A method for identifying the return location of artificial cloud seeding by aircraft based on wind field data

Abstract

This invention discloses a method for identifying the return-through location of aircraft-based artificial cloud seeding based on wind field data. The method includes determining wind field data based on the operational information of the aircraft to be evaluated; matching wind field data from adjacent time points to the flight trajectory using three-dimensional linear spatiotemporal interpolation based on latitude and longitude to obtain the wind speed and direction of the waypoints; extracting the seeding segment of the flight trajectory during the catalysis time and generating a waypoint mass mask; performing advection integration on the waypoint mass mask to obtain the latitude and longitude of the catalysis point after wind field transmission and the expected catalysis time period; matching the latitude and longitude of the catalysis point with the latitude and longitude of the return-through trajectory after the expected catalysis time period using a location identification algorithm; and verifying the return-through point based on cloud particle data using the actual catalysis point after matching to obtain the latitude, longitude, and time information of the effective return-through point. This method utilizes wind field data matched during flight catalysis operations to identify return-through detection points, improving the objectivity and efficiency of return-through location selection while also possessing good interpretability.

Description

Technical Field

[0001] This invention relates to the field of evaluating the effectiveness of artificial rainmaking aircraft catalytic operations, and in particular to a method for identifying the return location of aircraft artificial cloud seeding based on wind field data. Background Technology

[0002] Based on the principles of cloud seeding and weather modification, the effect of cloud seeding refers to the changes in the microphysical processes and precipitation processes of the target cloud after being influenced by artificial catalysis. These changes include two aspects: one is the alteration of the macroscopic and microscopic physical parameters of the cloud under the influence of catalysis, known as the direct effect of cloud seeding; the other is the change in ground precipitation caused by catalysis, known as the indirect effect of cloud seeding, which is also the main purpose of weather modification. Utilizing airborne detection equipment for in-situ cloud system detection is the most reliable way to obtain cloud particle distribution characteristics and is widely used in weather modification work. Aircraft in-situ physical verification uses high-precision meteorological observation data collected by cloud seeding aircraft, including probes measuring the microphysical characteristics of aerosols, clouds, and precipitation, to measure the cloud particle distribution characteristics before and after the seeding of the catalyst. Aircraft in-situ physical verification not only provides direct physical evidence of the rain enhancement effect but also allows for comparison and verification with other observation results of cloud seeding physical responses, thereby validating the accuracy and reliability of these observation results.

[0003] Traditional assessments of artificial rainmaking effectiveness rely on ground-based precipitation observation data. This data has limited spatial and temporal resolution, making it difficult to fully reflect subtle changes in clouds caused by operations. Furthermore, in-situ aircraft detection requires specific flight conditions, including specific altitudes and speeds. The accuracy, stability, and reliability of the detection equipment are also crucial factors for successful in-situ detection, and these conditions may be difficult to achieve or maintain. These constraints can lead to biases in the overall assessment of aircraft structure or performance. Additionally, the analysis process relies on subjectively selected pre- and post-operation times, making the accuracy and objectivity of the assessment results susceptible to human intervention. Research on methods for identifying the aircraft's return flight position after cloud seeding can objectively and quickly select the aircraft's position before and after cloud seeding, and rapidly and accurately determine whether the aircraft has conducted an effective return flight. This lays a foundation for further in-situ physical verification based on airborne detection. Due to the lack of scientific and effective methods and means, return flight position identification has long been one of the technical challenges hindering the development of cloud seeding operations. Summary of the Invention

[0004] The purpose of this invention is to provide a method for identifying the location of aircraft artificial cloud seeding back through wind field data.

[0005] To achieve the above objectives, the present invention is implemented according to the following technical solution:

[0006] This invention provides a method for identifying the location of aircraft-based artificial cloud seeding backtracking based on wind field data, comprising:

[0007] A acquires and preprocesses the operational information of the aircraft to be evaluated, including flight time, flight trajectory, catalytic time, and cloud particle data.

[0008] B determines wind field data based on the flight time and flight trajectory, and obtains wind speed and wind direction of the flight trajectory point by matching the wind field data of the nearest time point with the flight trajectory through three-dimensional linear spatiotemporal interpolation based on latitude and longitude.

[0009] C extracts the dispersal segment of the flight trajectory during the catalytic time, generates a track point mass mask based on the dispersal segment, performs advection integration on the track point mass mask, obtains the latitude and longitude of the catalytic point after wind field transmission and the expected catalytic time period, and matches the latitude and longitude of the catalytic point with the latitude and longitude of the return trajectory after the expected catalytic time period using a position recognition algorithm.

[0010] Based on the actual catalytic point after matching, D verifies the back-penetration point according to cloud particle data to obtain the latitude, longitude, and time information of the effective back-penetration point.

[0011] Furthermore, the method for obtaining the wind speed and direction of the said track point includes:

[0012] The time range of the wind field is set based on the flight time. The rectangular area extending outward from the trajectory boundary according to the extreme values ​​of latitude and longitude of the flight trajectory is taken as the spatial range of the wind field. Wind field data is obtained based on the time range and spatial range. The wind field data is selected from meteorological reanalysis data, airborne Doppler radar measured wind field data, ground radiosonde station observed wind profile data or wind field output from mesoscale numerical models. The wind field data includes the longitude and latitude of the wind field grid points, as well as the wind field information of the corresponding altitude layer.

[0013] Based on wind field data, the two nearest wind field time data points before and after each waypoint in the flight trajectory are extracted, and the initial wind field value at the corresponding time of the waypoint is calculated by linear interpolation. The formula for calculating the initial wind field value is as follows:

[0014]

[0015] in The wind field vector includes wind speed and direction, obtained from wind field information. t1 is the most recent time before the track point, t2 is the most recent time after the track point, and t... p Let t1 be the current time, and t1 ≤ t p ≤t2;

[0016] Based on the track point, the eight nearest grid points around the current wind field 3D grid are selected. The precise wind field vector of the track point is calculated using 3D linear interpolation based on the grid points and the initial wind field values. The formula for calculating the precise wind field vector is as follows:

[0017]

[0018] in Let w be the wind field vector of the l-th surrounding grid point. i The spatial interpolation weight for the l-th surrounding grid point is calculated based on the inverse ratio of the Euclidean distance between the track point and the grid point, and...

[0019] If the single-point wind speed error of the precise wind field vector in the track point exceeds 2 m / s or the single-point wind direction error exceeds 15°, then double the number of grid points around the track point are reselected for interpolation optimization to obtain the wind speed and wind direction of the track point. The corrected trajectory is then calculated based on the wind speed and wind direction of the track point.

[0020] Furthermore, the method for obtaining the corrected trajectory includes:

[0021] The average wind field vector of the region is calculated based on the wind speed and direction of the waypoints. A corrected trajectory is then calculated using the average wind field vector based on the flight path. The formula for calculating the corrected trajectory is as follows:

[0022]

[0023] in Here, N represents the region-averaged wind field vector, N is the number of wind field times selected during the flight time, and M is the number of grid points for the wind field within the area covered by the flight trajectory. Let t be the wind field vector at the d-th time and the k-th spatial grid point. d Let λ be the time of the d-th wind field occurrence. k Let be the longitude of the k-th grid point. Let k be the latitude of the k-th grid point. The corrected trajectory is the target time t. For the flight trajectory vector, Let be the definite integral from the flight start time t0 to the target time t. Let be the instantaneous wind field vector at time s.

[0024] Furthermore, the method for obtaining the trackpoint quality mask includes:

[0025] The time interval for the seeding segment is determined based on the catalytic time, which is from 30 seconds before the earliest catalytic time to 30 seconds after the latest catalytic time. Track points within the time interval are extracted based on the corrected trajectory to obtain the seeding segment trajectory. If the GPS positioning error of the track points in the seeding segment trajectory is less than or equal to 50 meters, the wind field data and cloud particle data corresponding to the track points are screened to see if they are complete. If they are complete, the wind field change rate between the track point and the five adjacent track points is screened to see if it is less than or equal to 15%. If the wind field change rate is less than or equal to 15%, the track point is marked as a high-quality track point, and the rest of the track points are marked as low-quality track points.

[0026] Based on the proportion of high-quality waypoints in the total number of waypoints in the broadcast segment, if the proportion is greater than or equal to 80%, a waypoint quality mask is obtained; if the proportion is less than 80%, the GPS positioning error is widened to twice and high-quality waypoints are re-selected.

[0027] Furthermore, the method for obtaining the latitude and longitude of the catalytic point after wind field transmission and the expected catalytic period includes:

[0028] Based on the trackpoint mass mask, the unit transmission wind speed of the trackpoint is converted from meters per second to degrees per second to obtain the east-west and north-south wind speeds in the latitude and longitude components. East and north winds are treated as positive values, and west and south winds as negative values. The latitude and longitude of the return point after wind field transmission are then calculated. The formula for calculating the latitude and longitude of the return point is as follows:

[0029] λ j =λ i +U ew ×(t j -t i )+Δd

[0030]

[0031] Where λ j For t j The longitude of the point where the time trajectory crosses back. For t j The latitude of the point of return of the time trajectory, λ i For t i Longitude of the track point in the time trajectory. For t i Latitude of the track point in the time trajectory, t j -t i ≥30 seconds,U ew U represents the east-west wind speed component at the track point. ew Δd represents the north-south wind speed component of the track point, and Δd is the distance correction between the return point and the nearest track point. Δd is less than 1 kilometer. The K-nearest neighbor algorithm is used to find the nearest distance to the return point in the corrected trajectory.

[0032] The latitude and longitude of the return point are used as the latitude and longitude of the catalytic point after the wind field transmission, and the corresponding time is used as the expected catalytic time. The expected catalytic period is determined to be 30 to 120 seconds after the expected catalytic time based on the vector modulus of the wind field transmission speed.

[0033] Furthermore, the method for matching the latitude and longitude of the catalytic point with the latitude and longitude of the return trajectory after the expected catalytic period using a location recognition algorithm includes:

[0034] Based on the catalytic point latitude and longitude sequence after wind field transmission, the actual backtrack latitude and longitude of the predicted catalytic period are extracted according to the track point mass mask. The latitude and longitude of the catalytic point latitude and longitude sequence and the corresponding latitude and longitude of the backtrack latitude and longitude, wind field vector, and cloud particle number concentration are combined to form a multi-dimensional feature vector. The trajectory similarity is calculated using a location recognition algorithm based on the multi-dimensional feature vector. The formula of the location recognition algorithm is as follows:

[0035]

[0036] Where D DTW-W (·,·) represents the similarity between the two trajectories, indicating the weighted distance. A is the latitude and longitude sequence of the catalyst point, and B is the latitude and longitude of the actual backtrack trajectory. Given a time warping function, find the optimal time alignment for the two trajectories. To match the wind field gradient vector at point x, where σ is the scale parameter of the wind field gradient, calibrated based on the mean wind shear of the wind field data. For multi-dimensional feature vectors, Trajectory A is time-normalized by φ A The point that is paired with point x. Trajectory B is time-normalized by φ B The point that is paired with point x;

[0037] If the trajectory similarity is less than 1 km of feature vector distance, then the catalyst point and the backtrack trajectory point are successfully matched.

[0038] Furthermore, the method for obtaining the latitude, longitude, and time information of the effective traverse point includes:

[0039] Based on the actual catalytic point and backtrack trajectory point of the matching, the timestamp and spatial latitude and longitude are extracted. Cloud particle detection data within 5 seconds before and after the backtrack point time are screened according to the timestamp, and the particle spectrum distribution is calculated. If the concentration of small cloud droplets with a particle size of less than 20 μm in the cloud particle detection data decreases by more than or equal to 20% compared with before catalysis, then it is verified whether the concentration of large particles with a particle size of more than or equal to 20 μm increases by more than or equal to 30% compared with before catalysis.

[0040] If the increase in large particle concentration is greater than 30% and the peak particle size distribution of the particle spectrum shifts radially to the direction of large particle size by 5 μm or more, then the backtrack trajectory point is taken as the effective backtrack point, and the latitude, longitude and time information of the effective backtrack point are obtained.

[0041] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0042] (1) By identifying the effective return point of the aircraft after cloud seeding operation, the present invention reduces unnecessary human screening of return point, improves the objectivity of cloud seeding operation evaluation, avoids the traditional aircraft observation data analysis which mainly relies on human and subjective selection of time periods before and after the operation, and then conducts qualitative case comparison analysis, and reduces the strong subjectivity of return point selection and identification.

[0043] (2) By accurately identifying the aircraft’s backflip behavior and combining it with real-time analysis of cloud seeding data, this invention effectively improves the monitoring capabilities of cloud layer changes and flight conditions. It can optimize the real-time cloud seeding position of the aircraft and generate the latitude, longitude and time of the backflip point quickly and efficiently, which is convenient for carrying out effect evaluation work. At the same time, based on multiple flight trajectory cases, this analysis method can also carry out quantitative analysis of cloud microphysical characteristics, providing a more comprehensive and systematic scientific basis for aircraft shadowing operations.

[0044] (3) In a single flight catalytic operation, this invention can make multiple and continuous observations of the cloud layer by setting multiple back-penetration detection points, thereby attempting to give quantitative analysis results, which advances the qualitative analysis of in-situ physical testing to quantitative analysis. Based on the large dataset accumulated by the weather modification aircraft, in-depth data analysis can be carried out. Compared with the case analysis of a single event, this big data analysis method can reveal the statistical laws of cloud microphysical characteristics over a long time series, providing a more comprehensive and systematic scientific basis for weather modification. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating the steps of an aircraft-based cloud seeding back-through location identification method based on wind field data, as described in an embodiment of the present invention. Detailed Implementation

[0046] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0047] Reference Figure 1As shown, this invention provides a method for identifying the location of aircraft-based artificial cloud seeding backtracking based on wind field data, including:

[0048] A acquires and preprocesses the operational information of the aircraft to be evaluated, including flight time, flight trajectory, catalytic time, and cloud particle data.

[0049] In the actual evaluation, the operational information of an artificial rainmaking aircraft was obtained and anomalies were removed by Kalman filtering. The flight time was from 15:20 to 18:40, with a total duration of 200 minutes, covering the entire period from aircraft takeoff, catalysis operation, and return. The flight trajectory was recorded by the airborne GPS system. The catalysis time was from 16:40 to 18:17, during which 24 silver iodide flames were continuously burned to provide artificial ice nuclei for the clouds. Cloud particle data was collected by the airborne cloud particle probe, with a particle size detection range of 2 micrometers to 6300 micrometers.

[0050] B determines wind field data based on the flight time and flight trajectory, and obtains wind speed and wind direction of the flight trajectory point by matching the wind field data of the nearest time point with the flight trajectory through three-dimensional linear spatiotemporal interpolation based on latitude and longitude.

[0051] In the actual assessment, based on the flight time of 15:20–18:40 as the core, extending outward by 1 hour, the wind field time range was determined to be 14:20–19:40. Using the extreme latitude and longitude values ​​of the flight trajectory as a basis, extending outward by 0.5° in each direction, the catalytic zone spatial range is between 89.8E-90.2E and 29.3N-29.7N, thus obtaining the wind field spatial range. CRA40 meteorological reanalysis data was selected, with a horizontal resolution of 0.25°×0.25° and a time resolution of 1 hour, including multiple altitude layers in the vertical direction. Specifically, the nearest wind field times before and after the 16:50 trackpoint were extracted, including times around 16:30. At 17:00, the initial wind field values ​​were calculated using a linear interpolation formula. At 16:30, the wind field vector was -3.2 m / s east-west and -1.5 m / s north-south. At 17:00, it was 4.1 m / s east-west and -1.2 m / s north-south, resulting in an initial wind field of (-3.8 m / s, -1.3 m / s). A track point with latitude and longitude of (89.95E, 29.48N) was selected, and the accurate wind field was calculated using a three-dimensional linear interpolation formula based on the eight nearest grid points. The wind speed at this point was -4.78 m / s east-west and -1.04 m / s north-south, which translates to a wind direction of 257.7° and a wind speed of 4.89 m / s.

[0052] C extracts the dispersal segment of the flight trajectory during the catalytic time, generates a track point mass mask based on the dispersal segment, performs advection integration on the track point mass mask, obtains the latitude and longitude of the catalytic point after wind field transmission and the expected catalytic time period, and matches the latitude and longitude of the catalytic point with the latitude and longitude of the return trajectory after the expected catalytic time period using a position recognition algorithm.

[0053] In the actual assessment, six wind field times (15:00, 16:00, 17:00, 18:00, 19:00, and 20:00) were selected within the flight time range. Four wind field grid points were selected within the flight trajectory coverage area. The regional average wind field vector was calculated as (3.0, -1.5) m / s by substituting the wind field vectors of each time point and grid point into the formula for the regional average wind field vector. The original flight trajectory vector at the target time of 16:50:00 corresponds to an eastward distance of 1200m and a northward distance of 2500m in the plane coordinate system. The integral term from the flight start time of 15:20:00 to the target time was calculated, and the integral result was eastward +1. 00m, northward +50m. Based on the corrected trajectory formula, the final corrected trajectory corresponds to approximately 1300m eastward and 2550m northward. The time interval for the seeding segment is determined to be from the earliest catalytic time 16:39:30 to the latest catalytic time 18:17:30. All track points within this time interval are extracted from the corrected trajectory. For each track point in the seeding segment trajectory, high-quality track points are sequentially verified and marked with "1", totaling approximately 986. High-quality track points are selected from the track point quality mask, and the wind speed is converted from m / s to degrees / s latitude and longitude components. These are then substituted into the latitude and longitude formula for the return point, where a λ is selected every 30 seconds. i , The latitude and longitude of the catalytic point are calculated, t j -t i For a time interval greater than or equal to 60 seconds, the search yields a Δd of 0.4 kilometers. The latitude and longitude of the first crossing point are calculated to be (89.93E, 29.49N). This latitude and longitude is used as the catalytic point after wind field transmission. The vector modulus of the wind field transmission speed is 3.2 m / s. The expected catalytic period is determined to be 30 to 120 seconds after the catalytic moment.

[0054] In the actual assessment, the sequence of catalytic points after wind field transmission was extracted. Based on the trackpoint mass mask, the latitude and longitude of the actual backtrack trajectory before and after the expected catalytic period were extracted. The latitude and longitude, wind field vector, and cloud particle number concentration were combined to form a multi-dimensional feature vector, which was then substituted into the location recognition algorithm formula. The wind field gradient in this area was calculated to be approximately 1 m / s / km using CRA40 wind field data. The calibration scale parameter was set to 1 based on the average wind shear of the wind field data. In the catalytic point A feature of the catalytic point sequence, the longitude was 89.9E, the latitude was 29.45N, the wind speed was -4.78 m / s, and the large particle concentration was 100 particles / cm³. 3 The characteristics of the return point B are: longitude 89.9°E, latitude 29.55°N, wind speed -4.8 m / s, and large particle concentration 130 particles / cm³. 3 The trajectory similarity between a catalyst point and a backtracking trajectory point was calculated to be 120.7m. Since 0.12 km is less than the eigenvector distance of 1 km, it was determined that the catalyst point and the backtracking trajectory point were successfully matched.

[0055] Based on the actual catalytic point after matching, D verifies the back-penetration point according to cloud particle data to obtain the latitude, longitude, and time information of the effective back-penetration point.

[0056] In the actual evaluation, for the crossover point with a successful match time of 16:51:30 and latitude and longitude of (89.93E, 29.49N), cloud particle detection data within 5 seconds before and after the time were screened. The cloud droplet number concentration at 16:50:00 before catalysis was 120 droplets / cm³. 3 The large particle number concentration is 30 particles / cm³. 3 The peak particle size of the particle spectrum is approximately 15 μm, and the droplet number concentration at the time of back-through at 16:51:30 is 90 droplets / cm³. 3 The large particle number concentration is 45 particles / cm³. 3 The peak particle size of the particle spectrum is about 22 μm, so this back-through point is an effective back-through point. A total of 27 catalytic points and effective back-through points are matched.

[0057] In this embodiment, the method for obtaining the wind speed and wind direction of the track point includes:

[0058] The time range of the wind field is set based on the flight time. The rectangular area extending outward from the trajectory boundary according to the extreme values ​​of latitude and longitude of the flight trajectory is taken as the spatial range of the wind field. Wind field data is obtained based on the time range and spatial range. The wind field data is selected from meteorological reanalysis data, airborne Doppler radar measured wind field data, ground radiosonde station observed wind profile data or wind field output from mesoscale numerical models. The wind field data includes the longitude and latitude of the wind field grid points, as well as the wind field information of the corresponding altitude layer.

[0059] Based on wind field data, the two nearest wind field time data points before and after each waypoint in the flight trajectory are extracted, and the initial wind field value at the corresponding time of the waypoint is calculated by linear interpolation. The formula for calculating the initial wind field value is as follows:

[0060]

[0061] in The wind field vector includes wind speed and direction, obtained from wind field information. t1 is the most recent time before the track point, t2 is the most recent time after the track point, and t... p Let t1 be the current time, and t1 ≤ t p ≤t2;

[0062] Based on the track point, the eight nearest grid points around the current wind field 3D grid are selected. The precise wind field vector of the track point is calculated using 3D linear interpolation based on the grid points and the initial wind field values. The formula for calculating the precise wind field vector is as follows:

[0063]

[0064] in Let w be the wind field vector of the l-th surrounding grid point. i The spatial interpolation weight for the l-th surrounding grid point is calculated based on the inverse ratio of the Euclidean distance between the track point and the grid point, and...

[0065] If the single-point wind speed error of the precise wind field vector in the track point exceeds 2 m / s or the single-point wind direction error exceeds 15°, then double the number of grid points around the track point are reselected for interpolation optimization to obtain the wind speed and wind direction of the track point. The corrected trajectory is then calculated based on the wind speed and wind direction of the track point.

[0066] In this embodiment, the method for obtaining the corrected trajectory includes:

[0067] The average wind field vector of the region is calculated based on the wind speed and direction of the waypoints. A corrected trajectory is then calculated using the average wind field vector based on the flight path. The formula for calculating the corrected trajectory is as follows:

[0068]

[0069] in Here, N represents the region-averaged wind field vector, N is the number of wind field times selected during the flight time, and M is the number of grid points for the wind field within the area covered by the flight trajectory. Let t be the wind field vector at the d-th time and the k-th spatial grid point. d Let λ be the time of the d-th wind field occurrence. k Let be the longitude of the k-th grid point. Let k be the latitude of the k-th grid point. The corrected trajectory is the target time t. For the flight trajectory vector, Let be the definite integral from the flight start time t0 to the target time t. Let be the instantaneous wind field vector at time s.

[0070] In this embodiment, the method for obtaining the trackpoint quality mask includes:

[0071] The time interval for the seeding segment is determined based on the catalytic time, which is from 30 seconds before the earliest catalytic time to 30 seconds after the latest catalytic time. Track points within the time interval are extracted based on the corrected trajectory to obtain the seeding segment trajectory. If the GPS positioning error of the track points in the seeding segment trajectory is less than or equal to 50 meters, the wind field data and cloud particle data corresponding to the track points are screened to see if they are complete. If they are complete, the wind field change rate between the track point and the five adjacent track points is screened to see if it is less than or equal to 15%. If the wind field change rate is less than or equal to 15%, the track point is marked as a high-quality track point, and the rest of the track points are marked as low-quality track points.

[0072] Based on the proportion of high-quality waypoints in the total number of waypoints in the broadcast segment, if the proportion is greater than or equal to 80%, a waypoint quality mask is obtained; if the proportion is less than 80%, the GPS positioning error is widened to twice and high-quality waypoints are re-selected.

[0073] In this embodiment, the method for obtaining the latitude and longitude of the catalytic point after wind field transmission and the expected catalytic period includes:

[0074] Based on the trackpoint mass mask, the unit transmission wind speed of the trackpoint is converted from meters per second to degrees per second to obtain the east-west and north-south wind speeds in the latitude and longitude components. East and north winds are treated as positive values, and west and south winds as negative values. The latitude and longitude of the return point after wind field transmission are then calculated. The formula for calculating the latitude and longitude of the return point is as follows:

[0075] λ j =λ i +U ew ×(t j -t i )+Δd

[0076]

[0077] Where λ j For t j The longitude of the point where the time trajectory crosses back. For t j The latitude of the point of return of the time trajectory, λ i For t i Longitude of the track point in the time trajectory. For t i Latitude of the track point in the time trajectory, t j -t i ≥30 seconds,U ew U represents the east-west wind speed component at the track point. ew Δd represents the north-south wind speed component of the track point, and Δd is the distance correction between the return point and the nearest track point. Δd is less than 1 kilometer. The K-nearest neighbor algorithm is used to find the nearest distance to the return point in the corrected trajectory.

[0078] The latitude and longitude of the return point are used as the latitude and longitude of the catalytic point after the wind field transmission, and the corresponding time is used as the expected catalytic time. The expected catalytic period is determined to be 30 to 120 seconds after the expected catalytic time based on the vector modulus of the wind field transmission speed.

[0079] In this embodiment, the method for matching the latitude and longitude of the catalytic point with the latitude and longitude of the return trajectory after the expected catalytic period using a location recognition algorithm includes:

[0080] Based on the catalytic point latitude and longitude sequence after wind field transmission, the actual backtrack latitude and longitude of the predicted catalytic period are extracted according to the track point mass mask. The latitude and longitude of the catalytic point latitude and longitude sequence and the corresponding latitude and longitude of the backtrack latitude and longitude, wind field vector, and cloud particle number concentration are combined to form a multi-dimensional feature vector. The trajectory similarity is calculated using a location recognition algorithm based on the multi-dimensional feature vector. The formula of the location recognition algorithm is as follows:

[0081]

[0082] Where D DTW-W (·,·) represents the similarity between the two trajectories, indicating the weighted distance. A is the latitude and longitude sequence of the catalyst point, and B is the latitude and longitude of the actual backtrack trajectory. Given a time warping function, find the optimal time alignment for the two trajectories. To match the wind field gradient vector at point x, where σ is the scale parameter of the wind field gradient, calibrated based on the mean wind shear of the wind field data. For multi-dimensional feature vectors, Trajectory A is time-normalized by φ A The point that is paired with point x. Trajectory B is time-normalized by φ B The point that is paired with point x;

[0083] If the trajectory similarity is less than 1 kilometer of feature vector distance, then the catalyst point and the backtrack trajectory point are successfully matched. In this embodiment, the method for obtaining the latitude, longitude, and time information of the effective backtrack point includes:

[0084] Based on the actual catalytic point and backtrack trajectory point of the matching, the timestamp and spatial latitude and longitude are extracted. Cloud particle detection data within 5 seconds before and after the backtrack point time are screened according to the timestamp, and the particle spectrum distribution is calculated. If the concentration of small cloud droplets with a particle size of less than 20 μm in the cloud particle detection data decreases by more than or equal to 20% compared with before catalysis, then it is verified whether the concentration of large particles with a particle size of more than or equal to 20 μm increases by more than or equal to 30% compared with before catalysis.

[0085] If the increase in large particle concentration is greater than 30% and the peak particle size distribution of the particle spectrum shifts radially to the direction of large particle size by 5 μm or more, then the backtrack trajectory point is taken as the effective backtrack point, and the latitude, longitude and time information of the effective backtrack point are obtained.

[0086] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

Claims

1. A method for identifying the location of aircraft-based artificial cloud seeding based on wind field data, characterized in that, Includes the following steps: A acquires and preprocesses the operational information of the aircraft to be evaluated, including flight time, flight trajectory, catalytic time, and cloud particle data. B determines wind field data based on the flight time and flight trajectory, and obtains wind speed and wind direction of the flight trajectory point by matching the wind field data of the nearest time point with the flight trajectory through three-dimensional linear spatiotemporal interpolation based on latitude and longitude. C extracts the dispersal segment of the flight trajectory during the catalytic time, generates a track point mass mask based on the dispersal segment, performs advection integration on the track point mass mask, obtains the latitude and longitude of the catalytic point after wind field transmission and the expected catalytic time period, and matches the latitude and longitude of the catalytic point with the latitude and longitude of the return trajectory after the expected catalytic time period using a position recognition algorithm. Based on the actual catalytic point after matching, D verifies the back-penetration point according to cloud particle data to obtain the latitude, longitude, and time information of the effective back-penetration point.

2. The method for identifying the location of aircraft-based artificial cloud seeding based on wind field data according to claim 1, characterized in that, Methods for obtaining the wind speed and direction of the said track points include: The time range of the wind field is set based on the flight time. The rectangular area extending outward from the trajectory boundary according to the extreme values ​​of latitude and longitude of the flight trajectory is taken as the spatial range of the wind field. Wind field data is obtained based on the time range and spatial range. The wind field data is selected from meteorological reanalysis data, airborne Doppler radar measured wind field data, ground radiosonde station observed wind profile data or wind field output from mesoscale numerical models. The wind field data includes the longitude and latitude of the wind field grid points, as well as the wind field information of the corresponding altitude layer. Based on wind field data, the two nearest wind field time data points before and after each waypoint in the flight trajectory are extracted, and the initial wind field value at the corresponding time of the waypoint is calculated by linear interpolation. The formula for calculating the initial wind field value is as follows: in The wind field vector, which includes wind speed and direction, is obtained from wind field information. The closest moment before the waypoint. The closest time after the waypoint. For the current moment, and ; Based on the track point, the eight nearest grid points around the current wind field 3D grid are selected. The precise wind field vector of the track point is calculated using 3D linear interpolation based on the grid points and the initial wind field values. The formula for calculating the precise wind field vector is as follows: in For the first Wind field vectors of surrounding grid points For the first The spatial interpolation weights of the surrounding grid points are calculated inversely proportional to the Euclidean distance between the track point and the grid point, and ; If the single-point wind speed error of the precise wind field vector in the track point exceeds 2 m / s or the single-point wind direction error exceeds 15°, then double the number of grid points around the track point are reselected for interpolation optimization to obtain the wind speed and wind direction of the track point. The corrected trajectory is then calculated based on the wind speed and wind direction of the track point.

3. The method for identifying the location of aircraft-based artificial cloud seeding based on wind field data according to claim 2, characterized in that, The method for obtaining the corrected trajectory includes: The average wind field vector of the region is calculated based on the wind speed and direction of the waypoints. A corrected trajectory is then calculated using the average wind field vector based on the flight path. The formula for calculating the corrected trajectory is as follows: in This is the regional average wind field vector. The number of wind field times selected during the flight time. This represents the number of grid points for the wind field within the area covered by the flight trajectory. For the first Each time, the first Wind field vector at each spatial grid point For the first The time of each wind field For the first Longitude of each grid point For the first The latitude of each grid point For the target time The corrected trajectory, For the flight trajectory vector, From the start of flight To the target time definite integral, For a moment The instantaneous wind field vector.

4. The method for identifying the location of aircraft-based artificial cloud seeding based on wind field data according to claim 1, characterized in that, The method for obtaining the trackpoint quality mask includes: Based on the catalytic time, the time interval for the seeding segment is determined to be 30 seconds before the earliest catalytic time to 30 seconds after the latest catalytic time. Track points within the time interval are extracted according to the corrected trajectory to obtain the seeding segment trajectory. If the GPS positioning error of the track points in the seeding segment trajectory is less than or equal to 50 meters, the wind field data and cloud particle data corresponding to the track points are screened to see if they are complete. If they are complete, the wind field change rate between the track point and the five adjacent track points is screened to see if it is less than or equal to 15%. If the wind field change rate is less than or equal to 15%, the track point is marked as a high-quality track point, and the rest of the track points are marked as low-quality track points. Based on the proportion of high-quality waypoints in the total number of waypoints in the broadcast segment, if the proportion is greater than or equal to 80%, a waypoint quality mask is obtained; if the proportion is less than 80%, the GPS positioning error is widened to twice and high-quality waypoints are re-selected.

5. The method for identifying the location of aircraft-based artificial cloud seeding based on wind field data according to claim 1, characterized in that, A method for obtaining the latitude and longitude of the catalytic point after wind field transmission and the expected catalytic period includes: Based on the trackpoint mass mask, the unit transmission wind speed of the trackpoint is converted from meters per second to degrees per second to obtain the east-west and north-south wind speeds in the latitude and longitude components. East and north winds are treated as positive values, and west and south winds as negative values. The latitude and longitude of the return point after wind field transmission are then calculated. The formula for calculating the latitude and longitude of the return point is as follows: in for The longitude of the point where the time trajectory crosses back. for The latitude of the point of return of the time trajectory. for Longitude of the track point in the time trajectory. for The latitude of the track points in the time trajectory. Second, The east-west wind speed component at the track point. The north-south wind speed component of the track point. This is the distance correction between the return point and the nearest waypoint, and For distances less than 1 kilometer, the K-nearest neighbor algorithm is used to find the closest distance to the return point in the corrected trajectory; The latitude and longitude of the return point are used as the latitude and longitude of the catalytic point after the wind field transmission, and the corresponding time is used as the expected catalytic time. The expected catalytic period is determined to be 30 to 120 seconds after the expected catalytic time based on the vector modulus of the wind field transmission speed.

6. The method for identifying the location of aircraft-based artificial cloud seeding backtracking based on wind field data according to claim 1, characterized in that, A method for matching the latitude and longitude of the catalytic point with the latitude and longitude of the return trajectory after the expected catalytic period using a location recognition algorithm includes: Based on the catalytic point latitude and longitude sequence after wind field transmission, the actual backtrack latitude and longitude of the predicted catalytic period are extracted according to the track point mass mask. The latitude and longitude of the catalytic point latitude and longitude sequence and the corresponding latitude and longitude of the backtrack latitude and longitude, wind field vector, and cloud particle number concentration are combined to form a multi-dimensional feature vector. The trajectory similarity is calculated using a location recognition algorithm based on the multi-dimensional feature vector. The formula of the location recognition algorithm is as follows: in Let represent the similarity between two trajectories, and let represent the weighted distance. The latitude and longitude sequence of the catalytic point. The actual latitude and longitude of the return journey. Given a time warping function, find the optimal time alignment for the two trajectories. For matching point pairs The wind field gradient vector at that location, The scale parameter for the wind field gradient is calibrated based on the mean wind shear of the wind field data. For multi-dimensional feature vectors, Trajectory A is time-normalized After and point Paired points, For trajectory After time regulation After and point Paired points; If the trajectory similarity is less than 1 km of feature vector distance, then the catalyst point and the backtrack trajectory point are successfully matched.

7. The method for identifying the location of aircraft-based artificial cloud seeding backtracking based on wind field data according to claim 1, characterized in that, The method for obtaining the latitude, longitude, and time information of the effective traverse point includes: Based on the timestamps and spatial latitude and longitude extracted from the actual catalytic point and backtrack trajectory points after matching, cloud particle detection data within 5 seconds before and after the backtrack point time are filtered according to the timestamps, and the particle spectrum distribution is calculated. If the concentration of small cloud droplets with a particle size of less than 20 μm in the cloud particle detection data decreases by more than or equal to 20% compared with before catalysis, then it is verified whether the concentration of large particles with a particle size of more than or equal to 20 μm increases by more than or equal to 30% compared with before catalysis. If the concentration of large particles increases by more than 30% and the peak particle size distribution of the particle spectrum shifts by more than or equal to 5 μm in the direction of the largest particle size, then the backtrack trajectory point is taken as the effective backtrack point, and the latitude, longitude and time information of the effective backtrack point is obtained.

Images