Lightning protection monitoring method and system

Abstract

The invention provides a lightning protection monitoring method and system, and the method comprises the steps: obtaining a previous atmospheric weighted temperature data set, a previous thundercloud observation data set and a previous radar echo data set, so as to train and construct an initial thundercloud prediction neural network, and further obtaining a final thundercloud prediction neural network; obtaining a plurality of predicted thundercloud pictures based on the final thundercloud prediction neural network, converting the predicted thundercloud pictures into a thundercloud grid chart, and obtaining a plurality of thunderstorm grid regions in the thundercloud grid chart; and obtaining a predicted thunderstorm moving path based on the thunderstorm grid region in the thundercloud grid chart under the adjacent time frames so as to carry out lightning protection. Through the final thundercloud prediction neural network, preliminary thunderstorm prediction is carried out, a thunderstorm moving path is obtained and predicted through the prediction thundercloud picture, an accurate prediction result of the position and time of the thunderstorm is completed, and then through combination with passive thunderstorm protection measures, the thunderstorm protection accuracy and the thunderstorm protection effect are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data prediction, and in particular to a lightning protection monitoring method and system. Background Art

[0002] Lightning is a type of discharge phenomenon accompanied by lightning and thunder. Its generation process often occurs inside cumulonimbus clouds under severe convective weather conditions. Lightning is a natural phenomenon with great destructive power and suddenness, which brings huge challenges and risks to the safe and stable operation of power systems.

[0003] For the power system, when lightning strikes directly or is induced into the transmission network, it will generate transient overvoltage, which is very likely to cause insulation breakdown, equipment damage, and even cause large-scale power outages.

[0004] Although traditional lightning protection measures can protect electrical equipment and transmission lines from damage caused by lightning to a certain extent, they are basically passive lightning protection measures, such as lightning rods, lightning arresters and grounding systems. Although passive protection methods can absorb and disperse the energy of lightning, they cannot predict the specific location and time of lightning in advance, making it difficult to provide accurate protection. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a lightning protection monitoring method and system, which aims to solve the technical problem that the prior art uses passive protection measures such as lightning rods, lightning arresters and grounding systems to protect against lightning, but it is impossible to predict the specific location and time of lightning in advance, making it difficult to provide accurate protection.

[0006] In order to achieve the above objectives, in a first aspect, an embodiment of the present application provides a lightning protection monitoring method, comprising the following steps: Obtain the past atmospheric weighted temperature dataset, past thundercloud observation dataset and past radar echo dataset corresponding to the prediction area; Constructing an initial thundercloud prediction neural network, and training the initial thundercloud prediction neural network into a final thundercloud prediction neural network based on the past atmospheric weighted temperature data set, the past thundercloud observation data set, and the past radar echo data set; Based on the final thundercloud prediction neural network, a plurality of predicted thundercloud images in continuous time frames corresponding to the prediction area are obtained, the predicted thundercloud images are converted into thundercloud grid images, and a plurality of thunderstorm grid areas are obtained in the thundercloud grid images; A predicted thunderstorm moving path is obtained based on the thunderstorm grid area in the thundercloud grid map in adjacent time frames, and lightning protection is performed based on the predicted thunderstorm moving path.

[0007] Compared with the prior art, the beneficial effects of the present invention are as follows: the final thundercloud prediction neural network is obtained through the past atmospheric weighted temperature data set, the past thundercloud observation data set and the past radar echo data set, and a number of the predicted thundercloud images are obtained based on the final thundercloud prediction neural network to complete a preliminary thunderstorm prediction for the prediction area, that is, the specific location and time of the lightning are predicted in advance, thereby improving the accuracy of lightning protection; the predicted thunderstorm movement path is obtained according to the predicted thundercloud image, and a more detailed prediction result of the location and time of the thunderstorm is completed, and the accuracy and effect of lightning protection are further improved by combining with passive lightning protection measures; the past atmospheric weighted temperature data set with a strong correlation with the generation of thunderstorms is introduced on the basis of the past thundercloud observation data set and the past radar echo data set, thereby effectively improving the prediction accuracy of the final thundercloud prediction neural network and improving the acquisition accuracy of the predicted thundercloud image.

[0008] Further, the past atmospheric weighted temperature dataset, the past thundercloud observation dataset and the past radar echo dataset respectively include past atmospheric weighted temperatures, past thundercloud observation images and past radar echo images in a number of continuous time frames, and the step of acquiring the past atmospheric weighted temperature dataset corresponding to the prediction area includes: Obtain the relative humidity and saturated water vapor pressure of different atmospheric layers in continuous time frames corresponding to the forecast area; In the same time frame, obtaining the average water vapor pressure of the corresponding atmosphere based on the relative humidity and the saturated water vapor pressure; The past atmospheric weighted temperatures are obtained through a plurality of the average water vapor pressures, and a plurality of the past atmospheric weighted temperatures in a continuous time frame are combined into a past atmospheric weighted temperature data set.

[0009] Furthermore, the formula for obtaining the average water vapor pressure is: , in, represents the average water vapor pressure of the ith atmospheric layer, represents the relative humidity of the ith atmospheric layer, represents the saturated water vapor pressure of the i-th atmospheric layer; The formula for obtaining the weighted atmospheric temperature in the past period is: , in, represents the past atmospheric weighted temperature in the jth time frame, represents the average temperature of the i-th atmospheric layer, represents the thickness of the i-th atmospheric layer.

[0010] Furthermore, the initial thundercloud prediction neural network includes a thundercloud prediction module, an echo prediction module and a temperature prediction module, and the step of training the initial thundercloud prediction neural network into a final thundercloud prediction neural network based on the past atmospheric weighted temperature data set, the past thundercloud observation data set and the past radar echo data set includes: Acquire a predicted radar echo atlas through the previous radar echo data set and the echo prediction module; Obtaining a predicted atmospheric weighted temperature set through the past atmospheric weighted temperature data set and the temperature prediction module; Acquire a predicted thundercloud atlas through the previous thundercloud observation data set, the predicted radar echo atlas, the predicted atmospheric weighted temperature set and the thundercloud prediction module; A loss function is constructed by using the predicted thundercloud atlas, the real thundercloud observation atlas, the predicted atmospheric weighted temperature set, the real atmospheric weighted temperature set, the predicted radar echo atlas and the real radar echo atlas. The thundercloud prediction module, the echo prediction module and the temperature prediction module are corrected based on the loss function to obtain a final thundercloud prediction neural network.

[0011] Furthermore, the acquisition formula of the predicted radar echo atlas is: , in, Indicates the predicted radar echo atlas from the next frame 1 to the next frame a at time t, It represents the set of past radar echo images from the past b frames to the current 0 frame at time t. represents the echo prediction module; The formula for obtaining the predicted atmospheric weighted temperature set is: , in, represents the atmospheric weighted temperature set from the next frame to the next frame a at time t, represents the set of past atmospheric weighted temperatures from the past b frames to the current 0 frame at time t, represents the temperature prediction module; The formula for obtaining the predicted thundercloud atlas is: , in, represents the predicted thundercloud atlas from 1 frame to a frame in the future at time t, represents the set of past thundercloud observation images from time t, from the past b frames to the current 0 frame. Represents the thundercloud prediction module; The formula of the loss function is: , in, It represents the real thundercloud observation atlas from the future frame 1 to the future frame a at time t, It represents the real radar echo atlas from the next frame 1 to the next frame a at time t. represents the real atmospheric weighted temperature set from the future frame 1 to the future frame a at time t, represents the loss function, , Both represent weight coefficients, , , They represent the module parameters of the echo prediction module, temperature prediction module and thundercloud prediction module respectively.

[0012] Furthermore, the thundercloud grid map includes a plurality of grid points of the same area, and the step of obtaining a plurality of thunderstorm grid areas in the thundercloud grid map includes: Determine whether the radar echo intensity of each of the grid points is higher than an intensity threshold, and select the grid points corresponding to the radar echo intensities higher than the intensity threshold as thunderstorm grid points; It is determined whether the thunderstorm grid points are adjacent to each other, so as to combine the adjacent thunderstorm grid points into a thunderstorm grid area.

[0013] Furthermore, the step of obtaining a predicted thunderstorm moving path based on the thunderstorm grid area in the thundercloud grid map in adjacent time frames includes: Obtaining a grid distance from each of the thunderstorm grid areas in the thundercloud grid map in a previous time frame to each of the thunderstorm grid areas in the thundercloud grid map in a subsequent time frame; Acquire a distance threshold based on a plurality of the grid distances, compare the grid distance with the distance threshold, and select the grid distance that is smaller than the distance threshold as a moving sub-path; A plurality of the moving sub-paths are connected in series to form a predicted thunderstorm moving path.

[0014] Furthermore, the calculation formula of the grid distance is: , in, Indicates the grid distance between the mth thunderstorm grid area in the previous time frame and the nth thunderstorm grid area in the next time frame. , Both represent weight functions, Indicates the distance between the confidence ellipse of the m-th thunderstorm grid area in the previous time frame and the center point of the n-th thunderstorm grid area in the next time frame. Represents the square root of the absolute value of the area difference between the confidence ellipse of the m-th thunderstorm grid area in the previous time frame and the confidence ellipse of the n-th thunderstorm grid area in the next time frame.

[0015] In a second aspect, an embodiment of the present application provides a lightning protection monitoring system, which is applied to the lightning protection monitoring method as described in the first aspect above, and the system includes: An acquisition module is used to acquire past atmospheric weighted temperature datasets, past thundercloud observation datasets, and past radar echo datasets corresponding to the prediction area; A construction module, used to construct an initial thundercloud prediction neural network, and train the initial thundercloud prediction neural network into a final thundercloud prediction neural network based on the past atmospheric weighted temperature data set, the past thundercloud observation data set and the past radar echo data set; A prediction module, used for obtaining a plurality of predicted thundercloud images in continuous time frames corresponding to the prediction area based on the final thundercloud prediction neural network, converting the predicted thundercloud images into thundercloud grid images, and obtaining a plurality of thunderstorm grid areas in the thundercloud grid images; The analysis module is used to obtain a predicted thunderstorm movement path based on the thunderstorm grid area in the thundercloud grid map under adjacent time frames, and perform lightning protection based on the predicted thunderstorm movement path.

[0016] In a third aspect, an embodiment of the present application provides a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the lightning protection monitoring method as described in the first aspect above is implemented.

[0017] In a fourth aspect, an embodiment of the present application provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the lightning protection monitoring method as described in the first aspect above is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 Flow chart of the lightning protection monitoring method in the first embodiment of the present invention; Figure 2 It is a structural block diagram of a lightning protection monitoring system in a second embodiment of the present invention; The following specific implementation manner will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0019] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0020] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0022] See also Figure 1 The lightning protection monitoring method provided by the first embodiment of the present invention comprises the following steps: S10: Acquire a past atmospheric weighted temperature dataset, a past thundercloud observation dataset, and a past radar echo dataset corresponding to the prediction area; The past atmospheric weighted temperature dataset, the past thundercloud observation dataset and the past radar echo dataset respectively include past atmospheric weighted temperatures, past thundercloud observation images and past radar echo images in a number of continuous time frames. It should be noted that the time frames in the past atmospheric weighted temperature dataset, the past thundercloud observation dataset and the past radar echo dataset are all the same time frames, and the past atmospheric weighted temperatures, the past thundercloud observation images and the past radar echo images in each same time frame correspond to each other.

[0023] The step S10 comprises: S110: Obtaining relative humidity and saturated water vapor pressure of different atmospheric layers in continuous time frames corresponding to the prediction area; In this embodiment, the relative humidity and the saturated water vapor pressure can be collected by a meteorological sensor. It can be understood that the relative humidity and the saturated water vapor pressure exist in each atmospheric layer.

[0024] S120: In the same time frame, obtaining an average water vapor pressure of a corresponding atmospheric layer based on the relative humidity and the saturated water vapor pressure; The formula for obtaining the average water vapor pressure is: , in, represents the average water vapor pressure of the ith atmospheric layer, represents the relative humidity of the ith atmospheric layer, represents the saturated water vapor pressure of the i-th atmospheric layer.

[0025] S130: acquiring the past atmospheric weighted temperatures through a plurality of the average water vapor pressures, and combining the past atmospheric weighted temperatures in a continuous time frame into a past atmospheric weighted temperature data set; The formula for obtaining the weighted atmospheric temperature in the past period is: , in, represents the past atmospheric weighted temperature in the jth time frame, represents the average temperature of the i-th atmospheric layer, represents the thickness of the i-th atmospheric layer.

[0026] S20: constructing an initial thundercloud prediction neural network, and training the initial thundercloud prediction neural network into a final thundercloud prediction neural network based on the past atmospheric weighted temperature dataset, the past thundercloud observation dataset, and the past radar echo dataset; The initial thundercloud prediction neural network includes a thundercloud prediction module, an echo prediction module and a temperature prediction module.

[0027] The step S20 comprises: S210: Acquire a predicted radar echo atlas through the previous radar echo data set and the echo prediction module; The acquisition formula of the predicted radar echo atlas is: , in, Indicates the predicted radar echo atlas from the next frame 1 to the next frame a at time t, It represents the set of past radar echo images from the past b frames to the current 0 frame at time t. Represents the echo prediction module.

[0028] In this embodiment, the past radar echo images in the past radar echo data set can be segmented into time series, the past radar echo images at the front of the time series are used as input values ​​of the echo prediction module, and the past radar echo images at the back of the time series are used as output values ​​of the echo prediction module to train the echo prediction module, and then the predicted radar echo atlas set is obtained through the echo prediction module.

[0029] S220: Obtaining a predicted atmospheric weighted temperature set through the past atmospheric weighted temperature data set and the temperature prediction module; The formula for obtaining the predicted atmospheric weighted temperature set is: , in, represents the atmospheric weighted temperature set from the next frame to the next frame a at time t, represents the set of past atmospheric weighted temperatures from the past b frames to the current 0 frame at time t, The training method of the temperature prediction module is the same as that of the echo prediction module, and will not be described in detail here.

[0030] S230: Acquire a predicted thundercloud atlas through the previous thundercloud observation data set, the predicted radar echo atlas, the predicted atmospheric weighted temperature set and the thundercloud prediction module; The formula for obtaining the predicted thundercloud atlas is: , in, represents the predicted thundercloud atlas from 1 frame to a frame in the future at time t, represents the set of past thundercloud observation images from time t, from the past b frames to the current 0 frame. It is understandable that when acquiring the predicted thundercloud atlas, the predicted radar echo atlas and the predicted atmospheric weighted temperature set are both used as input values ​​of the thundercloud prediction module, so as to serve as a reference for the thundercloud prediction module to make predictions, thereby improving the acquisition accuracy of the predicted thundercloud atlas.

[0031] S240: constructing a loss function through the predicted thundercloud atlas, the real thundercloud observation atlas, the predicted atmospheric weighted temperature set, the real atmospheric weighted temperature set, the predicted radar echo atlas and the real radar echo atlas, and calibrating the thundercloud prediction module, the echo prediction module and the temperature prediction module based on the loss function to obtain a final thundercloud prediction neural network; The formula of the loss function is: , in, It represents the real thundercloud observation atlas from the future frame 1 to the future frame a at time t, It represents the real radar echo atlas from the next frame 1 to the next frame a at time t. represents the real atmospheric weighted temperature set from the future frame 1 to the future frame a at time t, represents the loss function, , Both represent weight coefficients, , , Respectively represent the module parameters of the echo prediction module, the temperature prediction module and the thundercloud prediction module. It should be noted that, in the process of training the initial thundercloud prediction neural network, the real thundercloud observation atlas, the real atmospheric weighted temperature set and the real radar echo atlas are respectively formed by segmenting the past atmospheric weighted temperature data set, the past thundercloud observation data set and the past radar echo data set based on the time series.

[0032] S30: acquiring a plurality of predicted thundercloud images in continuous time frames corresponding to the predicted areas based on the final thundercloud prediction neural network, converting the predicted thundercloud images into thundercloud grid images, and acquiring a plurality of thunderstorm grid areas in the thundercloud grid images; It can be understood that by taking the acquired real-time atmospheric weighted temperature, real-time radar echo map and real-time thundercloud observation map as the input value of the final thundercloud prediction neural network, a plurality of the predicted thundercloud maps can be output. The essence of converting the predicted thundercloud map into a thundercloud grid map is to form a plurality of grid points in the predicted thundercloud map by setting separators with the same row spacing and column spacing in the predicted thundercloud map, thereby forming the thundercloud grid map. That is, the thundercloud grid map includes a plurality of grid points with the same area.

[0033] The step S30 comprises: S310: determining whether the radar echo intensity of each of the grid points is higher than an intensity threshold, and selecting the grid points corresponding to the radar echo intensities higher than the intensity threshold as thunderstorm grid points; S320: Determine whether the thunderstorm grid points are adjacent to each other, so as to combine the adjacent thunderstorm grid points into a thunderstorm grid area; It can be understood that if the two thunderstorm grid points are not adjacent, they are determined to be different thunderstorm grid areas.

[0034] S40: acquiring a predicted thunderstorm moving path based on the thunderstorm grid area in the thundercloud grid map in adjacent time frames, and performing lightning protection based on the predicted thunderstorm moving path; The step S40 comprises: S410: Obtaining a grid distance from each thunderstorm grid area in the thundercloud grid map in the previous time frame to each thunderstorm grid area in the thundercloud grid map in the next time frame; The calculation formula of the grid distance is: , in, Indicates the grid distance between the mth thunderstorm grid area in the previous time frame and the nth thunderstorm grid area in the next time frame. , Both represent weight functions, Indicates the distance between the confidence ellipse of the m-th thunderstorm grid area in the previous time frame and the center point of the n-th thunderstorm grid area in the next time frame. Represents the square root of the absolute value of the area difference between the confidence ellipse of the m-th thunderstorm grid area in the previous time frame and the confidence ellipse of the n-th thunderstorm grid area in the next time frame.

[0035] S420: acquiring a distance threshold based on a plurality of the grid distances, comparing the grid distance with the distance threshold, and selecting the grid distance that is smaller than the distance threshold as a moving sub-path; Assume that there are thunderstorm grid area a and thunderstorm grid area b in the thundercloud grid map in the previous time frame, and there are thunderstorm grid area c and thunderstorm grid area d in the thundercloud grid map in the next time frame, the grid distance between a and c is 10, the grid distance between a and d is 20, and the distance threshold is 15, then the grid distance between a and c is selected as the moving sub-path.

[0036] It should be noted that if multiple thunderstorm grid areas in the thundercloud grid map in the previous time frame and a thunderstorm grid area in the thundercloud grid map in the next time frame are all selected as the moving sub-paths, it is determined that thunderstorm merging has occurred, and if a thunderstorm grid area in the thundercloud grid map in the previous time frame and multiple thunderstorm grid areas in the thundercloud grid map in the next time frame are all selected as the moving sub-paths, it is determined that thunderstorm splitting has occurred. Preferably, if there is a thunderstorm grid area in the thundercloud grid map in the next time frame that does not match the path, it is selected as a new thunderstorm.

[0037] S430: connecting a plurality of the moving sub-paths in series to form a predicted thunderstorm moving path; It can be understood that by performing the above operation between a plurality of thundercloud grid images in continuous time frames, the connection work between the moving sub-paths can be completed.

[0038] The final thundercloud prediction neural network is obtained through the past atmospheric weighted temperature data set, the past thundercloud observation data set and the past radar echo data set, and a number of the predicted thundercloud images are obtained based on the final thundercloud prediction neural network to complete the preliminary thunderstorm prediction of the prediction area, that is, the specific location and time of lightning occurrence are predicted in advance, thereby improving the accuracy of lightning protection; the predicted thunderstorm movement path is obtained according to the predicted thundercloud image, and a more detailed prediction result of the location and time of the thunderstorm is completed, and through the combination with passive lightning protection measures, the accuracy and effect of lightning protection are further improved; by introducing the past atmospheric weighted temperature data set that has a strong correlation with the generation of thunderstorms on the basis of the past thundercloud observation data set and the past radar echo data set, the prediction accuracy of the final thundercloud prediction neural network is effectively improved, and the acquisition accuracy of the predicted thundercloud image is improved.

[0039] See also Figure 2 The second embodiment of the present invention provides a lightning protection monitoring system, which is applied to the lightning protection monitoring method described in the above embodiment, and will not be repeated here. As used below, the terms "module", "unit", "subunit", etc. can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0040] The system comprises: An acquisition module 10 is used to acquire a past atmospheric weighted temperature dataset, a past thundercloud observation dataset, and a past radar echo dataset corresponding to the prediction area; The acquisition module 10 includes: The first unit is used to obtain the relative humidity and saturated water vapor pressure of different atmospheric layers in a continuous time frame corresponding to the prediction area; A second unit is used for obtaining an average water vapor pressure of a corresponding atmospheric layer based on the relative humidity and the saturated water vapor pressure in a same time frame; The third unit is used to obtain the past atmospheric weighted temperatures through a plurality of the average water vapor pressures, and combine the plurality of past atmospheric weighted temperatures in a continuous time frame into a past atmospheric weighted temperature data set; A construction module 20 is used to construct an initial thundercloud prediction neural network, and train the initial thundercloud prediction neural network into a final thundercloud prediction neural network based on the past atmospheric weighted temperature data set, the past thundercloud observation data set and the past radar echo data set; The building block 20 comprises: A fourth unit is used to obtain a predicted radar echo atlas through the previous radar echo data set and the echo prediction module; A fifth unit is used to obtain a predicted atmospheric weighted temperature set through the past atmospheric weighted temperature data set and the temperature prediction module; A sixth unit is used to obtain a predicted thundercloud atlas through the previous thundercloud observation data set, the predicted radar echo atlas, the predicted atmospheric weighted temperature set and the thundercloud prediction module; The seventh unit is used to construct a loss function through the predicted thundercloud atlas, the real thundercloud observation atlas, the predicted atmospheric weighted temperature set, the real atmospheric weighted temperature set, the predicted radar echo atlas and the real radar echo atlas, and calibrate the thundercloud prediction module, the echo prediction module and the temperature prediction module based on the loss function to obtain a final thundercloud prediction neural network; A prediction module 30 is used to obtain a plurality of predicted thundercloud images in continuous time frames corresponding to the prediction area based on the final thundercloud prediction neural network, convert the predicted thundercloud images into thundercloud grid images, and obtain a plurality of thunderstorm grid areas in the thundercloud grid images; The prediction module 30 includes: An eighth unit is used to determine whether the radar echo intensity of each grid point is higher than an intensity threshold, and select the grid point corresponding to the radar echo intensity higher than the intensity threshold as a thunderstorm grid point; A ninth unit, configured to determine whether the thunderstorm grid points are adjacent to each other, so as to combine the adjacent thunderstorm grid points into a thunderstorm grid area; An analysis module 40 is used to obtain a predicted thunderstorm moving path based on the thunderstorm grid area in the thundercloud grid map in adjacent time frames, and perform lightning protection based on the predicted thunderstorm moving path; The analysis module 40 includes: A tenth unit is used to obtain a grid distance from each thunderstorm grid area in the thundercloud grid map in a previous time frame to each thunderstorm grid area in the thundercloud grid map in a next time frame; An eleventh unit is used to obtain a distance threshold based on a plurality of the grid distances, compare the grid distance with the distance threshold, and select the grid distance less than the distance threshold as a moving sub-path; The twelfth unit is used to connect several of the moving sub-paths in series to form a predicted thunderstorm moving path.

[0041] The present invention also provides a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the lightning protection monitoring method as described in the above technical solution when executing the computer program.

[0042] The present invention also provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the lightning protection monitoring method as described in the above technical solution is implemented.

[0043] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0044] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A lightning protection monitoring method, characterized in that: The following steps are involved: Obtain the past atmospheric weighted temperature dataset, past thundercloud observation dataset and past radar echo dataset corresponding to the prediction area; Constructing an initial thundercloud prediction neural network, and training the initial thundercloud prediction neural network into a final thundercloud prediction neural network based on the past atmospheric weighted temperature data set, the past thundercloud observation data set, and the past radar echo data set; Based on the final thundercloud prediction neural network, a plurality of predicted thundercloud images in continuous time frames corresponding to the prediction area are obtained, the predicted thundercloud images are converted into thundercloud grid images, and a plurality of thunderstorm grid areas are obtained in the thundercloud grid images; A predicted thunderstorm moving path is obtained based on the thunderstorm grid area in the thundercloud grid map in adjacent time frames, and lightning protection is performed based on the predicted thunderstorm moving path.

2. The lightning protection monitoring method according to claim 1, characterized in that: The past atmospheric weighted temperature dataset, the past thundercloud observation dataset and the past radar echo dataset respectively include past atmospheric weighted temperatures, past thundercloud observation images and past radar echo images in a number of continuous time frames. The step of obtaining the past atmospheric weighted temperature dataset corresponding to the prediction area includes: Obtain the relative humidity and saturated water vapor pressure of different atmospheric layers in continuous time frames corresponding to the forecast area; In the same time frame, obtaining the average water vapor pressure of the corresponding atmosphere based on the relative humidity and the saturated water vapor pressure; The past atmospheric weighted temperatures are obtained through a plurality of the average water vapor pressures, and a plurality of the past atmospheric weighted temperatures in a continuous time frame are combined into a past atmospheric weighted temperature data set.

3. The lightning protection monitoring method according to claim 2, characterized in that: The formula for obtaining the average water vapor pressure is: ， in, represents the average water vapor pressure of the ith atmospheric layer, represents the relative humidity of the ith atmospheric layer, represents the saturated water vapor pressure of the i-th atmospheric layer; The formula for obtaining the weighted atmospheric temperature in the past period is: ， in, represents the past atmospheric weighted temperature in the jth time frame, represents the average temperature of the i-th atmospheric layer, represents the thickness of the i-th atmospheric layer.

4. The lightning protection monitoring method according to claim 1, characterized in that: The initial thundercloud prediction neural network includes a thundercloud prediction module, an echo prediction module and a temperature prediction module. The step of training the initial thundercloud prediction neural network into a final thundercloud prediction neural network based on the past atmospheric weighted temperature data set, the past thundercloud observation data set and the past radar echo data set includes: Acquire a predicted radar echo atlas through the previous radar echo data set and the echo prediction module; Obtaining a predicted atmospheric weighted temperature set through the past atmospheric weighted temperature data set and the temperature prediction module; Acquire a predicted thundercloud atlas through the previous thundercloud observation data set, the predicted radar echo atlas, the predicted atmospheric weighted temperature set and the thundercloud prediction module; A loss function is constructed by using the predicted thundercloud atlas, the real thundercloud observation atlas, the predicted atmospheric weighted temperature set, the real atmospheric weighted temperature set, the predicted radar echo atlas and the real radar echo atlas. The thundercloud prediction module, the echo prediction module and the temperature prediction module are corrected based on the loss function to obtain a final thundercloud prediction neural network.

5. The lightning protection monitoring method according to claim 4, characterized in that: The acquisition formula of the predicted radar echo atlas is: ， in, Indicates the predicted radar echo atlas from the next frame 1 to the next frame a at time t, It represents the set of past radar echo images from the past b frames to the current 0 frame at time t. represents the echo prediction module; The formula for obtaining the predicted atmospheric weighted temperature set is: ， in, represents the atmospheric weighted temperature set from the next frame 1 to the next frame a at time t, represents the set of past atmospheric weighted temperatures from the past b frames to the current 0 frame at time t, represents the temperature prediction module; The formula for obtaining the predicted thundercloud atlas is: ， in, represents the predicted thundercloud atlas from 1 frame to a frame in the future at time t, represents the set of past thundercloud observation images from time t, from the past b frames to the current 0 frame. Represents the thundercloud prediction module; The formula of the loss function is: ， in, It represents the real thundercloud observation atlas from the future frame 1 to the future frame a at time t, It represents the real radar echo atlas from the next frame 1 to the next frame a at time t. represents the real atmospheric weighted temperature set from the future frame 1 to the future frame a at time t, represents the loss function, , Both represent weight coefficients, , , They represent the module parameters of the echo prediction module, temperature prediction module and thundercloud prediction module respectively.

6. The lightning protection monitoring method according to claim 1, characterized in that: The thundercloud grid map includes a plurality of grid points of the same area, and the step of obtaining a plurality of thunderstorm grid areas in the thundercloud grid map includes: Determine whether the radar echo intensity of each of the grid points is higher than an intensity threshold, and select the grid points corresponding to the radar echo intensities higher than the intensity threshold as thunderstorm grid points; It is determined whether the thunderstorm grid points are adjacent to each other, so as to combine the adjacent thunderstorm grid points into a thunderstorm grid area.

7. The lightning protection monitoring method according to claim 1, characterized in that: The step of acquiring a predicted thunderstorm moving path based on the thunderstorm grid area in the thundercloud grid map in adjacent time frames comprises: Obtaining a grid distance from each of the thunderstorm grid areas in the thundercloud grid map in a previous time frame to each of the thunderstorm grid areas in the thundercloud grid map in a subsequent time frame; Acquire a distance threshold based on a plurality of the grid distances, compare the grid distance with the distance threshold, and select the grid distance that is smaller than the distance threshold as a moving sub-path; A plurality of the moving sub-paths are connected in series to form a predicted thunderstorm moving path.

8. The lightning protection monitoring method according to claim 7, characterized in that: The calculation formula of the grid distance is: ， in, Indicates the grid distance between the mth thunderstorm grid area in the previous time frame and the nth thunderstorm grid area in the next time frame. , Both represent weight functions, Indicates the distance between the confidence ellipse of the m-th thunderstorm grid area in the previous time frame and the center point of the n-th thunderstorm grid area in the next time frame. Represents the square root of the absolute value of the area difference between the confidence ellipse of the m-th thunderstorm grid area in the previous time frame and the confidence ellipse of the n-th thunderstorm grid area in the next time frame.

9. A lightning protection monitoring system, applied to the lightning protection monitoring method according to any one of claims 1 to 8, characterized in that: The system comprises: An acquisition module is used to acquire past atmospheric weighted temperature datasets, past thundercloud observation datasets, and past radar echo datasets corresponding to the prediction area; A construction module is used to construct an initial thundercloud prediction neural network, and train the initial thundercloud prediction neural network into a final thundercloud prediction neural network based on the past atmospheric weighted temperature data set, the past thundercloud observation data set and the past radar echo data set; A prediction module, used for obtaining a plurality of predicted thundercloud images in continuous time frames corresponding to the prediction area based on the final thundercloud prediction neural network, converting the predicted thundercloud images into thundercloud grid images, and obtaining a plurality of thunderstorm grid areas in the thundercloud grid images; The analysis module is used to obtain a predicted thunderstorm movement path based on the thunderstorm grid area in the thundercloud grid map under adjacent time frames, and perform lightning protection based on the predicted thunderstorm movement path.

10. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the lightning protection monitoring method as described in any one of claims 1 to 8 is implemented.

Images