CN110221360A - A kind of power circuit thunderstorm method for early warning and system - Google Patents
A kind of power circuit thunderstorm method for early warning and system Download PDFInfo
- Publication number
- CN110221360A CN110221360A CN201910677961.9A CN201910677961A CN110221360A CN 110221360 A CN110221360 A CN 110221360A CN 201910677961 A CN201910677961 A CN 201910677961A CN 110221360 A CN110221360 A CN 110221360A
- Authority
- CN
- China
- Prior art keywords
- radar
- prediction
- radar complex
- reflectance map
- thunderstorm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/95—Radar or analogous systems specially adapted for specific applications for meteorological use
- G01S13/958—Theoretical aspects
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/10—Devices for predicting weather conditions
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- Environmental & Geological Engineering (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Electromagnetism (AREA)
- Life Sciences & Earth Sciences (AREA)
- Atmospheric Sciences (AREA)
- Biodiversity & Conservation Biology (AREA)
- Ecology (AREA)
- Environmental Sciences (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The embodiment of the present application discloses a kind of power circuit thunderstorm method for early warning and system, comprising: using channel cloud atlas as training sample training neural network, exports to predict radar complex reflectance map, obtains trained radar complex reflectivity prediction model；Channel cloud atlas to be measured is input in the radar complex reflectivity prediction model and obtains prediction radar complex reflectance map；Echo movement in the prediction radar complex reflectance map is tracked using tracking radar echoes by correlation TREC, predicts subsequent time radar complex reflectance map；Thunderstorm warning information is sent to the shaft tower in figure in different convection intensities region according to the obtained subsequent time radar complex reflectance map of the prediction.Solve radar observation point limited amount in the prior art, carry out thunderstorm detection that cannot comprehensively to all regions, and thunderstorm early warning is carried out to power circuit in advance.
Description
Technical field
This application involves meteorological detection technique field more particularly to a kind of power circuit thunderstorm method for early warning and systems.
Background technique
New Generation Doppler Weather Radar has bright in tracking strong convective weather, diastrous weather and acquisition wind field information
Aobvious advantage is preventing and reducing natural disasters and is playing indispensable effect in Weather-service, at present a new generation, China weather thunder
It is completed up to net (CINRAD Network), in total 158 Doppler radars, respectively S-band (east and littoral
Area) and C-band (northwest and the Northeast).
Thunder and lightning is the main reason for causing transmission line of electricity to trip, and weather radar can be to precipitation particles such as cloud and mist, rain, snow
There is good detectivity, strong convective weather is monitored and is tracked using radar in real time, provides and more improves finer bar
Tower power circuit weather monitoring early warning and diagnosis capability will be helpful to reduce influence of the Lightning Disaster to power grid.
The current radar observation point limited amount in China, and radar observation is limited in scope, therefore, it is difficult to comprehensively to all
The carry out thunderstorm detection in area, and thunderstorm early warning is carried out to power circuit in advance.
Summary of the invention
The embodiment of the present application provides a kind of power circuit thunderstorm method for early warning and system, solves radar in the prior art
Observation station limited amount, carry out thunderstorm detection that cannot comprehensively to all regions, and thunderstorm early warning is carried out to power circuit in advance
The problem of.
In view of this, the application first aspect provides a kind of power circuit thunderstorm method for early warning, which comprises
Using channel cloud atlas as training sample training neural network, exports to predict radar complex reflectance map, instructed
The radar complex reflectivity prediction model perfected；
Channel cloud atlas to be measured is input in the radar complex reflectivity prediction model and obtains prediction radar complex reflection
Rate figure；
Echo movement in the prediction radar complex reflectance map is tracked using tracking radar echoes by correlation TREC, in advance
Measure subsequent time radar complex reflectance map；
According to the obtained subsequent time radar complex reflectance map of the prediction in convection intensities different in figure region
Shaft tower send thunderstorm warning information.
Preferably, it exports described using channel cloud atlas as training sample training neural network to predict that radar complex is anti-
Rate figure is penetrated, before obtaining trained radar complex reflectivity prediction model further include: pretreatment fixed statellite channel data obtains
To channel cloud atlas.
Preferably, the pretreatment fixed statellite channel data, obtains channel cloud atlas and specifically includes:
The fixed statellite channel data is spliced；
Projection longitude and latitude range is determined according to the geographic range of the satellite channel data cover spliced；
Equidistant lattice point is generated according to the projection longitude and latitude range and photo resolution of the determination, obtains waiting longitudes and latitudes net
Lattice；
Interpolation is carried out to the equal longitudes and latitudes grid using Delaunay triangulation network interpolation method, obtains the channel cloud atlas.
Preferably, the pretreatment fixed statellite channel data, after obtaining channel cloud atlas further include:
To the obtained channel cloud atlas equalization processing；
And/or the obtained channel cloud atlas data normalization is handled.
Preferably, described using channel cloud atlas as the input of neural network, it exports to predict radar complex reflectance map, obtains
To trained radar complex reflectivity prediction model specifically:
The channel cloud atlas is divided into input of the picture element matrix as neural network of 5X5 size, by classifier point
After class, the corresponding radar emission rate value of each matrix is exported, the parameter of neural network is adjusted and iterates, until determination
The model parameter of radar complex reflectivity prediction model obtains radar complex reflectivity prediction model.
Preferably, described to be tracked in the prediction radar complex reflectance map using tracking radar echoes by correlation TREC
Echo movement, predicts subsequent time radar complex reflectance map specifically:
The prediction radar complex reflectance map is divided into the grid cell of same size；
Determine prediction described in a certain grid cell to current time in prediction radar complex reflectance map described in last moment
The motion vector of grid cell is corresponded in radar complex reflectance map；
The position that subsequent time corresponds to grid cell is predicted according to the motion vector and time difference；
Subsequent time radar complex reflectance map is obtained according to all grid cell positions that predict.
Preferably, the obtained subsequent time radar complex reflectance map according to the prediction is to convection current different in figure
Shaft tower in intensity area is sent before thunderstorm warning information further include:
Net is established according to the convective region and shaft tower position predicted in obtained subsequent time radar complex reflectance map
Lattice.
The application second aspect provides a kind of power circuit thunderstorm early warning system, the system comprises:
Model generation module, the model generation module are used to train neural network for channel cloud atlas as training sample,
Output is prediction radar complex reflectance map, obtains trained radar complex reflectivity prediction model；
First prediction module, first prediction module, which is used to for channel cloud atlas to be measured to be input to the radar complex, to reflect
Prediction radar complex reflectance map is obtained in rate prediction model；
Second prediction module, second prediction module are used to track using tracking radar echoes by correlation TREC described pre-
The echo movement in radar complex reflectance map is surveyed, subsequent time radar complex reflectance map is predicted；
Warning module, the warning module are used for the obtained subsequent time radar complex reflectance map according to the prediction
Thunderstorm warning information is sent to the shaft tower in figure in different convection intensities region.
Preferably, the system also includes preprocessing modules, and the preprocessing module is for pre-processing fixed statellite channel
Data obtain channel cloud atlas.
Preferably, the system also includes:
Grid establishes module, and the grid establishes subsequent time radar complex reflectivity of the module for obtaining according to prediction
Convective region and shaft tower position in figure establish grid.
As can be seen from the above technical solutions, the embodiment of the present application has the advantage that
In the embodiment of the present application, a kind of power circuit thunderstorm method for early warning is provided, including using channel cloud atlas as training
Sample training neural network exports to predict radar complex reflectance map, obtains trained radar complex reflectivity prediction mould
Type；Channel cloud atlas to be measured is input in the radar complex reflectivity prediction model and obtains prediction radar complex reflectance map；
Echo movement in the prediction radar complex reflectance map is tracked using tracking radar echoes by correlation TREC, is predicted next
Moment radar complex reflectance map；According to the obtained subsequent time radar complex reflectance map of the prediction to different right in figure
Shaft tower in intensity of flow region sends thunderstorm warning information.
Application scheme is input to trained neural network by using satellite channel cloud atlas and obtains radar complex emissivity figure,
Rather than thunderstorm early warning directly is carried out using radar reflectivity, and radar observation point limited amount in the prior art is avoided, it cannot
Comprehensively to the carry out thunderstorm detection of all regions the technical issues of, on the other hand, by using tracking radar echoes by correlation pair
The motion vector in strong convection region is tracked, and the tendency in strong convection region is predicted, so as in advance to strong convection region
Shaft tower in direction of advance carries out thunderstorm early warning.
Detailed description of the invention
Fig. 1 is a kind of method flow diagram of one embodiment of power circuit thunderstorm method for early warning of the application；
Fig. 2 is a kind of method flow diagram of another embodiment of power circuit thunderstorm method for early warning of the application；
Fig. 3 is that the application is a kind of system architecture of one embodiment of power circuit thunderstorm early warning system of the application；
Fig. 4 is radar complex emissivity spatial distribution and histogram point in a kind of power circuit thunderstorm method for early warning of the application
Cloth comparison diagram；
Fig. 5 is power grid shaft tower position and the superposition signal of strong convection region in a kind of power circuit thunderstorm method for early warning of the application
Figure.
Specific embodiment
In order to make those skilled in the art more fully understand application scheme, below in conjunction in the embodiment of the present application
Attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is only this
Apply for a part of the embodiment, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art exist
Every other embodiment obtained under the premise of creative work is not made, shall fall in the protection scope of this application.
Referring to Fig. 1, Fig. 1 is a kind of method flow of one embodiment of power circuit thunderstorm method for early warning of the application
Figure, as shown in Figure 1, including: in Fig. 1
101, it using channel cloud atlas as training sample training neural network, exports to predict radar complex reflectance map, obtains
To trained radar complex reflectivity prediction model.
It should be noted that passing through firstly the need of the corresponding relationship found between channel cloud atlas and radar complex reflectivity
The method of neural network selects optimal matching degree, so that radar complex reflectivity prediction model is completed in training；Wherein for every
One channel cloud atlas extracts input layer of the picture element matrix of 5x5 size as convolutional network, while output layer is imago in corresponding
The radar complex reflectivity by classifier classification processing of member, i.e. each matrix match a radar complex emissivity.
By largely training, model parameter is determined.
In the particular embodiment, each sample image of input layer is the array of 5x5x3, and 3 represent satellite channel number: B08
Three kinds of (6.2 μm)-B10 (7.3 μm), B13 (10.4 μm), B13 (10.4 μm)-B15 (12.4 μm) combinations, 5X5 representative image square
Battle array, each such matrix correspond to a radar reflectivity as output layer；Convolutional layer is that each convolution kernel size is
18 filters of 3x3；Down-sampling layer is the maximum value pond layer of 3x3；Layer of classifying uses softmax classifier.
1 radar inverse model parameter of table
Wherein using on May 16th~18,2018 and somewhere sample source on June 25~27, nerve net is built according to table 1
Network, training the neural network after store network between relationship.
After generating radar complex reflectivity prediction model, the model after training, which meets loss function by test sample, to be wanted
It asks, wherein the loss function based on softmax classification generally uses cross entropy loss function:
Wherein, yiIt is the true tag of classification i；piIt is the probability value of the calculated classification i of softmax above；K is classification
Number, N is total sample number.
102, channel cloud atlas to be measured is input in the radar complex reflectivity prediction model and obtains prediction radar complex
Reflectance map.
It should be noted that be input in established radar complex reflectivity prediction model according to channel cloud atlas obtain it is pre-
Radar complex reflectance map is surveyed, for the accuracy for the radar complex reflectance map that verifying prediction obtains, needs to predict to obtain
Radar complex reflectance map is compareed with actual radar complex reflectance map.
Again in specific embodiment, the radar complex reflectance map and the reflection of actual radar complex that prediction obtains can be calculated
The structural similarity SSIM of rate figure, specifically: using sliding window will the obtained radar complex reflectance map of prediction with it is actual
Radar complex reflectance map piecemeal, enabling piecemeal sum is N, it is contemplated that influence of the window shape to piecemeal is weighted using Gauss and counted
The mean value, variance and covariance of each window are calculated, the structural similarity SSIM of corresponding blocks is then calculated, finally makees average value
It is measured for the structural similarity of two images, i.e. average structure similitude MSSIM, structural similarity reaches pre-defined threshold value
It is spare to save network.
The measurement of SSIM similarity can be made of three kinds of contrast modules, be respectively as follows: brightness L, contrast C, and structure S is calculated
Expression formula difference is as follows:
SSIM (X, Y)=L (X, Y) * C (X, Y) * S (X, Y)
μX、μYRespectively indicate the mean value of image X and Y, σX、σYRespectively indicate the standard deviation of image X and Y, σX 2、σY 2Table respectively
Variance of the diagram as X and Y.σXσYRepresentative image X and Y covariance；C1, C2 and C3 are constant, are in order to avoid denominator is tieed up for 0
It is fixed to keep steady.C1=(K1*L) ^2, C2=(K2*L) ^2, C3=C2/2 is usually taken, generally K1=0.01, K2=0.03, L=
255。
Available radar reflectivity, which is compared, by mass data is greater than 20dBz and distribution than more continuous region, in advance
It surveys result and fact is corresponding relatively good.
103, the echo in the prediction radar complex reflectance map is tracked using tracking radar echoes by correlation TREC to transport
It is dynamic, predict subsequent time radar complex reflectance map.
It should be noted that being moved using tracking radar echoes by correlation TREC tracking echo, to the fortune in strong convection region
Dynamic rail mark is predicted that predict the radar complex reflectance map of subsequent time, wherein TREC is specifically included: will be previous
The radar complex reflectance map at moment is divided into the grid cell of same size, and onesize in the latter moment region of search
Grid cell carry out crosscorrelation calculating respectively.Crosscorrelation calculation method such as following formula:
Z in formula1And Z2Respectively former and later two when time reflectivity factor data, N is for the interior data for including of grid cell
Points.
TREC algorithm flow is as follows: (1) relevant calculating parameter is determined, to a certain in T1 moment radar complex reflectance map
All grid cells in grid cell and T2 moment radar complex reflectance map in region of search calculate crosscorrelation；(2) it searches
Rope maximum value therein and the terminal for thereby determining that motion vector.According to the variation of successive trellis cell position and time difference,
Determine motion vector；(3) (1) and (2) is repeated to all grid cells at T1 moment and obtains the motion vector point of all grids
Cloth.
104, according to the obtained subsequent time radar complex reflectance map of the prediction to different convection intensities area in figure
Shaft tower in domain sends thunderstorm warning information.
It should be noted that since there is certain corresponding relationship in radar complex reflectivity and convection intensity region, it can be true
It is also not identical to determine different zones its convection intensities, therefore, the region shaft tower can be sent different according to different convection intensities
The pre-warning signal of depth, such as high-risk early warning is sent for strong convection region.
Application scheme is input to trained neural network by using satellite channel cloud atlas and obtains radar complex emissivity figure,
Rather than thunderstorm early warning directly is carried out using radar reflectivity, and radar observation point limited amount in the prior art is avoided, it cannot
Comprehensively to the carry out thunderstorm detection of all regions the technical issues of, on the other hand, by using tracking radar echoes by correlation pair
The motion vector in strong convection region is tracked, and the tendency in strong convection region is predicted, so as in advance to strong convection region
Shaft tower in direction of advance carries out thunderstorm early warning.
In order to make it easy to understand, referring to Fig. 2, Fig. 2 is a kind of the another real of power circuit thunderstorm method for early warning of the application
The method flow diagram of example is applied, as shown in Fig. 2, specifically:
201, fixed statellite channel data is pre-processed, channel cloud atlas is obtained.
Preferably, fixed statellite channel data is pre-processed, channel cloud atlas is obtained and specifically includes:
Fixed statellite channel data is spliced；It is determined according to the geographic range of the satellite channel data cover spliced
Project longitude and latitude range；Equidistant lattice point is generated according to determining projection longitude and latitude range and photo resolution, obtains waiting longitudes and latitudes
Spend grid；Interpolation is carried out using Delaunay triangulation network interpolation method equity longitude and latitude grid, obtains the channel cloud atlas.
It should be noted that the fixed statellite channel data that satellite receiving system receives is strip data, therefore, it is necessary to
Strip data is spliced according to different spatial dimension demands, while carrying out quality of data inspection, on this basis logarithm
According to being projected, satellite data and radar data are subjected to spatial match in order to subsequent, projective parameter should be with radar complex
Reflectivity projection is consistent, it is therefore desirable to define for the satellite data spliced from original CGMS LRIT/HRIT global specifications
The longitudes and latitudes projections such as the nominal projection transform of stationary orbit is.
The detailed process of Delaunay triangulation network interpolation method among the above are as follows: obtain the corresponding longitude and latitude of data point set after splicing
Degree；Delaunay triangulation network is made based on longitude and latitude；The Interpolation of longitude and latitude such as satellite data are carried out based on Delaunay triangulation network.Its
Middle Delaunay triangulation network Interpolation Principle are as follows: the observation value based on triangular plate each in triangulation network vertex, it can be to region
Interior arbitrary point carries out linear interpolation.
Assuming that 3 vertex of triangular plate are A, B, C, indicated respectively with number 1,2,3, the longitude of vertex position is mi(i=
1,2,3) and latitude is ni(i=1,2,3), observation are respectively f (A), f (B), f (C), and any point x is relative to this triangle
Shape has unique barycentric coodinates to indicate:
X=μ A+B+ ω C (1)
Linear interpolation may be expressed as:
G=m1(n3-n2)+m2(n1-n3)+)m3(n2-n1) (6)
The result f (x) of any one position x triangular plate Nei can be obtained by the observation of formula (1)-(6) and three vertex.
Preferably due to which the Data Physical characteristic of each satellite channel acquisition is different from, it can according to need selection and defend
Star channel is acquired.
In a particular embodiment, final satellite channel is determined are as follows: B08 (6.2 μm)-B10 (7.3 μm), B13 (10.4 μm),
Three kinds of B13 (10.4 μm)-B15 (12.4 μm) combinations.
Preferably due to which radar complex reflectivity values are unevenly distributed weighing apparatus, most reflectivity are lower than 10dBz, such as Fig. 4
Shown, this makes prediction result concentrate on the maximum range 0~10 of sample probability, and in practical application and is not concerned with this part number
Value, so needing to carry out data balancing to sample.
Preferably due to which input data is from the bright temperature difference in satellite difference channel and different channels, data area distribution
Span is big, before carrying out network training, needs that data are normalized.
202, it using channel cloud atlas as training sample training neural network, exports to predict radar complex reflectance map, obtains
To trained radar complex reflectivity prediction model.
It should be noted that passing through firstly the need of the corresponding relationship found between channel cloud atlas and radar complex reflectivity
The method of neural network selects optimal matching degree, so that radar complex reflectivity prediction model is completed in training；Wherein for every
One channel cloud atlas extracts input layer of the picture element matrix of 5x5 size as convolutional network, while output layer is imago in corresponding
The radar complex reflectivity by classifier classification processing of member, i.e. each matrix match a radar complex emissivity.
By largely training, model parameter is determined.
In the particular embodiment, each sample image of input layer is the array of 5x5x3, and 3 represent satellite channel number: B08
Three kinds of (6.2 μm)-B10 (7.3 μm), B13 (10.4 μm), B13 (10.4 μm)-B15 (12.4 μm) combinations, 5X5 representative image square
Battle array, each such matrix correspond to a radar reflectivity as output layer；Convolutional layer is that each convolution kernel size is
18 filters of 3x3；Down-sampling layer is the maximum value pond layer of 3x3；Layer of classifying uses softmax classifier.
Wherein using on May 16th~18,2018 and somewhere sample source on June 25~27, nerve net is built according to table 1
Network, training the neural network after store network between relationship.
After generating radar complex reflectivity prediction model, the model after training, which meets loss function by test sample, to be wanted
It asks, wherein the loss function based on softmax classification generally uses cross entropy loss function:
Wherein, yiIt is the true tag of classification i；piIt is the probability value of the calculated classification i of softmax above；K is classification
Number, N is total sample number.
203, channel cloud atlas to be measured is input in the radar complex reflectivity prediction model and obtains prediction radar complex
Reflectance map.
It should be noted that be input in established radar complex reflectivity prediction model according to channel cloud atlas obtain it is pre-
Radar complex reflectance map is surveyed, for the accuracy for the radar complex reflectance map that verifying prediction obtains, needs to predict to obtain
Radar complex reflectance map is compareed with actual radar complex reflectance map.
Again in specific embodiment, the radar complex reflectance map and the reflection of actual radar complex that prediction obtains can be calculated
The structural similarity SSIM of rate figure, specifically: using sliding window will the obtained radar complex reflectance map of prediction with it is actual
Radar complex reflectance map piecemeal, enabling piecemeal sum is N, it is contemplated that influence of the window shape to piecemeal is weighted using Gauss and counted
The mean value, variance and covariance of each window are calculated, the structural similarity SSIM of corresponding blocks is then calculated, finally makees average value
It is measured for the structural similarity of two images, i.e. average structure similitude MSSIM, structural similarity reaches pre-defined threshold value
It is spare to save network.
The measurement of SSIM similarity can be made of three kinds of contrast modules, be respectively as follows: brightness L, contrast C, and structure S is calculated
Expression formula difference is as follows:
SSIM (X, Y)=L (X, Y) * C (X, Y) * S (X, Y)；
μX、μYRespectively indicate the mean value of image X and Y, σX、σYRespectively indicate the standard deviation of image X and Y, σX 2、σY 2Table respectively
Variance of the diagram as X and Y.σXσYRepresentative image X and Y covariance；C1, C2 and C3 are constant, are in order to avoid denominator is tieed up for 0
It is fixed to keep steady.C1=(K1*L) ^2, C2=(K2*L) ^2, C3=C2/2 is usually taken, generally K1=0.01, K2=0.03, L=
255。
Predicted portions are to input satellite channel data to predict radar complex reflectivity: obtaining any time satellite number
According to, for desired prediction target area extract B08-B10, tri- kinds of B13, B13-B15 combination 5x5x3 array set, according to
Training sample normalized parameter inputs trained model and carries out radar reflectivity prediction, prediction result after being normalized
Become radar complex albedo by weight of classification relation.
Available radar reflectivity, which is compared, by mass data is greater than 20dBz and distribution than more continuous region, in advance
It surveys result and fact is corresponding relatively good；And 30dBz area above corresponds to the drizzle or moderate rain of 5mm/h or so, be it is short face it is pre-
Report the range being primarily upon；The strong echo area of the radar of 45dBz or more corresponds to strong convective weather.
204, the echo in the prediction radar complex reflectance map is tracked using tracking radar echoes by correlation TREC to transport
It is dynamic, predict subsequent time radar complex reflectance map.
Preferably, step 204 specifically: will predict that radar complex reflectance map is divided into the grid cell of same size；
Determine last moment prediction radar complex reflectance map in a certain grid cell to current time predict radar complex reflectance map
The motion vector of middle corresponding grid cell；The position that subsequent time corresponds to grid cell is predicted according to motion vector and time difference
It sets；Subsequent time radar complex reflectance map is obtained according to all grid cell positions that predict.
It should be noted that being moved using tracking radar echoes by correlation TREC tracking echo, to the fortune in strong convection region
Dynamic rail mark is predicted that predict the radar complex reflectance map of subsequent time, wherein TREC is specifically included: will be previous
The radar complex reflectance map at moment is divided into the grid cell of same size, and onesize in the latter moment region of search
Grid cell carry out crosscorrelation calculating respectively.Crosscorrelation calculation method such as following formula:
Z in formula1And Z2Respectively former and later two when time reflectivity factor data, N is for the interior data for including of grid cell
Points.
TREC algorithm flow is as follows: (1) relevant calculating parameter is determined, to a certain in T1 moment radar complex reflectance map
All grid cells in grid cell and T2 moment radar complex reflectance map in region of search calculate crosscorrelation；(2) it searches
Rope maximum value therein and the terminal for thereby determining that motion vector.According to the variation of successive trellis cell position and time difference,
Determine motion vector；(3) (1) and (2) is repeated to all grid cells at T1 moment and obtains the motion vector point of all grids
Cloth.
205, the convective region in the subsequent time radar complex reflectance map obtained according to prediction and shaft tower position are built
Vertical grid.
It should be noted that and in order to which the shaft tower timely and accurately for strong convection region sends warning information, it needs really
The relationship of fixed pole tower and thunderstorm region, it is therefore desirable to according to power grid shaft tower data, the longitude and latitude of shaft tower is determined, thus by itself and thunder
Lattice relationship is established up to composite reflectivity figure, it is specific as shown in figure 5, wherein darker curve is shaft tower position.
206, according to the obtained subsequent time radar complex reflectance map of the prediction to different convection intensities area in figure
Shaft tower in domain sends thunderstorm warning information.
It should be noted that since there is certain corresponding relationship in radar complex reflectivity and convection intensity region, it can be true
It is also not identical to determine different zones its convection intensities, therefore, the region shaft tower can be sent different according to different convection intensities
The pre-warning signal of depth, such as high-risk early warning is sent for strong convection region.
The application third aspect provides a kind of system framework figure of one embodiment of power circuit thunderstorm early warning system,
As shown in figure 3, including: in Fig. 3
Model generation module 301, for exporting to predict radar using channel cloud atlas as training sample training neural network
Composite reflectivity figure obtains trained radar complex reflectivity prediction model.
First prediction module 302, for channel cloud atlas to be measured to be input in the radar complex reflectivity prediction model
Obtain prediction radar complex reflectance map.
Second prediction module 303, it is anti-for tracking the prediction radar complex using tracking radar echoes by correlation TREC
The echo movement in rate figure is penetrated, subsequent time radar complex reflectance map is predicted.
Warning module 304, for according to the obtained subsequent time radar complex reflectance map of the prediction in figure not
Thunderstorm warning information is sent with the shaft tower in convection intensity region.
Application scheme is input to trained neural network by using satellite channel cloud atlas and obtains radar complex emissivity figure,
Rather than thunderstorm early warning directly is carried out using radar reflectivity, and radar observation point limited amount in the prior art is avoided, it cannot
Comprehensively to the carry out thunderstorm detection of all regions the technical issues of, on the other hand, by using tracking radar echoes by correlation pair
The motion vector in strong convection region is tracked, and the tendency in strong convection region is predicted, so as in advance to strong convection region
Shaft tower in direction of advance carries out thunderstorm early warning.
The application fourth aspect provides a kind of another embodiment of power circuit thunderstorm early warning system, comprising:
Preprocessing module obtains channel cloud atlas for pre-processing fixed statellite channel data.
Model generation module, for exporting to predict radar group using channel cloud atlas as training sample training neural network
Reflectance map is closed, trained radar complex reflectivity prediction model is obtained.
First prediction module is obtained for channel cloud atlas to be measured to be input in the radar complex reflectivity prediction model
Predict radar complex reflectance map.
Second prediction module, for tracking the prediction radar complex reflectivity using tracking radar echoes by correlation TREC
Echo movement in figure, predicts subsequent time radar complex reflectance map.
Grid establishes module, for according to the convective region in the obtained subsequent time radar complex reflectance map of prediction with
And grid is established in shaft tower position.
Warning module, for the obtained subsequent time radar complex reflectance map according to the prediction to different right in figure
Shaft tower in intensity of flow region sends thunderstorm warning information.
The description of the present application and term " first " in above-mentioned attached drawing, " second ", " third ", " the 4th " etc. are (if deposited
) it is to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that use in this way
Data are interchangeable under appropriate circumstances, so that embodiments herein described herein for example can be in addition to illustrating herein
Or the sequence other than those of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that
Cover it is non-exclusive include, for example, containing the process, method, system, product or equipment of a series of steps or units need not limit
In step or unit those of is clearly listed, but may include be not clearly listed or for these process, methods, produce
The other step or units of product or equipment inherently.
It should be appreciated that in this application, " at least one (item) " refers to one or more, and " multiple " refer to two or two
More than a."and/or" indicates may exist three kinds of relationships, for example, " A and/or B " for describing the incidence relation of affiliated partner
It can indicate: only exist A, only exist B and exist simultaneously tri- kinds of situations of A and B, wherein A, B can be odd number or plural number.Word
Symbol "/" typicallys represent the relationship that forward-backward correlation object is a kind of "or"." at least one of following (a) " or its similar expression, refers to
Any combination in these, any combination including individual event (a) or complex item (a).At least one of for example, in a, b or c
(a) can indicate: a, b, c, " a and b ", " a and c ", " b and c ", or " a and b and c ", and wherein a, b, c can be individually, can also
To be multiple.
In several embodiments provided herein, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components
It can be combined or can be integrated into another system, or some features can be ignored or not executed.Another point, it is shown or
The mutual coupling, direct-coupling or communication connection discussed can be through some interfaces, the indirect coupling of device or unit
It closes or communicates to connect, can be electrical property, mechanical or other forms.
The above, above embodiments are only to illustrate the technical solution of the application, rather than its limitations；Although referring to before
Embodiment is stated the application is described in detail, those skilled in the art should understand that: it still can be to preceding
Technical solution documented by each embodiment is stated to modify or equivalent replacement of some of the technical features；And these
It modifies or replaces, the spirit and scope of each embodiment technical solution of the application that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of power circuit thunderstorm method for early warning characterized by comprising
Using channel cloud atlas as training sample training neural network, exports to predict radar complex reflectance map, trained
Radar complex reflectivity prediction model；
Channel cloud atlas to be measured is input in the radar complex reflectivity prediction model and obtains prediction radar complex reflectance map；
Echo movement in the prediction radar complex reflectance map is tracked using tracking radar echoes by correlation TREC, is predicted
Subsequent time radar complex reflectance map；
According to the obtained subsequent time radar complex reflectance map of the prediction to the bar in figure in different convection intensities region
Tower sends thunderstorm warning information.
2. a kind of power circuit thunderstorm method for early warning according to claim 1, which is characterized in that described by channel cloud atlas
It as training sample training neural network, exports to predict radar complex reflectance map, obtains trained radar complex reflection
Before rate prediction model further include:
Fixed statellite channel data is pre-processed, channel cloud atlas is obtained.
3. a kind of power circuit thunderstorm method for early warning according to claim 2, which is characterized in that the pretreatment is static to be defended
Star channel data obtains channel cloud atlas and specifically includes:
The fixed statellite channel data is spliced；
Projection longitude and latitude range is determined according to the geographic range of the satellite channel data cover spliced；
Equidistant lattice point is generated according to the projection longitude and latitude range and photo resolution of the determination, obtains waiting longitudes and latitudes grid；
Interpolation is carried out to the equal longitudes and latitudes grid using Delaunay triangulation network interpolation method, obtains the channel cloud atlas.
4. a kind of power circuit thunderstorm method for early warning according to claim 2, which is characterized in that the pretreatment is static to be defended
Star channel data, after obtaining channel cloud atlas further include:
To the obtained channel cloud atlas equalization processing；
And/or the obtained channel cloud atlas data normalization is handled.
5. a kind of power circuit thunderstorm method for early warning according to claim 1, which is characterized in that described to make channel cloud atlas
It for the input of neural network, exports to predict radar complex reflectance map, obtains trained radar complex reflectivity prediction mould
Type specifically:
The channel cloud atlas is divided into input of the picture element matrix as neural network of 5X5 size, after classifier is classified,
The corresponding radar emission rate value of each matrix is exported, the parameter of neural network is adjusted and iterates, until determining radar
The model parameter of composite reflectivity prediction model obtains radar complex reflectivity prediction model.
6. a kind of power circuit thunderstorm method for early warning according to claim 1, which is characterized in that described to use radar return
Correlation tracking method TREC tracks the echo movement in the prediction radar complex reflectance map, predicts subsequent time radar complex
Reflectance map specifically:
The prediction radar complex reflectance map is divided into the grid cell of same size；
It determines in prediction radar complex reflectance map described in last moment and predicts radar described in a certain grid cell to current time
The motion vector of grid cell is corresponded in composite reflectivity figure；
The position that subsequent time corresponds to grid cell is predicted according to the motion vector and time difference；
Subsequent time radar complex reflectance map is obtained according to all grid cell positions that predict.
7. a kind of power circuit thunderstorm method for early warning according to claim 1, which is characterized in that described according to the prediction
Obtained subsequent time radar complex reflectance map thunderstorm early warning letter is sent to the shaft tower in convection intensities different in figure region
Before breath further include:
Grid is established according to the convective region and shaft tower position predicted in obtained subsequent time radar complex reflectance map.
8. a kind of power circuit thunderstorm early warning system characterized by comprising
Model generation module, the model generation module are used for using channel cloud atlas as training sample training neural network, output
To predict radar complex reflectance map, trained radar complex reflectivity prediction model is obtained；
First prediction module, first prediction module are used to channel cloud atlas to be measured being input to the radar complex reflectivity pre-
It surveys in model and obtains prediction radar complex reflectance map；
Second prediction module, second prediction module are used to track the prediction thunder using tracking radar echoes by correlation TREC
Up to the echo movement in composite reflectivity figure, subsequent time radar complex reflectance map is predicted；
Warning module, the warning module are used for the obtained subsequent time radar complex reflectance map according to the prediction to figure
Shaft tower in middle difference convection intensity region sends thunderstorm warning information.
9. a kind of power circuit thunderstorm early warning system according to claim 8, which is characterized in that the system also includes:
Preprocessing module, the preprocessing module obtain channel cloud atlas for pre-processing fixed statellite channel data.
10. a kind of power circuit thunderstorm early warning system according to claim 8, which is characterized in that the system also includes:
Grid establishes module, and the grid is established in subsequent time radar complex reflectance map of the module for being obtained according to prediction
Convective region and shaft tower position establish grid.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910677961.9A CN110221360A (en) | 2019-07-25 | 2019-07-25 | A kind of power circuit thunderstorm method for early warning and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910677961.9A CN110221360A (en) | 2019-07-25 | 2019-07-25 | A kind of power circuit thunderstorm method for early warning and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110221360A true CN110221360A (en) | 2019-09-10 |
Family
ID=67813921
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910677961.9A Pending CN110221360A (en) | 2019-07-25 | 2019-07-25 | A kind of power circuit thunderstorm method for early warning and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110221360A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110764090A (en) * | 2019-10-22 | 2020-02-07 | 上海眼控科技股份有限公司 | Image processing method, image processing device, computer equipment and readable storage medium |
| CN110989040A (en) * | 2019-11-29 | 2020-04-10 | 张烨方 | Artificial intelligent lightning approach early warning method and system based on slice processing |
| CN111458769A (en) * | 2020-05-26 | 2020-07-28 | 南京大学 | Method and system for predicting environmental meteorological data of power transmission line |
| CN112198483A (en) * | 2020-09-28 | 2021-01-08 | 上海眼控科技股份有限公司 | Data processing method, device and equipment for satellite inversion radar and storage medium |
| CN112580779A (en) * | 2020-12-11 | 2021-03-30 | 上海眼控科技股份有限公司 | Radar data generation method, device, equipment and storage medium |
| CN114019514A (en) * | 2021-11-25 | 2022-02-08 | 浙江省气象台 | Thunderstorm strong wind early warning method, system, equipment and terminal |
| CN116047631A (en) * | 2023-03-31 | 2023-05-02 | 中科星图维天信(北京)科技有限公司 | Precipitation prediction method and device, electronic equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104849777A (en) * | 2014-12-08 | 2015-08-19 | 国家电网公司 | Power grid equipment area convection weather forecasting method based on cross correlation tracking method |
| CN107748942A (en) * | 2017-11-24 | 2018-03-02 | 清华大学 | Radar Echo Extrapolation Forecasting Methodology and system based on velocity field sensing network |
| CN108052734A (en) * | 2017-12-12 | 2018-05-18 | 中国电力科学研究院有限公司 | A kind of method and system predicted based on meteorologic parameter amplitude of lightning current |
| CN106019287B (en) * | 2016-07-29 | 2018-07-10 | 国网电力科学研究院武汉南瑞有限责任公司 | A kind of power transmission line lightning shielding monitoring and pre-alarming method based on miniradar |
| CN108445464A (en) * | 2018-03-12 | 2018-08-24 | 南京恩瑞特实业有限公司 | Satellite radar inverting fusion methods of the NRIET based on machine learning |
| CN108508505A (en) * | 2018-02-05 | 2018-09-07 | 南京云思创智信息科技有限公司 | Heavy showers and thunderstorm forecasting procedure based on multiple dimensioned convolutional neural networks and system |
| CN109001736A (en) * | 2018-06-12 | 2018-12-14 | 中国人民解放军国防科技大学 | Radar echo extrapolation method based on deep space-time prediction neural network |
-
2019
- 2019-07-25 CN CN201910677961.9A patent/CN110221360A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104849777A (en) * | 2014-12-08 | 2015-08-19 | 国家电网公司 | Power grid equipment area convection weather forecasting method based on cross correlation tracking method |
| CN106019287B (en) * | 2016-07-29 | 2018-07-10 | 国网电力科学研究院武汉南瑞有限责任公司 | A kind of power transmission line lightning shielding monitoring and pre-alarming method based on miniradar |
| CN107748942A (en) * | 2017-11-24 | 2018-03-02 | 清华大学 | Radar Echo Extrapolation Forecasting Methodology and system based on velocity field sensing network |
| CN108052734A (en) * | 2017-12-12 | 2018-05-18 | 中国电力科学研究院有限公司 | A kind of method and system predicted based on meteorologic parameter amplitude of lightning current |
| CN108508505A (en) * | 2018-02-05 | 2018-09-07 | 南京云思创智信息科技有限公司 | Heavy showers and thunderstorm forecasting procedure based on multiple dimensioned convolutional neural networks and system |
| CN108445464A (en) * | 2018-03-12 | 2018-08-24 | 南京恩瑞特实业有限公司 | Satellite radar inverting fusion methods of the NRIET based on machine learning |
| CN109001736A (en) * | 2018-06-12 | 2018-12-14 | 中国人民解放军国防科技大学 | Radar echo extrapolation method based on deep space-time prediction neural network |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110764090A (en) * | 2019-10-22 | 2020-02-07 | 上海眼控科技股份有限公司 | Image processing method, image processing device, computer equipment and readable storage medium |
| CN110989040A (en) * | 2019-11-29 | 2020-04-10 | 张烨方 | Artificial intelligent lightning approach early warning method and system based on slice processing |
| CN110989040B (en) * | 2019-11-29 | 2021-11-30 | 张烨方 | Artificial intelligent lightning approach early warning method and system based on slice processing |
| CN111458769A (en) * | 2020-05-26 | 2020-07-28 | 南京大学 | Method and system for predicting environmental meteorological data of power transmission line |
| CN112198483A (en) * | 2020-09-28 | 2021-01-08 | 上海眼控科技股份有限公司 | Data processing method, device and equipment for satellite inversion radar and storage medium |
| CN112580779A (en) * | 2020-12-11 | 2021-03-30 | 上海眼控科技股份有限公司 | Radar data generation method, device, equipment and storage medium |
| CN114019514A (en) * | 2021-11-25 | 2022-02-08 | 浙江省气象台 | Thunderstorm strong wind early warning method, system, equipment and terminal |
| US11802992B2 (en) | 2021-11-25 | 2023-10-31 | Zhejiang Meteorological Observatory | Thunderstorm gale early-warning method, system, equipment, and terminal |
| CN116047631A (en) * | 2023-03-31 | 2023-05-02 | 中科星图维天信(北京)科技有限公司 | Precipitation prediction method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110221360A (en) | A kind of power circuit thunderstorm method for early warning and system | |
| CN108596101B (en) | Remote sensing image multi-target detection method based on convolutional neural network | |
| CN106356757B (en) | A kind of power circuit unmanned plane method for inspecting based on human-eye visual characteristic | |
| Lang et al. | The severe thunderstorm electrification and precipitation study | |
| Hilburn et al. | Development and interpretation of a neural-network-based synthetic radar reflectivity estimator using GOES-R satellite observations | |
| Li et al. | Short-range quantitative precipitation forecasting in Hong Kong | |
| CN110083977B (en) | Atmospheric turbulence monitoring method based on deep learning | |
| CN108875905B (en) | Direct inversion method for visibility function of atmospheric temperature and humidity profile | |
| KR101258668B1 (en) | Korea local radar processing system | |
| Liang et al. | A composite approach of radar echo extrapolation based on TREC vectors in combination with model-predicted winds | |
| CN112764129A (en) | Method, system and terminal for thunderstorm short-term forecasting | |
| CN111339826B (en) | Landslide unmanned aerial vehicle linear sensor network frame detecting system | |
| ApRil | Automated storm tracking for terminal air traffic control | |
| CN103235890A (en) | System and method for satellite short-time approaching rainfall forecasting | |
| CN114966899A (en) | Regional visibility prediction method based on multi-source multi-element remote sensing technology cooperation | |
| KR101221755B1 (en) | Method for identifying reflectivity cells associated with severe weather | |
| Huang et al. | Locating parent lightning strokes of sprites observed over a mesoscale convective system in Shandong Province, China | |
| KR101221793B1 (en) | Method for tracking reflectivity cells associated with severe weather | |
| CN109117791A (en) | A kind of crowd density drawing generating method based on expansion convolution | |
| CN115691049A (en) | Convection birth early warning method based on deep learning | |
| Chen et al. | Proactive quality control: Observing system experiments using the NCEP global forecast system | |
| Yang et al. | Flood detection based on unmanned aerial vehicle system and deep learning | |
| Lewis et al. | Determination of spatial and temporal characteristics as an aid to neural network cloud classification | |
| CN109657679B (en) | Application satellite function type identification method | |
| Kwong et al. | An artificial neural network with chaotic oscillator for wind shear alerting |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190910 |
|
| RJ01 | Rejection of invention patent application after publication |