Search

CN-121997025-A - Rocket artificial precipitation influence area estimation method and system

CN121997025ACN 121997025 ACN121997025 ACN 121997025ACN-121997025-A

Abstract

The invention discloses a rocket artificial precipitation influence area estimation method and system, comprising the steps of A1, collecting rocket artificial precipitation operation parameter data, regional wind field data and atmospheric water vapor distribution data and respectively preprocessing, A2, respectively extracting structural diffusion characteristics, regional wind field characteristics and water vapor distribution characteristics, A3, obtaining multi-mode fusion characteristics and catalyst diffusion characteristics through cross-mode fusion, A4, extracting catalyst-water vapor interaction characteristics, A5, calculating rocket artificial precipitation operation basic influence area and basic grid range images, A6, calculating rocket artificial precipitation operation accurate influence area, and then carrying out grid statistics and scale conversion to obtain rocket artificial precipitation operation accurate influence area. The method can solve the problem that the traditional method simplifies the line source diffusion into a trapezoidal model and has insufficient estimation accuracy on the artificial precipitation influence area of the rocket caused by insufficient multi-factor consideration.

Inventors

  • LI DANXIANG
  • LI QIULIN

Assignees

  • 湖南逸尔科技服务有限公司

Dates

Publication Date
20260508
Application Date
20260128

Claims (9)

  1. 1. The method for estimating the artificial precipitation influence area of the rocket is characterized by comprising the following steps of: A1, acquiring rocket rain-increasing operation parameter data, regional wind field data and atmospheric water vapor distribution data, and respectively preprocessing to obtain preprocessed rocket rain-increasing operation parameter data, preprocessed regional wind field data and preprocessed atmospheric water vapor distribution data; a2, respectively extracting structural diffusion characteristics, regional wind field characteristics and water vapor distribution characteristics according to the preprocessed rocket rain-increasing operation parameter data, preprocessed regional wind field data and preprocessed atmospheric water vapor distribution data; a3, performing cross-modal fusion according to the structural diffusion characteristics, the regional wind field characteristics and the water vapor distribution characteristics to obtain multi-modal fusion characteristics and catalyst diffusion characteristics; A4, extracting catalyst-water vapor interaction characteristics according to catalyst diffusion characteristics, water vapor distribution characteristics, structural diffusion characteristics and regional wind field characteristics; a5, calculating a basic influence area and a basic grid range image of the artificial precipitation operation of the rocket according to the preprocessed rocket precipitation operation parameter data and the preprocessed regional wind field data; A6, fusing the basic influence area, the catalyst diffusion characteristic and the catalyst-water vapor interaction characteristic of the rocket artificial precipitation operation to obtain an accurate grid range image of the rocket artificial precipitation operation, and then carrying out grid statistics and scale conversion to obtain the accurate influence area of the rocket artificial precipitation operation.
  2. 2. The method for estimating the area of influence of artificial precipitation of rocket according to claim 1, wherein said A1 step comprises: A11, acquiring rocket rainfall-enhancing operation parameter data, wherein the data type is structured numerical data, and the data comprises an emission azimuth angle, rocket projectile scattering amount, projectile type identification and catalyst scattering distance; preprocessing rocket rainfall augmentation operation parameter data by adopting a method of linear interpolation complement of missing values, three-time standard difference constant value rejection and linear normalization in sequence to obtain preprocessed rocket rainfall augmentation operation parameter data; A12, collecting regional wind field data, wherein the data types are space-time sequence data, including near-ground wind speed, wind direction and turbulence intensity, and preprocessing the regional wind field data sequentially by adopting methods of Gaussian filtering denoising, linear normalization and time axis alignment to obtain preprocessed regional wind field data; A13, acquiring atmospheric vapor distribution data, wherein the data type is space raster data comprising the atmospheric vapor content and the relative humidity, and preprocessing the atmospheric vapor distribution data by adopting a grid resolution linear interpolation unification, linear normalization and time axis alignment preprocessing method to obtain preprocessed atmospheric vapor distribution data.
  3. 3. The method for estimating the area of influence of artificial precipitation of rocket according to claim 2, wherein the step A2 comprises: a21, extracting structural diffusion characteristics according to the preprocessed rocket rainfall augmentation operation parameter data; A22, extracting regional wind field characteristics according to the preprocessed regional wind field data; a23, extracting water vapor distribution characteristics according to the pretreated atmospheric water vapor distribution data.
  4. 4. A rocket artificial precipitation influence area estimation method according to claim 3, wherein said A2 step comprises: The extraction process of the structured diffusion feature comprises the steps of firstly processing the preprocessed rocket precipitation operation parameter data through a multilayer perceptron, sequentially processing the preprocessed rocket precipitation operation parameter data through a multilayer perceptron and a Sigmoid function, carrying out Hadamard product on the two processing results to obtain a structured diffusion intermediate feature, calculating the absolute value of the difference value between the launching azimuth angle in the preprocessed rocket precipitation operation parameter data and the wind direction angle during launching, carrying out cosine calculation on the absolute value, respectively calculating the ratio of the rocket projectile scattering dose in the preprocessed rocket precipitation operation parameter data to the standard scattering dose corresponding to a projectile type identifier, the ratio of the catalyst scattering distance to the standard scattering distance corresponding to the projectile type identifier, carrying out dimension splicing on the cosine calculation result and the two ratios to obtain an operation parameter field constraint vector, and carrying out Hadamard product on the structured diffusion intermediate feature and the operation parameter field constraint vector to obtain a structured association feature; The extraction process of the regional wind field characteristics comprises the steps of firstly processing the preprocessed regional wind field data through a multi-layer perceptron, dividing the processed regional wind field data by the square root of an attention dimension scaling factor, and then processing the processed regional wind field data through a Softmax function to obtain a wind speed time sequence node mask; The extraction process of the water vapor distribution characteristics comprises the steps of firstly processing the preprocessed atmospheric water vapor distribution data through a multi-scale convolution layer to obtain multi-scale grid characteristics, respectively carrying out average pooling and maximum pooling on the multi-scale grid characteristics, adding two pooling results, processing through a multi-layer perceptron, carrying out Hadamard product on the multi-scale grid characteristics to obtain space attention weighted characteristics, respectively carrying out average pooling and maximum pooling on the space attention weighted characteristics, and summing the two pooling results element by element to obtain the water vapor distribution characteristics.
  5. 5. The method for estimating a rocket artificial precipitation influence area according to claim 4, wherein said A4 step comprises: A41, according to the catalyst diffusion characteristics and the water vapor distribution characteristics, performing bidirectional modulation operation through a self-adaptive space-time modulation network, and extracting diffusion-water vapor modulation characteristics; a42, extracting diffusion-wind field modulation characteristics through a double-gating cross-attention mechanism according to the catalyst diffusion characteristics, the structured diffusion characteristics and the regional wind field characteristics; A43, carrying out residual enhancement and dynamic calibration operation on the diffusion-wind field modulation characteristic and the diffusion-water vapor modulation characteristic to obtain the catalyst-water vapor interaction characteristic.
  6. 6. The method for estimating a rocket artificial precipitation influence area according to claim 5, wherein said A4 step comprises: The self-adaptive space-time modulation network takes the catalyst diffusion characteristic and the water vapor distribution characteristic as input, obtains the diffusion attention characteristic through an attention mechanism, carries out Hadamard product operation on the diffusion attention characteristic after being processed by a multi-layer perceptron and activated by Sigmoid, carries out layer normalization, and then sums the diffusion attention characteristic with the catalyst diffusion characteristic after convolution treatment element by element to obtain the diffusion-water vapor modulation characteristic; The double-gating cross-attention mechanism firstly processes regional wind field characteristics and catalyst diffusion characteristics through a multi-layer perceptron and activates Sigmoid to obtain wind field gating weight and diffusion gating weight, then carries out Hadamard product operation on the wind field characteristics and the catalyst diffusion characteristics and attention association of the regional wind field characteristics respectively, sums the two operation results element by element, and carries out Hadamard product operation on the wind field characteristics and the catalyst diffusion characteristics after layer normalization after being processed through a Transformer network to obtain diffusion-wind field modulation characteristics.
  7. 7. The method for estimating the area of influence of artificial precipitation of rocket according to claim 6, wherein said A5 step comprises: A51, calculating the basic influence area of the artificial precipitation operation of the rocket under a traditional trapezoid model by adopting a line source diffusion method according to the preprocessed rocket precipitation operation parameter data and preprocessed regional wind field data, wherein in the traditional trapezoid model, the upper bottom of the trapezoid is a catalyst spreading distance, the lower bottom of the trapezoid is a turbulence diffusion width, the height is a advection diffusion length, and the extension direction is consistent with the wind direction; A52, setting the grid resolution by taking a rocket launching point as a grid origin; a53, dividing the advection diffusion length by the grid resolution along the trapezoid extending direction, and then rounding downwards to obtain the number of grids in the advection direction; A54, dividing the width of turbulent diffusion by 2 along the width direction of the trapezoid, dividing the width by the resolution of grids, and rounding downwards to obtain the number of grids at each direction on two sides; A55, according to the number of grids in the advection direction and the number of grids at each side, referring to the trapezoid boundary of the traditional trapezoid model, and performing grid arrangement to obtain a basic grid range image.
  8. 8. The method for estimating a rocket artificial precipitation influence area according to claim 7, wherein said A6 step comprises: a61, carrying out feature fusion according to the basic influence area, the catalyst diffusion feature and the catalyst-water vapor interaction feature of the rocket artificial precipitation operation to obtain an accurate grid range image of the rocket artificial precipitation operation; A62, carrying out grid statistics and scale conversion according to the accurate grid range image of the rocket artificial precipitation operation to obtain the accurate influence area of the rocket artificial precipitation operation.
  9. 9. A rocket artificial precipitation impact area estimation system, comprising: The rocket operation data acquisition module acquires rocket rain-increasing operation parameter data, regional wind field data and atmospheric water vapor distribution data and respectively performs pretreatment to obtain pretreated rocket rain-increasing operation parameter data, pretreated regional wind field data and pretreated atmospheric water vapor distribution data; The basic feature extraction module is used for respectively extracting structural diffusion features, regional wind field features and water vapor distribution features according to the preprocessed rocket rain-increasing operation parameter data, preprocessed regional wind field data and preprocessed atmospheric water vapor distribution data; The cross-modal feature fusion module is used for carrying out cross-modal fusion according to the structural diffusion feature, the regional wind field feature and the water vapor distribution feature to obtain a multi-modal fusion feature and a catalyst diffusion feature; The catalyst-water vapor interaction module extracts catalyst-water vapor interaction characteristics according to catalyst diffusion characteristics, water vapor distribution characteristics, structural diffusion characteristics and regional wind field characteristics; The basic influence area calculation module is used for calculating basic influence area and basic grid range images of the artificial precipitation operation of the rocket according to the preprocessed rocket precipitation operation parameter data and the preprocessed regional wind field data; The accurate influence area calculation module is used for fusing basic influence area, catalyst diffusion characteristics and catalyst-water vapor interaction characteristics of the rocket artificial precipitation operation to obtain accurate influence area of the rocket artificial precipitation operation, and carrying out grid statistics and scale conversion to obtain accurate influence area of the rocket artificial precipitation operation so as to realize the rocket artificial precipitation influence area estimation method according to any one of claims 1-8.

Description

Rocket artificial precipitation influence area estimation method and system Technical Field The invention relates to the technical field of artificial precipitation, in particular to a method and a system for estimating the influence area of artificial precipitation of a rocket. Background The core flow of the technology is to collect basic information such as rocket operation parameters, regional wind field data, atmospheric water vapor distribution data and the like, simulate a catalyst diffusion range by means of a diffusion model, further determine the influence area of the precipitation operation, and provide data basis for subsequent precipitation operation layout, resource allocation and effect recombination. The artificial precipitation enhancement operation scene of the rocket has remarkable specificity and complexity, and is characterized in that firstly, influence factors are tightly interwoven, the precipitation enhancement effect is simultaneously limited by three core factors of rocket operation parameters, regional wind fields and atmospheric water vapor, wherein the rocket operation parameters are required to be matched with the inherent characteristics of the shell, and are also required to be matched with the wind direction during emission, otherwise, the initial diffusion direction of the catalyst is directly influenced, the regional wind fields are always in dynamic evolution, the advection diffusion track of the catalyst is changed by the fine changes of wind speed and turbulence intensity, the characteristics are key factors in the dominant diffusion process, the atmospheric water vapor is in multi-scale space distribution characteristics, the water vapor concentration difference between the center and the edge of the cloud cluster is obvious, and whether the catalyst can be effectively activated to form precipitation is directly influenced. Secondly, the diffusion process is complex and special, the rocket catalyst is scattered by a line source, the coupling effect of advection and turbulence exists in actual diffusion, dynamic changes such as water vapor adsorption and condensation can be accompanied, and the actual diffusion rule is difficult to accurately re-etch through a simple mathematical model. In the field, the existing research focuses on the quantification of the influence area of an airplane rain-increasing scene, systematic research on the influence range of rocket operation is insufficient, the existing estimation model mostly simplifies the line source diffusion process of rocket operation into a trapezoidal model to carry out linear estimation, the actual rocket-broadcasted catalyst is advection-turbulence coupling diffusion in the atmosphere environment and is accompanied with the dynamic change of water vapor adsorption, the linear simplification model is difficult to describe the actual diffusion rule, and the interaction among meteorological elements, operation parameters, water vapor conditions and other factors is ignored, so that the deviation between the estimation result and the actual influence range is larger, and the accurate rain-increasing operation requirement is difficult to meet. At present, although some technologies attempt to correct a traditional artificial precipitation estimation model by adopting a deep learning model to improve the precision, the existing deep learning application scheme is designed aiming at the non-point source or point source diffusion characteristics of airplane precipitation, does not adapt the diffusion characteristics of rocket line source operation in a targeted manner, and is difficult to make up the linear simplification defect of the traditional trapezoidal model, and the high-precision quantification of the influence area of rocket artificial precipitation cannot be realized. Disclosure of Invention In view of the above, the present invention aims to provide a method and a system for estimating the artificial precipitation influence area of a rocket, so as to solve the problem of insufficient accuracy in estimating the artificial precipitation influence area of a rocket caused by insufficient consideration of multiple factors by simplifying the line source diffusion into a trapezoidal model in the conventional method. A rocket artificial precipitation influence area estimation method comprises the following steps: A1, acquiring rocket rain-increasing operation parameter data, regional wind field data and atmospheric water vapor distribution data, and respectively preprocessing to obtain preprocessed rocket rain-increasing operation parameter data, preprocessed regional wind field data and preprocessed atmospheric water vapor distribution data; a2, respectively extracting structural diffusion characteristics, regional wind field characteristics and water vapor distribution characteristics according to the preprocessed rocket rain-increasing operation parameter data, preprocessed regional wind field data and preproces