Drainage basin-oriented sound wave precipitation enhancement meteorological condition response digital twinborn simulation platform

Abstract

The invention discloses a drainage basin-oriented acoustic precipitation meteorological condition response digital twinborn simulation platform. The method comprises the following steps of: constructing a three-dimensional space lower boundary condition with physical authenticity; continuously receiving a multi-source meteorological observation data stream, and generating a meteorological initial field and a boundary condition through data preprocessing and multi-source data fusion; coupling the non-static mesoscale meteorological model and the two-parameter cloud microphysical parameterization scheme, and outputting four-dimensional space-time distribution data; the two-parameter cloud microphysical parameterization scheme is dynamically corrected by solving the output of the sound wave propagation equation based on the microphysical model; and iteratively adjusting the sound wave parameter combination according to a preset rainfall enhancement objective function by adopting a global optimization algorithm and a deep neural network agent model, and outputting an optimal sound wave parameter combination and a corresponding predicted rainfall enhancement effect. According to the method, high-fidelity whole-process simulation can be realized, a sound wave-cloud interaction mechanism is disclosed, and the applicability and economic benefits of a sound wave precipitation enhancement technology in actual drainage basin water resource management are improved.

Description

Technical Field

[0001] This invention belongs to the field of meteorology and hydrology, and specifically relates to a digital twin simulation platform for meteorological condition response of acoustic rain enhancement for watersheds. Background Technology

[0002] Artificial rain enhancement, as an important technological means to alleviate regional water shortages, has received widespread attention in recent years. Among them, acoustic rain enhancement, as a non-chemical and environmentally friendly intervention method, promotes precipitation formation by stimulating microphysical processes in clouds with specific sound waves, showing promising application prospects. This technology involves multiple disciplines such as atmospheric physics, acoustics, fluid mechanics, and meteorological numerical simulation. Its core lies in understanding and controlling the propagation characteristics of sound waves in complex atmospheric media and their influence mechanisms on key microphysical processes such as cloud droplet condensation, ice crystal growth, and collision-coalescence. However, due to the highly nonlinear, spatiotemporally heterogeneous, and sensitive nature of cloud precipitation systems, the actual effect of acoustic rain enhancement is easily modulated by multidimensional meteorological factors such as temperature stratification, humidity distribution, wind field structure, and turbulence intensity at the watershed scale, posing a severe challenge to related research and applications.

[0003] Among the key challenges for watershed-scale acoustic rain enhancement technology, the crucial issue lies in achieving precise adaptation between acoustic parameters and dynamic meteorological conditions. Current technologies primarily rely on empirical settings or small-scale field experiments to determine key parameters such as acoustic frequency, intensity, emission height, and beam direction, lacking the ability to systematically simulate the entire acoustic-cloud interaction process under different meteorological scenarios. This crude parameter selection approach is not only inefficient but also fails to reveal the cumulative effects and feedback mechanisms of acoustic disturbances within the context of macroscopic hydrological cycles, severely hindering a deeper understanding of the technology's physical mechanisms.

[0004] The shortcomings of existing technologies in the field of acoustic rain enhancement are mainly manifested in the following aspects: First, they lack the ability to integrate high-resolution meteorological data with cloud microphysical processes for multi-scale coupled simulation, making it impossible to accurately reproduce the propagation attenuation, refraction, and deflection of sound waves in the real atmospheric environment and their dynamic impact on the evolution of cloud particle spectra. For example, Chinese patent CN114626417A discloses a rainfall detection method and system based on acoustic signal feature analysis. This invention relies solely on the acoustic signal itself for feature extraction and classification, without integrating high-resolution meteorological data, cloud microphysical process data, and watershed characteristic information. The lack of multi-scale coupled simulation support easily leads to insufficient accuracy and reliability in rainfall prediction. Second, an iteratively optimizable virtual experimental framework has not yet been established, resulting in… The adjustment of acoustic parameters is highly dependent on expensive and time-consuming field operations, making it difficult to support rapid and precise rain enhancement decisions. Secondly, rain enhancement effect assessments are mostly based on statistical comparison methods, which are easily affected by random fluctuations in natural precipitation and lack quantitative prediction capabilities based on process mechanisms. Finally, existing platforms generally do not consider the modulation effect of watershed topography, climate zones, and seasonal precipitation characteristics on the effectiveness of acoustic waves, making it difficult to design regionally customized rain enhancement schemes. For example, Chinese patent CN111276157B discloses a method and device for rainfall intensity identification and model training based on rain sound. Although this invention combines rainfall audio signals with environmental data, it does not solve the adaptation problem for different watershed topography, climate zones, and seasonal precipitation characteristics. These shortcomings collectively constitute the main bottlenecks in the current development of acoustic rain enhancement technology towards scientific, intelligent, and efficient applications, urgently requiring a digital twin platform that integrates meteorological sensing, process simulation, parameter optimization, and effect prediction to overcome these limitations. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention proposes a digital twin simulation platform for meteorological condition response of acoustic rain enhancement in watersheds.

[0006] The technical solution of the present invention is as follows: On the one hand, this invention provides a digital twin simulation method for meteorological condition response of acoustic rain enhancement in watersheds, comprising the following steps: A high-precision digital elevation model of the target watershed is constructed, and land use classification data and watershed hydrological network topology are integrated to form a three-dimensional lower boundary condition with physical realism. It continuously receives multi-source meteorological observation data streams, and generates high-resolution and physically consistent meteorological initial fields and boundary conditions through data preprocessing and multi-source data fusion. Based on the generated initial meteorological field and boundary conditions, a non-hydrostatic mesoscale meteorological model is coupled with a two-parameter cloud microphysical parameterization scheme to output four-dimensional spatiotemporal distribution data including the updated meteorological field and the content of various cloud water substances. In the updated meteorological field, the acoustic wave propagation equation considering atmospheric refraction, absorption attenuation and turbulent scattering is solved, and based on the microphysical model, the rate coefficients of related processes in the two-parameter cloud microphysical parameterization scheme are dynamically corrected using the output of the acoustic wave propagation equation. Using a global optimization algorithm and a deep neural network proxy model, the combination of acoustic parameters is iteratively adjusted according to a preset rain enhancement objective function, and the optimal combination of acoustic parameters and its corresponding predicted rain enhancement effect are output.

[0007] Preferably, the method further includes: The simulation process will present the updated meteorological field, cloud water content, sound wave propagation path, sound wave parameter combination iteration process, and predicted rain enhancement effect to the user in the form of two-dimensional planar diagrams, three-dimensional stereoscopic rendering diagrams, and dynamic time-series animations, and receive interactive adjustment instructions from the user on simulation parameters or optimization targets. The optimal combination of acoustic parameters is compiled to generate a quantitative prediction map of the predicted rain enhancement effect, and a comprehensive evaluation report is issued, which includes analysis of the physical rationality of the process, uncertainty assessment, and comparison with historical cases.

[0008] Preferably, the integration of land use classification data and watershed hydrological network topology specifically involves: integrating land use classification data, clearly distinguishing underlying surface types including urban built-up areas, farmland cultivation areas, forest cover areas, and lake and river water bodies, and constructing a hydrological network topology based on water level analysis algorithms; The specific method for constructing the hydrological network topology based on the water level analysis algorithm is as follows: Depressions were filled in the high-precision digital elevation model, and a priority queue algorithm was used to eliminate depressions in the data to ensure the continuity of water flow. The direction of water flow in each grid cell of a high-precision digital elevation model is determined based on the D8 algorithm. The cumulative runoff is calculated based on the flow direction matrix, and an initial river network is generated based on the cumulative runoff. The Strahler classification system is then used to classify the river channels. The hydrological network topology is constructed by representing the river channel connections using an adjacency matrix, and the automatically extracted hydrological network topology is then verified and corrected.

[0009] Preferably, the continuous reception of multi-source meteorological observation data streams, through data preprocessing and multi-source data fusion, generates high-resolution and physically consistent meteorological initial fields and boundary conditions, specifically as follows: It continuously receives multi-source meteorological observation data streams, including: infrared and visible light channel radiation data from geostationary meteorological satellites, reflectivity factor and radial wind data detected by weather radar, near-surface temperature, humidity, air pressure and wind field data collected by automatic weather stations, and temperature, humidity and wind vertical profile data acquired by radiosondes. Cloud detection and calibration are performed on infrared and visible light channel radiation data from geostationary meteorological satellites; ground clutter suppression and velocity de-blurring are performed on reflectivity factor and radial wind data detected by weather radar; extreme value removal and time consistency checks are performed on near-surface temperature, humidity, air pressure and wind field data collected by automatic weather stations; and system error correction and vertical interpolation are performed on temperature, humidity and wind vertical profile data acquired by radiosondes. Based on the above data preprocessing, a three-dimensional variational assimilation algorithm is used to fuse multi-source data to obtain high-resolution and physically consistent meteorological initial fields and boundary conditions.

[0010] Preferably, the non-hydrostatic mesoscale meteorological model adopts the ARPS model, which includes the x-direction momentum equation, continuity equation, thermodynamic energy equation, moist air state equation, water conservation equation, cloud microphysical transformation equation, and vertical motion equation. The dual-parameter cloud microphysics parameterization scheme predicts ten types of forecast variables, including cloud water mass mixing ratio and number concentration, rainwater mass mixing ratio and number concentration, cloud ice mass mixing ratio and number concentration, snow mass mixing ratio and number concentration, and graupel mass mixing ratio and number concentration, using the mass conservation equation and number concentration equation. The dual-parameter cloud microphysics parameterization scheme transforms microphysical mechanisms into computable source and sink terms, which are then fed back into a non-hydrostatic mesoscale meteorological model to simulate the entire life cycle evolution of cloud systems over the target watershed, including the generation, development, maturation, and dissipation. The output includes updated meteorological fields and four-dimensional spatiotemporal distribution data of various cloud water content.

[0011] Preferably, the sound wave propagation equation includes the basic sound wave equation, the ray tracing form equation, the sound intensity attenuation equation, and the correction term equation considering turbulent scattering. Based on the microphysical model, the improvement in cloud droplet collision efficiency under the action of the sound field is first calculated by using the output of the sound wave propagation equation. Then, the promoting effect of the sound field on the freezing process of supercooled cloud droplets is simulated to dynamically correct the rate coefficients of relevant processes in the two-parameter cloud microphysical parameterization scheme.

[0012] Preferably, the global optimization algorithm adopts a genetic algorithm or a particle swarm optimization algorithm, and the input individual of the global optimization algorithm is a combination of acoustic parameters including four decision variables: frequency, acoustic intensity, emission height, and beam angle. The deep neural network proxy model adopts an improved ConvLSTM-Attention network model, which includes an input layer, a 3D convolutional layer, a ConvLSTM layer, a spatiotemporal attention layer, a connection layer, and an output layer. The preset rain enhancement objective function is to maximize the ratio of the increase in surface precipitation within the target watershed to the operating cost; For each individual in each iteration, a deep neural network surrogate model is used to update the meteorological field and acoustic parameter combination as input features, and outputs the predicted cloud microphysical process evolution sequence and the final ground precipitation. Based on the predicted cloud microphysical process evolution sequence and final ground precipitation output by the deep neural network surrogate model, the global optimization algorithm updates and iterates until the preset convergence condition is met, and then outputs the optimal combination of acoustic parameters and its corresponding predicted rainfall enhancement effect.

[0013] On the other hand, the present invention provides a digital twin simulation platform for meteorological condition response of acoustic rain enhancement for watersheds, including a watershed three-dimensional geometric model module, a meteorological data input and assimilation module, a multi-scale meteorological and cloud microphysics model module, an acoustic wave propagation and action mechanism module, and a parameter optimization and effect prediction module. The watershed 3D geometric model module is used to construct a high-precision digital elevation model of the target watershed and integrate land use classification data and watershed hydrological network topology to form a physically realistic 3D spatial lower boundary condition. The meteorological data input and assimilation module is used to continuously receive multi-source meteorological observation data streams, and generate high-resolution and physically consistent meteorological initial fields and boundary conditions through data preprocessing and multi-source data fusion. The multi-scale meteorological and cloud microphysics model module is used to couple a non-hydrostatic mesoscale meteorological model with a two-parameter cloud microphysics parameterization scheme based on the generated meteorological initial field and boundary conditions, and output four-dimensional spatiotemporal distribution data including updated meteorological fields and the content of various cloud water substances. The acoustic wave propagation and action mechanism module is used to solve the acoustic wave propagation equation considering atmospheric refraction, absorption attenuation and turbulent scattering in the updated meteorological field, and based on the microphysical model, dynamically correct the rate coefficients of related processes in the two-parameter cloud microphysical parameterization scheme using the output of the acoustic wave propagation equation. The parameter optimization and effect prediction module is used to iteratively adjust the combination of acoustic parameters according to the preset rain enhancement objective function using a global optimization algorithm and a deep neural network proxy model, and output the optimal combination of acoustic parameters and its corresponding predicted rain enhancement effect.

[0014] In another aspect, the present invention also provides an electronic device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any embodiment of the present invention.

[0015] In another aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any embodiment of the present invention.

[0016] Compared with the prior art, the present invention has the following technical effects: This invention achieves accurate prediction and automatic parameter optimization of the entire process of acoustic rain enhancement through high-fidelity simulation and intelligent optimization, significantly improving operational efficiency and success rate. At the same time, it constructs a low-cost, zero-risk virtual experimental environment, reducing R&D costs while providing strong support for in-depth research on the mechanism of acoustic-cloud interaction and the formulation of regional customized rain enhancement schemes, greatly enhancing the applicability and economic benefits of this technology in water resource management. Attached Figure Description

[0017] Figure 1 This is an overall flowchart of the digital twin simulation method for meteorological condition response of acoustic rain enhancement oriented to watersheds, as described in this invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present application and with reference to the accompanying drawings.

[0019] Example 1 This embodiment provides a digital twin simulation method for the meteorological condition response of acoustic rain enhancement in a watershed, see reference. Figure 1 As shown, it includes the following steps: A high-precision digital elevation model (DEM) of the target watershed is constructed, integrating land use classification data and the watershed's hydrological network topology to form physically realistic three-dimensional lower boundary conditions. Specifically, the high-precision DEM model used depends on the actual application scenario. For example, in a typical watershed water resources management and weather modification operational scenario in North China, a DEM with a horizontal resolution of no less than 90 meters and a vertical accuracy strictly controlled within 5 meters is used. This is simultaneously fused with a 1:50,000 scale topographic map and multispectral remote sensing imagery to accurately depict the direction of ridgelines, the development characteristics of valleys, and the slope and aspect distribution in different areas within the watershed.

[0020] As a preferred embodiment of this practice, the integration of land use classification data and watershed hydrological network topology specifically involves: integrating land use classification data, clearly distinguishing underlying surface types such as urban built-up areas, farmland cultivation areas, forest cover areas, and lake and river water bodies, constructing a hydrological network topology based on water level analysis algorithms, and clearly defining the spatial connection relationship and confluence path between the main river channel and tributaries at all levels.

[0021] The specific method for constructing the hydrological network topology based on the water level analysis algorithm is as follows: The high-precision digital elevation model is filled with depressions, and a priority queue algorithm is used to eliminate depressions in the data to ensure the continuity of water flow.

[0022] Determine the flow direction of each grid cell in the high-precision digital elevation model based on the D8 algorithm:

[0023] In the formula, is the flow direction; is the elevation of the central grid, is the elevation of the adjacent grid, is the grid spacing.

[0024] Calculate the cumulative flow volume according to the flow direction matrix, calculate the upstream catchment area of each grid according to the flow direction matrix, and define it as the starting point of the river channel when the cumulative volume exceeds the threshold T (usually set to 0.5% of the catchment area).

[0025] Generate an initial river network based on the cumulative flow volume, and classify the river channels using the Strahler classification system.

[0026] Construct the hydrological network topology by representing the river channel connection relationship through an adjacency matrix. As defined, nodes: river channel confluence points, source points, outlet points; edges: river channel segments connecting nodes; attributes: river channel length, slope, width, etc.

[0027] Fuse the water system information in the topographic map to verify and correct the automatically extracted hydrological network topology, ensuring the accurate spatial connection relationship between the main river channel and the tributaries.

[0028] Furthermore, the catchment hydrological network topology structure can be represented by a directed acyclic graph (DAG), and its mathematical expression is G=(V,E), where V is the set of nodes, E is the set of edges, and it satisfies: for any edge e=(v_i,v_j)∈E, there is i<j (to ensure no cycle); each node v_k has an attribute set A_k={catchment area, slope, land use type,...}; and the river channel connection relationship satisfies the mass conservation equation.

[0029] Continuously receive multi-source meteorological observation data streams, and generate a high-resolution and physically consistent meteorological initial field and boundary conditions through data preprocessing such as data quality control and spatio-temporal interpolation and multi-source data fusion.

[0030] As a preferred implementation manner of this embodiment, the continuous reception of multi-source meteorological observation data streams and the generation of a high-resolution and physically consistent meteorological initial field and boundary conditions through data preprocessing and multi-source data fusion are specifically as follows: Continuously receive multi-source meteorological observation data streams, including: infrared and visible channel radiation data of stationary meteorological satellites, reflectivity factor and radial wind data detected by weather radars, near-surface temperature, humidity, air pressure and wind field data collected by automatic weather stations, and vertical profiles of temperature, humidity and wind obtained by radiosondes.

[0031] Cloud detection and calibration are performed on infrared and visible light channel radiation data from geostationary meteorological satellites. Ground clutter suppression and velocity de-blurring are performed on reflectivity factor and radial wind data detected by weather radar. Extreme value removal and time consistency checks are performed on near-surface temperature, humidity, air pressure and wind field data collected by automatic weather stations. System error correction and vertical interpolation are performed on temperature, humidity and wind vertical profile data acquired by radiosondes.

[0032] Specifically, the cloud detection and calibration verification includes: radiometric calibration of the visible light channel and infrared channel; preliminary identification based on the threshold method of visible light reflectance threshold and infrared brightness temperature difference; multispectral feature analysis to eliminate desert interference through cross-channel comparison (shortwave infrared, water vapor channel); and texture analysis to determine clouds by calculating the standard deviation of the visible light channel; cloud phase identification; and calibration verification including: cross-validation by comparing with synchronous measurements from ground observation stations and time series consistency checks based on the rate of change of adjacent orbit data.

[0033] The ground clutter suppression can employ a dual filtering technique, combining time-domain and frequency-domain filtering. The time-domain filtering can utilize adaptive recursive filtering, and the frequency-domain filtering process can be as follows: the time-domain signal is converted to the frequency domain using Fourier transform; after the Fourier transform, a Gaussian window function is applied to selectively filter the frequency components of the frequency-domain signal. The velocity de-ambiguity processing can employ a continuity constraint algorithm, which includes: initial ambiguity point identification, two-dimensional phase expansion, and quality control: applying meteorological consistency checks to exclude velocity values ​​that do not conform to atmospheric dynamics laws.

[0034] Based on the above data preprocessing, a three-dimensional variational assimilation algorithm is used to fuse multi-source data to obtain a high-resolution and physically consistent meteorological initial field and boundary conditions. The objective function of the three-dimensional variational assimilation algorithm is defined as the weighted sum of squares of the deviations between the analysis field and the background field, and the differences between observed and simulated values: ,in Let be the objective function. For the analysis field, As background scene, The background error covariance matrix, For the observation vector, For the observation operator, Let be the observation error covariance matrix. By solving the minimization problem of this objective function, physically consistent high-resolution meteorological initial fields and boundary conditions are obtained, providing reliable driving data for subsequent numerical simulations.

[0035] Based on the generated initial meteorological field and boundary conditions, a non-hydrostatic mesoscale meteorological model is coupled with a two-parameter cloud microphysical parameterization scheme to output four-dimensional spatiotemporal distribution data, including updated meteorological fields (including three-dimensional wind, temperature, and humidity fields) and the content of various cloud water substances.

[0036] In a preferred embodiment of this practice, the non-hydrostatic mesoscale meteorological model employs the ARPS model, preferably with a horizontal grid spacing of 1 to 3 kilometers, a vertical number of at least 50 layers, and a model top height of 20 kilometers. Its core equations include the momentum equation in the x-direction:

[0037] In the formula, Three-dimensional wind field; For time; Spatial coordinates; It refers to air pressure; For Coriolis parameters; Wind speed in the y direction; This is the turbulent diffusion term.

[0038] Continuity equation:

[0039] In the formula, air density; This is the wind speed vector.

[0040] Thermodynamic energy equation:

[0041] In the formula, Potential temperature; For heat source items; It is a specific heat at constant pressure; This is the reference potential temperature.

[0042] Equation of state for moist air:

[0043] In the formula, It refers to air pressure; air density; The gas constant for dry air; It is a false fever.

[0044] The equations of vertical motion are extended to non-static terms through static equilibrium corrections:

[0045] In the formula, Wind speed in the z-direction; This is the wind speed vector; For horizontal gradient operators; It is the acceleration due to gravity; For reference density; This is the buoyancy term; This represents the vertical turbulent diffusion term.

[0046] Water conservation equation:

[0047] In the formula, The total water-material mixing ratio; This represents the total source and sink terms of microphysical processes.

[0048] Cloud microphysical transformation equation:

[0049]

[0050] In the formula, The cloud-water mixing ratio; The self-conversion rate; For collision and collection rate; Condensation / evaporation rate; Cloud droplet number concentration; The nucleation rate; The collision and consumption rate.

[0051] The dual-parameter cloud microphysics parameterization scheme predicts ten types of forecast variables, including cloud water mass mixing ratio and number concentration, rainwater mass mixing ratio and number concentration, cloud ice mass mixing ratio and number concentration, snow mass mixing ratio and number concentration, and graupel mass mixing ratio and number concentration, using the mass conservation equation and number concentration equation.

[0052] In one specific embodiment, this embodiment employs the two-parameterization method of the Purdue-Lin microphysics scheme, whose core equation is for each type of hydrogel. (Cloud water, rainwater, cloud ice, snow, sleet), solve for the mass mixing ratio. Sum of concentrations The prediction equation is as follows. The mass conservation equation is expressed as:

[0053] In the formula, Source and sink terms, including condensation / evaporation terms. , Collision and Collection Items Freezing process item .

[0054] The number concentration equation is expressed as:

[0055] in For generation rate, This represents the consumption rate.

[0056] The specific process includes: Cloud droplet nucleation: , For supersaturation, Cloud droplet number concentration, The concentration of cloud condensation nuclei. It is an experience index.

[0057] Cloud droplet collision and merging: collecting collision and merging kernel efficiency , For turbulent kinetic energy, This is an empirical coefficient.

[0058] Ice crystals proliferate: ,in Ice crystal number concentration, Based on the basic ice crystal concentration, For temperature sensitivity coefficient, This is the freezing temperature. This is a reference temperature.

[0059] Freezing process: ,in For freezing rate, B is an empirical coefficient, which is related to the type of aerosol. This refers to the content of supercooled water.

[0060] The above numerical results are achieved using a step-by-step integration method. First, the source and sink terms of the microphysical process are calculated, and then the advection is calculated. A time step (e.g., 5s) is set to ensure stability.

[0061] The dual-parameter cloud microphysical parameterization scheme describes in detail the transformation mechanism of cloud droplets into raindrops through water vapor condensation and turbulent collision, simulates the evolution path of ice crystals into snowflakes through sublimation growth and crystal aggregation, and accurately calculates the frost formation process caused by the collision of ice phase particles with supercooled water droplets and its contribution to precipitation formation.

[0062] The dual-parameter cloud microphysics parameterization scheme transforms microphysical mechanisms into computable source and sink terms, which are then fed back into a non-hydrostatic mesoscale meteorological model to simulate the entire life cycle evolution of cloud systems over the target watershed, including the generation, development, maturation, and dissipation. The output includes updated meteorological fields and four-dimensional spatiotemporal distribution data of various cloud water content.

[0063] Specifically, the implementation process of the dual-parameter cloud microphysics parameterization scheme, which transforms microphysical mechanisms into computable source and sink terms and feeds them back to the non-hydrostatic mesoscale meteorological model, is as follows: read the initial meteorological field and boundary conditions; calculate the source and sink terms of processes such as condensation, collision and freezing in sequence; update the water mixture ratio and number concentration; feed them back to the non-hydrostatic mesoscale meteorological model to affect atmospheric motion, update the meteorological field, and feed back the topographic influence.

[0064] In updating the meteorological field, the acoustic propagation equation considering atmospheric refraction, absorption attenuation, and turbulent scattering is solved. Based on a microphysical model, the rate coefficients of relevant processes in the two-parameter cloud microphysical parameterization scheme are dynamically corrected using the output of the acoustic propagation equation. This step uses the acoustic ray tracing method to solve the acoustic propagation equation considering the effects of wind shear and temperature gradient, calculating the three-dimensional propagation path of the sound wave from the ground source to the target cloud region to determine the boundary of the effective action area, and accurately quantifying the spatial distribution of sound intensity and wavefront morphology changes under the effects of atmospheric refraction, absorption attenuation, and turbulent scattering.

[0065] In a preferred embodiment of this invention, the sound wave propagation equation includes: Basic acoustic wave equation:

[0066] In the formula, Sound pressure (Pa) is the pressure exerted by sound waves. Speed ​​of sound; For wind fields.

[0067] Ray tracing formal equation:

[0068] in For the arc length of the ray, For the speed of sound, This is a position vector.

[0069] Sound intensity attenuation equation:

[0070] in Atmospheric absorption coefficient, Thermal conductivity, For isobaric specific heat, For constant volume specific heat, For dynamic viscosity, For volume viscosity, Thermal conductivity, For sound intensity, This refers to air density.

[0071] The corrected equation considering turbulent scattering is as follows:

[0072] in For correction items, The refractive index is the structural constant. For wave number, For transmission distance, The scattering angle is... For reference sound intensity.

[0073] Based on a microphysical model, the improvement in cloud droplet collision efficiency under the influence of a sound field is first calculated using the output of the sound wave propagation equation. Then, the promoting effect of the sound field on the freezing process of supercooled cloud droplets is simulated. The ice nucleus activation rate exhibits an exponential growth characteristic with every 10 dB increase in sound pressure level. This is used to dynamically correct the rate coefficients of relevant processes in the two-parameter cloud microphysical parameterization scheme, and to directly reflect the influence of sound waves on cloud droplet collision efficiency and ice crystal nucleation rate in the numerical simulation process.

[0074] Specifically, the core formula of the microscopic physical model includes the improvement of cloud droplet collision efficiency:

[0075] in To improve collision efficiency, For characteristic frequencies, For reference sound intensity, For Weber numbers, For sound wave frequency, For sound intensity, It is an environmental correction factor.

[0076] Freezing-promoting effect:

[0077] in, To freeze the nuclear concentration, Based on the freezing rate, The barometric pressure sensitivity coefficient This is the actual air pressure. Standard atmospheric pressure Sound pressure level, The sound field enhancement coefficient is experimentally determined to be 0.15 ± 0.03.

[0078] The application process of the microphysical model is as follows: input the sound wave propagation equation and output the result; calculate the impact of the sound field on cloud microphysical processes, including correcting the collision and merger efficiency. , To correct the efficiency of collision-merging kernels, Based on collision and core efficiency; update freeze rate , To update the freeze rate, Based on the freezing rate, To update the frozen nucleus concentration, The nuclear concentration is frozen based on this.

[0079] In a specific embodiment, the boost amplitude, the value of the sound wave frequency, and the sound intensity range can be dynamically adjusted according to the actual application scenario, preferably within a frequency range of 20-2000Hz and a sound intensity range of 100-140dB. For example, the frequency range adjustment can be based on: low-frequency extension (<20Hz): suitable for deep convective cloud systems (cloud top height > 8km), where the sound wave wavelength matches the cloud scale better. ,in For the minimum frequency, For the speed of sound, Cloud thickness. High-frequency extension (>2000Hz): Applicable to shallow, thin stratiform clouds (cloud thickness <1km), must meet the following requirements: , For the maximum frequency, This refers to the minimum cloud droplet diameter. Sound intensity threshold adjustment methods: Lower limit adjustment: When the relative humidity is less than the set percentage (e.g., 60%), the lower limit of sound intensity needs to be increased. Upper limit adjustment: For example, in turbulent conditions... At speeds of m / s, the upper limit of sound intensity needs to be lowered to prevent excessive energy scattering. Alternatively, an adaptive adjustment mechanism can be employed: by monitoring cloud base height in real time. and cloud and water path , dynamically calculate the optimal parameters.

[0080] Using a global optimization algorithm and a deep neural network proxy model, the combination of acoustic parameters is iteratively adjusted according to a preset rain enhancement objective function, and the optimal combination of acoustic parameters and its corresponding predicted rain enhancement effect are output.

[0081] As a preferred embodiment of this example, the global optimization algorithm adopts a genetic algorithm or a particle swarm optimization algorithm. The input individual of the global optimization algorithm is a combination of acoustic parameters including four decision variables: frequency, acoustic intensity, emission height, and beam angle.

[0082] The operational mechanisms when using genetic algorithms include: Selection: tournament selection can be used. Crossover: simulated binary crossover (SBX). Mutation: polynomial mutation.

[0083] When using the particle swarm optimization algorithm, the velocity update formula can be improved:

[0084] In the formula, Let be the d-dimensional velocity of the i-th particle at time t+1; For time-varying inertial weights; For the first Individual particles Moment Dimensional velocity represents the current state of motion; The cognitive coefficient (usually taken as 2.0); The random numbers are uniformly distributed in [0,1], and randomness is introduced to avoid premature convergence. For the first The best position in the history of each particle Dimensional components; For the first Individual particles Moment Dimensional components; The social coefficient (usually taken as 2.0) controls the intensity of social learning; The random numbers are uniformly distributed in [0,1], and randomness is introduced to avoid premature convergence. For guiding position Dimensional components; The learning rate; The objective function is... for Dimensional components.

[0085] The inertial weight Non-linear decrease:

[0086] In the formula, As initial inertia weights, emphasizing global search, The final inertia weight emphasizes local search and can be set according to actual needs, such as... , , This represents the maximum number of iterations.

[0087] The particle swarm optimization algorithm can incorporate adaptive mesh technology to manage the archive and dynamically adjust the mesh accuracy.

[0088] The specific calculation of the acoustic wave parameter combination is as follows: Frequency determined: This represents the cloud thickness (output from the meteorological model).

[0089] Sound intensity determination: Atmospheric attenuation coefficient, For transmission distance, The intensity of the emitted sound source.

[0090] Launch altitude: , The height of the inversion layer. This refers to the height of the cloud base.

[0091] Beam angle: , The diameter of the transmitting array (default 20m). The wavelength of the sound wave.

[0092] The deep neural network proxy model adopts an improved ConvLSTM-Attention network model, including an input layer, a 3D convolutional layer, a ConvLSTM layer, a spatiotemporal attention layer, a connection layer, and an output layer. In a specific embodiment, the deep neural network proxy model can be set as follows: Input layer: meteorological update field (64×64×10) + acoustic parameter combination (4); 3D convolutional layer: 16 3×3×3 convolutional kernels, ReLU activation; ConvLSTM layer: 32 channels, memory unit; spatiotemporal attention layer: captures long-range dependencies; Output layer: cloud microphysical sequence (24×64×64×10) + precipitation (64×64).

[0093] The evaluation process of the deep neural network agent model is as follows: all individuals in each generation are evaluated; evaluation process: input parameters → network forward propagation → output prediction results; cloud microphysical sequence generation: the hidden state is converted into a spatiotemporal sequence through a decoder; precipitation calculation.

[0094] The preset rain enhancement objective function maximizes the ratio of the increase in surface precipitation within the target watershed to the operational cost. The increase in precipitation is obtained through numerical simulation, while the operational cost, taking into account factors such as equipment energy consumption, personnel allocation, and time windows, can be expressed as:

[0095] In the formula, Operating costs; energy costs: Equipment characteristic coefficients). For sound intensity, For distance; labor costs: (Staffing requirements increasing with distance) Time window cost: (Adjusted to 2.0h for high-altitude areas); Weighting coefficient: Yuan / kWh Yuan / person·h Yuan.

[0096] For each individual in each iteration, a deep neural network surrogate model is used to update the meteorological field and acoustic parameter combination as input features, and outputs the predicted cloud microphysical process evolution sequence and the final ground precipitation.

[0097] Based on the predicted cloud microphysical process evolution sequence and final ground precipitation output by the deep neural network surrogate model, a global optimization algorithm is used for iterative updates until a preset convergence condition is met, at which point the optimal combination of acoustic parameters and its corresponding predicted rainfall enhancement effect are output. Specifically, the deep neural network surrogate model outputs the difference in final ground precipitation calculations for scenarios with and without acoustic waves. Then, the objective function value of the global optimization algorithm is calculated. A global optimization algorithm generates a new generation of population.

[0098] In a preferred embodiment of this invention, the method further includes: The system presents the updated meteorological field, cloud water content, sound wave propagation path, sound wave parameter combination iteration process, and predicted rainfall enhancement effect to users in the form of two-dimensional planar diagrams, three-dimensional stereoscopic rendering diagrams, and dynamic time-series animations during the simulation process, and receives interactive adjustment instructions from users for simulation parameters or optimization targets.

[0099] Specifically, the three-dimensional wind field in the simulation process is dynamically presented as a two-dimensional streamline diagram and a three-dimensional wind plume. The distribution of cloud water content is displayed through isosurfaces and vertical profiles with different color codes. The sound wave propagation path is superimposed on the three-dimensional scene as a ray beam. The optimization process of sound wave parameter combinations is updated in real time with convergence curves and population distribution maps. The final rainfall enhancement effect is clearly displayed in the form of a spatial distribution map of the increase in watershed rainfall. Users can adjust simulation parameters in real time through the graphical user interface, including modifying the simulation duration, setting different cloud microphysical parameterization schemes, selecting specific sound wave action areas, and redefining the weight coefficients in the optimization objective function. All interactive commands are transmitted to each calculation module in real time through an event-driven mechanism and trigger dynamic updates of the simulation process.

[0100] The optimal combination of acoustic parameters is compiled to generate a quantitative prediction map of the predicted rainfall enhancement effect, and a comprehensive evaluation report is produced, including process physics rationality analysis, uncertainty assessment, and comparison with historical cases. Furthermore, a spatial distribution map of the predicted watershed rainfall increase is generated, stored as raster data with a spatial resolution consistent with the meteorological model grid, and accompanied by complete metadata. The evaluation report includes a process physics rationality analysis, verifying the physical consistency of the simulation results by comparing the evolution paths of cloud microphysical processes under two scenarios: with and without acoustic effects, and without external field intervention, using multi-scale meteorological and cloud microphysical model steps. and , The concentration of frozen nuclei in a sound wave scenario. To freeze the nucleus concentration in a non-interventional scenario, a sensitivity experiment based on ensemble forecasting technology was conducted. This experiment involved generating 20 perturbation initial fields in the meteorological data input and assimilation steps, calculating the precipitation forecast standard deviation, quantifying the impact of initial field errors and parameterization scheme uncertainties on the prediction results, and comparing and verifying the results with measured precipitation data from similar historical weather cases. Quantitative evaluation indicators such as prediction bias, correlation coefficient, and skill score were also calculated.

[0101] In a more specific embodiment, in the application scenario of ecological water replenishment in the arid Northwest watershed, the method described in this embodiment is adapted to specific terrain and climatic conditions. The watershed 3D geometric modeling step focuses on enhancing the accuracy of depicting mountainous terrain and valley topography, employing digital elevation model data with a horizontal resolution increased to 30 meters and a vertical accuracy further improved to 2 meters. It also refines the topography of windward and leeward slopes in mountainous areas to accurately capture the impact of orographic lifting on cloud development. The meteorological data input and assimilation step enhances the assimilation capability of observational data specific to plateau regions, particularly strengthening the quality control of boundary layer structure in radiosonde data, including boundary layer identification using the Richardson number criterion and outlier detection using the sliding window 3σ criterion. Furthermore, it improves the accuracy of wind field analysis under complex terrain by adapting to a terrain-adapted coordinate system, decomposing the wind field, and correcting the terrain.

[0102] The multi-scale meteorological and cloud microphysics model module has been adjusted to suit the characteristics of cloud systems in arid regions. The horizontal grid spacing has been increased to 500 meters for better analysis of convective cloud clusters, and the number of vertical layers has been increased to 70 to finely characterize the boundary layer structure and cloud top height. Horizontal grid: Meeting cloud resolution conditions Vertical layer distribution: , This refers to the height of the cloud top.

[0103] The cloud microphysics parameterization scheme has been optimized to describe the ice phase process, and the parameterization of the freezing of supercooled cloud droplets and the propagation of ice crystals has been strengthened, so as to more accurately simulate the mixed phase cloud precipitation mechanism commonly found in arid areas.

[0104] The sound wave propagation and action mechanism module optimizes the sound wave propagation model for high-altitude atmospheric conditions, specifically considering the impact of low-pressure environments on sound wave attenuation and revising the formula for calculating the sound wave absorption coefficient. The microscopic physical model enhances the description of the ice nucleus activation process, establishing a sound wave-ice nucleus activation response relationship applicable to aerosol conditions in arid regions, thus improving the accuracy of predicting catalytic effects. Among these improvements, the revised sound wave absorption coefficient... Represented as: , This is the actual air pressure. Standard atmospheric pressure. Arid zone correction: , The sensitivity coefficient for dust aerosols. This refers to the concentration of dust.

[0105] The parameter optimization and effect prediction steps incorporate ecological benefit indicators into the objective function, and perform correlation analysis between rainfall enhancement effects and the degree of improvement in vegetation restoration water requirements and soil moisture. The ecological benefit indicators are defined as follows: , As an indicator of ecological benefits, Soil moisture increase (mm) : Vegetation index improved (NDVI units). Runoff recharge (m³). , , These are the corresponding weight factors. The objective function of the global optimization algorithm is then reconstructed as: ,in , Weighted by economic benefits, Weighting based on ecological benefits.

[0106] Historical cases from arid regions are added to the training dataset of the surrogate model to improve the model's predictive generalization ability under special climatic conditions. The global optimization algorithm can employ multi-objective particle swarm optimization or multi-objective genetic algorithm to simultaneously optimize rainfall enhancement and operational costs, outputting a set of parameter combinations that satisfy different decision-making preferences.

[0107] The visualization and interactive interface steps have added an ecological benefit assessment layer, overlaying rainfall prediction results with ecological indicators such as vegetation index and soil moisture to support comprehensive decision-making by water resource management departments. The results output and evaluation steps have expanded the assessment indicator system, adding quantitative assessment of the ecological water replenishment benefits, and generating a comprehensive benefit assessment report that includes expected increases in rainfall, soil moisture, and vegetation response.

[0108] Example 2 Accordingly, this embodiment provides a watershed-oriented digital twin simulation platform for acoustic rain enhancement meteorological condition response, used to implement the method described in Embodiment 1, including a watershed three-dimensional geometric model module, a meteorological data input and assimilation module, a multi-scale meteorological and cloud microphysics model module, an acoustic wave propagation and action mechanism module, a parameter optimization and effect prediction module, a visualization and interactive interface module, and a result output and evaluation module.

[0109] The watershed 3D geometric model module is used to construct a high-precision digital elevation model of the target watershed and integrate land use classification data and watershed hydrological network topology to form physically realistic 3D spatial lower boundary conditions.

[0110] The meteorological data input and assimilation module is used to continuously receive multi-source meteorological observation data streams, and generate high-resolution and physically consistent meteorological initial fields and boundary conditions through data preprocessing and multi-source data fusion.

[0111] The multi-scale meteorological and cloud microphysics model module is used to couple a non-hydrostatic mesoscale meteorological model with a two-parameter cloud microphysics parameterization scheme based on the generated initial meteorological field and boundary conditions, and output four-dimensional spatiotemporal distribution data including updated meteorological fields and the content of various cloud water substances.

[0112] The Sound Wave Propagation and Mechanism Module is used to solve the sound wave propagation equation considering atmospheric refraction, absorption attenuation, and turbulent scattering in the updated meteorological field, and to dynamically correct the rate coefficients of related processes in the two-parameter cloud microphysical parameterization scheme based on the output of the sound wave propagation equation using a microphysical model.

[0113] The parameter optimization and effect prediction module is used to iteratively adjust the combination of acoustic parameters according to the preset rain enhancement objective function using a global optimization algorithm and a deep neural network proxy model, and output the optimal combination of acoustic parameters and its corresponding predicted rain enhancement effect.

[0114] The visualization and interactive interface module is used to present the updated meteorological field, cloud water content, sound wave propagation path, sound wave parameter combination iteration process, and predicted rainfall enhancement effect to the user in the form of two-dimensional planar diagrams, three-dimensional stereoscopic rendering diagrams, and dynamic time-series animations during the simulation process, and to receive interactive adjustment instructions from the user on simulation parameters or optimization targets.

[0115] The results output and evaluation module is used to compile the optimal combination of acoustic parameters, generate a quantitative prediction map of the predicted rain enhancement effect, and produce a comprehensive evaluation report that includes process physical rationality analysis, uncertainty assessment, and comparison with historical cases.

[0116] Example 3 This embodiment provides an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method described in Embodiment 1 of the present invention.

[0117] Example 4 This embodiment provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in Embodiment 1 of the present invention.

[0118] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, A and B simultaneously, or B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c can be single or multiple.

[0119] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0120] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0121] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a 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 to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0122] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A digital twin simulation method for meteorological condition response of acoustic rain enhancement in a watershed, characterized in that, Includes the following steps: A high-precision digital elevation model of the target watershed is constructed, and land use classification data and watershed hydrological network topology are integrated to form a three-dimensional lower boundary condition with physical realism. It continuously receives multi-source meteorological observation data streams, and generates high-resolution and physically consistent meteorological initial fields and boundary conditions through data preprocessing and multi-source data fusion. Based on the generated initial meteorological field and boundary conditions, a non-hydrostatic mesoscale meteorological model is coupled with a two-parameter cloud microphysical parameterization scheme to output four-dimensional spatiotemporal distribution data including the updated meteorological field and the content of various cloud water substances. In the updated meteorological field, the acoustic wave propagation equation considering atmospheric refraction, absorption attenuation and turbulent scattering is solved, and based on the microphysical model, the rate coefficients of related processes in the two-parameter cloud microphysical parameterization scheme are dynamically corrected using the output of the acoustic wave propagation equation. Using a global optimization algorithm and a deep neural network proxy model, the combination of acoustic parameters is iteratively adjusted according to a preset rain enhancement objective function, and the optimal combination of acoustic parameters and its corresponding predicted rain enhancement effect are output.

2. The digital twin simulation method for meteorological condition response of acoustic rain enhancement oriented to watersheds as described in claim 1, characterized in that, The method further includes: The simulation process will present the updated meteorological field, cloud water content, sound wave propagation path, sound wave parameter combination iteration process, and predicted rain enhancement effect to the user in the form of two-dimensional planar diagrams, three-dimensional stereoscopic rendering diagrams, and dynamic time-series animations, and receive interactive adjustment instructions from the user on simulation parameters or optimization targets. The optimal combination of acoustic parameters is compiled to generate a quantitative prediction map of the predicted rain enhancement effect, and a comprehensive evaluation report is issued, which includes analysis of the physical rationality of the process, uncertainty assessment, and comparison with historical cases.

3. The digital twin simulation method for meteorological condition response of acoustic rain enhancement oriented to watersheds as described in claim 1, characterized in that, The integrated land use classification data and watershed hydrological network topology are specifically as follows: integrate land use classification data, clearly distinguish underlying surface types including urban built-up areas, farmland cultivated areas, forest cover areas and lakes and rivers, and construct a hydrological network topology based on water level analysis algorithms; The specific method for constructing the hydrological network topology based on the water level analysis algorithm is as follows: Depressions were filled in the high-precision digital elevation model, and a priority queue algorithm was used to eliminate depressions in the data to ensure the continuity of water flow. The direction of water flow in each grid cell of a high-precision digital elevation model is determined based on the D8 algorithm. The cumulative runoff is calculated based on the flow direction matrix, and an initial river network is generated based on the cumulative runoff. The Strahler classification system is then used to classify the river channels. The hydrological network topology is constructed by representing the river channel connections using an adjacency matrix, and the automatically extracted hydrological network topology is then verified and corrected.

4. The digital twin simulation method for meteorological condition response of acoustic rain enhancement oriented to watersheds as described in claim 1, characterized in that, The continuous reception of multi-source meteorological observation data streams, through data preprocessing and multi-source data fusion, generates high-resolution and physically consistent meteorological initial fields and boundary conditions, specifically as follows: It continuously receives multi-source meteorological observation data streams, including: infrared and visible light channel radiation data from geostationary meteorological satellites, reflectivity factor and radial wind data detected by weather radar, near-surface temperature, humidity, air pressure and wind field data collected by automatic weather stations, and temperature, humidity and wind vertical profile data acquired by radiosondes. Cloud detection and calibration are performed on infrared and visible light channel radiation data from geostationary meteorological satellites; ground clutter suppression and velocity de-blurring are performed on reflectivity factor and radial wind data detected by weather radar; extreme value removal and time consistency checks are performed on near-surface temperature, humidity, air pressure and wind field data collected by automatic weather stations; and system error correction and vertical interpolation are performed on temperature, humidity and wind vertical profile data acquired by radiosondes. Based on the above data preprocessing, a three-dimensional variational assimilation algorithm is used to fuse multi-source data to obtain high-resolution and physically consistent meteorological initial fields and boundary conditions.

5. The digital twin simulation method for meteorological condition response of acoustic rain enhancement oriented to watersheds as described in claim 1, characterized in that, The non-hydrostatic mesoscale meteorological model adopts the ARPS model, which includes the momentum equation in the x-direction, the continuity equation, the thermodynamic energy equation, the moist air state equation, the water conservation equation, the cloud microphysical transformation equation, and the vertical motion equation. The dual-parameter cloud microphysics parameterization scheme predicts ten types of forecast variables, including cloud water mass mixing ratio and number concentration, rainwater mass mixing ratio and number concentration, cloud ice mass mixing ratio and number concentration, snow mass mixing ratio and number concentration, and graupel mass mixing ratio and number concentration, using the mass conservation equation and number concentration equation. The dual-parameter cloud microphysics parameterization scheme transforms microphysical mechanisms into computable source and sink terms, which are then fed back into a non-hydrostatic mesoscale meteorological model to simulate the entire life cycle evolution of cloud systems over the target watershed, including the generation, development, maturation, and dissipation. The output includes updated meteorological fields and four-dimensional spatiotemporal distribution data of various cloud water content.

6. The digital twin simulation method for meteorological condition response of acoustic rain enhancement oriented to watersheds as described in claim 1, characterized in that, The sound wave propagation equations include the basic sound wave equation, the ray tracing form equation, the sound intensity attenuation equation, and the correction term equation considering turbulent scattering. Based on the microphysical model, the improvement in cloud droplet collision efficiency under the action of the sound field is first calculated by using the output of the sound wave propagation equation. Then, the promoting effect of the sound field on the freezing process of supercooled cloud droplets is simulated to dynamically correct the rate coefficients of relevant processes in the two-parameter cloud microphysical parameterization scheme.

7. The digital twin simulation method for meteorological condition response of acoustic rain enhancement oriented to watersheds as described in claim 1, characterized in that, The global optimization algorithm adopts a genetic algorithm or a particle swarm algorithm. The input individual of the global optimization algorithm is a combination of acoustic parameters including four decision variables: frequency, acoustic intensity, emission height, and beam angle. The deep neural network proxy model adopts an improved ConvLSTM-Attention network model, which includes an input layer, a 3D convolutional layer, a ConvLSTM layer, a spatiotemporal attention layer, a connection layer, and an output layer. The preset rain enhancement objective function is to maximize the ratio of the increase in surface precipitation within the target watershed to the operating cost; For each individual in each iteration, a deep neural network surrogate model is used to update the meteorological field and acoustic parameter combination as input features, and outputs the predicted cloud microphysical process evolution sequence and the final ground precipitation. Based on the predicted cloud microphysical process evolution sequence and final ground precipitation output by the deep neural network surrogate model, the global optimization algorithm updates and iterates until the preset convergence condition is met, and then outputs the optimal combination of acoustic parameters and its corresponding predicted rainfall enhancement effect.

8. A digital twin simulation platform for meteorological condition response of acoustic rain enhancement for watersheds, characterized in that, The platform is used to implement the method as described in any one of claims 1 to 7, including a watershed three-dimensional geometric model module, a meteorological data input and assimilation module, a multi-scale meteorological and cloud microphysics model module, a sound wave propagation and action mechanism module, and a parameter optimization and effect prediction module; The watershed 3D geometric model module is used to construct a high-precision digital elevation model of the target watershed and integrate land use classification data and watershed hydrological network topology to form a physically realistic 3D spatial lower boundary condition. The meteorological data input and assimilation module is used to continuously receive multi-source meteorological observation data streams, and generate high-resolution and physically consistent meteorological initial fields and boundary conditions through data preprocessing and multi-source data fusion. The multi-scale meteorological and cloud microphysics model module is used to couple a non-hydrostatic mesoscale meteorological model with a two-parameter cloud microphysics parameterization scheme based on the generated meteorological initial field and boundary conditions, and output four-dimensional spatiotemporal distribution data including updated meteorological fields and the content of various cloud water substances. The acoustic wave propagation and action mechanism module is used to solve the acoustic wave propagation equation considering atmospheric refraction, absorption attenuation and turbulent scattering in the updated meteorological field, and based on the microphysical model, dynamically correct the rate coefficients of related processes in the two-parameter cloud microphysical parameterization scheme using the output of the acoustic wave propagation equation. The parameter optimization and effect prediction module is used to iteratively adjust the combination of acoustic parameters according to the preset rain enhancement objective function using a global optimization algorithm and a deep neural network proxy model, and output the optimal combination of acoustic parameters and its corresponding predicted rain enhancement effect.

9. An electronic device, the electronic device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images