CN113253363A - Lightning activity path prediction method and system - Google Patents
Lightning activity path prediction method and system Download PDFInfo
- Publication number
- CN113253363A CN113253363A CN202110540378.0A CN202110540378A CN113253363A CN 113253363 A CN113253363 A CN 113253363A CN 202110540378 A CN202110540378 A CN 202110540378A CN 113253363 A CN113253363 A CN 113253363A
- Authority
- CN
- China
- Prior art keywords
- thunderstorm
- lightning
- training
- evaluated
- path prediction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/10—Devices for predicting weather conditions
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention discloses a method for predicting a lightning activity path, which comprises the following steps: acquiring atmospheric electric field data of a project area to be evaluated, calculating a cumulative difference value and judging whether a thunderstorm occurs in the project area to be evaluated; acquiring lightning positioning data of a project area to be evaluated, and analyzing thunderstorm information in the project area to be evaluated by adopting a cluster analysis method; establishing and training a thunderstorm movement relative speed recognition model, inputting atmospheric electric field time sequence data of a project area to be evaluated into the trained thunderstorm movement relative speed recognition model to judge the thunderstorm relative movement information, and obtaining a judgment result; and constructing and training a thunderstorm path prediction model, processing lightning positioning data, inputting the lightning positioning data into the trained thunderstorm path prediction model to predict the thunderstorm movement track, and correcting the predicted thunderstorm movement track by using a judgment result. The method realizes real-time lightning early warning monitoring and real-time hidden danger risk assessment, and can accurately predict the movement track of the thunderstorm.
Description
Technical Field
The invention relates to the technical field of lightning activity monitoring, in particular to a method and a system for predicting a lightning activity path.
Background
Lightning is common disastrous weather, and can cause damage to buildings, power supply and distribution systems and low-voltage electrical systems due to huge current, hot high temperature and strong electromagnetic radiation to generate huge damage in a moment, so that great economic loss and adverse social effects are caused. For a long time, the thunder and lightning disasters bring serious casualties and economic losses to China, thousands of people are killed by lightning each year, and the economic losses reach billions of yuan. Therefore, the thunder and lightning prediction and early warning is one of the important subjects in the meteorological field, and scholars at home and abroad carry out a great deal of research and obtain great results. At present, equipment used for thunder and lightning early warning in China mainly comprises a Doppler radar, an atmospheric electric field instrument and the like. Researches show that the reflectivity of a weather radar has a certain relation with the occurrence of thunder, and partial scholars collect real-time data by using the Doppler radar and realize the identification, tracking and prediction of the thunderstorm by using methods such as linear extrapolation, mode identification and the like; lightning generating fields generally have 2-way characteristics: the jumping and the increase of the absolute value of the electric field, so that some scholars can early warn thunder according to the change rule of the atmospheric electric field curve, the polarity reversal of the atmospheric electric field and other characteristics. However, radar echoes reflect storm processes, lightning cannot be accurately reflected, the atmospheric electric field instrument can only perform small-range lightning early warning, and the moving path of the lightning storm cannot be tracked in real time.
Disclosure of Invention
Aiming at the defects in the prior art, the embodiment of the invention provides a lightning activity path prediction method and a lightning activity path prediction system, which can accurately predict the thunderstorm movement track and realize real-time lightning early warning and monitoring.
In a first aspect, a method for predicting a lightning activity path provided in an embodiment of the present invention includes the following steps:
acquiring atmospheric electric field data of a project area to be evaluated, calculating a cumulative difference value and judging whether a thunderstorm occurs in the project area to be evaluated;
acquiring lightning positioning data of a project area to be evaluated, and analyzing thunderstorm information in the project area to be evaluated by adopting a cluster analysis method;
establishing and training a thunderstorm movement relative speed recognition model, inputting atmospheric electric field time sequence data of a project area to be evaluated into the trained thunderstorm movement relative speed recognition model to judge the thunderstorm relative movement information, and obtaining a judgment result;
and constructing and training a thunderstorm path prediction model, processing the lightning positioning data, inputting the lightning positioning data into the trained thunderstorm path prediction model to predict the thunderstorm movement track, and correcting the predicted thunderstorm movement track by using the judgment result.
In a second aspect, a lightning activity path prediction system provided by an embodiment of the present invention includes: an atmospheric electric field data acquisition module, a lightning location data acquisition module, a thunderstorm motion analysis module and a thunderstorm path prediction module, wherein,
the atmospheric electric field data acquisition module is used for acquiring atmospheric electric field data of a project area to be evaluated, calculating a cumulative difference value and judging whether a thunderstorm occurs in the project area to be evaluated;
the lightning positioning data acquisition module is used for acquiring lightning positioning data of a project area to be evaluated and analyzing thunderstorm information in the project area to be evaluated by adopting a cluster analysis method;
the thunderstorm motion analysis module is used for constructing and training a thunderstorm movement relative speed recognition model, inputting the atmospheric electric field time sequence data of the project area to be evaluated into the trained thunderstorm movement relative speed recognition model to judge the thunderstorm relative motion information, and obtaining a judgment result;
and the thunderstorm path prediction module builds and trains a thunderstorm path prediction model, processes the lightning positioning data, inputs the lightning positioning data into the trained thunderstorm path prediction model to predict a thunderstorm movement track, and corrects the predicted thunderstorm movement track by using the judgment result.
The invention has the beneficial effects that:
according to the thunder and lightning activity path prediction method and system provided by the embodiment of the invention, the thunderstorm movement track is predicted by utilizing a spatial clustering analysis method based on atmospheric electric field data and lightning positioning data, so that real-time fixed-point thunder and lightning early warning monitoring is realized, more refined thunder and lightning activity prediction abstinence is provided, the thunderstorm movement track can be accurately predicted, and important technical guidance is provided for intelligent lightning protection.
Drawings
In order to more clearly illustrate the detailed description of the invention or the technical solutions in the prior art, the drawings that are needed in the detailed description of the invention or the prior art will be briefly described below. Throughout the drawings, like elements or portions are generally identified by like reference numerals. In the drawings, elements or portions are not necessarily drawn to scale.
FIG. 1 illustrates a flow chart of a lightning activity path prediction method provided by a first embodiment of the invention;
FIG. 2 shows a clustering flow chart of the lightning location data in the first embodiment of the invention;
FIG. 3 is a diagram showing the clustering effect in the first embodiment of the present invention;
fig. 4 shows a block diagram of a GRU unit in a first embodiment of the invention;
FIG. 5 is a block diagram illustrating a GRU recurrent neural network model in a first embodiment of the present invention;
fig. 6 shows a block diagram of a lightning activity path prediction system according to a second embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should be further understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.
As used in this specification and the appended claims, the term "if" may be interpreted contextually as "when", "upon" or "in response to a determination" or "in response to a detection". Similarly, the phrase "if it is determined" or "if a [ described condition or event ] is detected" may be interpreted contextually to mean "upon determining" or "in response to determining" or "upon detecting [ described condition or event ]" or "in response to detecting [ described condition or event ]".
It is to be noted that, unless otherwise specified, technical or scientific terms used herein shall have the ordinary meaning as understood by those skilled in the art to which the invention pertains.
As shown in fig. 1, a flowchart of a lightning activity path prediction method according to a first embodiment of the present invention is shown, and the method includes the following steps:
and A, acquiring atmospheric electric field data of the project area to be evaluated, and calculating the accumulated difference value to judge whether a thunderstorm occurs in the project area to be evaluated.
And B: the lightning positioning data of the project area to be evaluated is obtained, and thunderstorm information in the project area to be evaluated is analyzed by adopting a cluster analysis method.
And C: and constructing and training a thunderstorm movement relative speed recognition model, inputting the atmospheric electric field time sequence data of the project area to be evaluated into the trained thunderstorm movement relative speed recognition model to judge the thunderstorm relative movement information, and obtaining a judgment result.
Step D: and constructing and training a thunderstorm path prediction model, processing the lightning positioning data, inputting the lightning positioning data into the trained thunderstorm path prediction model to predict the thunderstorm movement track, and correcting the predicted thunderstorm movement track by using the judgment result.
The method for predicting the lightning activity path solves the problem of refinement of lightning approach prediction, lightning early warning signals sent by meteorological departments are on-plane lightning early warnings, coverage is large, administrative areas are generally taken as the minimum units, in other words, the refinement is not enough, points cannot be refined, and the guiding value of specific units is low. Meanwhile, the lightning early warning signal is a time scale from two hours to six hours, and the delay is too large for a specific unit. The method utilizes atmospheric electric field data and lightning positioning data to develop a spatial clustering lightning path prediction algorithm, lightning prediction and early warning are accurate to project points, and time is fine to 1 hour.
In the step a, the method specifically comprises: the electric field jumping phenomenon before the lightning is generated is captured by an accumulative difference method, and the calculation formula is as follows.
Wherein E is the atmospheric electric field value, E (t)0) Is t0Value of atmospheric electric field at time, E (t)1) Is t1Value of atmospheric electric field at time t0And t1When two adjacentThe interval is the sampling time interval of the atmospheric electric field instrument, t1-t015 min. When the value E (t)' reaches a set threshold value, the thunderstorm can be considered to occur in the early warning area, and the selection of the threshold value needs to be determined according to local weather and geographic conditions.
Specifically, collecting atmospheric electric field data in the early warning area, calculating an atmospheric electric field accumulated difference value within 15min according to a formula (1), and considering that thunderstorm activity exists in the early warning area when an early warning threshold value is reached.
In the step B, the lightning location data is monitoring data of the lightning location system of the meteorological department, and includes information such as time, longitude and latitude, strength, and steepness, and the time and longitude and latitude are used in this embodiment. The method specifically comprises the steps of arranging the lightning location data into sequence data arranged at intervals of 10min, and converting the sequence data into a CSV comma separation file.
The Clustering analysis of the lightning positioning data in the step B is Based on DBSCAN (sensitivity-Based Spatial Clustering of Applications with Noise, Density-Based Clustering method), which is a Density-Based Spatial Clustering algorithm, has high Clustering speed, can effectively process Noise points and find Spatial clusters of any shape; the number of clusters to be divided is not required to be input; the method has the advantages of being capable of finding out noise points in data, insensitive to noise and the like, and is a method widely applied to the data mining technology. Assume that the desired sample set is D ═ x1,x2,...,xm) And m is an integer, the following concepts and notations are introduced:
1) epsilon neighborhood: for xjE.g. D, whose e neighborhood contains the sum x in the sample set DjA set of subsamples having a distance of not more than epsilon, i.e.
The number of this subsample set is denoted as | N ∈ (x)i)|。
2) Core object: for any sample xjE.g. D, if N e (x) corresponding to epsilon neighborhoodj) Containing at least MinPts samples, i.e. if | N ∈ (x)j) | is not less than MinPts, then xjIs the core object.
3) The density is up to: if xiAt xjIn the neighborhood of epsilon, and xjIs a core object, then called xiFrom xjThe density is up to. Note that the opposite does not necessarily hold, i.e., x cannot be said at this timejFrom xiDensity is direct, unless and xiIs also a core object.
4) The density can reach: for xiAnd xjIf the sample sequence p1, p2, …, pT exists, the requirement that p1 is xi,pt＝xjAnd pt +1 is directly reached from pt density, then x is calledjFrom xiThe density can be reached. That is, the density can be achieved to satisfy transitivity. The transfer samples p1, p2, …, pT-1 in the sequence are all core objects at this time, because only the core objects can make the other sample densities go through. Note that the density can be achieved without satisfying the symmetry, which can be derived from the asymmetry of the density through.
5) Density connection: for xiAnd xjIf there is a core object sample xkLet x beiAnd xjAre all xkWhen the density is up, it is called xiAnd xjThe densities are connected. Note that the density connectivity is such that symmetry is satisfied.
The clustering flow chart of the lightning location data is shown in fig. 2, and the clustering effect is shown in fig. 3. The step B is realized by the following steps:
step B1: inputting a parameter field radius epsilon, a threshold MinPts and a lightning positioning data set D;
step B2: randomly selecting a point p in the data set D, and judging whether the point p is a core point;
step B3: if the point p is a core point, finding out all points with reachable direct density in the neighborhood;
step B4: judging whether all points in the lightning positioning data set D are searched completely, if not, repeatedly executing the steps B2-B4;
step B5: if yes, combining the points with the reachable density and expanding the clusters;
step B6: and outputting the target cluster set.
The step C specifically comprises the following steps:
step C1, collecting atmospheric electric field data, combining lightning location data, manually marking data of thunderstorms close to the project area to be evaluated and far from the project area to be evaluated, setting distance threshold values according to the standard of the close and far according to the actual situation, dividing training sets and verification sets according to the proportion of 7:3, and carrying out normalization processing on the data, wherein the normalization aims to prevent neuron output saturation caused by overlarge net input absolute value and enable a neural network to quickly converge, the calculation formula is shown in formula (2),
wherein, x [ n ]]Is the nth value, x, of the data sequencemin、xmaxAre the minimum and maximum values of the data sequence,
is x [ n ]]The normalized value of (a).
And step C2, establishing a thunderstorm movement relative speed recognition model by utilizing the one-dimensional convolution neural network, setting super parameters such as iteration rounds, learning rate and the like, inputting training set data into the neural network model for training, obtaining a trained thunderstorm movement relative speed recognition model after the training rounds are reached, inputting a verification set into the trained thunderstorm movement relative speed recognition model for verification, and evaluating and fixing the model.
Step C3: and normalizing the atmospheric electric field time sequence data of the project area to be evaluated, and inputting the thunderstorm movement relative speed identification model to obtain the movement direction of the thunderstorm relative to the atmospheric electric field instrument.
The step D specifically comprises the following steps:
step D1: b, calculating the mass center of the lightning cluster by using the target cluster set output in the step B and using a formula (3), and establishing a thunderstorm movement path data set;
wherein (x)i，yi) The coordinates of the ith lightning in the lightning cluster and (x, y) the coordinates of the centroid of the lightning cluster.
And removing the short-duration thunderstorm path data, and arranging the short-duration thunderstorm path data into data required by a thunderstorm path prediction model, wherein the 6 continuous thunderstorm position data is one sample. The data set is normalized and the maximum and minimum values are recorded. And dividing the normalized data into a training set and a verification set according to the proportion of 7: 3.
The thunderstorm path prediction model is based on a GRU (gated Recurrent Unit) gated cyclic unit structure network, is a variant of a traditional RNN (cyclic neural network), can effectively capture semantic association between Long sequences like an LSTM (Long Short-Term Memory), and relieves gradient disappearance or explosion phenomena, and meanwhile, the structure and the calculation of the model are simpler than those of the LSTM. As shown in fig. 4, a block diagram of a GRU unit is shown, the GRU unit having two gates, a reset gate and an update gate, the reset gate determining how to combine the new input with the previous memory, the update gate determining how much of the previous memory has acted, the GRU being calculated as follows:
zt＝σ(Wz·[yt-1,xt]+bu) (4)
rt＝σ(Wr·[yt-1,xt]+br) (5)
y't＝tanh(W·[rt*yt-1,xt]) (6)
yt＝(1-zt)*yt-1+zt*y't (7)
wherein σ is sigmoid activation function, tanh is tanh activation function, Wz、WrW is the weight, bu、brTo be offset, ytAs output at the current time, yt-1The output of the last moment. Equation (4) is called the update gate and equation (5) is called the reset gate.
Step D2: and establishing and training a thunderstorm path prediction model based on the GRU recurrent neural network.
A three-layer GRU recurrent neural network is established by utilizing paddlepaddlele, the structure of a GRU recurrent neural network model is shown in figure 5, the network is set to be a training mode, and hyper-parameters such as iteration turns and learning rate are set. Training and learning the thunderstorm movement path prediction model read in by the training set data processed in the step D1 to obtain a trained thunderstorm movement path prediction model, and predicting the last 3 data by using the first 3 data; and after the number of training rounds is reached, inputting the verification set data into the trained thunderstorm movement path prediction model for verification and evaluation, fixing the model, and finishing the training and learning of the thunderstorm movement path prediction model.
When the thunderstorm path prediction model of the GRU recurrent neural network is trained in the step D2, the optimizer is an Adam optimizer, the activation function is relu, dropout is adopted for improving the generalization of the network, and the discarding rate is 0.3.
The loss function adopted in step D2 is an MSE (mean square error) loss function, and the calculation formula is as follows:
wherein y _ is a predicted value of the neural network, i.e. y in formula (7)tAnd y is the tag value, i.e. the true location of the thunderstorm at that moment in the dataset.
And D3, processing the lightning positioning data, inputting the processed lightning positioning data into a thunderstorm path prediction model, and predicting the thunderstorm movement track.
Processing the lightning positioning data according to the methods in the steps B and D1 to obtain thunderstorm position data; normalizing the thunderstorm position data by using the maximum value and the minimum value recorded in the step D1; and setting the thunderstorm path prediction model as a prediction mode, and inputting the data after normalization processing to obtain a prediction result.
Step D4: and correcting the predicted thunderstorm movement track by using the calculation result of the step C3.
The software and hardware environment used in step C, D is: the operating system is windows 10; the CPU is AMDR 7-2700; the display card is NVIDIA RTX-2060; the AI frame is hundred degrees paddlepaddle; the memory is 32G; the programming language is python.
According to the lightning activity path prediction method provided by the embodiment of the invention, the thunderstorm movement track is predicted by utilizing the atmospheric electric field data, the lightning positioning data and the spatial clustering analysis method, so that real-time lightning early warning monitoring and real-time hidden danger risk assessment are realized, more refined lightning activity prediction abstinence is provided, the thunderstorm movement track can be accurately predicted, and the lightning prediction early warning is accurately carried out to the project area to be assessed.
In the first embodiment, a lightning activity path prediction method is provided, and correspondingly, a lightning activity path prediction system is also provided. Please refer to fig. 6, which is a block diagram illustrating a lightning activity path prediction system according to a second embodiment of the present invention. Since the apparatus embodiments are substantially similar to the method embodiments, they are described in a relatively simple manner, and reference may be made to some of the descriptions of the method embodiments for relevant points. The device embodiments described below are merely illustrative.
As shown in fig. 6, a block diagram of a lightning activity path prediction system according to a second embodiment of the present invention is shown, and the system includes: an atmospheric electric field data acquisition module, a lightning location data acquisition module, a thunderstorm motion analysis module and a thunderstorm path prediction module, wherein,
the atmospheric electric field data acquisition module is used for acquiring atmospheric electric field data of a project area to be evaluated, calculating a cumulative difference value and judging whether a thunderstorm occurs in the project area to be evaluated;
the lightning positioning data acquisition module is used for acquiring lightning positioning data of a project area to be evaluated and analyzing thunderstorm information in the project area to be evaluated by adopting a cluster analysis method;
the thunderstorm motion analysis module is used for constructing and training a thunderstorm movement relative speed recognition model, inputting the atmospheric electric field time sequence data of the project area to be evaluated into the trained thunderstorm movement relative speed recognition model to judge the thunderstorm relative motion information, and obtaining a judgment result;
and the thunderstorm path prediction module builds and trains a thunderstorm path prediction model, processes the lightning positioning data, inputs the lightning positioning data into the trained thunderstorm path prediction model to predict a thunderstorm movement track, and corrects the predicted thunderstorm movement track by using the judgment result.
In this embodiment, the lightning location data acquisition module comprises a cluster analysis unit for inputting parameter neighborhood radius, threshold value and lightning location data set;
randomly selecting a point in the lightning location data set, and judging whether the point is a core point;
if the point is a core point, finding out all points with reachable direct density in the neighborhood;
judging whether all points in the lightning location data set are searched completely, if not, returning to the step of repeatedly executing the search;
if yes, combining the points with the reachable density and expanding the clusters;
and outputting the target cluster set.
In this embodiment, the thunderstorm motion analysis module includes an identification model construction and training unit, the identification model construction and training unit is used for marking data of the thunderstorm close to the project area to be evaluated and far from the project area to be evaluated from the atmospheric electric field data and the lightning positioning data, and dividing a training set and a verification set according to a set ratio;
establishing a thunderstorm movement relative speed recognition model by adopting a one-dimensional convolutional neural network, setting training parameters, inputting a training set into the thunderstorm movement relative speed recognition model for training, inputting a verification set for verification after the number of training rounds is reached, and obtaining the trained thunderstorm movement relative speed recognition model.
The thunderstorm path prediction module comprises a prediction module construction and training unit, wherein the prediction module construction and training unit is used for calculating the mass center of a lightning cluster according to the output target cluster set, establishing a thunderstorm moving path data set, and dividing the data set into a training set and a verification set according to a set proportion;
and establishing a thunderstorm path prediction model based on the GRU recurrent neural network, inputting the training set into the thunderstorm path prediction model for training, and inputting the training set into the thunderstorm path prediction model for verification by using the verification set after the set number of training rounds is reached to obtain the trained thunderstorm path prediction model.
The above is a description of an embodiment of a lightning activity path prediction system according to a second embodiment of the present invention.
The thunder and lightning activity path prediction system and the thunder and lightning activity path prediction method provided by the invention have the same inventive concept and the same beneficial effects, and are not repeated herein.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the present invention, and they should be construed as being included in the following claims and description.
Claims (8)
1. A lightning activity path prediction method is characterized by comprising the following steps:
acquiring atmospheric electric field data of a project area to be evaluated, calculating a cumulative difference value and judging whether a thunderstorm occurs in the project area to be evaluated;
acquiring lightning positioning data of a project area to be evaluated, and analyzing thunderstorm information in the project area to be evaluated by adopting a cluster analysis method;
establishing and training a thunderstorm movement relative speed recognition model, inputting atmospheric electric field time sequence data of a project area to be evaluated into the trained thunderstorm movement relative speed recognition model to judge the thunderstorm relative movement information, and obtaining a judgment result;
and constructing and training a thunderstorm path prediction model, processing the lightning positioning data, inputting the lightning positioning data into the trained thunderstorm path prediction model to predict the thunderstorm movement track, and correcting the predicted thunderstorm movement track by using the judgment result.
2. The method for predicting the lightning activity path according to claim 1, wherein the analyzing the thunderstorm information in the project area to be evaluated by using the cluster analysis method specifically comprises:
inputting parameter neighborhood radius, threshold and lightning location data set;
randomly selecting a point in the lightning location data set, and judging whether the point is a core point;
if the point is a core point, finding out all points with reachable direct density in the neighborhood;
judging whether all points in the lightning location data set are searched completely, if not, returning to the step of repeatedly executing the search;
if yes, combining the points with the reachable density and expanding the clusters;
and outputting the target cluster set.
3. The method for predicting lightning activity path according to claim 2, wherein the constructing and training a thunderstorm movement relative speed recognition model specifically comprises:
marking data of a thunderstorm close to a project area to be evaluated and data of a thunderstorm far from the project area to be evaluated from atmospheric electric field data and lightning positioning data, and dividing a training set and a verification set according to a set ratio;
establishing a thunderstorm movement relative speed recognition model by adopting a one-dimensional convolutional neural network, setting training parameters, inputting a training set into the thunderstorm movement relative speed recognition model for training, inputting a verification set for verification after the number of training rounds is reached, and obtaining the trained thunderstorm movement relative speed recognition model.
4. The lightning activity path prediction method according to claim 3, wherein the building and training of the thunderstorm path prediction model specifically comprises:
calculating the mass center of the lightning cluster according to the output target cluster set, establishing a thunderstorm movement path data set, and dividing the data set into a training set and a verification set according to a set proportion;
and establishing a thunderstorm path prediction model based on the GRU recurrent neural network, inputting the training set into the thunderstorm path prediction model for training, and inputting the training set into the thunderstorm path prediction model for verification by using the verification set after the set number of training rounds is reached to obtain the trained thunderstorm path prediction model.
5. A lightning activity path prediction system, comprising: an atmospheric electric field data acquisition module, a lightning location data acquisition module, a thunderstorm motion analysis module and a thunderstorm path prediction module, wherein,
the atmospheric electric field data acquisition module is used for acquiring atmospheric electric field data of a project area to be evaluated, calculating a cumulative difference value and judging whether a thunderstorm occurs in the project area to be evaluated;
the lightning positioning data acquisition module is used for acquiring lightning positioning data of a project area to be evaluated and analyzing thunderstorm information in the project area to be evaluated by adopting a cluster analysis method;
the thunderstorm motion analysis module is used for constructing and training a thunderstorm movement relative speed recognition model, inputting the atmospheric electric field time sequence data of the project area to be evaluated into the trained thunderstorm movement relative speed recognition model to judge the thunderstorm relative motion information, and obtaining a judgment result;
and the thunderstorm path prediction module builds and trains a thunderstorm path prediction model, processes the lightning positioning data, inputs the lightning positioning data into the trained thunderstorm path prediction model to predict a thunderstorm movement track, and corrects the predicted thunderstorm movement track by using the judgment result.
6. The lightning activity path prediction system of claim 5, wherein the lightning location data acquisition module comprises a cluster analysis unit to input a parameter neighborhood radius, a threshold value, and a set of lightning location data;
randomly selecting a point in the lightning location data set, and judging whether the point is a core point;
if the point is a core point, finding out all points with reachable direct density in the neighborhood;
judging whether all points in the lightning location data set are searched completely, if not, returning to the step of repeatedly executing the search;
if yes, combining the points with the reachable density and expanding the clusters;
and outputting the target cluster set.
7. The lightning activity path prediction system of claim 6, wherein the thunderstorm motion analysis module comprises a recognition model construction and training unit for marking data of the thunderstorm close to the project area to be evaluated and far from the project area to be evaluated from the atmospheric electric field data and lightning location data, and dividing the training set and the verification set according to a set ratio;
establishing a thunderstorm movement relative speed recognition model by adopting a one-dimensional convolutional neural network, setting training parameters, inputting a training set into the thunderstorm movement relative speed recognition model for training, inputting a verification set for verification after the number of training rounds is reached, and obtaining the trained thunderstorm movement relative speed recognition model.
8. The lightning activity path prediction system of claim 7, wherein the thunderstorm path prediction module comprises a prediction module construction and training unit, the prediction module construction and training unit is configured to calculate a centroid of a lightning cluster from the output target cluster sets, establish a thunderstorm movement path data set, and divide the data set into a training set and a validation set according to a set proportion;
and establishing a thunderstorm path prediction model based on the GRU recurrent neural network, inputting the training set into the thunderstorm path prediction model for training, and inputting the training set into the thunderstorm path prediction model for verification by using the verification set after the set number of training rounds is reached to obtain the trained thunderstorm path prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110540378.0A CN113253363A (en) | 2021-05-18 | 2021-05-18 | Lightning activity path prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110540378.0A CN113253363A (en) | 2021-05-18 | 2021-05-18 | Lightning activity path prediction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113253363A true CN113253363A (en) | 2021-08-13 |
Family
ID=77182535
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110540378.0A Withdrawn CN113253363A (en) | 2021-05-18 | 2021-05-18 | Lightning activity path prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113253363A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115792437A (en) * | 2021-11-18 | 2023-03-14 | 苏州工业园区科佳自动化有限公司 | Digital monitoring method and system for lightning arrester |
-
2021
- 2021-05-18 CN CN202110540378.0A patent/CN113253363A/en not_active Withdrawn
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115792437A (en) * | 2021-11-18 | 2023-03-14 | 苏州工业园区科佳自动化有限公司 | Digital monitoring method and system for lightning arrester |
| CN115792437B (en) * | 2021-11-18 | 2024-01-26 | 苏州工业园区科佳自动化有限公司 | Digital monitoring method and system for lightning arrester |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20200160061A1 (en) | Automatic ship tracking method and system based on deep learning network and mean shift | |
| Zhang et al. | Modeling pedestrians’ near-accident events at signalized intersections using gated recurrent unit (GRU) | |
| Yazdi et al. | Flight delay prediction based on deep learning and Levenberg-Marquart algorithm | |
| CN113327022B (en) | Lightning protection safety risk management system and method | |
| CN109117883B (en) | SAR image sea ice classification method and system based on long-time memory network | |
| CN114220271A (en) | Traffic flow prediction method, equipment and storage medium based on dynamic space-time graph convolution cycle network | |
| CN109357747B (en) | A kind of identification of online train and speed estimation method based on fiber-optic vibration signal | |
| Jing et al. | HPRNN: A hierarchical sequence prediction model for long-term weather radar echo extrapolation | |
| CN115220133B (en) | Rainfall prediction method, device and equipment for multiple meteorological elements and storage medium | |
| Peng et al. | CNGAT: A graph neural network model for radar quantitative precipitation estimation | |
| CN115761534A (en) | Method for detecting and tracking small target of infrared unmanned aerial vehicle under air background | |
| CN114802303A (en) | Obstacle trajectory prediction method, obstacle trajectory prediction device, electronic device, and storage medium | |
| Wong et al. | Automatic tropical cyclone eye fix using genetic algorithm | |
| Han et al. | Short-term 4D trajectory prediction based on LSTM neural network | |
| Girimurugan et al. | Application of Deep Learning to the Prediction of Solar Irradiance through Missing Data | |
| CN113253363A (en) | Lightning activity path prediction method and system | |
| CN113836808A (en) | PM2.5 deep learning prediction method based on heavy pollution feature constraint | |
| CN112884222B (en) | Time-period-oriented LSTM traffic flow density prediction method | |
| Kosasih et al. | Travel time estimation for destination In Bali using KNN-regression method with Tensorflow | |
| Nie et al. | Self-attention UNet model for radar based precipitation nowcasting | |
| Ma et al. | PPDTSA: Privacy-preserving deep transformation self-attention framework for object detection | |
| CN115909717A (en) | Expressway short-term traffic flow prediction method based on deep learning | |
| Su et al. | Trajectory prediction based on machine learning | |
| CN112667763B (en) | Track prediction method based on self-adaptive timestamp and multi-scale feature extraction | |
| Yoon et al. | Very short-term rainfall prediction based on radar image learning using deep neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20210813 |