped to 32km / h (preset value ≥ 60km / h), and the measured path curvature was 4.1 radians / km. Warehousing inventory: The raw material inventory in Tianjin Manufacturing Park dropped to 52 (Decision target value 65 Insurance Transactions: Hebei's traffic delay insurance transactions surged by 218 in a single day, with the peak coinciding with a period of heavy snow. Output: Business execution status streams reveal unusual supply chain fluctuations.

[0166] Steps: Weather-decision consistency verification, deviation matrix calculation:

[0167]

[0168] Trigger condition: curvature deviation > 25 Triggers reinitialization.

[0169] Steps: Dynamic reinitialization of vortex field, update parameters: Supply chain network status: Shijiazhuang power plant inventory drops to 45 , Beijing-Tianjin logistics flux decreased by 40 Re-execute step 201: input the updated supply chain parameters to generate the modified vorticity field (the vorticity of the Shijiazhuang blocking high pressure is increased from 3.8×10 -4 s -1 Increased to 4.2×10 -4 s -1 ).

[0170] Steps: Real-time update of decision manifold, manifold reconstruction: Reconstruct the decision manifold based on the new vorticity field, and adjust the curvature of the Shijiazhuang area from -0.09 to -0.12 km -2(Blocking high pressure is enhanced). The peak value of the entropy generation rate field increases from 0.85 to 0.93 h -1 .

[0171] Steps: Parameter set rolling optimization, new decision parameters: logistics curvature coefficient: 3.8 radians / km (originally 3.2); inventory elasticity factor: 0.61 (originally 0.73); risk lead time: 7.2 hours (originally 8.5);

[0172] Implementation effect (08:00 on February 14): The raw material inventory of Tianjin Manufacturing Park is stable at 63 (The deviation is narrowed to 5 Shijiazhuang coal transportation delay was shortened to 7.5 hours (error 4.2 ).

[0173] In the embodiment of the present invention, a fractional-order Laplace operator is introduced to realize multi-scale coupling, and the dynamic fractional-order exponent is used to adaptively balance large-scale smoothing and small-scale detail retention, thereby solving the forecast error caused by scale discontinuity in traditional numerical models. A meteorological phase change feature extraction method based on continuous homology analysis is used to accurately locate the cyclone generation position through the H1 hole generation event, and the frontal break point is identified using the connected component splitting event. A Riemannian metric field construction method for meteorological risk transmission is proposed, and the Gaussian curvature constraint is coupled with the fractional-order entropy change rate to form a 12-dimensional symplectic manifold decision space. This system represents the paradigm shift of meteorological services from experience-driven to physical constraint + data-driven in the era of big data, and provides a solution for urban resilience construction in the context of frequent extreme weather events around the world. Its technical architecture is scalable and can be quickly migrated to meteorologically sensitive industries such as agriculture, aviation, and electricity.

[0174] The above describes in detail the intelligent weather forecast system simulated by numerical models in the embodiment of the present invention from the perspective of modular functional entities. The following describes in detail the intelligent weather forecast device simulated by numerical models in the embodiment of the present invention from the perspective of hardware processing.

[0175] Figure 3 This is a schematic diagram of the structure of an intelligent weather forecast device simulated by a numerical model provided by an embodiment of the present invention. The intelligent weather forecast device 300 simulated by the numerical model may have relatively large differences due to different configurations or performances. The device 300 includes a transmitter 301, a receiver 302 and a processor 303. The processor 303 may also be a controller. Figure 3 denoted as “controller / processor 303 ”. Optionally, the device 300 may further include a modem processor 305 , wherein the modem processor 305 may include an encoder 306 , a modulator 307 , a decoder 308 , and a demodulator 309 .

[0176] In one example, transmitter 301 conditions (e.g., performs analog-to-analog conversion, filtering, amplification, and frequency upconversion) the output samples and generates an uplink signal, which is transmitted via an antenna to an access network device. On the downlink, the antenna receives the downlink signal transmitted by the access network device. Receiver 302 conditions (e.g., performs filtering, amplification, frequency downconversion, and digitization) the signal received from the antenna and provides input samples. Within modem processor 305, encoder 306 receives traffic data and signaling messages to be transmitted on the uplink and processes them (e.g., formats, encodes, and interleaves them). Modulator 307 further processes (e.g., performs symbol mapping and modulation) the encoded traffic data and signaling messages and provides output samples. Demodulator 309 processes (e.g., demodulates) the input samples and provides symbol estimates. Decoder 308 processes (e.g., deinterleaves and decodes) the symbol estimates and provides decoded data and signaling messages for transmission to device 300. The encoder 306, modulator 307, demodulator 309, and decoder 308 can be implemented by the combined modem processor 305. These units perform processing based on the radio access technology (e.g., LTE and other evolved system access technologies) used by the radio access network. It should be noted that when the device 300 does not include the modem processor 305, the above functions of the modem processor 305 can also be performed by the processor 303.

[0177] The processor 303 controls and manages the actions of the device 300, and is configured to execute the processing performed by the device 300 in the above-described embodiments of the present disclosure. For example, the processor 303 is also configured to execute the various steps of the sending device or the receiving device in the above-described method embodiments, and / or other steps of the technical solutions described in the embodiments of the present disclosure.

[0178] Furthermore, the device 300 may further include a memory 304 , and the memory 304 is used to store program codes and data for the device 300 .

[0179] It is understandable that Figure 3 Only a simplified design of the device 300 is shown. In actual applications, the device 300 may include any number of transmitters, receivers, processors, modem processors, memories, etc., and all devices that can implement the embodiments of the present disclosure are within the scope of protection of the embodiments of the present disclosure.

[0180] The present invention also provides an intelligent weather forecasting device simulated by a numerical model, which includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor executes the steps of the intelligent weather forecasting system simulated by the numerical model in the above-mentioned embodiments.

[0181] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the steps of the intelligent weather forecasting system simulated by the numerical model.

[0182] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0183] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0184] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent weather forecast system simulated by numerical model, characterized in that: The intelligent weather forecast system simulated by the numerical model includes: The correction module is used to perform spatial fractional Laplace operation on the original vorticity field output by the meteorological numerical model to generate a corrected vorticity field with cross-scale continuity, including: Based on the original vorticity field data of the meteorological numerical model, the different scale components of the vorticity field are separated by wavelet multi-scale decomposition to obtain the vorticity scale separation matrix. According to the vorticity scale separation matrix, the fractional order index β∈(0.5,1) is dynamically determined according to the energy spectral density distribution of each scale component, and an adaptive fractional order correction operator is output; According to the vorticity scale separation matrix and the adaptive fractional-order correction operator, the corresponding fractional-order operator operation is applied to each scale component, and the cross-scale continuous field is reconstructed through non-local integral fusion to output the cross-scale continuous corrected vorticity field; Based on the cross-scale continuous correction of the vorticity field, it is verified that the field satisfies the potential vorticity conservation law and energy flux continuity, and a corrected vorticity field that complies with physical constraints is obtained; An identification module is used to perform continuous coherence analysis based on the modified vorticity field, identify and output a three-dimensional meteorological phase change coordinate set, including: Input the corrected vorticity field that complies with physical constraints, extract the closed isosurfaces in the vorticity field that meet the cyclone formation threshold, and obtain the three-dimensional coordinate set of the cyclone candidate area; Input the three-dimensional coordinate set of the cyclone candidate area, construct the α-complex topology structure and perform edge simplification to obtain the simplified meteorological topology complex; Based on the simplified meteorological topological complex, three types of meteorological phase change events are identified and marked, and the original meteorological phase change event set is output; Based on the original meteorological phase change event set, invalid events are filtered based on meteorological business rules, and a business-compliant meteorological phase change coordinate set is output; A mapping module, configured to map the supply chain node coordinates to the meteorological phase change coordinate set through conformal differential transformation to generate a Riemannian metric field of meteorological risk transmission; A construction module is used to construct a decision manifold for meteorological-commercial coupling by taking the Gaussian curvature of the Riemannian metric field as a constraint condition and combining the fractional entropy change rate of the vorticity field; The allocation module is used to solve the minimum entropy generation path on the decision manifold and output an executable decision parameter set.

2. The intelligent weather forecasting system based on numerical model simulation according to claim 1, characterized in that: The identification and marking of three types of meteorological phase change events include connected component splitting events marked as frontal breakpoints, H1 hole formation events marked as cyclone formation locations, and high persistence critical points marked as blocking high pressure centers; The filtering of invalid events based on meteorological business rules includes removing cyclone events with an altitude of more than 10 km, removing frontal breakpoints with a duration of less than 6 hours, and retaining blocking high pressure with an intensity greater than the 90th percentile.

3. The intelligent weather forecasting system based on numerical model simulation according to claim 1, characterized in that: include: Based on the business compliance meteorological phase change coordinate set, the three types of meteorological phase change events are spatially associated with the key nodes of the preset supply chain network, and a benchmark point pair mapping table is output; Input the benchmark point mapping table, construct a differential homeomorphism transformation that maintains the Gaussian curvature, ensure that the geodesic in meteorological space is mapped to the shortest path in commercial space, and that the curvature of the cyclone path is equivalent to the curvature of the logistics path. Output the curvature conservation transformation function. According to the curvature conservation transformation function and the reference point pair mapping table, the Jacobian matrix of the transformation function is calculated to generate the metric tensor component of meteorological risk transmission and output the initial Riemann metric tensor field; The initial Riemann metric tensor field is input, the metric components are adjusted according to the supply chain topological connectivity, and the Riemann metric field of meteorological risk transmission is output.

4. The intelligent weather forecasting system based on numerical model simulation according to claim 3, characterized in that: The spatial association of the three types of meteorological phase change events with key nodes of the preset supply chain network includes: associating the cyclone generation location with the coordinates of the logistics hub, associating the frontal breakpoint with the coordinates of the production base, and associating the blocking high pressure center with the coordinates of the energy supply point; The metric tensor components of the meteorological risk transmission include g 11 is the transmission intensity of meteorological risk along longitude, g 22 is the transmission intensity of meteorological risk along latitude, g 12 is the meteorological-commercial coupling correlation; The adjustment of the metric component according to the supply chain topology connectivity includes enhancing the g of the connectivity path 12 component, weakening the g of the interrupt node 11 / g 22 Quantity.

5. The intelligent weather forecasting system based on numerical model simulation according to claim 3, characterized in that: include: Based on the Riemannian metric field of meteorological risk transmission, the Gaussian curvature scalar distribution of the metric field is calculated and the meteorological risk curvature field is output; According to the corrected vorticity field that complies with physical constraints, the fractional entropy change rate of the vorticity field is calculated and the vorticity entropy change rate field is output; According to the meteorological risk curvature field and the vorticity entropy change rate field, the symplectic form basis is generated through Poisson bracket operation, and the initial symplectic structure tensor is output; According to the initial symplectic structure tensor, the supply chain state variables are embedded into the symplectic structure and the coupled basic manifold is output; Based on the meteorological risk curvature field and the coupled basic manifold, the manifold geometry is optimized with Gaussian curvature as a hard constraint, and the meteorological-commercial coupled decision manifold is output.

6. The intelligent weather forecasting system based on numerical model simulation according to claim 5, characterized in that: include: According to the decision-making manifold of meteorological-commercial coupling, the scalar value of entropy generation rate at each point on the manifold is calculated to obtain the scalar field of entropy generation rate; Input the entropy production rate scalar field, generate the initial geodesic path connecting the initial state and target state of the supply chain on the decision manifold, and output the initial decision path; According to the entropy generation rate scalar field and the initial decision path, the initial path is optimized with the goal of minimizing the entropy generation rate integral, and the minimum entropy generation path is output; According to the minimum entropy generation path, key parameters are extracted along the path: the curvature coefficient of the flow path is extracted from the maximum curvature value of the path; The tangent vector change rate is used to extract the inventory elasticity factor; Path length / risk transmission speed extracts meteorological risk lead time; Output the original decision parameter set; Based on the original decision parameter set, the parameters are calibrated according to the physical constraints of the supply chain, the logistics curvature coefficient is limited to [0.1, 5.0] radians / km, the inventory elasticity factor is normalized to the interval [0, 1], the meteorological risk lead time is adjusted according to the logistics speed, and the executable decision parameter set is output.

7. The intelligent weather forecasting system based on numerical model simulation according to claim 6, characterized in that: It also includes a feedback module for real-time meteorological decision-making feedback loop: Collect GPS trajectory data of logistics system, warehouse inventory change data and insurance transaction time series data in real time, and output business execution status flow; Input executable decision parameter sets and business execution status streams to verify three key consistency indicators: the actual curvature of the logistics path and the decision curvature coefficient; Inventory elasticity actual response and inventory elasticity factor; The actual occurrence time of the risk event and the meteorological risk lead time; Output weather-decision deviation matrix; When any indicator in the weather-decision deviation matrix exceeds a threshold: trigger the re-execution of the Laplace operation; use the current business execution state flow as the new supply chain network parameter; output a re-initialization trigger signal; responding to the re-initialization trigger signal; Reconstruct the decision manifold based on the updated modified vorticity field and output the adaptively updated decision manifold; Re-execute the output optimized decision parameter set on the updated decision manifold.

Images