CN109710595A - The construction method of transmission of electricity corridor bird pest hotspot graph based on limited information - Google Patents
The construction method of transmission of electricity corridor bird pest hotspot graph based on limited information Download PDFInfo
- Publication number
- CN109710595A CN109710595A CN201811455816.8A CN201811455816A CN109710595A CN 109710595 A CN109710595 A CN 109710595A CN 201811455816 A CN201811455816 A CN 201811455816A CN 109710595 A CN109710595 A CN 109710595A
- Authority
- CN
- China
- Prior art keywords
- bird pest
- bird
- information
- pest
- data
- 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.)
- Granted
Links
Classifications
-
- 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
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The construction method for the transmission of electricity corridor bird pest hotspot graph based on limited information that the invention discloses a kind of, it include: that regional geography space lattice divides, finite region bird pest information is collected, bird pest analysis of Influential Factors, finite region bird pest influence factor correlation analysis, global association degree are deduced, the building of bird pest situation information, bird pest hot spot value calculates, graphical treatment.The present invention merges the multi-source datas such as geography information, weather information, birds distributed intelligence, the variation of a wide range of bird trouble on transmission line distributed areas and time of origin is estimated in time, effective foundation is provided for transmission line of electricity O&M, improve bird trouble on transmission line inspection and efficiency of getting rid of the danger, improve the safe operation ability of transmission line of electricity, and then pushes the development of power grid high quality.
Description
Technical field
The present invention relates to electric intelligent O&M technical fields, and in particular to a kind of transmission of electricity corridor bird pest based on limited information
The construction method of hotspot graph.
Background technique
In recent years, as the continuous complete of the laws and regulations of animal is protected in the continuous improvement of China's natural environment, and correlation
Kind, the procreation of birds is gradually accelerated, and birds activity is increasingly frequent, and the safety for having seriously affected the occasions such as transmission line of electricity and airport is high
Effect operation.For example, birds can pollute transmission line of electricity in the birds droppings that flight course falls in transmission line of electricity, flashover is formed；Together
When, birds are nested on shaft tower, and falling for material of nesting also results in the failures such as transmission line of electricity short circuit, tripping.Therefore, it is necessary to
Building transmission of electricity corridor harmful bird (relating to harmful bird class) is distributed and bird pest (related birds cause damages or may cause the case where endangering)
Hotspot graph.
The safe and highly efficient operation of transmission line of electricity directly determines the stabilization of overall grid, is the premise that user stablizes electricity consumption,
It is also the guarantee of power grid high quality development.In recent years, line tripping disconnection fault caused by bird pest becomes the main event of power transmission network
One of barrier.For bird pest because its type is more, quantity is big, and distribution is wide, and variation is fast, influences vulnerable to geography and climate, existing bird pest prevention and treatment
With monitoring mainly by manual inspection, unmanned plane inspection or fixed point video acquisition, judges and mark, blindness is big, anti-bird
Not in time, anti-bird effect is unobvious.Electric power operation maintenance personnel is difficult timely and effectively to grasp bird pest situation, wastes huge manpower object
Power, but difficulty has preferable control efficiency.
Current bird pest scene and time Estimate are mainly determined according to operation/maintenance data over the years, since bird pest is movable not
Certainty, bird pest region and time difference every year cause that O&M difficulty is big, effect is poor.If can be transported in bird pest over the years
On the basis of dimension record, the information such as fusion geography, meteorology, ecology, the bird trouble on transmission line that timely updates distributed intelligence improves bird
The validity and directive significance of evil distribution map, are urgent problems to be solved during current power transmission O&M.There is an urgent need to structures as a result,
A kind of method for building automation highlights bird pest situation hot spot, guide line inspection.
Summary of the invention
For practical problem present in the above power grid inspection, the present invention proposes a kind of transmission of electricity corridor based on limited information
The construction method of bird pest hotspot graph can effectively promote the quality and efficiency of bird pest inspection in electric power O&M.
In order to realize above-mentioned task, the invention adopts the following technical scheme:
A kind of construction method of the transmission of electricity corridor bird pest hotspot graph based on limited information, includes the following steps:
Step 1, regional geography space lattice divides
The bird pest region for determining hotspot graph to be constructed first carries out gridding division by geographical space to region；
Step 2, finite region bird pest information is collected
The finite region in the bird pest region is chosen as sample and carries out data collection, is acquired in the finite region
Bird pest information, cleans data, is processed into the data mode that can be excavated, then data prediction, generates for data mining
The target data that algorithm uses；
Step 3, bird pest analysis of Influential Factors
To step 2 treated data are read out and carry out data aggregation, extensive, standardization, attribute construction, fusion point
Analysis, identifies frequent item set from data, then using the rule of these frequent item sets creation description incidence relation, finds all
Support is greater than or equal to given minimum support frequent item set, finds generation and is not less than Minimum support4 rule association,
Birds distribution space-time and the informational linkages such as geographical weather environment get up, and therefrom analysis selects bird pest Correlative Influence Factors；
Step 4, finite region bird pest and influence factor correlation analysis
Feature is extracted and set up from step 2 treated data, carries out bird pest sequence information data, including birds distribution
The data mining of time series, distribution space sequence, Relevant Factors of Distribution sequence；Fractional bits are used to bird pest sequence information data
Homing method quantitative analysis bird pest influence factor obtains the correlation rule set between birds distribution space-time and bird pest information；Benefit
With correlation rule set training deeply learning network, target depth intensified learning network model is constructed；
Step 5, global association degree is deduced
The bird pest information for collecting other regions in bird pest region other than the finite region described in the step 2, by other areas
The bird pest information data in domain carries out data cleansing and pretreatment according to the identical method of step 2, and is based on Pearson correlation coefficient
Method analyzes the similarity of other region bird pest information and finite region bird pest information；
Then based on target depth intensified learning network model, the net in other regions is obtained using transfer learning method
Network model, to know the bird pest distribution situation of other regional predictions；Finally by the distribution of finite region bird pest and other regions
Bird pest distribution is integrated to obtain global bird pest distributed intelligence；
Step 6, bird pest situation information constructs
Global bird pest distributed intelligence is obtained according to prediction, constructs bird pest situation map；In order to quantify the degree of danger of birds, adopt
The harm that bird pest is indicated with bird pest hot spot value indicates that bird pest danger is higher when the hot spot value the high；
Step 7, bird pest hot spot value calculates
The bird pest hot spot value in each grid is calculated, bird pest hot spot value matrix is constructed；
Step 8, graphical treatment
Color mapping table is constructed, the mapping of bird pest hot spot value to color is realized, so as to form hotspot graph figure layer；According to bird pest
Hot spot value matrix and color mapping table convert bird pest hot spot value matrix to the color image distinguished by different colours.
Further, the bird pest information include: existing literature research, cabling investigation on the spot, bird pest Tripping data,
Bird pest distributed intelligence, environmental landscape data, GIS data, history meteorological data, electric power O&M record, shaft tower, line construction are special
Sign, real-time detection apparatus, video image, radar monitoring and bird damage prevention device information, bird ecology model, bird ecology scholarship and moral conduct
It is characterized.
Further, the data mining algorithm includes K-Means algorithm, support vector machines, Apriori algorithm, most
Big Expectation Algorithm, AdaBoost, K arest neighbors sorting algorithm.
Further, fractional bits homing method quantitative analysis bird pest influence factor, regression model are used described in step 4
Are as follows:
yit=x 'itβθ+μθit
Quantθ(yit|xit)=x 'itβθ
In formula, yitCollect cluster value, x for i-th of sample data birds of t momentitFor bird pest influence factor, βθFor under residual error θ
Coefficient to be estimated, μθitFor i-th of sample data of t moment at residual error θ residual error variable, Quantθ(yit|xit) indicate given xitItem
Y under partitResidual error θ condition quantile, the value range of residual error θ are as follows: 0≤θ≤1；
Solve factor beta to be estimatedθObjective function be residual absolute value weighted average it is minimum, formula is as follows:
Work as yit≥x′itβθ, i.e. residual error is timing, assigns residual error θ weight；Work as yit< x 'itβ when that is, residual error is negative, is assigned
Residual error 1- θ weight；Solve factor beta to be estimatedθ, i.e. correlation rule, and then thus obtain between birds distribution space-time and bird pest information
Correlation rule set.
Further, similarity described in step 5 includes: geographical environment, landscape information similarity calculation；Weather meteorology letter
Cease similarity calculation；Bird ecology environmental model similarity calculation；Transmission line of electricity history O&M situation similarity calculation.
Further, global bird pest distributed intelligence is obtained according to prediction described in step 6, constructs bird pest situation map, comprising:
If D={ D1,D2,…,DnBe bird pest situation map in n situation harmful bird set, U=[u1,u2,…,um] it is to retouch
State the set of the m influence factor index of situation harmful bird；The summation at current time, bird pest situation information can use original situation information
Matrix A=(aij)n×mIt indicates:
Matrix A=(aij)n×mIn every a line ui=[ai1,…,aim] indicate situation harmful bird DiM influence factor index
Under situation information, situation harmful bird DiInfluence factor index aijBigger expression situation harmful bird DiActivity to the net where it
The transmission line of electricity harm of lattice is bigger；
Column element each in matrix A is standardized, Standard Process is obtainedWherein:
Further, the bird pest hot spot value in each grid of calculating described in step 7, constructs bird pest hot spot value matrix, comprising:
By situation harmful bird DiIt is set as point target, current influence factor index uiUnder the standard value of each bird pest situation information bediBy situation harmful bird to the distance for investigating grid s, h is distance threshold；Solve net
The formula of the bird pest hot spot value f (s) of lattice s is as follows:
G in formulaiSituation harmful bird i is indicated to the density contribution value of grid s, h > 0 in formula indicates distance threshold,For
Situation harmful bird DiInfluence factor index weights item, initially take
All grids are traversed, the bird pest hot spot value of each grid is found out, bird pest hot spot value can be constructed by, which being then combined with, adds up
Matrix:
F=(fab)p×q
Wherein, p, q are size of mesh opening, fabIndicates coordinate be (a, b) grid s before current bird pest hotspot graph t duration model
Enclose interior bird pest hot spot value accumulated value, f under original stateab=0；
Bird pest hot spot value accumulated matrix F is standardized, bird pest hot spot value matrix is finally obtained:
It is as follows to standardize formula:
It is then each in bird pest hot spot value matrix F*Value is equal to the bird pest hot spot value R (s) in (a, the b) grid.
The present invention has following technical characterstic compared with prior art:
1. the multi-source datas such as present invention fusion geography information, weather information, birds distributed intelligence, in time estimation are a wide range of defeated
The variation of electric line bird pest distributed areas and time of origin provides effective foundation for transmission line of electricity O&M, improves transmission line of electricity bird
Evil inspection and efficiency of getting rid of the danger improve the safe operation ability of transmission line of electricity, and then push the development of power grid high quality.
2. the present invention constructs a kind of method of automation, bird pest situation hot spot is highlighted, guide line inspection can
Effectively promote the quality and efficiency of bird pest inspection in electric power O&M.
Detailed description of the invention
Fig. 1 is the overall flow schematic diagram of the method for the present invention；
Fig. 2 is bird pest data mining process model schematic disclosed by the embodiments of the present invention；
Fig. 3 is color mapping table disclosed by the embodiments of the present invention；
Fig. 4 is hotspot graph figure layer schematic diagram disclosed by the embodiments of the present invention.
Specific embodiment
The construction method for the transmission of electricity corridor bird pest hotspot graph based on limited information that embodiment of the invention discloses a kind of, such as
Shown in Fig. 1, comprising: regional geography space lattice divides, and finite region bird pest information is collected, and bird pest analysis of Influential Factors is limited
Region bird pest influence factor correlation analysis, global association degree are deduced, and bird pest situation information building, bird pest hot spot value calculates, figure
Shape handles these steps, and step of the invention is described in detail below.
Step 1, regional geography space lattice divides
The bird pest region for determining hotspot graph to be constructed first carries out gridding division by geographical space to the region, forms p
The grid of × q, sizing grid are selected as c × c；Sizing grid can determine by hotspot graph precision size, the more intensive then state of grid dividing
Gesture is finer, but can processing speed be slowed down, and can suitably adjust sizing grid to reach optimum efficiency.
50m*50m, 100m*100m, 200m*200m, the multi gears such as 500m*500m, 1km*1km may be selected in the grid；
In general, number of grid increases, and computational accuracy can increase, but calculation scale also will increase simultaneously, so determining net
Two combined factors should be weighed when lattice quantity to consider.When being unable to satisfy the precision of actual requirement such as the grid precision currently selected,
Sizing grid is then adjusted to small grid to meet required precision, reaches the balance of bird pest display precision and computing capability, makes to meet
Best bird pest state under computing capability.
Step 2, finite region bird pest information is collected
The finite region in the bird pest region of the hotspot graph to be constructed is chosen as sample and carries out data collection, is acquired
Bird pest information, cleans data in the finite region, will come from different operating system, the data (multi-source of different-format
Data) it is processed into the data mode that can be excavated, data prediction is carried out to data, generates the target used for data mining algorithm
Data, as shown in Fig. 2, being bird pest data mining process model.
There are two types of methods for the selection of the finite region, one is choosing representative region (such as lake, field, high mountain, mound
Typical case's geomorphic province such as mound) as sample progress bird pest data collection, the second is randomly selecting, to reach covering all as far as possible
More bird pest situation, so that result is more accurate.The typical sample and random sample size in general each transmission of electricity corridor take length 2-
Stripe region in 5km power transmission line corridor or so 200m-500m, every hundred kms transmission line of electricity need to choose at least 10% sample
Line, according further to transmission line of electricity the number through region topography and geomorphology type, changeable line-transect number.For example, 100km long transmits electricity
3km line-transect is then chosen respectively in representative region, and in addition midway through lake, field, three kinds of high mountain typical landforms in corridor
The two sections of 3km randomly selected acquire data as line-transect.
The bird pest information includes but is not limited to existing literature research, cabling investigation on the spot, bird pest Tripping data, bird
Evil distributed intelligence, environmental landscape data (such as land use, water body, culture, Urban Changes), GIS data, history are meteorological
Data, electric power O&M record, shaft tower, line construction feature, real-time detection apparatus, video image, radar monitoring and anti-bird pest dress
The information such as confidence breath, bird ecology model, bird ecology behavioural characteristic.
The data mode excavated refers to the file system format being widely used in Hadoop system, example
Such as SequenceFile, MapFile, RCFile, ORCFile and Parque；According to the different characteristic of file when practical application
Select optimal file format.The preprocessing process use data mining universal method, including classification, recurrence, cluster,
Data Dimensionality Reduction, model selection etc..
Due to data source multiplicity, data mining algorithm is also not fixed, and available core algorithm includes but is not limited to K-
Means algorithm, support vector machines, Apriori algorithm, greatest hope (EM) algorithm, AdaBoost, K arest neighbors (k-Nearest
Neighbor, KNN) sorting algorithm etc..
Since bird pest area distribution is extensive, (data collection, we can not be carried out to entire hotspot graph bird pest to be constructed region
The method that case uses is that selected part region (finite region i.e. in step 2) carries out bird pest data collection, limited by analyzing
The incidence relation of the factors such as bird pest and environment in region, and then migrated to the region of not detailed bird pest data (i.e. by promoting
Other regions of step 5), the incidence relation of the factors such as bird pest and the environment in other regions is obtained, is finally obtained global (i.e. entire
Bird pest region) bird pest and the factors such as environment incidence relation, and then the hotspot graph in entire bird pest region can be constructed.
Step 3, bird pest analysis of Influential Factors
Using Hadoop distributed file system HDFS, to step 2, treated that data are read out, and passes through automaticdata
Analytical technology is big to this tittle, the data more than type carry out data aggregation, extensive, standardization, attribute construction, convergence analysis, sharp
Frequent item set is identified from data set with Apriori and FP-Growth algorithm, then utilizes the creation description of these frequent item sets
The rule of incidence relation finds all supports and is greater than or equal to given minimum support frequent item set, it is not small to find generation
In Minimum support4 rule association, birds distribution space-time get up with informational linkages such as geography weather environments, and therefrom analyze
Select bird pest Correlative Influence Factors.
The bird pest Correlative Influence Factors refer to the factor for having correlativity with bird pest, such as the lower birds activity of temperature
More infrequently, then temperature and birds activity are negative correlativing relations, and temperature is the Correlative Influence Factors of bird pest.
Step 4, finite region bird pest and influence factor correlation analysis
Feature is extracted and set up from step 2 treated data, uses MapReduce data mining framework and described
Core algorithm carries out birds distribution time sequence, distribution space sequence, Relevant Factors of Distribution sequence and carries out data mining；To bird pest
Sequence information data, including birds distribution time sequence, distribution space sequence, Relevant Factors of Distribution sequence are returned using fractional bits
Method quantitative analysis bird pest influence factor, regression model are as follows:
yit=x 'itβθ+μθit
Quantθ(yit|xit)=x 'itβθ
In formula, yitCollect cluster value namely regression estimates value, x for i-th of sample data birds of t momentitFor bird pest influence because
Element, i.e. the actual observation value of i-th of influence factor of t moment, βθFor the coefficient to be estimated under residual error θ, μθitFor i-th of sample of t moment
Data residual error variable, Quant at residual error θθ(yit|xit) indicate given xitUnder the conditions of yitResidual error θ condition quantile, residual error θ
Value range are as follows: 0≤θ≤1.
Solve factor beta to be estimatedθObjective function be residual absolute value weighted average it is minimum, formula is as follows:
Work as yit≥x′itβθ, i.e. residual error is timing, assigns residual error θ weight；Work as yit< x 'itβ when that is, residual error is negative, is assigned
Residual error 1- θ weight.Solve factor beta to be estimatedθ, it is that regression optimization obtains the weighted average minimum that objective function is residual absolute value,
Effect is xitAnd yitRelated coefficient, i.e., relationship namely correlation rule between the described bird pest correlative factor.
Birds, which are obtained, by solution above collects cluster value yitWith bird pest influence factor xitBetween correlation rule:
yit=x 'itβθ
And then thus obtain birds distribution space-time (i.e. birds collection cluster value yitSet) with bird pest information (i.e. bird pest influence because
Plain xitSet) between correlation rule set；Using correlation rule set training deeply learning network, building target is deep
Spend intensified learning network model.
The correlation rule set, for determining the relative influence of bird pest；When residual values very little, it may be considered that should
Factor can ignore bird pest influence.Meanwhile the input port of deep neural network cannot direct plunge into all data, close
Connection rule can filter out more valuable factor and be trained.Obtained effect is easier to ensure that, and time consumption for training will
It greatly shortens.
Step 5, global association degree is deduced
Collect the bird pest in other regions in the bird pest region of hotspot graph to be constructed other than the finite region described in the step 2
The limited multi-source bird pest information data in other regions (is obtained data herein and is less than or equal to bird pest Information Number described in step 2 by information
According to) according to the identical method progress data cleansing of step 2 and pretreatment, and other regions are analyzed based on Pearson correlation coefficient method
The similarity of bird pest information and finite region bird pest information.
The similarity includes 1) geographical environment, landscape information similarity calculation；2) weather weather information similarity meter
It calculates；3) bird ecology environmental model similarity calculation；4) transmission line of electricity history O&M situation similarity calculation.
The multi-source bird pest information of the finite region, refers to the data that can be directly obtained such as weather, meteorology, this
A little basic datas even can also obtain in other regions of not particulars.
The Pearson correlation coefficient of other regions bird pest influence factor X and finite region bird pest influence factor Y is to use other
What the covariance of region bird pest influence factor X and finite region bird pest influence factor Y was obtained divided by the standard deviation of the two:
ρ in formulaX,YIndicate the Pearson correlation coefficient between X and Y, μXIndicate the average value of X, μYIndicate the average value of Y, σX
Indicate the standard deviation of X, σYIndicate the standard deviation of Y, the value range of related coefficient is [- 1,1], works as ρX,YX and Y are then indicated when=0
Dissmilarity works as ρX,YIndicate that two variables are positively correlated when > 0, conversely, working as ρX,YTwo variable negative correlation are indicated when < 0.
Then the influence factor of other region bird pests can use the bird pest influence factor of Pearson correlation coefficient and finite region
Instead of:
X=ρX,YY
Then based on target depth intensified learning network model, the net in other regions is obtained using transfer learning method
Network model, by other regions bird pest influence factor X input network model, the bird pest information in other regions that can be predicted, then
It can know the bird pest distribution situation of other regional predictions.Finally the bird pest by the distribution of finite region bird pest and other regions is distributed
It is integrated to obtain global (the bird pest region of hotspot graph to be constructed) bird pest distributed intelligence.
Step 6, bird pest situation information constructs
Global bird pest distributed intelligence is obtained according to prediction, bird pest situation map is constructed, if D={ D1,D2,…,DnIt is bird pest state
The set of n situation harmful bird in gesture figure, U=[u1,u2,…,um] it is the set for describing the m influence factor index of situation harmful bird；
The summation at current time, bird pest situation information can use original situation information matrix A=(aij)n×mIt indicates:
Matrix A=(aij)n×mIn every a line ui=[ai1,…,aim] indicate situation harmful bird DiM influence factor index
Under situation information, situation harmful bird DiInfluence factor index aijBigger expression situation harmful bird DiActivity to the net where it
The transmission line of electricity harm of lattice is bigger, i.e., for transmission line of electricity, aijIt is the smaller the better；The harm refers to birds activity to transmission line of electricity
Caused by degree of danger.
Each element information dimension is different inside original situation information matrix A, and Meaning of Information is different, needs to carry out standard to it
The processing of change can just be compared and treated.Therefore column element each in matrix A is standardized, obtains standardization square
Battle arrayWherein:
In order to quantify the degree of danger of birds, we indicate the harm of bird pest using bird pest hot spot value, when hot spot value is got over
It is high then indicate that bird pest danger is higher.The calculation method introduced below that bird pest hot spot value is calculated according to influence factor index.
Step 7, bird pest hot spot value calculates
This programme obtains harmful bird density in grid by improved weighting kernel density estimation method and embodies influence factor index
Influence to bird pest hot spot value.By situation harmful bird DiIt is set as point target, current influence factor index uiUnder each bird pest situation information
Standard value beBy situation harmful bird to the distance for investigating grid s, h is distance
Threshold value；Thus the formula for the bird pest hot spot value f (s) for solving grid s with improved weighting kernel density estimation method is as follows:
This formula can be regarded as: the bird pest hot spot value f (s) of grid s is equal to all situation harmful birds pair for being less than h away from its distance
Its density contribution value giLinear superposition.G in formulaiIndicate situation harmful bird i to the density contribution value of grid s, h > 0 in formula, indicate away from
From threshold value,For situation harmful bird DiInfluence factor index weights item, initially takeAdjust influence factor
Influence of the adjustable influence factor weight of the functional form of index weights item to bird pest hot spot value size.Here k (di, h) and choosing
For classical quartic polynomial kernel function, classical quartic polynomial kernel function meet it is assumed hereinafter that: situation harmful bird distance is remoter,
Influence to hot spot value is smaller, exists simultaneously a distance threshold h, and after distance is more than threshold value h, situation harmful bird is to investigated grid
Bird pest hot spot value does not have an impact.
Similarly, all grids are traversed by improved weighting kernel density estimation method, finds out the bird pest hot spot value of each grid, so
Bird pest hot spot value accumulated matrix can be constructed by merging afterwards:
F=(fab)p×q
Wherein, p, q are size of mesh opening, fabIndicates coordinate be (a, b) grid s before current bird pest hotspot graph t duration model
Enclose interior bird pest hot spot value accumulated value, f under original stateab=0.
Bird pest hot spot value accumulated matrix F is standardized, bird pest hot spot value matrix is finally obtained:
It is as follows to standardize formula:
It is then each in bird pest hot spot value normalized matrix F*Value is equal to the bird pest hot spot value R in (a, the b) grid
(s)。
Step 8, graphical treatment
Including constructing color mapping table, as shown in figure 3, the mapping of bird pest hot spot value to color is realized, so as to form hot spot
Figure figure layer.According to current time bird pest hot spot value matrixAnd color mapping table, hot spot value matrix is converted
For the color image distinguished by different colours, as shown in Figure 4.
As it can be seen that the construction method of the transmission of electricity bird pest hotspot graph in corridor proposed by the present invention based on limited information, highlights
Bird pest situation hot spot, guide line inspection can effectively promote the quality and efficiency of bird pest inspection in electric power O&M.
Claims (7)
1. a kind of construction method of the transmission of electricity corridor bird pest hotspot graph based on limited information, which comprises the steps of:
Step 1, regional geography space lattice divides
The bird pest region for determining hotspot graph to be constructed first carries out gridding division by geographical space to region；
Step 2, finite region bird pest information is collected
The finite region in the bird pest region is chosen as sample and carries out data collection, acquires bird pest in the finite region
Information cleans data, is processed into the data mode that can be excavated, then data prediction, generates for data mining algorithm
The target data used；
Step 3, bird pest analysis of Influential Factors
To step 2, treated that data are read out and carry out data aggregation, extensive, standardization, attribute construction, convergence analysis,
Frequent item set is identified from data, then using the rule of these frequent item sets creation description incidence relation, finds all
Degree of holding is greater than or equal to given minimum support frequent item set, finds generation and is not less than Minimum support4 rule association,
Birds are distributed the informational linkages such as space-time and geographical weather environment and get up, and therefrom analysis selection bird pest Correlative Influence Factors；
Step 4, finite region bird pest and influence factor correlation analysis
Feature is extracted and set up from step 2 treated data, carries out bird pest sequence information data, including birds distribution time
The data mining of sequence, distribution space sequence, Relevant Factors of Distribution sequence；Bird pest sequence information data is returned using fractional bits
Method quantitative analysis bird pest influence factor obtains the correlation rule set between birds distribution space-time and bird pest information；Utilize this
Correlation rule set trains deeply learning network, constructs target depth intensified learning network model；
Step 5, global association degree is deduced
The bird pest information for collecting other regions in bird pest region other than the finite region described in the step 2, by other regions
Bird pest information data carries out data cleansing and pretreatment according to the identical method of step 2, and based on Pearson correlation coefficient method point
Analyse the similarity of other region bird pest information and finite region bird pest information；
Then based on target depth intensified learning network model, the network mould in other regions is obtained using transfer learning method
Type, to know the bird pest distribution situation of other regional predictions；Finally by the bird pest of finite region bird pest distribution and other regions
Distribution is integrated to obtain global bird pest distributed intelligence；
Step 6, bird pest situation information constructs
Global bird pest distributed intelligence is obtained according to prediction, constructs bird pest situation map；In order to quantify the degree of danger of birds, using bird
Evil hot spot value indicates the harm of bird pest, when the hot spot value the high, indicates that bird pest danger is higher；
Step 7, bird pest hot spot value calculates
The bird pest hot spot value in each grid is calculated, bird pest hot spot value matrix is constructed；
Step 8, graphical treatment
Color mapping table is constructed, the mapping of bird pest hot spot value to color is realized, so as to form hotspot graph figure layer；According to bird pest hot spot
Value matrix and color mapping table convert bird pest hot spot value matrix to the color image distinguished by different colours.
2. the construction method of the transmission of electricity corridor bird pest hotspot graph based on limited information as described in claim 1, which is characterized in that
The bird pest information includes: existing literature research, cabling investigation on the spot, bird pest Tripping data, bird pest distributed intelligence, environment
Landscape data, GIS data, history meteorological data, electric power O&M record, shaft tower, line construction feature, real-time detection apparatus, view
Frequency image, radar monitoring and bird damage prevention device information, bird ecology model, bird ecology behavioural characteristic.
3. the construction method of the transmission of electricity corridor bird pest hotspot graph based on limited information as described in claim 1, which is characterized in that
The data mining algorithm include K-Means algorithm, support vector machines, Apriori algorithm, EM algorithm,
AdaBoost, K arest neighbors sorting algorithm.
4. the construction method of the transmission of electricity corridor bird pest hotspot graph based on limited information as described in claim 1, which is characterized in that
Fractional bits homing method quantitative analysis bird pest influence factor, regression model are used described in step 4 are as follows:
yit=x 'itβθ+μθit
Quantθ(yit|xit)=x 'itβθ
In formula, yitCollect cluster value, x for i-th of sample data birds of t momentitFor bird pest influence factor, βθFor under residual error θ wait estimate
Coefficient, μθitFor i-th of sample data of t moment at residual error θ residual error variable, Quantθ(yit|xit) indicate given xitUnder the conditions of
yitResidual error θ condition quantile, the value range of residual error θ are as follows: 0≤θ≤1；
Solve factor beta to be estimatedθObjective function be residual absolute value weighted average it is minimum, formula is as follows:
Work as yit≥x′itβθ, i.e. residual error is timing, assigns residual error θ weight；Work as yit< x 'itβ when that is, residual error is negative, assigns residual error
1- θ weight；Solve factor beta to be estimatedθ, i.e. correlation rule, and then thus obtain the pass between birds distribution space-time and bird pest information
Join regular collection.
5. the construction method of the transmission of electricity corridor bird pest hotspot graph based on limited information as described in claim 1, which is characterized in that
Similarity described in step 5 includes: geographical environment, landscape information similarity calculation；Weather weather information similarity calculation；Birds
Ecological environment distortion calculates；Transmission line of electricity history O&M situation similarity calculation.
6. the construction method of the transmission of electricity corridor bird pest hotspot graph based on limited information as described in claim 1, which is characterized in that
Global bird pest distributed intelligence is obtained according to prediction described in step 6, constructs bird pest situation map, comprising:
If D={ D1,D2,…,DnBe bird pest situation map in n situation harmful bird set, U=[u1,u2,…,um] it is description state
The set of the m influence factor index of gesture harmful bird；The summation at current time, bird pest situation information can use original situation information matrix
A=(aij)n×mIt indicates:
Matrix A=(aij)n×mIn every a line ui=[ai1,…,aim] indicate situation harmful bird DiM influence factor index under
Situation information, situation harmful bird DiInfluence factor index aijBigger expression situation harmful bird DiActivity to the grid where it
Transmission line of electricity harm is bigger；
Column element each in matrix A is standardized, Standard Process is obtainedWherein:
7. the construction method of the transmission of electricity corridor bird pest hotspot graph based on limited information as described in claim 1, which is characterized in that
Bird pest hot spot value in each grid of calculating described in step 7, constructs bird pest hot spot value matrix, comprising:
By situation harmful bird DiIt is set as point target, current influence factor index uiUnder the standard value of each bird pest situation information bediBy situation harmful bird to the distance for investigating grid s, h is distance threshold；Solve net
The formula of the bird pest hot spot value f (s) of lattice s is as follows:
G in formulaiSituation harmful bird i is indicated to the density contribution value of grid s, h > 0 in formula indicates distance threshold,For situation evil
Bird DiInfluence factor index weights item, initially take
All grids are traversed, the bird pest hot spot value of each grid is found out, bird pest hot spot value accumulated matrix can be constructed by being then combined with:
F=(fab)p×q
Wherein, p, q are size of mesh opening, fabIndicates coordinate is the grid s of (a, b) before current bird pest hotspot graph within the scope of t duration
Bird pest hot spot value accumulated value, f under original stateab=0；
Bird pest hot spot value accumulated matrix F is standardized, bird pest hot spot value matrix is finally obtained:
It is as follows to standardize formula:
It is then each in bird pest hot spot value matrix F*Value is equal to the bird pest hot spot value R (s) in (a, the b) grid.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811455816.8A CN109710595B (en) | 2018-11-30 | 2018-11-30 | Method for constructing bird damage hotspot diagram of power transmission corridor based on limited information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811455816.8A CN109710595B (en) | 2018-11-30 | 2018-11-30 | Method for constructing bird damage hotspot diagram of power transmission corridor based on limited information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109710595A true CN109710595A (en) | 2019-05-03 |
| CN109710595B CN109710595B (en) | 2022-03-25 |
Family
ID=66254458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811455816.8A Active CN109710595B (en) | 2018-11-30 | 2018-11-30 | Method for constructing bird damage hotspot diagram of power transmission corridor based on limited information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109710595B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110188832A (en) * | 2019-05-08 | 2019-08-30 | 中国民航科学技术研究院 | Based on the airport resident bird activity rhythm statistical analysis technique for visiting bird radar |
| CN111860945A (en) * | 2020-06-10 | 2020-10-30 | 广东电网有限责任公司 | Power transmission line bird distribution prediction method based on multi-source data |
| CN112669970A (en) * | 2020-12-30 | 2021-04-16 | 华南师范大学 | Infectious disease space-time prediction method based on big data deep learning and robot |
| CN115912183A (en) * | 2023-03-09 | 2023-04-04 | 国网湖北省电力有限公司经济技术研究院 | High-voltage transmission line ecological measure inspection method and system and readable storage medium |
| CN117271959A (en) * | 2023-11-21 | 2023-12-22 | 中南大学 | Uncertainty evaluation method and equipment for PM2.5 concentration prediction result |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008136362A (en) * | 2006-11-29 | 2008-06-19 | Chugoku Electric Power Co Inc:The | Method for preventing bird damage, utility pole capable of preventing bird damage, transmission steel tower capable of preventing bird damage, and method for repelling birds |
| CN102393879A (en) * | 2011-07-21 | 2012-03-28 | 广东电网公司佛山供电局 | Method and device for drawing bird damage distribution graph of power grid |
| CN104268794A (en) * | 2014-09-19 | 2015-01-07 | 国家电网公司 | Drawing method of power transmission line bird damage distribution diagram |
| CN105335898A (en) * | 2015-11-11 | 2016-02-17 | 国网内蒙古东部电力有限公司电力科学研究院 | Bird-caused damage risk distribution map drawing system |
| US20160055400A1 (en) * | 2014-08-21 | 2016-02-25 | Boulder Imaging, Inc. | Avian detection systems and methods |
| CN106405339A (en) * | 2016-11-11 | 2017-02-15 | 中国南方电网有限责任公司 | Power transmission line fault reason identification method based on high and low frequency wavelet feature association |
| CN107092993A (en) * | 2017-06-05 | 2017-08-25 | 国网陕西省电力公司电力科学研究院 | Reclosing success rate association analysis method based on Disasters Type and line information |
-
2018
- 2018-11-30 CN CN201811455816.8A patent/CN109710595B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008136362A (en) * | 2006-11-29 | 2008-06-19 | Chugoku Electric Power Co Inc:The | Method for preventing bird damage, utility pole capable of preventing bird damage, transmission steel tower capable of preventing bird damage, and method for repelling birds |
| CN102393879A (en) * | 2011-07-21 | 2012-03-28 | 广东电网公司佛山供电局 | Method and device for drawing bird damage distribution graph of power grid |
| US20160055400A1 (en) * | 2014-08-21 | 2016-02-25 | Boulder Imaging, Inc. | Avian detection systems and methods |
| CN104268794A (en) * | 2014-09-19 | 2015-01-07 | 国家电网公司 | Drawing method of power transmission line bird damage distribution diagram |
| CN105335898A (en) * | 2015-11-11 | 2016-02-17 | 国网内蒙古东部电力有限公司电力科学研究院 | Bird-caused damage risk distribution map drawing system |
| CN106405339A (en) * | 2016-11-11 | 2017-02-15 | 中国南方电网有限责任公司 | Power transmission line fault reason identification method based on high and low frequency wavelet feature association |
| CN107092993A (en) * | 2017-06-05 | 2017-08-25 | 国网陕西省电力公司电力科学研究院 | Reclosing success rate association analysis method based on Disasters Type and line information |
Non-Patent Citations (2)
| Title |
|---|
| JUAN MANUEL PÉREZ-GARCÍA ET AL: "Using risk prediction models and species sensitivity maps for large-scale", 《BIOLOGICAL CONSERVATION》 * |
| 卢恩泽 等: "基于灾害类型与线路信息的电网重合闸成功率关联分析", 《高压电器》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110188832A (en) * | 2019-05-08 | 2019-08-30 | 中国民航科学技术研究院 | Based on the airport resident bird activity rhythm statistical analysis technique for visiting bird radar |
| CN110188832B (en) * | 2019-05-08 | 2020-12-08 | 中国民航科学技术研究院 | Airport bird-remaining activity rhythm statistical analysis method based on bird-detecting radar |
| CN111860945A (en) * | 2020-06-10 | 2020-10-30 | 广东电网有限责任公司 | Power transmission line bird distribution prediction method based on multi-source data |
| CN112669970A (en) * | 2020-12-30 | 2021-04-16 | 华南师范大学 | Infectious disease space-time prediction method based on big data deep learning and robot |
| CN115912183A (en) * | 2023-03-09 | 2023-04-04 | 国网湖北省电力有限公司经济技术研究院 | High-voltage transmission line ecological measure inspection method and system and readable storage medium |
| CN117271959A (en) * | 2023-11-21 | 2023-12-22 | 中南大学 | Uncertainty evaluation method and equipment for PM2.5 concentration prediction result |
| CN117271959B (en) * | 2023-11-21 | 2024-02-20 | 中南大学 | Uncertainty evaluation method and equipment for PM2.5 concentration prediction result |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109710595B (en) | 2022-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109710595A (en) | The construction method of transmission of electricity corridor bird pest hotspot graph based on limited information | |
| CN109376605B (en) | Electric power inspection image bird-stab-prevention fault detection method | |
| CN107728234B (en) | Thunder and lightning strength value prediction method based on atmospheric electric field data | |
| CN102096825A (en) | Graph-based semi-supervised high-spectral remote sensing image classification method | |
| CN116625327B (en) | Big data driven marine environment monitoring and early warning method | |
| Saxena et al. | A review study of weather forecasting using artificial neural network approach | |
| CN109917394B (en) | Short-term intelligent extrapolation method based on weather radar | |
| CN102999927A (en) | Fine partition method of soil pollutant content spatial distribution | |
| Elbeltagi et al. | Optimizing hyperparameters of deep hybrid learning for rainfall prediction: a case study of a Mediterranean basin | |
| Wang et al. | Projection Pursuit Method Based on Connection Cloud Model for Assessment of Debris Flow Disasters. | |
| Lu et al. | Lightning strike location identification based on 3D weather radar data | |
| Yao et al. | CIKM AnalytiCup 2017: short-term precipitation forecasting based on radar reflectivity images | |
| Noor et al. | Prediction map of rainfall classification using random forest and inverse distance weighted (IDW) | |
| CN117033923A (en) | Method and system for predicting crime quantity based on interpretable machine learning | |
| CN103268416A (en) | Wind speed prediction method based on STCP-BP | |
| Liu et al. | Integration transformer for ground-based cloud image segmentation | |
| Socaci et al. | XNow: A deep learning technique for nowcasting based on radar products’ values prediction | |
| Aranay et al. | Deep active genetic learning-based assessment of lakes’ water quality using climate data | |
| CN114063063A (en) | Geological disaster monitoring method based on synthetic aperture radar and point-like sensor | |
| Qin et al. | Sea surface temperature clustering based on type-2 fuzzy theory | |
| Jaseena et al. | An improved multivariate weather prediction model using deep neural networks and particle swarm optimisation | |
| Ghaderi et al. | Improving AMSR2 total precipitable water vapour model using metaheuristic algorithms | |
| Li et al. | Research on application of graph neural network in water quality prediction | |
| Zhang | Large-scale ship fault data retrieval algorithm supporting complex query in cloud computing | |
| Fior et al. | Estimating the incidence of adverse weather effects on road traffic safety using time series embeddings |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |