CN109710595B - Method for constructing bird damage hotspot diagram of power transmission corridor based on limited information - Google Patents
Method for constructing bird damage hotspot diagram of power transmission corridor based on limited information Download PDFInfo
- Publication number
- CN109710595B CN109710595B CN201811455816.8A CN201811455816A CN109710595B CN 109710595 B CN109710595 B CN 109710595B CN 201811455816 A CN201811455816 A CN 201811455816A CN 109710595 B CN109710595 B CN 109710595B
- Authority
- CN
- China
- Prior art keywords
- bird
- bird damage
- information
- data
- damage
- 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.)
- Expired - Fee Related
Links
- 230000006378 damage Effects 0.000 title claims abstract description 208
- 238000000034 method Methods 0.000 title claims abstract description 45
- 230000005540 biological transmission Effects 0.000 title claims abstract description 44
- 238000010586 diagram Methods 0.000 title claims abstract description 21
- 238000009826 distribution Methods 0.000 claims abstract description 48
- 238000012423 maintenance Methods 0.000 claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 9
- 238000004458 analytical method Methods 0.000 claims abstract description 5
- 238000010276 construction Methods 0.000 claims abstract description 5
- 241000607479 Yersinia pestis Species 0.000 claims description 50
- 239000011159 matrix material Substances 0.000 claims description 33
- 238000004422 calculation algorithm Methods 0.000 claims description 18
- 238000007418 data mining Methods 0.000 claims description 13
- 230000000694 effects Effects 0.000 claims description 13
- 238000013507 mapping Methods 0.000 claims description 10
- 230000002787 reinforcement Effects 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000009825 accumulation Methods 0.000 claims description 6
- 238000004140 cleaning Methods 0.000 claims description 6
- 239000003086 colorant Substances 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000011160 research Methods 0.000 claims description 6
- 230000002265 prevention Effects 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 230000002776 aggregation Effects 0.000 claims description 4
- 238000004220 aggregation Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 238000012544 monitoring process Methods 0.000 claims description 4
- 230000006399 behavior Effects 0.000 claims description 3
- 238000007635 classification algorithm Methods 0.000 claims description 3
- 230000007613 environmental effect Effects 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims description 3
- 238000011897 real-time detection Methods 0.000 claims description 3
- 238000012706 support-vector machine Methods 0.000 claims description 3
- 238000013526 transfer learning Methods 0.000 claims description 3
- 238000007689 inspection Methods 0.000 abstract description 9
- 238000004364 calculation method Methods 0.000 abstract description 8
- 238000000556 factor analysis Methods 0.000 abstract description 4
- 230000008859 change Effects 0.000 abstract description 3
- 238000011161 development Methods 0.000 abstract description 3
- 230000008030 elimination Effects 0.000 abstract description 2
- 238000003379 elimination reaction Methods 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 6
- 108010014173 Factor X Proteins 0.000 description 3
- 230000002596 correlated effect Effects 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 241001465754 Metazoa Species 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 235000005770 birds nest Nutrition 0.000 description 1
- 244000240635 birds nest Species 0.000 description 1
- 238000009395 breeding Methods 0.000 description 1
- 230000001488 breeding effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 210000003608 fece Anatomy 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 235000005765 wild carrot Nutrition 0.000 description 1
Images
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 invention discloses a method for constructing a bird damage hot spot diagram of a power transmission corridor based on limited information, which comprises the following steps: the method comprises the steps of regional geographic space grid division, limited region bird damage information collection, bird damage influence factor analysis, limited region bird damage influence factor association degree analysis, global association degree deduction, bird damage situation information construction, bird damage hot point value calculation and graphical processing. According to the invention, multi-source data such as geographic information, meteorological information and bird distribution information are fused, the change of bird damage distribution areas and occurrence time of the power transmission line in a large range is estimated in time, an effective basis is provided for operation and maintenance of the power transmission line, bird damage routing inspection and danger elimination efficiency of the power transmission line is improved, the safe operation capacity of the power transmission line is improved, and high-quality development of a power grid is further promoted.
Description
Technical Field
The invention relates to the technical field of intelligent power operation and maintenance, in particular to a method for constructing a bird damage hot spot diagram of a power transmission corridor based on limited information.
Background
In recent years, along with the continuous improvement of natural environment in China and the continuous improvement of laws and regulations for protecting animals, the breeding of birds is gradually accelerated, the birds move more and more frequently, and the safe and efficient operation of occasions such as power transmission lines, airports and the like is seriously influenced. For example, in the power transmission line, bird droppings of birds falling in the flying process pollute the power transmission line and form flashover; meanwhile, birds nest on the tower, and the falling of nesting materials can cause faults such as short circuit and tripping of the power transmission line. Therefore, it is necessary to construct a distribution of power transmission corridor pest birds (bird-related) and a heat map of bird pest (where the relevant birds pose a hazard or are likely to pose a hazard).
The safe and efficient operation of the power transmission line directly determines the stability of the whole power grid, is the premise of stable power utilization of users, and is the guarantee of high-quality development of the power grid. In recent years, a line trip power-off fault caused by bird damage has become one of the main faults of a power transmission network. The bird pest is because its kind is many, and is in large quantity, and the distribution is wide, and the change is fast, easily receives geographical climate influence, and current bird pest prevention and control and monitoring mainly make the judgement and mark through artifical patrolling and examining, unmanned aerial vehicle patrols and examines or fixed point video acquisition, and the blindness is big, prevents that the bird is untimely, prevents that the bird effect is not obvious. The electric power operation and maintenance personnel can hardly timely and effectively master the bird damage condition, huge manpower and material resources are wasted, and a better prevention and control effect is difficult to achieve.
The current bird damage occurrence place and time estimation is mainly determined according to operation and maintenance data of all the years, and due to uncertainty of bird damage activities, bird damage areas and time are different every year, operation and maintenance difficulty is large, and the effect is poor. If the information such as geography, weather and ecology can be fused on the basis of bird damage operation and maintenance records in the past year, the bird damage distribution information of the power transmission line can be updated in time, the effectiveness and the guiding significance of the bird damage distribution map are improved, and the problem to be solved urgently in the current power transmission operation and maintenance process is solved. Therefore, an automatic method is urgently needed to be constructed, so that hot spots of bird pest situations are highlighted, and route inspection is guided.
Disclosure of Invention
Aiming at the practical problems existing in the power grid inspection, the invention provides the method for constructing the bird damage hot spot diagram of the power transmission corridor based on the limited information, and the quality and the efficiency of bird damage inspection in power operation and maintenance can be effectively improved.
In order to realize the task, the invention adopts the following technical scheme:
a method for constructing a bird damage hotspot graph of a power transmission corridor based on limited information comprises the following steps:
step 1, regional geospatial meshing
Firstly, determining a bird damage area to be constructed with a heat point diagram, and carrying out gridding division on the area according to a geographic space;
step 2, collecting bird damage information in limited area
Selecting a limited region in the bird damage region as a sample, collecting data, collecting bird damage information in the limited region, cleaning the data, processing the data into a data form capable of being mined, and then preprocessing the data to generate target data for a data mining algorithm;
step 3, bird damage influence factor analysis
Reading the data processed in the step 2, performing data aggregation, generalization, normalization, attribute construction and fusion analysis, identifying frequent item sets from the data, creating rules describing association relations by using the frequent item sets, searching all frequent item sets with the support degree greater than or equal to a given minimum support degree, searching and generating association with the minimum reliability rule, associating bird distribution space-time with information such as geographic meteorological environment and the like, and analyzing and selecting bird damage related influence factors;
step 4, analyzing the association degree of bird damage and influence factors in the limited area
Extracting and constructing characteristics from the data processed in the step 2, and performing data mining on bird damage sequence information data including bird distribution time sequences, distribution space sequences and distribution related factor sequences; quantitatively analyzing bird damage influence factors by adopting a fractional bit regression method on bird damage sequence information data to obtain an association rule set between bird distribution space-time and bird damage information; training a deep reinforcement learning network by using the association rule set to construct a target deep reinforcement learning network model;
step 5, global relevance deduction
Collecting bird damage information of other areas except the limited area in the step 2 in the bird damage area, carrying out data cleaning and preprocessing on the bird damage information data of the other areas according to the same method in the step 2, and analyzing the similarity between the bird damage information of the other areas and the bird damage information of the limited area based on a Pearson correlation coefficient method;
then, on the basis of the target deep reinforcement learning network model, obtaining network models of other areas by adopting a transfer learning method, thereby obtaining the bird damage distribution condition predicted by the other areas; finally, integrating the bird damage distribution of the limited area and the bird damage distribution of other areas to obtain overall bird damage distribution information;
step 6, building bird pest situation information
Obtaining global bird damage distribution information according to prediction, and constructing a bird damage situation map; in order to quantify the danger degree of the birds, a bird damage hot spot value is adopted to represent the damage of the bird damage, and when the hot spot value is higher, the bird damage risk is higher;
step 7, calculating bird damage hot point values
Calculating bird damage hot point values in each grid, and constructing a bird damage hot point value matrix;
step 8, graphical processing
Constructing a color mapping table to realize mapping from bird damage hot point values to colors so as to form a hot point diagram layer; and converting the bird damage hot spot value matrix into a color image distinguished by different colors according to the bird damage hot spot value matrix and the color mapping table.
Further, the bird damage information includes: the system comprises the following components of the existing literature research, on-site wiring research, bird trouble tripping data, bird trouble distribution information, environmental landscape data, GIS data, historical meteorological data, electric power operation and maintenance records, towers, line structure characteristics, real-time detection devices, video images, radar monitoring and bird trouble prevention device information, bird ecology models and bird ecology behavior characteristics.
Further, the data mining algorithm comprises a K-Means algorithm, a support vector machine, an Apriori algorithm, a maximum expectation algorithm, AdaBoost and a K nearest neighbor classification algorithm.
Further, the quantitative analysis of bird damage influence factors by using the fractional bit regression method in the step 4 is characterized in that a regression model is as follows:
yit＝x′itβθ+μθit
Quantθ(yit|xit)＝x′itβθ
in the formula, yitFor the ith sample data bird cluster value at time t, xitBeta is a factor affecting bird damageθIs the coefficient to be estimated under the residual thetaθitResidual variable, Quant, of the ith sample data at time t under residual thetaθ(yit|xit) Denotes a given xitUnder the condition of yitThe conditional quantile of the residual theta is as follows: theta is more than or equal to 0 and less than or equal to 1;
solving the coefficient beta to be estimatedθThe objective function of (2) is the weighted average minimum of the absolute values of the residuals, and the formula is as follows:
when y isit≥x′itβθWhen the residual is positive, giving the weight of the residual theta; when y isit＜x′itBeta, namely when the residual error is negative, giving the weight of 1-theta to the residual error; solving the coefficient beta to be estimatedθNamely, association rules, and then an association rule set between bird distribution space-time and bird damage information is obtained.
Further, the similarity in step 5 includes: calculating the similarity of the geographic environment and the landscape information; calculating the similarity of weather and meteorological information; calculating the similarity of the bird ecological environment model; and calculating the similarity of the historical operation and maintenance conditions of the power transmission line.
Further, the step 6 of obtaining global bird damage distribution information according to the prediction and constructing a bird damage situation map includes:
let D ═ D1,D2,…,DnThe situation is the set of n situation pest birds in the bird pest situation map, U ═ U1,u2,…,um]A set of m influence factor indexes for describing situation pest birds; at the current moment, the sum of bird pest situation information can be represented by an original situation information matrix A ═ aij)n×mRepresents:
the matrix A ═ aij)n×mEach row u ini＝[ai1,…,aim]Representing situation pest DiSituation information under the m influence factor indexes, situation pest DiIndex of influence factor ofijThe larger the size of the pest bird DiThe greater the damage of the activity of (2) to the transmission line of the grid in which it is located;
standardizing each row of elements in the matrix A to obtain a standardized matrix
Wherein:
further, the step 7 of calculating the bird damage hot point values in each grid to construct a bird damage hot point value matrix includes:
will be in situation to pest birds DiSet as point target, current influence factor index uiThe standard value of each bird pest situation information is
diThe distance from the situation pest to the checked grid s is defined, and h is a distance threshold value; the formula for solving the bird damage hot point value f(s) of the grid s is as follows:
in the formula giRepresenting the density contribution value of the situation pest i to the grid s, wherein h>0, representing a distance threshold value,
for pest birds DiShadow ofThe weight term of the sound factor index is initially taken
Traversing all grids, solving the bird damage hot point values of all the grids, and then combining to construct a bird damage hot point value accumulation matrix:
F＝(fab)p×q
wherein p and q are the grid size, fabRepresenting the accumulated value of bird damage hot point values of grids s with coordinates (a, b) in the time length range t before the current bird damage hot point diagram, and f in the initial stateab＝0；
Standardizing the bird damage hot point value accumulation matrix F to finally obtain a bird damage hot point value matrix:
the normalization formula is as follows:
then each in the matrix F of bird damage hot spot values
The value is equal to the bird pest hotspot value r(s) in the (a, b) grid.
Compared with the prior art, the invention has the following technical characteristics:
1. according to the invention, multi-source data such as geographic information, meteorological information and bird distribution information are fused, the change of bird damage distribution areas and occurrence time of the power transmission line in a large range is estimated in time, an effective basis is provided for operation and maintenance of the power transmission line, bird damage routing inspection and danger elimination efficiency of the power transmission line is improved, the safe operation capacity of the power transmission line is improved, and high-quality development of a power grid is further promoted.
2. The invention constructs an automatic method, highlights the situation hot spot of the bird damage, guides the line inspection and can effectively improve the quality and efficiency of the bird damage inspection in the electric power operation and maintenance.
Drawings
FIG. 1 is a schematic overall flow diagram of the process of the present invention;
FIG. 2 is a schematic diagram of a bird damage data mining process model disclosed in an embodiment of the present invention;
FIG. 3 is a color mapping table disclosed in an embodiment of the present invention;
fig. 4 is a schematic diagram of a hot spot map layer disclosed in the embodiment of the present invention.
Detailed Description
The embodiment of the invention discloses a method for constructing a bird damage hotspot graph of a power transmission corridor based on limited information, which comprises the following steps of: the method comprises the steps of regional geographic space grid division, limited region bird damage information collection, bird damage influence factor analysis, limited region bird damage influence factor association degree analysis, global association degree deduction, bird damage situation information construction, bird damage hot point value calculation and graphical processing, and the steps of the method are explained in detail below.
Step 1, regional geospatial meshing
Firstly, determining a bird damage area to be constructed with a heat point diagram, and carrying out gridding division on the area according to a geographic space to form a p × q grid, wherein the size of the grid is selected to be c × c; the size of the grid can be determined according to the accuracy of the hot spot map, the situation is finer when the grid division is denser, but the processing speed is reduced, and the size of the grid can be adjusted appropriately to achieve the best effect.
The grid can select multiple gears such as 50m × 50m, 100m × 100m, 200m × 200m, 500m × 500m, 1km × 1km and the like; generally, the number of grids is increased, the calculation accuracy is improved, but the calculation scale is also increased, so that the number of grids is determined by balancing two factors. If the precision of the currently selected grid can not meet the precision of the actual requirement, the grid size is adjusted to be the small grid to meet the precision requirement, the balance between the bird damage display precision and the calculation capability is achieved, and the best bird damage state under the calculation capability is met.
Step 2, collecting bird damage information in limited area
Selecting a limited area in the bird damage area of the hot spot diagram to be constructed as a sample, collecting data, collecting bird damage information in the limited area, cleaning the data, processing data (multi-source data) from different operating systems and in different formats into a data form capable of being mined, preprocessing the data, and generating target data for a data mining algorithm, wherein the target data is a bird damage data mining process model as shown in fig. 2.
The limited area selection has two methods, one is to select typical areas (such as typical landform areas of lakes, fields, mountains, hills and the like) as samples to collect bird damage data, and the other is to select randomly so as to cover all bird damage conditions as much as possible and enable results to be more accurate. The typical sample and the random sample size of each power transmission corridor are generally long strip areas within 200m-500m around a power transmission line corridor with the length of 2-5km, at least 10% of sample lines are required to be selected for each hundred kilometers of power transmission lines, and in addition, the number of the sample lines can be increased or decreased according to the number of the topographic features of the areas where the power transmission lines pass. For example, a 100km long power transmission corridor passes through three typical landforms of a lake, a field and a mountain midway, 3km sample lines are respectively selected in a typical area, and two sections of randomly selected 3km are used as sample line acquisition data.
The bird damage information comprises but is not limited to existing literature research, on-site wiring research, bird damage trip data, bird damage distribution information, environmental landscape data (such as land utilization, water bodies, artificial buildings, city transition and the like), GIS data, historical meteorological data, electric power operation and maintenance records, towers, line structure characteristics, real-time detection devices, video images, radar monitoring and bird damage prevention device information, bird ecology models, bird ecology behavior characteristics and the like.
The mineable data form refers to a file system format widely used on a Hadoop system, such as sequence File, MapFile, RCFile, ORCFile and Parque; in practical application, the optimal file format is selected according to different characteristics of the file. The preprocessing process adopts a general method of data mining, and comprises classification, regression, clustering, data dimension reduction, model selection and the like.
Data mining algorithms are also not fixed due to the variety of data sources, and available core algorithms include, but are not limited to, K-Means algorithms, support vector machines, Apriori algorithms, maximum Expectation (EM) algorithms, AdaBoost, K-Nearest Neighbor (KNN) classification algorithms, and the like.
Because the bird damage areas are widely distributed, the whole hot spot graph bird damage area to be constructed (data collection can not be carried out, the method adopted by the scheme is that a part of the area (namely the limited area in the step 2) is selected for carrying out bird damage data collection, the incidence relation between the bird damage in the limited area and factors such as the environment is analyzed, then the area without detailed bird damage data (namely other areas in the step 5) is popularized and moved, the incidence relation between the bird damage in other areas and the factors such as the environment is obtained, the incidence relation between the bird damage in the whole area (namely the whole bird damage area) and the factors such as the environment is finally obtained, and then the hot spot graph of the whole bird damage area can be constructed.
Step 3, bird damage influence factor analysis
Reading the data processed in the step 2 by adopting a Hadoop Distributed File System (HDFS), performing data aggregation, generalization, normalization, attribute construction and fusion analysis on the large and various data by an automatic data analysis technology, identifying frequent item sets from the data sets by using Apriori and FP-Growth algorithms, creating rules for describing association relations by using the frequent item sets, searching all frequent item sets with the support degree greater than or equal to a given minimum support degree, searching for association not less than a minimum credibility rule, associating bird distribution space-time with information such as a geographic meteorological environment and the like, and analyzing and selecting bird damage related influence factors.
The bird damage related influence factor refers to a factor having a relation with bird damage, for example, the lower the temperature, the more frequent the bird activity is, the negative relation is between the temperature and the bird activity, and the temperature is the bird damage related influence factor.
Step 4, analyzing the association degree of bird damage and influence factors in the limited area
Extracting and constructing features from the data processed in the step 2, and performing data mining on a bird distribution time sequence, a distribution space sequence and a distribution related factor sequence by using a MapReduce data mining frame and the core algorithm; quantitatively analyzing bird damage influence factors of bird damage sequence information data including bird distribution time sequences, bird distribution space sequences and bird distribution related factor sequences by adopting a fractional bit regression method, wherein a regression model is as follows:
yit＝x′itβθ+μθit
Quantθ(yit|xit)＝x′itβθ
in the formula, yitFor the ith sample data bird cluster value at time t, i.e. regression estimation value, xitIs the actual observed value, beta, of the bird damage influencing factor, i.e. the ith influencing factor at time tθIs the coefficient to be estimated under the residual thetaθitResidual variable, Quant, of the ith sample data at time t under residual thetaθ(yit|xit) Denotes a given xitUnder the condition of yitThe conditional quantile of the residual theta is as follows: theta is more than or equal to 0 and less than or equal to 1.
Solving the coefficient beta to be estimatedθThe objective function of (2) is the weighted average minimum of the absolute values of the residuals, and the formula is as follows:
when y isit≥x′itβθWhen the residual is positive, giving the weight of the residual theta; when y isit＜x′itBeta, i.e., the residual is negative, the residual is given a weight of 1-theta. Solving the coefficient beta to be estimatedθThe weighted average of the absolute value of the residual error obtained by regression optimization is minimum, and the function is xitAnd yitI.e. the relationship between the bird damage related factors, i.e. the association rules.
By solving the aboveTo bird collecting value yitInfluence factor x of bird damageitThe association rules between:
yit＝x′itβθ
thereby obtaining bird distribution space-time (namely bird aggregation value y)itSet of) and bird damage information (i.e., bird damage influencing factor x)itSet) of association rules between the sets; and training a deep reinforcement learning network by using the association rule set to construct a target deep reinforcement learning network model.
The association rule set is used for determining the related influence of bird damage; when the residual value is small, it can be considered that the effect of the factor on bird damage is negligible. Meanwhile, all data cannot be directly input into an input port of the deep neural network, and valuable factors can be screened out for training through association rules. The obtained effect is easier to guarantee, and the training time is greatly shortened.
Step 5, global relevance deduction
Collecting bird damage information of other areas except the limited area in the step 2 in the bird damage area to be constructed with the hot spot diagram, carrying out data cleaning and preprocessing on limited multi-source bird damage information data (the obtained data is less than or equal to the bird damage information data in the step 2) of the other areas according to the same method in the step 2, and analyzing the similarity between the bird damage information of the other areas and the bird damage information of the limited area based on a Pearson correlation coefficient method.
The similarity comprises 1) geographic environment and landscape information similarity calculation; 2) calculating the similarity of weather and meteorological information; 3) calculating the similarity of the bird ecological environment model; 4) and calculating the similarity of the historical operation and maintenance conditions of the power transmission line.
The multi-source bird damage information of the limited area refers to data such as weather, weather and the like which can be directly obtained, and the basic data can be obtained even in other areas without detailed information.
The Pearson correlation coefficient of the bird damage influence factor X in the other areas and the bird damage influence factor Y in the limited areas is obtained by dividing the covariance of the bird damage influence factor X in the other areas and the bird damage influence factor Y in the limited areas by the standard deviation of the bird damage influence factor X in the other areas and the bird damage influence factor Y in the limited areas:
where rhoX,YDenotes the Pearson correlation coefficient, μ, between X and YXDenotes the mean value of X,. mu.YDenotes the mean value of Y,. sigmaXDenotes the standard deviation, σ, of XYThe standard deviation of Y is shown, and the value range of the correlation coefficient is [ -1,1 [)]When rhoX,YWhen 0, X and Y are not similar, and when rhoX,Y>When 0, it means that the two variables are positively correlated, whereas when pX,YA value of < 0 indicates that the two variables are inversely correlated.
Then the influence factors of bird damage in other areas can be replaced by Pearson's correlation coefficient and the influence factors of bird damage in limited areas:
X＝ρX,YY
and then, on the basis of the target deep reinforcement learning network model, obtaining network models of other areas by adopting a transfer learning method, inputting bird damage influence factors X of the other areas into the network models, obtaining predicted bird damage information of the other areas, and obtaining the predicted bird damage distribution conditions of the other areas. And finally integrating the bird damage distribution of the limited area and the bird damage distribution of other areas to obtain global (the bird damage area of the hot spot diagram to be constructed) bird damage distribution information.
Step 6, building bird pest situation information
Obtaining global bird damage distribution information according to prediction, constructing a bird damage situation map, and setting D to { D ═ D1,D2,…,DnThe situation is the set of n situation pest birds in the bird pest situation map, U ═ U1,u2,…,um]A set of m influence factor indexes for describing situation pest birds; at the current moment, the sum of bird pest situation information can be represented by an original situation information matrix A ═ aij)n×mRepresents:
the matrix A ═ aij)n×mEach row u ini＝[ai1,…,aim]Representing situation pest DiSituation information under the m influence factor indexes, situation pest DiIndex of influence factor ofijThe larger the size of the pest bird DiThe greater the damage of the activity of (a) to the transmission line of the grid in which it is located, i.e. for the transmission line, aijThe smaller the better; the harm refers to the danger degree of bird activities to the power transmission line.
The information dimensions and information meanings of all elements in the original situation information matrix A are different, and the elements need to be standardized to be compared and processed. Therefore, each column element in the matrix A is standardized to obtain a standardized matrix
Wherein:
in order to quantify the degree of danger of birds, a bird damage hot spot value is used for representing the damage of birds, and the higher the hot spot value is, the higher the risk of bird damage is. The following describes a calculation method for calculating the bird damage hot spot value according to the influence factor index.
Step 7, calculating bird damage hot point values
According to the scheme, the density of the harmful birds in the grid is obtained through an improved weighted kernel density estimation method, and the influence of the influence factor indexes on the bird damage hot point value is reflected. Will be in situation to pest birds DiSet as point target, current influence factor index uiThe standard value of each bird pest situation information is
The distance from the situation pest to the checked grid s is defined, and h is a distance threshold value; the formula for solving the bird damage hot spot values f(s) of the grid s using the modified weighted kernel density estimation is thus as follows:
this formula can be understood as: the bird damage hot point value f(s) of the grid s is equal to the density contribution g of all situation pest birds with a distance less than h from the gridiLinear superposition of (2). In the formula giRepresenting the density contribution value of the situation pest i to the grid s, wherein h>0, representing a distance threshold value,
for pest birds DiThe weight term of the influence factor index of (2) is initially taken
The influence of the weight of the influence factor on the magnitude of the bird damage hotspot value can be adjusted by adjusting the functional form of the index weight term of the influence factor. Where k (d)iAnd h) a classical quartic polynomial kernel is selected, which satisfies the following assumptions: the farther the situation pest birds are away, the smaller the influence on the hot point value is, and meanwhile, a distance threshold h exists, and when the distance exceeds the threshold h, the situation pest birds do not influence the hot point value of the examined grid birds.
Similarly, all grids are traversed by an improved weighted kernel density estimation method, bird damage hot point values of all grids are solved, and then bird damage hot point value accumulation matrixes can be constructed by combination:
F＝(fab)p×q
wherein p and q are the grid size, fabRepresenting the accumulated value of bird damage hot point values of grids s with coordinates (a, b) in the time length range t before the current bird damage hot point diagram, and f in the initial stateab＝0。
Standardizing the bird damage hot point value accumulation matrix F to finally obtain a bird damage hot point value matrix:
the normalization formula is as follows:
then each in the hot point value standardization matrix F of bird damage
The value is equal to the bird pest hotspot value r(s) in the (a, b) grid.
Step 8, graphical processing
The method comprises the steps of constructing a color mapping table, and realizing mapping of bird damage hot point values to colors as shown in FIG. 3, so as to form a hot point map layer. According to the hot spot value matrix of the bird damage at the current moment
And a color mapping table converting the hot spot value matrix into a color image distinguished by different colors, as shown in fig. 4.
Therefore, the method for constructing the power transmission corridor bird damage hotspot graph based on the limited information highlights bird damage situation hotspots, guides line routing inspection and can effectively improve the quality and efficiency of bird damage inspection in power operation and maintenance.
Claims (6)
1. A method for constructing a bird damage hotspot graph of a power transmission corridor based on limited information is characterized by comprising the following steps:
step 1, dividing regional geographical space grids;
determining a bird damage area to be constructed with a heat point diagram, and carrying out gridding division on the area according to a geographic space;
step 2, collecting bird damage information in a limited area;
selecting a limited region in the bird damage region as a sample, collecting data, collecting bird damage information in the limited region, cleaning the data, processing the data into a data form capable of being mined, and then preprocessing the data to generate target data for a data mining algorithm;
step 3, analyzing bird damage influence factors;
reading the data processed in the step 2, performing data aggregation, generalization, normalization, attribute construction and fusion analysis, identifying frequent item sets from the data, then creating rules for describing association relations by using the frequent item sets, searching all frequent item sets with the support degree greater than or equal to a given minimum support degree, searching and generating rule association not less than minimum reliability, associating bird distribution space-time with geographical meteorological environment information, and analyzing and selecting bird damage related influence factors;
step 4, analyzing the association degree of bird damage and influence factors in the limited area;
extracting and constructing characteristics from the data processed in the step 2, and performing data mining on bird damage sequence information data, wherein the bird damage sequence information data comprises a bird distribution time sequence, a distribution space sequence and a distribution related factor sequence; quantitatively analyzing bird damage influence factors by adopting a fractional bit regression method on bird damage sequence information data to obtain an association rule set between bird distribution space-time and bird damage information; training a deep reinforcement learning network by using the association rule set to construct a target deep reinforcement learning network model;
step 5, deducing the global relevance;
collecting bird damage information of other areas except the limited area in the step 2 in the bird damage area, carrying out data cleaning and preprocessing on the bird damage information data of the other areas according to the same method in the step 2, and analyzing the similarity between the bird damage information of the other areas and the bird damage information of the limited area based on a Pearson correlation coefficient method;
then, on the basis of the target deep reinforcement learning network model, obtaining network models of other areas by adopting a transfer learning method, thereby obtaining the bird damage distribution condition predicted by the other areas; finally, integrating the bird damage distribution of the limited area and the bird damage distribution of other areas to obtain overall bird damage distribution information;
step 6, constructing bird pest situation information;
obtaining global bird damage distribution information according to prediction, and constructing a bird damage situation map; in order to quantify the danger degree of the birds, a bird damage hot spot value is adopted to represent the damage of the bird damage, and when the hot spot value is higher, the bird damage risk is higher;
step 7, calculating bird damage hot point values;
calculating bird damage hot point values in each grid, and constructing a bird damage hot point value matrix, wherein the method comprises the following steps:
let D ═ D1，D2，...，DnThe situation pest D is the set of n situation pest birds in the bird pest situation mapiSet as point target, current influence factor index uiThe standard value of each bird pest situation information is
diThe distance from the situation pest to the checked grid s is defined, and h is a distance threshold value; the formula for solving the bird damage hot point value f(s) of the grid s is as follows:
in the formula giRepresenting the density contribution value of the situation pest i to the grid s, wherein h is more than 0, representing the distance threshold value,
for pest birds DiThe weight term of the influence factor index of (2) is initially taken
Traversing all grids, solving the bird damage hot point values of all the grids, and then combining to construct a bird damage hot point value accumulation matrix:
F＝(fab)p×q
wherein p and q are the grid size, fabRepresenting the accumulated value of bird damage hot point values of grids s with coordinates (a, b) in the time length range t before the current bird damage hot point diagram, and f in the initial stateab＝0；
Standardizing the bird damage hot point value accumulation matrix F to finally obtain a bird damage hot point value matrix:
the normalization formula is as follows:
the matrix F of hot spot values of bird damage*Each of
Taking values equal to bird damage hotspot values r(s) within the grid (a, b);
step 8, graphical processing;
constructing a color mapping table to realize mapping from bird damage hot point values to colors so as to form a hot point diagram layer; and converting the bird damage hot spot value matrix into a color image distinguished by different colors according to the bird damage hot spot value matrix and the color mapping table.
2. The method for constructing a limited information based power transmission corridor bird trouble heat spot map as claimed in claim 1, wherein said bird trouble information includes: the system comprises the following components of the existing literature research, on-site wiring research, bird trouble tripping data, bird trouble distribution information, environmental landscape data, GIS data, historical meteorological data, electric power operation and maintenance records, towers, line structure characteristics, real-time detection devices, video images, radar monitoring and bird trouble prevention device information, bird ecology models and bird ecology behavior characteristics.
3. The method for constructing a limited information based bird pest hotspot graph of a power transmission corridor as claimed in claim 1, wherein the data mining algorithm comprises K-Means algorithm, support vector machine, Apriori algorithm, max-expectation algorithm, AdaBoost, K nearest neighbor classification algorithm.
4. The method for constructing the finite information-based power transmission corridor bird trouble heat point diagram according to claim 1, wherein the bird trouble influence factors are quantitatively analyzed by adopting a fractional bit regression method in the step 4, and the regression model is as follows:
ykt＝xktβθ+μθkt
Quantθ(ykt|xkt)＝xktβθ
in the formula, yktFor the k sample data bird cluster value at time t, xktBeta is a factor affecting bird damageθIs the coefficient to be estimated under the residual thetaθktResidual variable, Quant, of the kth sample data at time t under residual thetaθ(ykt|xkt) Denotes a given xktUnder the condition of yktThe conditional quantile of the residual theta is as follows: theta is more than or equal to 0 and less than or equal to 1;
solving the coefficient beta to be estimatedθThe objective function of (2) is the weighted average minimum of the absolute values of the residuals, and the formula is as follows:
when y iskt≥xktβθWhen the residual is positive, giving the weight of the residual theta; when y iskt＜xktBeta, namely when the residual error is negative, giving the weight of 1-theta to the residual error; solving the coefficient beta to be estimatedθNamely, association rules, and then an association rule set between bird distribution space-time and bird damage information is obtained.
5. The method for constructing a finite information-based transmission corridor bird trouble heat spot map as claimed in claim 1, wherein the similarity of step 5 includes: calculating the similarity of the geographic environment and the landscape information; calculating the similarity of weather and meteorological information; calculating the similarity of the bird ecological environment model; and calculating the similarity of the historical operation and maintenance conditions of the power transmission line.
6. The method for constructing a finite information-based transmission corridor bird trouble heat point diagram according to claim 1, wherein the step 6 of constructing a bird trouble map according to the global bird trouble distribution information obtained by prediction comprises the following steps:
U＝[u1，u2，...，um]a set of m influence factor indexes for describing situation pest birds; at the current moment, the sum of bird pest situation information is represented by an original situation information matrix A ═ aij)n×mRepresents:
the matrix A ═ aij)n×mEach row u ini＝[ai1，...，aim]Representing situation pest DiSituation information under the m influence factor indexes, situation pest DiIndex of influence factor ofijThe larger the size of the pest bird DiThe greater the damage of the activity of (2) to the transmission line of the grid in which it is located;
standardizing each row of elements in the matrix A to obtain a standardized matrix
Wherein:
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 CN109710595A (en) | 2019-05-03 |
| CN109710595B true CN109710595B (en) | 2022-03-25 |
Family
ID=66254458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811455816.8A Expired - Fee Related 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) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN115912183B (en) * | 2023-03-09 | 2023-05-26 | 国网湖北省电力有限公司经济技术研究院 | Ecological measure inspection method and system for high-voltage transmission line and readable storage medium |
| CN117271959B (en) * | 2023-11-21 | 2024-02-20 | 中南大学 | Uncertainty evaluation method and equipment for PM2.5 concentration prediction result |
Citations (6)
| 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 |
| 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 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| ES2821735T3 (en) * | 2014-08-21 | 2021-04-27 | Identiflight Int Llc | Bird detection system and procedure |
-
2018
- 2018-11-30 CN CN201811455816.8A patent/CN109710595B/en not_active Expired - Fee Related
Patent Citations (6)
| 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 |
| 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 |
|---|
| Using risk prediction models and species sensitivity maps for large-scale;Juan Manuel Pérez-García et al;《Biological Conservation》;20170509;全文 * |
| 基于灾害类型与线路信息的电网重合闸成功率关联分析;卢恩泽 等;《高压电器》;20170516;第53卷(第5期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109710595A (en) | 2019-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109710595B (en) | Method for constructing bird damage hotspot diagram of power transmission corridor based on limited information | |
| Liu et al. | Predicting aircraft trajectories: A deep generative convolutional recurrent neural networks approach | |
| US12026440B1 (en) | Optimizing aircraft flows at airports using data driven predicted capabilities | |
| CN116625327B (en) | Big data driven marine environment monitoring and early warning method | |
| Buckland et al. | Aerial surveys of seabirds: the advent of digital methods | |
| CN110033051B (en) | Fishing trawler behavior discrimination method based on multi-step clustering | |
| CN109376605B (en) | Electric power inspection image bird-stab-prevention fault detection method | |
| Zhang et al. | Mining photo-sharing websites to study ecological phenomena | |
| CN109117883B (en) | SAR image sea ice classification method and system based on long-time memory network | |
| CN111311081B (en) | Ocean ecological abnormity danger identification method and device based on multi-source heterogeneous data | |
| CN114444378A (en) | Short-term power prediction method for regional wind power cluster | |
| CN114936332A (en) | Method and device for analyzing ship traffic flow in water area of wind power plant and storage medium | |
| CN112215416B (en) | Intelligent planning inspection route system and method | |
| Dutta et al. | Image analysis-based automatic detection of transmission towers using aerial imagery | |
| CN110992307A (en) | Insulator positioning and identifying method and device based on YOLO | |
| CN107038902B (en) | Unmanned aerial vehicle cruising route optimization method based on road network physical structure | |
| CN110287995B (en) | Multi-feature learning network model method for grading all-day overhead traffic jam conditions | |
| CN113536944A (en) | Distribution line inspection data identification and analysis method based on image identification | |
| CN111949840B (en) | Method and device for constructing topological graph structure based on data of Internet of things | |
| Noor et al. | Prediction map of rainfall classification using random forest and inverse distance weighted (IDW) | |
| Socaci et al. | XNow: A deep learning technique for nowcasting based on radar products’ values prediction | |
| Yang et al. | A data-driven method for flight time estimation based on air traffic pattern identification and prediction | |
| Pavlove et al. | Efficient Deep Learning Methods for Automated Visibility Estimation at Airports | |
| CN117458450B (en) | Power data energy consumption prediction analysis method and system | |
| KR102675220B1 (en) | Method and system for providing floating waste source detection service based on satellite data |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220325 |