Intelligent early warning system based on thunder and lightning data

Abstract

The invention discloses an intelligent early warning system based on thunder and lightning data, and relates to the technical field of thunder and lightning early warning, and the early warning system provided by the invention comprehensively collects radar meteorological data, atmospheric electric field data and thunder and lightning activity data through thunder and lightning detection terminals distributed in each target area; real-time processing and storage are performed through the cloud data platform, so that the integrity and timeliness of the data are ensured; the adaptability and continuity of data are improved by dynamically setting historical and current thunder and lightning activity periods and flexibly adjusting preset time; the feature data acquisition module accurately extracts data related to the lightning intensity of the ground-to-ground lightning type, and the lightning prediction module provides scientific lightning activity prediction by training an intensity prediction probability model and an intensity prediction model; the early warning system can efficiently evaluate and trigger early warning information by calculating lightning risk parameters, marking potential risk areas and combining safety level parameters, and helps related departments to take measures in time to reduce loss caused by lightning disasters.

Description

Technical Field

[0001] The present invention relates to the technical field of lightning early warning, and in particular to an intelligent early warning system based on lightning data. Background Art

[0002] Lightning is a discharge phenomenon that occurs between clouds or between clouds and the ground in nature. With the continuous development of the economy and the rapid progress of society, the personal safety and property losses caused by lightning have attracted people's attention. Lightning warning, as an important measure in active lightning protection, is of great significance in reducing the harm caused by lightning.

[0003] The lightning warning system existing in the prior art continuously monitors the atmospheric electric field strength, identifies and determines the characteristics of the lightning electric field in real time, and provides effective warning information for the approach of lightning. However, the formation of lightning is affected by many factors, such as the geographical and geological characteristics of the occurrence area, climate characteristics, changes in the surrounding environment, and the instantaneous characteristics of lightning. These bring difficulties to the accurate warning of lightning, resulting in the traditional warning system having problems of low applicability and low warning accuracy, which is difficult to meet the actual thunderstorm warning needs.

[0004] To solve this problem, we propose an intelligent early warning system based on lightning data. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an intelligent early warning system based on lightning data, which can effectively solve the problems in the above-mentioned background technology.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] Intelligent early warning system based on lightning data, including:

[0008] Lightning detection terminals are distributed in each target area and are used to obtain radar meteorological data, atmospheric electric field data and lightning activity data of different lightning activity events in each target area. Each lightning detection terminal is assigned a unique identifier. The lightning detection device is connected to the cloud data platform through a local area network;

[0009] The cloud data platform is used to obtain radar meteorological data, atmospheric electric field data and lightning activity data of different lightning activity events in each target area collected in real time by the lightning terminal, dynamically update in real time based on the collected data, and store the collected historical data; it is also used to divide the regions according to each target area, obtain the geographical feature data of each divided area, and set the safety level parameters of the divided areas on each target area based on the geographical feature data;

[0010] The characteristic data acquisition module is used to obtain historical data from the cloud data platform, calculate the correlation between radar meteorological data and atmospheric electric field data and ground-to-ground lightning, and screen out relevant data with a high correlation with the intensity of ground-to-ground lightning in the radar meteorological data and atmospheric electric field data, and record them as lightning-related characteristic data;

[0011] The lightning prediction module is used to train the lightning probability prediction model to predict the lightning probability of each divided area in the current lightning activity cycle; to train the lightning intensity prediction model to predict the lightning intensity of ground-to-ground lightning in the short term; based on the lightning prediction probability of each divided area in the target area and the lightning prediction intensity of ground-to-ground lightning, calculate the lightning risk parameters of each divided area, and mark the potential risk areas based on the lightning risk parameters;

[0012] The lightning warning module is used to calculate the lightning warning parameters of the potential risk area according to the safety level parameters, lightning prediction probability and lightning prediction intensity of the divided areas in the potential risk area. The set parameter threshold is based on the lightning warning parameter assessment to determine whether to trigger the alarm information; if the alarm is triggered, the alarm information of the target area is published and the lightning warning image of the target area is generated and published to the administrator of the target area; if the alarm is not triggered, continuous monitoring is performed.

[0013] Preferably, the lightning detection terminal specifically comprises:

[0014] A meteorological data detection unit, which captures radar meteorological data in the target area through a meteorological radar, wherein the radar meteorological data includes reflectivity, radial velocity, spectrum width information, vertical liquid water content, echo top height and echo intensity;

[0015] An electric field data detection unit, which captures the atmospheric electric field data in the target area in real time through an atmospheric electric field instrument, wherein the atmospheric electric field data includes electric field time series differential data, electric field intensity change rate and charge density;

[0016] The Fengyun satellite detection unit captures lightning activity data of single lightning through the satellite's lightning imager, and uses a clustering algorithm to classify single lightning that are close in time and space as the same lightning activity event, so as to classify the lightning activity data of single lightning into the corresponding lightning activity event, and obtain lightning activity data in different lightning activity events in the target area, wherein the lightning activity data includes time information, lightning location data, lightning intensity data and lightning type data;

[0017] A lightning data storage unit, used to store radar meteorological data, atmospheric electric field data and lightning activity data collected by the meteorological data detection unit, the electric field data detection unit and the Fengyun satellite detection unit;

[0018] The lightning data extraction unit is used to record the most recent lightning activity event as the current lightning activity event, and record all lightning activity events before the current lightning activity event as historical lightning activity events. When the Fengyun satellite detection unit collects the latest lightning activity data, it uses a clustering algorithm to perform cluster analysis on it and the lightning activity data in the current lightning activity event to determine whether the latest lightning activity data belongs to the current lightning activity event, and updates the current lightning activity event and the historical lightning activity event according to the clustering result; it is used to set the current lightning activity cycle and the historical lightning activity cycle, and the current lightning activity cycle and the historical lightning activity cycle are used to reflect the current lightning activity respectively. The relevant time period of the lightning activity event and the historical activity event; set the collection frequency, extract the radar meteorological data and atmospheric electric field data in the historical lightning activity cycle and the current lightning activity cycle from the lightning data storage unit according to the collection frequency, and align the radar meteorological data and the atmospheric electric field data through the timestamp; extract the lightning activity data collected in the historical lightning activity cycle and the current lightning activity cycle, find the closest timestamp through the nearest neighbor method based on the time information of the lightning activity data, and align the lightning activity data with the timestamp. If there is no lightning activity data on other timestamps, use the flag bit to indicate that the lightning activity data of the timestamp is missing and no lightning activity has occurred;

[0019] The lightning data transmission unit is used to package the radar meteorological data, atmospheric electric field data and lightning activity data in the historical lightning activity cycle extracted by the lightning data extraction unit into a historical data set, and to package the radar meteorological data, atmospheric electric field data and lightning activity data in the current lightning activity cycle extracted by the lightning data extraction unit into a real-time data set, which is updated in real time based on the latest data obtained.

[0020] Preferably, the cloud data platform specifically includes:

[0021] The cloud data acquisition module is used to acquire the historical data set and the real-time data set collected by the lightning detection terminal in each target area; and the same identifier is assigned to the historical data set and the real-time data set collected by the lightning detection terminal in the same target area;

[0022] The cloud data processing module is used to clean the radar meteorological data, atmospheric electric field data and lightning activity data in the historical data set and real-time data set acquired by the cloud data acquisition module, remove abnormal values ​​and noise data, and perform real-time standardization processing to ensure the consistency and comparability of the data;

[0023] A cloud data storage module, used to store historical data sets processed by the cloud data processing module;

[0024] The security level setting module divides the regions based on each target area, obtains the geographical feature data of each divided region, and sets the security level parameters of the divided regions on each target area based on the geographical feature data.

[0025] Preferably, the feature data acquisition module specifically includes:

[0026] A historical data acquisition unit, used to acquire historical data sets in the cloud data storage module;

[0027] The feature data extraction unit is used to select multiple historical data sets as samples, calculate the correlation between the radar meteorological data and the atmospheric electric field data and the lightning intensity of the ground-to-ground type based on the Pearson correlation coefficient, and screen out the relevant data with a high correlation with the lightning intensity of the ground-to-ground type in the radar meteorological data and the atmospheric electric field data, and record them as lightning-related feature data.

[0028] Preferably, the lightning prediction module specifically includes:

[0029] The coverage degree calculation unit is used to obtain the historical lightning activity cycle of the historical lightning activity events corresponding to different historical data sets in the cloud data storage module; obtain the atmospheric electric field parameters of various points in the target area within the historical lightning activity cycle, and obtain the relevant areas affected by the lightning activity events in the target area based on the changes in the atmospheric electric field parameters within the historical activity cycle; calculate the coverage degree of each divided area based on the coverage area of ​​the relevant area in each divided area in the target area;

[0030] The lightning probability prediction model training unit is used to obtain the altitude and soil resistivity of each divided area; obtain the coverage of each divided area in different historical lightning activity cycles; obtain the frequency of lightning activity in each divided area in different historical lightning activity cycles; pre-process the above-obtained data and combine them into a feature vector set; train the logistic regression model through the feature vector set, adjust the model parameters to optimize the prediction performance, and use the trained model for lightning probability prediction of each divided area in the current lightning activity cycle;

[0031] The lightning intensity prediction model training unit is used to obtain the historical data set in the cloud data storage module, and extract the lightning-related characteristic data and lightning activity data in the historical data set; the above-extracted data is used as the data set, and the data set is divided into a training set and a test set, and the training set is used to train the selected neural network model, and the trained model is used to predict the lightning intensity of ground-to-ground lightning in the short term; during the training process of the lightning intensity prediction model, cross-validation and grid search are used to perform hyperparameter tuning to optimize the performance of the lightning intensity prediction model;

[0032] A lightning prediction unit is used to receive the real-time data set from the cloud data processing module, input the radar meteorological data, atmospheric electric field data and lightning activity data in the real-time data set into the lightning probability prediction model and the lightning intensity prediction model, so as to output the lightning prediction probability of each divided area in the target area and the lightning prediction intensity of ground-to-ground lightning in real time;

[0033] The potential risk marking unit is used to obtain the lightning prediction probability and lightning prediction intensity of ground-to-ground lightning in each divided area of ​​the target area, calculate the lightning risk parameters of each divided area based on the obtained prediction data, set the risk parameter threshold, and when the lightning risk parameter is greater than the risk parameter threshold, mark the divided area as a potential risk area and send a risk signal to the lightning warning module.

[0034] Preferably, the security level setting module specifically includes:

[0035] Regional division unit, based on the target area, uses GIS tools to generate grids, divides the area, and assigns a unique ID to each divided area;

[0036] A geographic feature data acquisition unit, which acquires geographic feature data of each divided area, wherein the geographic feature data includes soil resistivity, building type information, protection measures information, population density information, and power grid layout information;

[0037] The security level calculation unit is used to calculate the security level parameters of each divided area based on the geographical feature data.

[0038] Preferably, the lightning warning module specifically comprises:

[0039] A threshold management unit, which sets a first parameter threshold and a second parameter threshold of a lightning warning based on historical data and expert experience, and dynamically adjusts the first parameter threshold and the second parameter threshold according to actual conditions;

[0040] The warning trigger unit is used to obtain the risk signal sent by the potential risk marking unit; based on the safety level parameters, lightning prediction probability and lightning prediction intensity of the divided areas in the potential risk area, the lightning warning parameters of the potential risk area are calculated, and the parameter threshold is set based on the lightning warning parameter evaluation to determine whether to trigger the alarm information;

[0041] The warning image generation unit is used to obtain the map of the target area, the lightning prediction probability of the corresponding divided area, the lightning prediction intensity parameters and the warning information, and generate a lightning warning image through Matplotlib with the above data, using color coding to indicate the warning level of the grid where the potential risk area is located;

[0042] The warning information issuing unit sends the warning information initiated by the warning triggering unit and the lightning warning image generated by the warning image generating unit to the administrator end of each target area.

[0043] Preferably, the security level calculation unit is used to calculate the security level parameters of each divided area according to the geographical feature data, including:

[0044] The obtained geographical feature data were scored using a scoring system to obtain the following scoring data: building type score, building material score, building height score, lightning protection facility score, lightning protection facility maintenance status score, population density score, power grid layout score, and power grid equipment aging score;

[0045] The calculation formula of the security level parameter is as follows:

[0046]

[0047] In the formula, S x represents the security level parameter for dividing area x, j represents the total number of scoring data, η i represents the weight parameter of the i-th scoring data, k i Represents the i-th rating data.

[0048] Preferably, in the warning trigger unit, a first parameter threshold and a second parameter threshold are set, and the lightning warning parameter determines the alarm level through the first parameter threshold and the second parameter threshold; when the warning parameter is greater than the first parameter threshold, a red thunderstorm warning message is triggered; when the warning parameter is less than the second parameter threshold, the warning message is not triggered; when the warning parameter is greater than / equal to the second parameter threshold and less than / equal to the first parameter threshold, a yellow thunderstorm warning message is triggered.

[0049] Preferably, the setting of the current lightning activity cycle and the historical lightning activity cycle specifically includes:

[0050] In the lightning data extraction unit, the historical lightning activity cycle and the current lightning activity cycle are used to represent the relevant time periods of the historical lightning activity events and the current lightning activity events respectively;

[0051] For the historical lightning activity cycle, extract the time information of all lightning activity data therein, extend the earliest time information forward by a preset time as the start time of the historical lightning activity cycle, and extend the latest time information backward by a preset time as the end time of the historical lightning activity cycle;

[0052] For the current lightning activity cycle, the time information of all lightning activity data in the current lightning activity event is extracted, and the earliest time information is extended forward by a preset time as the start time of the current lightning activity cycle. The start time of the current lightning activity cycle and the current actual time are all the current lightning activity cycle, and the current lightning activity cycle is updated and adjusted based on the latest lightning activity data obtained.

[0053] Compared with the prior art, the present invention provides an intelligent early warning system based on lightning data, which has the following beneficial effects:

[0054] The early warning system collects radar meteorological data, atmospheric electric field data and lightning activity data in real time through lightning detection terminals distributed in various target areas. The cloud data platform efficiently stores, processes and dynamically updates the collected data, and sets safety level parameters for different divided areas through regional divisions. The lightning prediction module trains the lightning intensity prediction model through machine learning algorithms, and outputs the lightning prediction trajectory of ground-to-ground lightning, the lightning prediction probability and intensity of the divided area in real time. The lightning warning module calculates the lightning warning parameters of the target area based on the safety level parameters, lightning prediction probability and intensity of the divided area, so as to trigger warning information of different levels and generate intuitive warning images, ensuring the timeliness and effectiveness of the warning information. The system realizes data collection, processing, analysis and warning issuance through the collaborative work of various modules, thereby improving the accuracy and response speed of lightning warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The figure is a structural block diagram of the intelligent early warning system based on lightning data of the present invention.

[0056] Figure 2 It is a structural block diagram of the security level setting module of the present invention. DETAILED DESCRIPTION

[0057] In order to make the technical means, creative features, objectives and effects of the present invention easy to understand, the present invention will be further described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] In order to address the shortcomings of the existing technology, such as Figure 1 As shown, the present invention provides an intelligent early warning system based on lightning data, comprising:

[0059] Lightning detection terminals are distributed in each target area and are used to obtain radar meteorological data, atmospheric electric field data and lightning activity data of different lightning activity events in each target area. Each lightning detection terminal is assigned a unique identifier. The lightning detection device is connected to the cloud data platform through a local area network;

[0060] The cloud data platform is used to obtain radar meteorological data, atmospheric electric field data and lightning activity data of different lightning activity events in each target area collected in real time by the lightning terminal, dynamically update in real time based on the collected data, and store the collected historical data; it is also used to divide the regions according to each target area, obtain the geographical feature data of each divided area, and set the safety level parameters of the divided areas on each target area based on the geographical feature data;

[0061] The characteristic data acquisition module is used to obtain historical data from the cloud data platform, calculate the correlation between radar meteorological data and atmospheric electric field data and ground-to-ground lightning, and screen out relevant data with a high correlation with the intensity of ground-to-ground lightning in the radar meteorological data and atmospheric electric field data, and record them as lightning-related characteristic data;

[0062] The lightning prediction module is used to train the lightning probability prediction model to predict the lightning probability of each divided area in the current lightning activity cycle; to train the lightning intensity prediction model to predict the lightning intensity of ground-to-ground lightning in the short term; based on the lightning prediction probability of each divided area in the target area and the lightning prediction intensity of ground-to-ground lightning, calculate the lightning risk parameters of each divided area, and mark the potential risk areas based on the lightning risk parameters;

[0063] The lightning warning module is used to calculate the lightning warning parameters of the potential risk area according to the safety level parameters, lightning prediction probability and lightning prediction intensity of the divided areas in the potential risk area. The set parameter threshold is based on the lightning warning parameter assessment to determine whether to trigger the alarm information; if the alarm is triggered, the alarm information of the target area is published and the lightning warning image of the target area is generated and published to the administrator of the target area; if the alarm is not triggered, continuous monitoring is performed.

[0064] Specifically, the lightning detection terminal specifically includes:

[0065] A meteorological data detection unit, which captures radar meteorological data in the target area through a meteorological radar, wherein the radar meteorological data includes reflectivity, radial velocity, spectrum width information, vertical liquid water content, echo top height and echo intensity;

[0066] An electric field data detection unit, which captures the atmospheric electric field data in the target area in real time through an atmospheric electric field instrument, wherein the atmospheric electric field data includes electric field time series differential data, electric field intensity change rate and charge density;

[0067] The Fengyun satellite detection unit captures lightning activity data of single lightning through the satellite's lightning imager, and uses a clustering algorithm to classify single lightning that are close in time and space as the same lightning activity event, so as to classify the lightning activity data of single lightning into the corresponding lightning activity event, and obtain lightning activity data in different lightning activity events in the target area, wherein the lightning activity data includes time information, lightning location data, lightning intensity data and lightning type data;

[0068] A lightning data storage unit, used to store radar meteorological data, atmospheric electric field data and lightning activity data collected by the meteorological data detection unit, the electric field data detection unit and the Fengyun satellite detection unit;

[0069] The lightning data extraction unit is used to record the most recent lightning activity event as the current lightning activity event, and record all lightning activity events before the current lightning activity event as historical lightning activity events. When the Fengyun satellite detection unit collects the latest lightning activity data, it uses a clustering algorithm to perform cluster analysis on it and the lightning activity data in the current lightning activity event to determine whether the latest lightning activity data belongs to the current lightning activity event, and updates the current lightning activity event and the historical lightning activity event according to the clustering result; it is used to set the current lightning activity cycle and the historical lightning activity cycle, and the current lightning activity cycle and the historical lightning activity cycle are used to reflect the current lightning activity respectively. The relevant time period of the lightning activity event and the historical activity event; set the collection frequency, extract the radar meteorological data and atmospheric electric field data in the historical lightning activity cycle and the current lightning activity cycle from the lightning data storage unit according to the collection frequency, and align the radar meteorological data and the atmospheric electric field data through the timestamp; extract the lightning activity data collected in the historical lightning activity cycle and the current lightning activity cycle, find the closest timestamp through the nearest neighbor method based on the time information of the lightning activity data, and align the lightning activity data with the timestamp. If there is no lightning activity data on other timestamps, use the flag bit to indicate that the lightning activity data of the timestamp is missing and no lightning activity has occurred;

[0070] A lightning data transmission unit is used to package the radar meteorological data, atmospheric electric field data and lightning activity data in the historical lightning activity cycle extracted by the lightning data extraction unit into a historical data set, and to package the radar meteorological data, atmospheric electric field data and lightning activity data in the current lightning activity cycle extracted by the lightning data extraction unit into a real-time data set, and the real-time data set is updated in real time based on the latest data obtained;

[0071] It should be noted that the setting of the current lightning activity cycle and the historical lightning activity cycle specifically includes:

[0072] In the lightning data extraction unit, the historical lightning activity cycle and the current lightning activity cycle are used to represent the relevant time periods of the historical lightning activity events and the current lightning activity events respectively;

[0073] For the historical lightning activity cycle, extract the time information of all lightning activity data therein, extend the earliest time information forward by a preset time as the start time of the historical lightning activity cycle, and extend the latest time information backward by a preset time as the end time of the historical lightning activity cycle;

[0074] For the current lightning activity cycle, extract the time information of all lightning activity data in the current lightning activity event, extend the earliest time information forward by a preset time as the start time of the current lightning activity cycle, and the start time of the current lightning activity cycle until the current actual time is the current lightning activity cycle, and the current lightning activity cycle is updated and adjusted based on the latest lightning activity data obtained;

[0075] Specifically, the preset time can be set to 2 hours, and the preset time can be dynamically adjusted according to meteorological conditions, geographical location and seasonal changes;

[0076] In the above-mentioned lightning data extraction unit, by setting the current lightning activity cycle and the historical lightning activity cycle, the timeliness and continuity of the data are ensured through a dynamic update mechanism, which provides a guarantee for the continuous operation of the early warning system.

[0077] Specifically, the cloud data platform includes:

[0078] The cloud data acquisition module is used to acquire the historical data set and the real-time data set collected by the lightning detection terminal in each target area; and the same identifier is assigned to the historical data set and the real-time data set collected by the lightning detection terminal in the same target area;

[0079] The cloud data processing module is used to clean the radar meteorological data, atmospheric electric field data and lightning activity data in the historical data set and real-time data set acquired by the cloud data acquisition module, remove abnormal values ​​and noise data, and perform real-time standardization processing to ensure the consistency and comparability of the data;

[0080] A cloud data storage module, used to store historical data sets processed by the cloud data processing module;

[0081] The security level setting module divides the regions based on each target area, obtains the geographical feature data of each divided region, and sets the security level parameters of the divided regions on each target area based on the geographical feature data.

[0082] Specifically, the feature data acquisition module includes:

[0083] A historical data acquisition unit, used to acquire historical data sets in the cloud data storage module;

[0084] The feature data extraction unit is used to select multiple historical data sets as samples, calculate the correlation between the radar meteorological data and the atmospheric electric field data and the lightning intensity of the ground-to-ground type based on the Pearson correlation coefficient, and screen out the relevant data with a high correlation with the lightning intensity of the ground-to-ground type in the radar meteorological data and the atmospheric electric field data, and record them as lightning-related feature data.

[0085] Specifically, the lightning prediction module specifically includes:

[0086] The coverage degree calculation unit is used to obtain the historical lightning activity cycle of the historical lightning activity events corresponding to different historical data sets in the cloud data storage module; obtain the atmospheric electric field parameters of various points in the target area within the historical lightning activity cycle, and obtain the relevant areas affected by the lightning activity events in the target area based on the changes in the atmospheric electric field parameters within the historical activity cycle; calculate the coverage degree of each divided area based on the coverage area of ​​the relevant area in each divided area in the target area;

[0087] It should be noted that the relevant area refers to the area affected by lightning activity in the target area; when calculating the coverage area of ​​each divided area in the target area, the spatial overlay analysis tool can be used to calculate the intersection area of ​​each divided area and the relevant area, and the obtained intersection area is the coverage area; calculate the coverage degree, which is the ratio of the coverage area to the area of ​​the divided area;

[0088] The lightning probability prediction model training unit is used to obtain the altitude and soil resistivity of each divided area; obtain the coverage of each divided area in different historical lightning activity cycles; obtain the frequency of lightning activity in each divided area in different historical lightning activity cycles; pre-process the above-obtained data and combine them into a feature vector set; train the logistic regression model through the feature vector set, adjust the model parameters to optimize the prediction performance, and use the trained model for lightning probability prediction of each divided area in the current lightning activity cycle;

[0089] The lightning intensity prediction model training unit is used to obtain the historical data set in the cloud data storage module, and extract the lightning-related characteristic data and lightning activity data in the historical data set; the above-extracted data is used as the data set, and the data set is divided into a training set and a test set, and the training set is used to train the selected neural network model, and the trained model is used to predict the lightning intensity of ground-to-ground lightning in the short term; during the training process of the lightning intensity prediction model, cross-validation and grid search are used to perform hyperparameter tuning to optimize the performance of the lightning intensity prediction model;

[0090] A lightning prediction unit is used to receive the real-time data set from the cloud data processing module, input the radar meteorological data, atmospheric electric field data and lightning activity data in the real-time data set into the lightning probability prediction model and the lightning intensity prediction model, so as to output the lightning prediction probability of each divided area in the target area and the lightning prediction intensity of ground-to-ground lightning in real time;

[0091] A potential risk marking unit is used to obtain the lightning prediction probability and lightning prediction intensity of ground-to-ground lightning in each divided area of ​​the target area, calculate the lightning risk parameter of each divided area based on the obtained prediction data, set the risk parameter threshold, and when the lightning risk parameter is greater than the risk parameter threshold, mark the divided area as a potential risk area, and send a risk signal to the lightning warning module;

[0092] It should be noted that the calculation formula for lightning risk parameters is:

[0093] X n =ρ 1 A n ·ρ 2 B;

[0094] Where, X n represents the lightning risk parameter of the nth divided area, A n represents the lightning prediction probability of the nth divided area, B represents the lightning prediction intensity, ρ 1 , 2 They represent weight parameters respectively.

[0095] Specifically, Figure 2 As shown, the security level setting module includes:

[0096] Regional division unit, based on the target area, uses GIS tools to generate grids, divides the area, and assigns a unique ID to each divided area;

[0097] A geographic feature data acquisition unit, which acquires geographic feature data of each divided area, wherein the geographic feature data includes soil resistivity, building type information, protection measures information, population density information, and power grid layout information;

[0098] A safety level calculation unit is used to score the obtained geographical feature data according to the scoring system to obtain the following scoring data: building type score, building material score, building height score, lightning protection facility score, lightning protection facility maintenance status score, population density score, power grid layout score, and power grid equipment aging degree score;

[0099] The calculation formula of the security level parameter is as follows:

[0100]

[0101] In the formula, S x represents the security level parameter for dividing area x, j represents the total number of scoring data, η i represents the weight parameter of the i-th scoring data, k i Represents the i-th rating data;

[0102] It should be noted that the above rating range is 1-10, where 1 represents the lowest impact and 10 represents the highest impact;

[0103] Specifically, the lightning warning module includes:

[0104] A threshold management unit, which sets a first parameter threshold and a second parameter threshold of a lightning warning based on historical data and expert experience, and dynamically adjusts the first parameter threshold and the second parameter threshold according to actual conditions;

[0105] The warning trigger unit is used to obtain the risk signal sent by the potential risk marking unit; based on the safety level parameters, lightning prediction probability and lightning prediction intensity of the divided areas in the potential risk area, the lightning warning parameters of the potential risk area are calculated, and the parameter threshold is set based on the lightning warning parameter evaluation to determine whether to trigger the alarm information;

[0106] The warning image generation unit is used to obtain the map of the target area, the lightning prediction probability of the corresponding divided area, the lightning prediction intensity parameters and the warning information, and generate a lightning warning image through Matplotlib with the above data, using color coding to indicate the warning level of the grid where the potential risk area is located;

[0107] The warning information issuing unit sends the warning information sent by the warning triggering unit and the lightning warning image generated by the warning image generating unit to the administrator end of each target area;

[0108] Specifically, in the warning trigger unit, the lightning warning parameter is the safety level parameter of the divided area, the lightning prediction probability of the divided area, and the weighted lightning prediction intensity; in the warning trigger unit, a first parameter threshold and a second parameter threshold are set, and the lightning warning parameter determines the alarm level through the first parameter threshold and the second parameter threshold; when the warning parameter is greater than the first parameter threshold, a red thunderstorm warning message is triggered; when the warning parameter is less than the second parameter threshold, the warning message is not triggered; when the warning parameter is greater than / equal to the second parameter threshold and less than / equal to the first parameter threshold, a yellow thunderstorm warning message is triggered.

[0109] In summary, the advantages of the present invention are as follows: the early warning system provided by the present invention comprehensively collects radar meteorological data, atmospheric electric field data and lightning activity data through lightning detection terminals distributed in each target area; real-time processing and storage are performed through a cloud data platform to ensure the integrity and timeliness of the data; the historical and current lightning activity cycles are dynamically set and the preset time is flexibly adjusted to improve the adaptability and continuity of the data; the feature data acquisition module accurately extracts data related to the intensity of ground-to-ground lightning, and the lightning prediction module provides scientific lightning activity predictions by training intensity prediction probability models and intensity prediction models; the early warning system calculates lightning risk parameters, marks potential risk areas, and combines safety level parameters to efficiently evaluate and trigger warning information, generate intuitive lightning warning images, and ensure the timely transmission and effective use of information, thereby helping relevant departments to take timely measures to reduce the losses caused by lightning disasters.

[0110] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. Intelligent early warning system based on lightning data, characterized by: include: Lightning detection terminals are distributed in each target area and are used to obtain radar meteorological data, atmospheric electric field data and lightning activity data of different lightning activity events in each target area. Each lightning detection terminal is assigned a unique identifier. The lightning detection device is connected to the cloud data platform through a local area network; The cloud data platform is used to obtain radar meteorological data, atmospheric electric field data and lightning activity data of different lightning activity events in each target area collected in real time by the lightning terminal, dynamically update in real time based on the collected data, and store the collected historical data; it is also used to divide the regions according to each target area, obtain the geographical feature data of each divided area, and set the safety level parameters of the divided areas on each target area based on the geographical feature data; The characteristic data acquisition module is used to obtain historical data from the cloud data platform, calculate the correlation between radar meteorological data and atmospheric electric field data and ground-to-ground lightning, and screen out relevant data with a high correlation with the intensity of ground-to-ground lightning in the radar meteorological data and atmospheric electric field data, and record them as lightning-related characteristic data; The lightning prediction module is used to train the lightning probability prediction model to predict the lightning probability of each divided area in the current lightning activity cycle; it is used to train the lightning intensity prediction model to predict the lightning intensity of ground-to-ground lightning in the short term; Based on the lightning prediction probability of each divided area in the target area and the lightning prediction intensity of ground-to-ground lightning, the lightning risk parameters of each divided area are calculated, and the potential risk areas are marked based on the lightning risk parameters; The lightning warning module is used to calculate the lightning warning parameters of the potential risk area according to the safety level parameters, lightning prediction probability and lightning prediction intensity of the divided areas in the potential risk area. The parameter threshold is set based on the lightning warning parameter evaluation to determine whether to trigger the alarm information; If an early warning is triggered, the warning information of the target area is released and a lightning early warning image of the target area is generated and released to the administrator of the target area; If no warning is triggered, continue monitoring.

2. The intelligent early warning system based on lightning data according to claim 1 is characterized in that: The lightning detection terminal specifically comprises: A meteorological data detection unit, which captures radar meteorological data in the target area through a meteorological radar, wherein the radar meteorological data includes reflectivity, radial velocity, spectrum width information, vertical liquid water content, echo top height and echo intensity; An electric field data detection unit, which captures the atmospheric electric field data in the target area in real time through an atmospheric electric field instrument, wherein the atmospheric electric field data includes electric field time series differential data, electric field intensity change rate and charge density; The Fengyun satellite detection unit captures lightning activity data of single lightning through the satellite's lightning imager, and uses a clustering algorithm to classify single lightning that are close in time and space as the same lightning activity event, so as to classify the lightning activity data of single lightning into the corresponding lightning activity event, and obtain lightning activity data in different lightning activity events in the target area, wherein the lightning activity data includes time information, lightning location data, lightning intensity data and lightning type data; A lightning data storage unit, used to store radar meteorological data, atmospheric electric field data and lightning activity data collected by the meteorological data detection unit, the electric field data detection unit and the Fengyun satellite detection unit; The lightning data extraction unit is used to record the most recent lightning activity event as the current lightning activity event, and record all lightning activity events before the current lightning activity event as historical lightning activity events. When the Fengyun satellite detection unit collects the latest lightning activity data, it uses a clustering algorithm to perform cluster analysis on it and the lightning activity data in the current lightning activity event to determine whether the latest lightning activity data belongs to the current lightning activity event, and updates the current lightning activity event and the historical lightning activity event according to the clustering result; it is used to set the current lightning activity cycle and the historical lightning activity cycle, and the current lightning activity cycle and the historical lightning activity cycle are used to reflect the current lightning activity respectively. The relevant time period of the lightning activity event and the historical activity event; set the collection frequency, extract the radar meteorological data and atmospheric electric field data in the historical lightning activity cycle and the current lightning activity cycle from the lightning data storage unit according to the collection frequency, and align the radar meteorological data and the atmospheric electric field data through the timestamp; extract the lightning activity data collected in the historical lightning activity cycle and the current lightning activity cycle, find the closest timestamp through the nearest neighbor method based on the time information of the lightning activity data, and align the lightning activity data with the timestamp. If there is no lightning activity data on other timestamps, use the flag bit to indicate that the lightning activity data of the timestamp is missing and no lightning activity has occurred; The lightning data transmission unit is used to package the radar meteorological data, atmospheric electric field data and lightning activity data in the historical lightning activity cycle extracted by the lightning data extraction unit into a historical data set, and to package the radar meteorological data, atmospheric electric field data and lightning activity data in the current lightning activity cycle extracted by the lightning data extraction unit into a real-time data set, which is updated in real time based on the latest data obtained.

3. The intelligent early warning system based on lightning data according to claim 2 is characterized in that: The cloud data platform specifically includes: The cloud data acquisition module is used to acquire the historical data set and the real-time data set collected by the lightning detection terminal in each target area; and the same identifier is assigned to the historical data set and the real-time data set collected by the lightning detection terminal in the same target area; The cloud data processing module is used to clean the radar meteorological data, atmospheric electric field data and lightning activity data in the historical data set and real-time data set acquired by the cloud data acquisition module, remove abnormal values ​​and noise data, and perform real-time standardization processing; A cloud data storage module, used to store historical data sets processed by the cloud data processing module; The security level setting module divides the regions based on each target area, obtains the geographical feature data of each divided region, and sets the security level parameters of the divided regions on each target area based on the geographical feature data.

4. The intelligent early warning system based on lightning data according to claim 3 is characterized in that: The feature data acquisition module specifically includes: A historical data acquisition unit, used to acquire historical data sets in the cloud data storage module; The feature data extraction unit is used to select multiple historical data sets as samples, calculate the correlation between the radar meteorological data and the atmospheric electric field data and the lightning intensity of the ground-to-ground type based on the Pearson correlation coefficient, and screen out the relevant data with a high correlation with the lightning intensity of the ground-to-ground type in the radar meteorological data and the atmospheric electric field data, and record them as lightning-related feature data.

5. The intelligent early warning system based on lightning data according to claim 4 is characterized in that: The lightning prediction module specifically includes: The coverage degree calculation unit is used to obtain the historical lightning activity cycle of the historical lightning activity events corresponding to different historical data sets in the cloud data storage module; obtain the atmospheric electric field parameters of various points in the target area within the historical lightning activity cycle, and obtain the relevant areas affected by the lightning activity events in the target area based on the changes in the atmospheric electric field parameters within the historical activity cycle; calculate the coverage degree of each divided area based on the coverage area of ​​the relevant area in each divided area in the target area; The lightning probability prediction model training unit is used to obtain the altitude and soil resistivity of each divided area; obtain the coverage of each divided area in different historical lightning activity cycles; obtain the frequency of lightning activity in each divided area in different historical lightning activity cycles; pre-process the above-obtained data and combine them into a feature vector set; train the logistic regression model through the feature vector set, adjust the model parameters to optimize the prediction performance, and use the trained model for lightning probability prediction of each divided area in the current lightning activity cycle; The lightning intensity prediction model training unit is used to obtain the historical data set in the cloud data storage module, and extract the lightning-related characteristic data and lightning activity data in the historical data set; the above-extracted data is used as the data set, and the data set is divided into a training set and a test set, and the training set is used to train the selected neural network model, and the trained model is used to predict the lightning intensity of ground-to-ground lightning in the short term; during the training process of the lightning intensity prediction model, cross-validation and grid search are used to perform hyperparameter tuning to optimize the performance of the lightning intensity prediction model; A lightning prediction unit is used to receive the real-time data set from the cloud data processing module, input the radar meteorological data, atmospheric electric field data and lightning activity data in the real-time data set into the lightning probability prediction model and the lightning intensity prediction model, so as to output the lightning prediction probability of each divided area in the target area and the lightning prediction intensity of ground-to-ground lightning in real time; The potential risk marking unit is used to obtain the lightning prediction probability and lightning prediction intensity of ground-to-ground lightning in each divided area of ​​the target area, calculate the lightning risk parameters of each divided area based on the obtained prediction data, set the risk parameter threshold, and when the lightning risk parameter is greater than the risk parameter threshold, mark the divided area as a potential risk area and send a risk signal to the lightning warning module.

6. The intelligent early warning system based on lightning data according to claim 5 is characterized in that: The security level setting module specifically includes: Regional division unit, based on the target area, uses GIS tools to generate grids, divides the area, and assigns a unique ID to each divided area; A geographic feature data acquisition unit, which acquires geographic feature data of each divided area, wherein the geographic feature data includes building type information, protection measures information, population density information, and power grid layout information; The security level calculation unit is used to calculate the security level parameters of each divided area based on the geographical feature data.

7. The intelligent early warning system based on lightning data according to claim 6 is characterized in that: The lightning warning module specifically includes: A threshold management unit, which sets a first parameter threshold and a second parameter threshold of a lightning warning based on historical data and expert experience, and dynamically adjusts the first parameter threshold and the second parameter threshold according to actual conditions; The warning trigger unit is used to obtain the risk signal sent by the potential risk marking unit; based on the safety level parameters, lightning prediction probability and lightning prediction intensity of the divided areas in the potential risk area, the lightning warning parameters of the potential risk area are calculated, and the parameter threshold is set based on the lightning warning parameter evaluation to determine whether to trigger the alarm information; The warning image generation unit is used to obtain the map of the target area, the lightning prediction probability of the corresponding divided area, the lightning prediction intensity parameters and the warning information, and generate a lightning warning image with Matplotlib using the above data, using color coding to indicate the warning level of the grid where the potential risk area is located; The warning information issuing unit sends the warning information initiated by the warning triggering unit and the lightning warning image generated by the warning image generating unit to the administrator end of each target area.

8. The intelligent early warning system based on lightning data according to claim 6 is characterized in that: include: The security level calculation unit is used to calculate the security level parameters of each divided area according to the geographical feature data, including: The obtained geographical feature data were scored using a scoring system to obtain the following scoring data: building type score, building material score, building height score, lightning protection facility score, lightning protection facility maintenance status score, population density score, power grid layout score, and power grid equipment aging score; The calculation formula of the security level parameter is as follows: In the formula, S x represents the security level parameter for dividing area x, j represents the total number of scoring data, η i represents the weight parameter of the i-th scoring data, k i Represents the i-th rating data.

9. The intelligent early warning system based on lightning data according to claim 8 is characterized in that: include: In the warning trigger unit, a first parameter threshold and a second parameter threshold are set, and the lightning warning parameter determines the alarm level through the first parameter threshold and the second parameter threshold; When the warning parameter is greater than the first parameter threshold, a red thunderstorm warning message is triggered; When the warning parameter is less than the second parameter threshold, no alarm information is triggered; When the warning parameter is greater than / equal to the second parameter threshold and less than / equal to the first parameter threshold, a yellow thunderstorm warning message is triggered.

10. The intelligent early warning system based on lightning data according to claim 9 is characterized in that: The setting of the current lightning activity cycle and the historical lightning activity cycle specifically includes: In the lightning data extraction unit, the historical lightning activity cycle and the current lightning activity cycle are used to represent the relevant time periods of the historical lightning activity events and the current lightning activity events respectively; For the historical lightning activity cycle, extract the time information of all lightning activity data therein, extend the earliest time information forward by a preset time as the start time of the historical lightning activity cycle, and extend the latest time information backward by a preset time as the end time of the historical lightning activity cycle; For the current lightning activity cycle, the time information of all lightning activity data in the current lightning activity event is extracted, and the earliest time information is extended forward by a preset time as the start time of the current lightning activity cycle. The start time of the current lightning activity cycle and the current actual time are all the current lightning activity cycle, and the current lightning activity cycle is updated and adjusted based on the latest lightning activity data obtained.

Images