CN116523299A - Grading evaluation method for lightning damage characteristic parameters of distribution line pole tower - Google Patents
Grading evaluation method for lightning damage characteristic parameters of distribution line pole tower Download PDFInfo
- Publication number
- CN116523299A CN116523299A CN202310368062.7A CN202310368062A CN116523299A CN 116523299 A CN116523299 A CN 116523299A CN 202310368062 A CN202310368062 A CN 202310368062A CN 116523299 A CN116523299 A CN 116523299A
- Authority
- CN
- China
- Prior art keywords
- lightning
- data
- sparrow
- tower
- current amplitude
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000011156 evaluation Methods 0.000 title claims abstract description 23
- 241000287127 Passeridae Species 0.000 claims abstract description 93
- 238000013528 artificial neural network Methods 0.000 claims abstract description 37
- 238000000034 method Methods 0.000 claims abstract description 24
- 238000013210 evaluation model Methods 0.000 claims abstract description 17
- 230000008859 change Effects 0.000 claims description 48
- 230000006870 function Effects 0.000 claims description 48
- 239000013598 vector Substances 0.000 claims description 39
- 208000025274 Lightning injury Diseases 0.000 claims description 23
- 238000012876 topography Methods 0.000 claims description 23
- 238000012549 training Methods 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 14
- 238000012360 testing method Methods 0.000 claims description 13
- 238000010845 search algorithm Methods 0.000 claims description 12
- 230000006698 induction Effects 0.000 claims description 10
- 230000003044 adaptive effect Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000035772 mutation Effects 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 7
- 238000010586 diagram Methods 0.000 claims description 6
- 239000012212 insulator Substances 0.000 claims description 5
- 238000011160 research Methods 0.000 claims description 5
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 3
- 108010063499 Sigma Factor Proteins 0.000 claims description 3
- 230000001174 ascending effect Effects 0.000 claims description 3
- 238000013523 data management Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 230000003094 perturbing effect Effects 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 210000002364 input neuron Anatomy 0.000 claims description 2
- 210000004205 output neuron Anatomy 0.000 claims description 2
- 238000013145 classification model Methods 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 8
- 230000001537 neural effect Effects 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 239000004020 conductor Substances 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 208000003098 Ganglion Cysts Diseases 0.000 description 2
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 208000005400 Synovial Cyst Diseases 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012502 risk assessment Methods 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 102100037091 Exonuclease V Human genes 0.000 description 1
- 101000881977 Homo sapiens Exonuclease V Proteins 0.000 description 1
- 239000004568 cement Substances 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 229910052742 iron Inorganic materials 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- APTZNLHMIGJTEW-UHFFFAOYSA-N pyraflufen-ethyl Chemical compound C1=C(Cl)C(OCC(=O)OCC)=CC(C=2C(=C(OC(F)F)N(C)N=2)Cl)=C1F APTZNLHMIGJTEW-UHFFFAOYSA-N 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- 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
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Artificial Intelligence (AREA)
- Mathematical Analysis (AREA)
- Tourism & Hospitality (AREA)
- Probability & Statistics with Applications (AREA)
- Mathematical Physics (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Pure & Applied Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Optimization (AREA)
- Evolutionary Computation (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Computational Mathematics (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Algebra (AREA)
- Quality & Reliability (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
Abstract
The invention discloses a grading evaluation method for lightning damage characteristic parameters of a distribution line tower, which comprises the steps of firstly collecting longitude and latitude data of a typical tower, lightning current latitude data, lightning current amplitude data and line typical scene factor data; calculating the lightning loss density, and statistically analyzing the lightning current amplitude distribution rule under the condition of typical line scene factors; according to the lightning current amplitude distribution rule and lightning resistance level, providing a lightning current amplitude influence coefficient; and finally, multiplying the lightning current amplitude influence coefficient by the lightning current density to obtain lightning damage intensity indexes, and constructing a risk classification evaluation model to classify the risk of the tower by using the improved sparrow search algorithm-radial basis neural network based on the lightning damage intensity indexes. According to the lightning protection method, lightning loss density and lightning current amplitude distribution rules under different typical line scene factors are analyzed, lightning damage risk classification is carried out by utilizing the improved classification model, the rate and accuracy of risk classification are improved, and technical support is provided for differential lightning protection configuration of the distribution line towers.
Description
Technical Field
The invention relates to the technical field of lightning protection of power distribution network lines, in particular to a grading evaluation method for lightning damage characteristic parameters of a power distribution line pole tower.
Background
In recent years, the lightning protection technology of the power transmission main network is becoming perfect, and as the scale of the power distribution network is enlarged, the frequency of lightning strokes on power distribution network lines and equipment is increased, the influence is increased, and the lightning protection requirement on the power distribution lines is also continuously improved. And the condition of lightning distribution lines is different in different scenes. The distribution of the topography can influence the probability distribution of lightning ground flash and lightning current amplitude, and the distribution line in mountain areas is obviously influenced by the topography. The frequency of lightning strikes to the ground in mountain areas is higher than that in plain areas. The mountain top and the mountain foot flat land with larger span are more prone to lightning stroke faults. Multiple sections of high thunderstorm areas exist in areas with complex topography on the same line. In plain areas, however, lightning strike faults are more likely to occur near water areas, and the soil in water areas such as rivers and lakes is relatively moist, low in resistivity, sufficient in charge quantity and easy to gather, so that the lightning strike tripping rate of lines near the area is increased. If the actual situation of the site is not considered, the same lower lightning protection standard design is adopted for all lines, so that the lightning protection capability of the lines is weak, and lightning trip is easy to cause.
The lightning positioning system is widely applied to a main network, a large amount of lightning activity data is accumulated, meanwhile, the lightning positioning system is also analyzed by combining tripping data of a power transmission line, a large amount of researches on lightning risks under different topography and topography are carried out, and an evaluation standard of the lightning risks is formed. But the distribution line still lacks relevant standard at present, and this is because current thunder and lightning monitoring system positioning accuracy is limited, and the distance is farther between the transmission main network iron tower, does benefit to the division region, and the environmental difference that different towers are located is bigger, and its voltage level is high moreover, and the response thunder can not cause the tripping operation accident, and the lightning current amplitude that the lightning is required of direct lightning initiation thunderbolt trouble is very big, and the fault probability is low. And to distribution lines, response thunder and direct lightning all can cause the tripping operation accident, the shaft tower interval is little, a certain lightning activity probably all leads to the fact the influence to a slice multistage shaft tower, under the limited circumstances of positioning accuracy, the correlation of shaft tower tripping operation and lightning activity is difficult to confirm, moreover because voltage level is low, response thunder just can arouse the tripping operation, and most do not have the lightning conductor, the required lightning current amplitude of fault causes is less, lack standard basis to the lightning stroke fault risk of a certain shaft tower of distribution lines, moreover because distribution lines shaft tower quantity is many, the scene that the shaft tower is located is various, scene factor data is huge, utilize ordinary classification model to carry out risk classification, classification efficiency is not high, the accuracy is not enough.
Aiming at the problems, considering not the relevance of the tower fault and the lightning activity in the lightning positioning system, but the characteristics of the tower model and the environment where the tower is positioned, analyzing the lightning loss density and the lightning current amplitude distribution rule under different typical line scene factors through a large amount of statistical data, and then directly selecting characteristic parameters according to the scene factors where the power distribution tower is positioned, and carrying out lightning damage characteristic grading evaluation by utilizing an improved sparrow search algorithm-radial basis neural network. According to the invention, the distribution condition is considered, the comprehensive analysis is carried out by combining the amplitude value and the influence area of the distribution tower due to lightning trip, the lightning current amplitude value change under the condition of different typical line scene factors is calculated, the characteristic coefficient of the lightning current amplitude value under the condition of different line scene factors is provided, the lightning hazard risk intensity index obtained by the product of the lightning current amplitude value influence coefficient and the lightning density is used as the basis, the improved sparrow search algorithm-radial basis neural network classifier is utilized to carry out risk classification on the line tower, the classification model after improvement improves the classification efficiency and accuracy, and the basis is provided for the lightning protection measure configuration of the distribution tower
Disclosure of Invention
According to the technical problems, the invention provides a grading evaluation method for lightning damage characteristic parameters of a distribution line pole tower, according to lightning activity statistical data in different typical distribution line scenes, lightning drop densities in different typical scene factors are analyzed and calculated, lightning trip calculation is carried out on the typical distribution line to obtain line lightning resistance levels, lightning current amplitude changes in different typical line scene factors are calculated, lightning current amplitude influence coefficients are provided, lightning damage risk intensity indexes are obtained by utilizing products of the lightning current amplitude influence coefficients and the lightning drop densities, the characteristic indexes are used as basis, and the improved sparrow search algorithm-radial basis neural network classifier is utilized to classify the lightning risk of the pole tower, so that technical support is provided for differential lightning protection configuration of the line pole tower.
The technical scheme adopted by the invention is as follows: a grading evaluation method for lightning damage characteristic parameters of a distribution line pole tower comprises the following steps:
step 1: collecting and extracting longitude and latitude data of a tower in a typical region, scene factor data of a typical distribution line, lightning current latitude data and lightning current amplitude data, and calculating lightning loss density according to the data;
step 2: the lightning current amplitude distribution rule under different scene factors is statistically analyzed, the direct lightning overvoltage and the induced lightning overvoltage of the tower are calculated, and the lightning resistance level of the line is obtained, wherein the lightning resistance level of the line comprises a direct lightning Lei Nailei level and an induced Lei Nailei level;
step 3: determining lightning current amplitude influence coefficients according to the lightning current amplitude distribution rules and the line lightning resistance level under the conditions of different scene factors obtained in the step 2, and multiplying the determined lightning current amplitude influence coefficients by lightning loss density to obtain lightning damage risk intensity indexes;
step 4: grading lightning risk according to lightning risk intensity indexes, and forming a data set with corresponding typical distribution line scene factor data;
step 5: improving a sparrow search algorithm, optimizing a radial basis function neural network by using the sparrow search algorithm, and constructing an improved sparrow search algorithm-radial basis function neural network classifier;
Step 6: training and testing the data set obtained in the step 4 by utilizing an improved sparrow searching algorithm-radial basis function neural network to obtain a risk grading evaluation model; and grading lightning damage risks of the towers by using the risk grading evaluation model after the testing is completed.
Further preferably, the step 1 includes the steps of:
step 1.1, dividing a research area into grids, and calculating typical distribution line scene factor data in the grids:
step 1.2, calculating the thunder density: converting the statistical lightning current latitude and longitude data into an Excel data format file with the extension of xls, defining X-axis and Y-axis data by Arcgis, selecting map projection, importing the selected lightning current latitude and longitude data, dividing the data into grids consistent with the step 1.1, counting the grid area and the number of lightning drops in the grids, and calculating the lightning drop density by using a field calculator through a density function.
Further preferably, the process of step 1.1 is:
1) Gradient vector: extracting a gradient layer from the DEM elevation data by using a gradient tool, and extracting a gradient from the gradient layer for the second time by using the gradient tool again to obtain gradient change rate data;
2) Slope vector: obtaining the maximum elevation value of DEM elevation data, obtaining a DEM data layer opposite to the original topography by using a grid calculator tool, obtaining slope data I by using a slope tool to extract a slope from the DEM data layer opposite to the original topography, obtaining slope change rate data II of the opposite topography by using a slope tool to extract a slope from the slope data I, obtaining slope data II by using the slope tool again, obtaining slope change rate data I of the original DEM data by extracting a slope from the original DEM data, obtaining a slope change rate data I of the original DEM data by using a slope tool to extract a slope from the slope data II, adding the slope change rate data I of the original DEM data and the slope change rate data II of the opposite topography by using a grid calculator again, and subtracting one half of the absolute value of the difference between the slope change rate I of the original DEM data and the slope change rate data II of the opposite topography to obtain an error-free DEM slope change rate;
3) Terrain relief vector: using a focus statistics tool, respectively selecting a maximum elevation value of the elevation data of the DEM and a minimum elevation value of the elevation data of the DEM according to the statistics type, obtaining a data layer A and a data layer B respectively without changing other parameters, and subtracting the data layer B from the data layer A to obtain the extremely poor in the same neighborhood range in the image layer;
4) Surface coverage type: splicing the images after loading the original data, embedding the images into the new grids by using tools in a tool box according to the sequence of the data management tool, the grids and the grid data set, extracting and analyzing by using a space analysis tool, and finally extracting according to a mask to obtain the tower surface coverage classification data.
Further preferably, the step 2 includes the steps of:
step 2.1, converting the lightning current latitude data into an Excel data format file with the extension name of xls, defining X-axis and Y-axis data by using Arcgis, selecting map projection, importing the selected lightning current latitude data, equally determining the size of each grid layer, dividing the grid layer into grids, counting the number of lightning drops in the grids, and obtaining a lightning drop density grid vector diagram through density function calculation processing by using a field calculator; extracting the magnitude of lightning current of each lightning point by adopting a multi-value extraction to point tool;
Step 2.2, analyzing lightning current amplitude rules under different distribution line scene factor conditions: the lightning current on the typical scene factor condition is selected for analysis, the corresponding lightning current amplitude value is extracted, and parameter fitting is carried out on the lightning current amplitude value;
step 2.3, carrying out lightning-proof level calculation on a typical overhead distribution line: the direct impact Lei Nailei level comprises a lightning stroke wire lightning resistance level and a lightning stroke tower lightning resistance level, and is calculated by using a lightning stroke wire lightning resistance level formula (3) and a lightning stroke tower lightning resistance level formula (4):
wherein I is G The lightning-proof level of the lightning strike wire is; i D The lightning resistance level of the lightning stroke tower is; u (U) 50％ 50% flashover voltage for the insulator; r is R ch The grounding resistor of the tower; l (L) gt Equivalent inductance of the tower; h is a d Is the average height of the line;
the induction Lei Nai is calculated by using the formula (5) from the magnitude of the current and the vertical distance between the ground flash point and the protected line to obtain the induction Lei Nailei level:
wherein I is g Sensing Lei Nailei level for distribution lines; s is the distance from the lightning strike point to the nearest point of the wire.
Further preferably, step 3 comprises the following process:
step 3.1: calculating the tripping condition of a typical distribution line to obtain the level of direct impact Lei Nailei, directly tripping the line, and adjusting the risk grade of a tower with a statistical result of direct impact on the basis of the evaluation of the model by one level; analyzing and comparing the induction Lei Nailei level calculated in the step 2 with the median number of the lightning current amplitude value fitted by the lightning current amplitude value and the change rate of the lightning current amplitude value, so as to determine the lightning current amplitude value coefficient;
Step 3.2: multiplying the lightning current amplitude influence coefficient obtained in the step 3.1 with the lightning intensity to obtain lightning hazard risk intensity indexes, and classifying the risks of a typical line scene according to the lightning hazard risk intensity indexes.
Further preferably, step 4 comprises the following process: :
step 4.1: in Arcgis, generating a shapefile by using the risk level obtained in the step 3.2 and the corresponding scene factor in a corresponding way, defining the element type as a line type, opening an advanced editor, and performing surface construction on the line by using a construction surface tool;
step 4.2: in matlab, reading the surface data in Arcgis into matlab by shape, judging data attribution by using an index command in matlab, and correspondingly generating a data set Q.
Further preferably, the specific process of step 5 is as follows:
step 5.1, improving a sparrow search algorithm: (1) reverse learning strategy initializing population: first randomly initializing the positions x of N sparrow individuals in a search space i Wherein i is the sparrow individual number as the initial population RP; then constructing a reverse population OP, wherein the OP is formed by each sparrow individual x in the initial population RP i Is the inverse of individual x i ' composition; finally merging the population RP and OP, arranging the adaptive values of 2N sparrow individuals according to ascending order, and selecting the first N sparrow individuals with the adaptive values as an initial population; (2) Sine and cosine algorithm idea improvement The present position: adding a nonlinear sine learning factor, wherein the nonlinear sine learning factor and the improved finder position formula are as follows:
w＝w min +(w max -w min )·sin(tπ/iter max ) (7)
X(t+1)＝X(t)+r 1 ×sinr 2 |r 3 ×P t -X(t)|，r 4 <0.5 (8)
X(t+1)＝X(t)+r 1 ×cosr 2 |r 3 ×P t -X(t)| r 4 ≥0.5 (9)
wherein w is a nonlinear sine learning factor, w min Is nonlinear sine learning factor minimum value, w max Is the maximum value of nonlinear sine learning factors, item is a degree of freedom parameter, t is the current iteration number, and P t R is the current optimal solution position 2 、r 3 、r 4 Is a random parameter, X is a randomly generated solution, t represents the iteration number, and the parameter r 1 Is responsible for balancing the whole algorithm detection and exploitation process, and has the expression r 1 =h-h (u/v), h being a constant, u, v being the current iteration number and the maximum iteration number, respectively; (3) initializing a population by a reverse learning strategy: the levy flight strategy improves follower position: the follower position is updated according to the following formula:
wherein,,for the previous finder's occupied position, < >>For the location occupied by the current finder, +.>For the worst position occupied by the previous finder, n is the maximum of the expandable spatial dimension, +.>Is the best place for the current finder to occupy, and the Levy flight mechanism is as follows:
wherein epsilon is a flight factor, sigma is a flight smoothness parameter, epsilon takes 1.5, and sigma is calculated as follows:
(4) Adaptive t distribution variation strategy: updating formulas (8), (9) and (10) for target positions, improving global exploration performance by using disturbance capability of t distribution mutation, and updating sparrow positions by adopting a self-adaptive t distribution mutation strategy, wherein the method specifically comprises the following steps:
in the formula (I)Indicating the position of sparrow after mutation, x i For the position of sparrow individual, t (iter) represents the t distribution regarding the iteration number as the degree of freedom of the parameter;
step 5.2, constructing an improved sparrow search algorithm-radial basis function neural network classifier: initializing parameters for improving sparrow search algorithm, and carrying out central vector C on jth node of hidden layer of radial basis function neural network algorithm j Width vector D of the j-th node of the hidden layer j Coding, generating an initialized sparrow population, receiving information transmitted by the sparrow population by a radial basis function neural network algorithm, and decoding the sparrow population to obtain a corresponding optimized center vector C' j And width vector D' j Calculating the fitness value of each sparrow, finding out the current optimal value, the worst value and the corresponding position, selecting part from sparrows with better fitness value as a finder, updating the positions according to the formulas (8) and (9), taking the rest sparrows as followers, and introducingUpdating the position according to the formula (10) by using a Levy flight strategy; and (3) carrying out self-adaptive t distribution variation according to a formula (13), perturbing the current position, generating a new solution, calculating an fitness value and updating the individual position of the sparrow, and constructing an improved sparrow search algorithm-radial basis function neural network classifier.
Further preferably, the specific process of the step 6 is as follows:
step 6.1: according to the data set Q obtained in the step 5.2, establishing the data set Q as a sequence sample, and obtaining a risk classification evaluation model by the improved sparrow search algorithm-radial basis function network classifier through collecting the sequence sample, preprocessing data, determining input and output neurons, constructing the improved sparrow search algorithm-radial basis function network, training the improved sparrow search algorithm-radial basis function network and testing the improved sparrow search algorithm-radial basis function network;
step 6.2: carrying out lightning hazard risk classification on the towers: and selecting an actual power distribution network line of the lightning accident, and classifying and identifying scenes of each base tower by using a risk classification evaluation model.
The invention has the advantages that: the distribution lines towers are large in data, corresponding scene factor data are huge, and if manual work is utilized to classify the risks of the towers, the workload is huge. And the interval of the distribution towers is small, a certain lightning activity can affect a piece of multi-stage towers, and when the existing risk assessment method adopting the combination of fault statistical analysis and lightning activity is applied to distribution, the risk assessment method can only assess high risk to the line with faults, so that the risk towers in the same environment do not obtain due protection configuration. According to the lightning protection device, characteristics of a tower model and an environment are considered, lightning loss density and lightning current amplitude distribution rules under different typical line scene factors are analyzed through a large amount of statistical data, lightning risk intensity indexes are provided, the lightning risk of the type of line under different line scene factors is determined by using an improved sparrow search algorithm-radial basis neural network classifier based on the characteristic indexes, the evaluation efficiency and the evaluation accuracy are improved, the evaluation result is wider in applicability, lightning risk evaluation can be performed on the same configuration line under similar working condition environments, and the installation configuration of the lightning protection device is guided.
At present, for the influence of environmental factors on lightning activities, most researches only analyze the correlation of lightning density and topography, but the invention considers not only the lightning density under different typical line scenes, but also the distribution rule of lightning current amplitude under different typical line scenes, and meanwhile considers the lightning current amplitude as probability characteristic.
Drawings
FIG. 1 is a graph of lightning current magnitude fitting under mountain conditions for a line scenario.
Fig. 2 is a schematic diagram of training convergence iteration number of the improved sparrow search algorithm-radial basis function neural network algorithm.
FIG. 3 is a graph of training set accuracy for improved sparrow search algorithm-radial basis function neural network algorithm training.
FIG. 4 is a graph of test set accuracy for improved sparrow search algorithm-radial basis function neural network algorithm training.
Fig. 5 shows a distribution diagram of a section of line corridor and lightning trip in the Nanchang city.
Fig. 6 is a line tower risk classification prediction graph.
Detailed Description
The invention is described in further detail below with reference to the attached drawings in conjunction with specific embodiments:
a grading method for lightning damage characteristic parameters of a distribution line pole tower area comprises the following steps:
step 1: collecting and extracting longitude and latitude data of a tower in a typical region, scene factor data of a typical distribution line, lightning current latitude data and lightning current amplitude data, and calculating lightning loss density according to the data;
step 2: the lightning current amplitude distribution rule under different scene factors is statistically analyzed, the direct lightning overvoltage and the induced lightning overvoltage of the tower are calculated, and the lightning resistance level of the line is obtained, wherein the lightning resistance level of the line comprises a direct lightning Lei Nailei level and an induced Lei Nailei level;
step 3: determining lightning current amplitude influence coefficients according to the lightning current amplitude distribution rules and the line lightning resistance level under the conditions of different scene factors obtained in the step 2, and multiplying the determined lightning current amplitude influence coefficients by lightning loss density to obtain lightning damage risk intensity indexes;
step 4: grading lightning risk according to lightning risk intensity indexes, and forming a data set with corresponding typical distribution line scene factor data;
Step 5: improving a sparrow search algorithm, optimizing a radial basis function neural network by using the sparrow search algorithm, and constructing an improved sparrow search algorithm-radial basis function neural network classifier;
step 6: training and testing the data set obtained in the step 4 by utilizing an improved sparrow searching algorithm-radial basis function neural network to obtain a risk grading evaluation model; and grading lightning damage risks of the towers by using the risk grading evaluation model after the testing is completed.
Step 1 of this embodiment includes the following steps:
step 1.1, lightning current and latitude data and lightning current amplitude data are obtained by calculation from 2017-2021 Nanchang city lightning data, typical distribution line scene factor data are obtained from Nanchang city administrative division data and Nanchang city elevation data, wherein soil coverage rate image data extracted through nationwide remote sensing image data are obtained, and meanwhile, in order to determine the position of a typical line tower, longitude and latitude data of a typical region tower are obtained.
Dividing a research area into 184364 grids by Arcgis, and calculating typical distribution line scene factor data in the grids:
1) Gradient vector: and extracting slope of the slope layer by using a slope tool for DEM elevation data in Nanchang city, and extracting the slope of the slope layer by using the slope tool again to obtain slope change rate data SOS.
2) Slope vector: obtaining the maximum elevation value of DEM elevation data of Nanchang city, subtracting the DEM data layer (input formula is FDEM=MAX-DEM) from the maximum elevation value of the DEM elevation data of Nanchang city by using a grid calculator tool to obtain DEM data layer ' FDEM ' opposite to the original topography, extracting slope direction from the DEM data layer ' FDEM ' opposite to the original topography by using a slope direction tool to obtain slope direction data I ' FDEM1 ', extracting slope from the slope direction data I ' FDEM1 ' by using a slope direction tool to obtain slope direction change rate data II ' SOA2 ' of the opposite topography, extracting slope direction from the original DEM data by using a slope direction tool again to obtain slope direction data II ' DEM1 ', extracting slope direction from the slope direction data II ' DE M1 ' by using a slope direction tool to obtain slope direction change rate data I ' SOA1 ' of the original DEM data, adding the slope direction change rate data I of the original DEM data and the slope direction change rate II ' of the opposite topography by using a grid calculator again, subtracting the slope direction change rate I of the original DEM data and obtaining a slope direction change rate I of the opposite topography as a half of the absolute value, and obtaining a slope change rate of the slope error of the absolute value: soa= ((soa1+soa2) -abs (soa1-soa2))/2 ].
3) Terrain relief vector: and (3) obtaining the maximum elevation value of the DEM elevation data and the minimum elevation value of the DEM elevation data by using a focus statistics tool (the statistics type is respectively selected from MAX and MIN), obtaining a data layer A and a data layer B respectively without changing other parameters, and obtaining the range in the same neighborhood range in the image layer by subtracting the data layer B from the data layer A (using a grid calculator input formula: A-B).
4) Surface coverage type: splicing the images after loading the original data, embedding the images into the new grids by using tools in a tool box according to the sequence of the data management tool, the grids and the grid data set, extracting and analyzing by using a space analysis tool, and finally extracting according to a mask to obtain the tower surface coverage classification data.
Step 1.2, calculating the thunder density: converting 5-year lightning current latitude and longitude data counted in the step 1 into an Excel data format file, expanding the Excel data format file to be named as xls, defining X-axis and Y-axis data by using Arcgis, selecting proper map projection, automatically importing the selected lightning current latitude and longitude data, dividing the selected lightning current latitude and longitude data into 184364 grids unified with the step 2.1, counting the grid area s and the number f of lightning drops in the grids, and obtaining the lightning drop density through calculation processing of a density function (1) by using a field calculator.
Wherein N is g For the mine-down density, N is the number of grids.
In this embodiment, step 2.1, using the lightning current latitude data of the south-Chang city in 2017-2021 counted in step 1, converting the lightning current latitude data into Excel data format file, with the extension name of. Xls, using Arcgis to define X-axis and Y-axis data, selecting appropriate map projection, automatically importing the selected lightning current latitude data, determining the size of each grid layer as 200m×200m, dividing the grid layer into 184364 grids, and then counting the grid area as 0.04km 2 The number of the thunder in the grid is calculated and processed by a field calculator through a density function to obtain a thunder density grid vector diagram; and extracting the magnitude of the lightning current of each lightning point by adopting a multi-value extraction to point tool.
Step 2.2, analyzing lightning current amplitude rules under different distribution line scene factor conditions: the lightning current under 5 typical scene factors of mountain land, plain, hard land, farmland and river in Nanchang city is selected for analysis, the corresponding lightning current amplitude value is extracted, and parameter fitting is carried out on the lightning current amplitude value according to the formula (2):
wherein P is a lightning current amplitude rule, I is a lightning current amplitude, a is a median number of the lightning current amplitude, and b is a rate of change of the lightning current amplitude.
Fitting to obtain the median value of the positive polarity lightning current amplitude value of 30.2483, the change rate of the positive polarity lightning current amplitude value of 2.682, the median value of the negative polarity lightning current amplitude value of 39.7793, the change rate of the negative polarity lightning current amplitude value of 3.250, the median value of the overall lightning current amplitude value of 37.885 and the change rate of the overall lightning current amplitude value of 3.073 under the characteristic condition of mountain land, wherein the fitting result is shown in figure 1. Analysis was performed taking the mountain negative polarity current cumulative probability as an example, with a median lightning current magnitude of 37.779kA, i.e., 50% of the probability that the lightning current magnitude was higher than 37.779 kA. Fitting is carried out in the same way to obtain the median value of the positive polarity lightning current amplitude value of 26.4202, the change rate of the positive polarity lightning current amplitude value of 2.581, the median value of the negative polarity lightning current amplitude value of 37.489, the change rate of the negative polarity lightning current amplitude value of 2.393, the median value of the overall lightning current amplitude value of 34.0333 and the change rate of the overall lightning current amplitude value of 2.464 under farmland characteristic conditions; under the characteristic conditions of plain, the median number of the positive polarity lightning current amplitude is 22.78, the rate of change of the positive polarity lightning current amplitude is 2.231, the median number of the negative polarity lightning current amplitude is 42.110, the rate of change of the negative polarity lightning current amplitude is 3.135, the median number of the overall lightning current amplitude is 33.563, and the rate of change of the overall lightning current amplitude is 2.674; under the hard ground characteristic condition, the median value of the positive polarity lightning current amplitude is 16.391, the rate of change of the positive polarity lightning current amplitude is 2.431, the median value of the negative polarity lightning current amplitude is 27.845, the rate of change of the negative polarity lightning current amplitude is 2.437, the median value of the overall lightning current amplitude is 24.589, and the rate of change of the overall lightning current amplitude is 2.311; under river characteristic conditions, the median value of the positive polarity lightning current amplitude is 20.041, the rate of change of the positive polarity lightning current amplitude is 2.800, the median value of the negative polarity lightning current amplitude is 27.198, the rate of change of the negative polarity lightning current amplitude is 2.659, the median value of the overall lightning current amplitude is 25.453, and the rate of change of the overall lightning current amplitude is 2.620.
Step 2.3, carrying out lightning-proof level calculation on a typical overhead distribution line: under the condition of no lightning conductor assumption, the direct lightning strike has two modes, namely, the direct lightning strike is conducted on the lead, and the impact on the tower is conducted on the lead to form an impact. The direct impact Lei Nailei level comprises a lightning stroke wire lightning resistance level and a lightning stroke tower lightning resistance level, and is calculated by using a lightning stroke wire lightning resistance level formula (3) and a lightning stroke tower lightning resistance level formula (4).
Wherein I is G The lightning resistance level of the lightning strike wire is kA; i D Is the lightning resistance level of the lightning stroke tower and kA; u (U) 50％ 50% flashover voltage of the insulator is kV; r is R ch Is the grounding resistance of the tower, omega; l (L) gt The equivalent inductance of the tower is H; h is a d The average height of the line, m.
The induction Lei Nai is calculated as the current level and the vertical distance between the ground flash point and the protected line using equation (5) to obtain the induction Lei Nailei level.
Wherein I is g Sensing Lei Nailei level, kA for distribution lines; s is the distance from the lightning strike point to the nearest point of the wire, and m.
A typical single-circuit non-lightning-protection line frame empty distribution line in Nanchang areas adopts a non-pull-line single-pole cement tower, the pole tower wave impedance is 250 omega, the inductance is 0.84uH/m, and the pole tower impact grounding resistance is 10 omega. The line insulator adopts a P-15 pin insulator, U 50％ The main conductor is LGJ-120 with a radius of 1.52cm, a span of 50m, a line sag of 0.233m and a line wave impedance of 400 omega.
The lightning-proof level of the top of the direct impact Lei Zhiji tower and the lightning-proof level of the lead are respectively 4.5 and 2kA, and the induction Lei Nailei level is respectively 71kA, 56kA, 43kA and 32kA when the position of the lightning stroke point is respectively 100m, 75m, 50m and 30m from the tower.
And 2.2, the fitted median is larger than the line direct impact Lei Nailei level, the line can be judged to be directly tripped, and the risk grade of the tower with the statistical result of direct impact is adjusted by one level on the basis of the model evaluation. Analyzing and comparing the calculated induction Lei Nailei level with the median value of the lightning current amplitude fitted by the lightning current amplitude and the speed of the change of the lightning current amplitude, and when the median value of the lightning current amplitude fitted is (34, 35), the unit is kA; the rate of change of the magnitude of the lightning current is at (2.6,2.8), which is a coefficient of change, dimensionless. The lightning current amplitude influence coefficient under the condition is 1.1, and the specific calculation result is shown in table 1. The lightning current amplitude influence coefficients under the conditions of 5 typical line scene factors of mountain land, plain land, hard land, farmland and river are 1.1, 1.0, 1.2, 1.3 and 1.8.
TABLE 1
Step 3.2: multiplying the lightning current amplitude influence coefficient obtained in the step 3.1 with the lightning intensity to obtain a lightning hazard risk intensity index L, classifying the typical line scene according to the lightning hazard risk intensity index L, namely classifying the distribution line scene factor data corresponding to the interval into class 2 risks according to the lightning hazard risk intensity index in the interval (0.936,3.336), and classifying the lightning hazard risks of class 5 typical line tower scenes into class 5 classes to form a data set of the typical distribution line scene factor data and the lightning hazard risk classes, wherein the lightning hazard risk classification table is shown in table 2.
TABLE 2
| Risk level | Lightning hazard risk intensity index |
| Level 1 | L g <0.936 |
| Level 2 | 0.936≤L g <3.336 |
| 3 grade | 3.336≤L g <9.576 |
| Grade 4 | 9.576≤L g <18.6 |
| Grade 5 | 18.6≤L g |
The specific process of step 4 in this embodiment is:
step 4.1: creating a shape in Arcgis by using the data set obtained in the step 3.2, defining the element type as a line type, opening an advanced editor, and performing surface construction on the line by using a construction surface tool;
step 4.2: in matlab, reading the surface data in Arcgis into matlab by shape, judging data attribution by using an index command in matlab, and correspondingly generating a data matrix Q.
Q＝[x 1 ,x 2 ,x 3 ,x 4 ,x 5 ，y] (6)
Wherein x is 1 ,x 2 ,x 3 ,x 4 ,x 5 The method is characterized in that 5 scene factor vectors, namely an altitude Cheng Shiliang, a gradient vector, a slope vector, a topography relief vector and a surface coverage type vector, are respectively adopted, y is a corresponding classification level, and the values are 1,2, 3, 4 and 5.
The specific process of step 5 in this embodiment is:
step 5.1, improving a sparrow search algorithm: (1) reverse learning strategy initializing population: first randomly initializing the positions x of N sparrow individuals in a search space i (i=1, 2, … …, N), where i is the sparrow individual number, which is also the inverse dot-expandable spatial dimension, as the initial population RP; then constructing a reverse population OP, wherein the OP is formed by each sparrow individual x in the initial population RP i Is the inverse of individual x i ' composition; finally, mergingAnd (3) the population RP and OP are obtained by arranging the adaptive values of 2N sparrow individuals in an ascending order, and the N sparrow individuals with the front adaptive value are selected as the initial population. (2) sine and cosine algorithm ideas improve finder position: adding a nonlinear sine learning factor, wherein the nonlinear sine learning factor and the improved finder position formula are as follows:
w＝w min +(w max -w min )·sin(tπ/iter max ) (7)
X(t+1)＝X(t)+r 1 ×sinr 2 |r 3 ×P t -X(t)|，r 4 <0.5 (8)
X(t+1)＝X(t)+r 1 ×cosr 2 |r 3 ×P t -X(t)| r 4 ≥0.5 (9)
where w is a nonlinear sine learning factor, w indicates the area (or direction of movement) of the next position, w min Is nonlinear sine learning factor minimum value, w max Is the maximum value of nonlinear sine learning factors, item is a degree of freedom parameter, t is the current iteration number, and P t R is the current optimal solution position 2 r 3 ，r 4 Is a random parameter, r 2 ～U(0,2π)，r 3 ～U(0,2)，r 4 U (0, 1), X is a randomly generated solution, t represents the number of iterations, parameter r 1 Is responsible for balancing the whole algorithm detection and exploitation process, and has the expression r 1 =h-h (u/v), h is a constant, u, v are the current iteration number and the maximum iteration number, respectively. (3) initializing a population by a reverse learning strategy: the levy flight strategy improves follower position: the follower position is updated according to the following formula:
wherein,,for the previous finder's occupied position, < >>Is occupied by the current finder Position (S)>For the worst position occupied by the previous finder, n is the maximum of the expandable spatial dimension, +.>Is the best place for the current finder to occupy, and the Levy flight mechanism is as follows:
wherein epsilon is a flight factor, sigma is a flight smoothness parameter, epsilon takes 1.5, and sigma is calculated as follows:
(4) Adaptive t distribution variation strategy: updating formulas (8), (9) and (10) for target positions, improving global exploration performance by using disturbance capability of t distribution mutation, and updating sparrow positions by adopting a self-adaptive t distribution mutation strategy, wherein the method specifically comprises the following steps:
in the formula (I)Indicating the position of sparrow after mutation, x i For the location of the sparrow individual, t (iter) represents the t distribution regarding the number of iterations as a degree of freedom of the parameters.
Step 5.2, constructing an improved sparrow search algorithm-radial basis function neural network classifier: initializing parameters for improving sparrow search algorithm, and carrying out central vector C on jth node of hidden layer of radial basis function neural network algorithm j Width vector D of the j-th node of the hidden layer j Coding, generating an initialized sparrow population, receiving information transmitted by the sparrow population by a radial basis function neural network algorithm, and decoding the sparrow population to obtain a corresponding sparrow populationIs the optimized center vector C' j And width vector D' j Calculating the fitness value (error generated by a radial basis function neural network model) of each sparrow, finding out the current optimal value, the worst value and the corresponding position, selecting a part from the sparrows with the better fitness value as a finder, updating the position according to the formulas (8) and (9), taking the rest sparrows as followers, introducing a Levy flight strategy, and updating the position according to the formula (10); and (3) carrying out self-adaptive t distribution variation according to a formula (13), perturbing the current position, generating a new solution, calculating an fitness value and updating the individual position of the sparrow, and constructing an improved sparrow search algorithm-radial basis function neural network classifier.
The specific process of step 6 in this embodiment is:
step 6.1: according to the data matrix Q obtained in the step 5.2, the data matrix Q is established as a sequence sample, and preprocessing is carried out on data to determine an input vector X:
X＝[x 1 ,x 2 ,x 3 ,x 4 ,x 5 ] T (14)
wherein x is 1 ,x 2 ,x 3 ,x 4 ,x 5 The scene factors are 5 types respectively, and the node number of the input layer is 5;
determining input-output neurons: determining an output vector Y and a desired output vector O:
Y＝[y 1 ,y 2 ,y 3 ,y 4 ,y 5 ] T (15)
O＝[o 1 ,o 2 ,o 3 ,o 4 ,o 5 ] T (16)
wherein y is 1 ,y 2 ,y 3 ,y 4 ,y 5 Respectively 5 outputs o 1 ,o 2 ,o 3 ,o 4 ,o 5 Respectively 5 expected outputs, T representing the transpose;
construction of an improved sparrow search algorithm-radial basis function neural network: dividing the risk of the tower into 5 types, initializing the connection weight value from the hidden layer to the output layer by using the node number of the input layer as 5, and calculating and improving central parameters of each neuron of the hidden layer by using a sparrow search algorithm-radial basis function neural network:
C’ j ＝[c’ j1 ,c’ j2 ,c’ j3 ,c’ j4 ,c’ j5 ] T (17)
D’ j ＝[d’ j1 ,d’ j2 ,d’ j3 ,d’ j4 ,d’ j5 ] T (18)
Wherein C' j Is the center vector, D 'of the j-th node of the improved hidden layer' j For the width vector of the j-th node of the improved hidden layer, j=1, 2,3 … 195, 195 total nodes, c' j1 ,c’ j2 ,c’ j3 ,c’ j4 ,c’ j5 1 st, 2 nd, 3 rd, 4 th, 5 th central parameters, d 'of the j-th node of the hidden layer' j1 ,d’ j2 ,d’ j3 ,d’ j4 ,d’ j5 The 1 st, 2 nd, 3 rd, 4 th and 5 th width parameters of the j-th node of the hidden layer are respectively. Then, the output value of each neuron of the hidden layer is calculated, and the output of the neurons of the output layer is calculated. Using a radial basis function neural network evaluation function E:
wherein j is the number of input layer ganglion points, k is the number of output layer ganglion points, y jk Output of layer neural node k path for input layer neural node j, o jk The desired output of the layer neural node k path is output for the input layer neural node j.
Training improved sparrow search algorithm-radial basis function neural network: and carrying out iterative operation self-adaptive adjustment on the weight, the center and the width parameters to an optimal value according to an improved sparrow searching algorithm to obtain a risk classification evaluation model. And training the established improved sparrow search algorithm-radial basis function neural network by taking 160 typical distribution line pole tower scene factor data as training samples. After 10 times of loop iteration, the training set converges and the training is terminated, and the improved sparrow search algorithm-radial basis function neural network parameter is obtained. A schematic of the number of iterations of training convergence is shown in fig. 2.
Checking the established radial basis function neural network risk classification evaluation model: and (3) selecting 35 sample data in a data set of typical distribution line scene factor data and lightning hazard risk grades as unknown samples to be tested, and checking the established risk classification evaluation model to obtain the accuracy of a training set of the risk classification evaluation model of 95.625%, the accuracy of a testing set of 94.2857%, and the accuracy of the testing set of more than 90% of the accuracy threshold value of a classification algorithm, wherein the training of the risk classification evaluation model is completed. The training set and the accuracy result of the test set after the test are shown in fig. 3 and fig. 4.
Step 6.2: carrying out risk classification on the towers: an actual distribution network line with more lightning accidents is selected, the total length of the line is about 6.991km, 61 basic towers are added, and a line corridor diagram is shown in fig. 5. And classifying and identifying the scene of each base tower by using a risk classification evaluation model to obtain 10 base towers of lines #7- #10, #53- #55, #59- #61 and the like as high lightning stroke risk sections, wherein the classification and prediction results are shown in fig. 6. In recent years, the lightning stroke tower sections of the line are in the identified high lightning stroke risk sections, and the identification result is consistent with the historical lightning occurrence condition of the line.
According to the lightning protection method, lightning damage intensities under different distribution line scene factors are established through statistical data, so that lightning stroke fault risks of a certain tower of the distribution line are estimated, the lightning strike density is considered, the distribution rule of lightning current amplitude under different distribution line scene factors is considered, the lightning resistance level of a specific distribution line is combined, and meanwhile, the improved sparrow search algorithm-radial basis function neural network is utilized, so that the grading speed is improved, and the accuracy of estimating grading is ensured.
Claims (7)
1. The grading evaluation method for the lightning damage characteristic parameters of the distribution line pole tower is characterized by comprising the following steps of:
step 1: collecting and extracting longitude and latitude data of a tower in a typical region, scene factor data of a typical distribution line, lightning current latitude data and lightning current amplitude data, and calculating lightning loss density according to the data;
step 2: the lightning current amplitude distribution rule under different scene factors is statistically analyzed, the direct lightning overvoltage and the induced lightning overvoltage of the tower are calculated, and the lightning resistance level of the line is obtained, wherein the lightning resistance level of the line comprises a direct lightning Lei Nailei level and an induced Lei Nailei level;
step 3: determining lightning current amplitude influence coefficients according to the lightning current amplitude distribution rules and the line lightning resistance level under the conditions of different scene factors obtained in the step 2, and multiplying the determined lightning current amplitude influence coefficients by lightning loss density to obtain lightning damage risk intensity indexes;
Step 4: grading lightning risk according to lightning risk intensity indexes, and forming a data set with corresponding typical distribution line scene factor data;
step 5: improving a sparrow search algorithm, optimizing a radial basis function neural network by using the sparrow search algorithm, and constructing an improved sparrow search algorithm-radial basis function neural network classifier;
step 6: training and testing the data set obtained in the step 4 by utilizing an improved sparrow searching algorithm-radial basis function neural network to obtain a risk grading evaluation model; and grading lightning damage risks of the towers by using the risk grading evaluation model after the testing is completed.
2. The grading evaluation method for lightning protection feature parameters of a distribution line tower according to claim 1, wherein the step 1 comprises the following steps:
1.1, dividing a research area into grids by utilizing Arcgis, and calculating typical distribution line scene factor vector data in the grids, wherein the typical distribution line scene factor vector data comprises gradient vectors, slope vectors, topography relief vectors and earth surface coverage types;
gradient vector: extracting a gradient layer from the DEM elevation data by using a gradient tool, and extracting a gradient from the gradient layer for the second time by using the gradient tool again to obtain gradient change rate data;
Slope vector: obtaining the maximum elevation value of DEM elevation data, obtaining a DEM data layer opposite to the original topography by using a grid calculator tool, obtaining slope data I by using a slope tool to extract a slope from the DEM data layer opposite to the original topography, obtaining slope change rate data II of the opposite topography by using a slope tool to extract a slope from the slope data I, obtaining slope data II by using the slope tool again, obtaining slope change rate data I of the original DEM data by extracting a slope from the original DEM data, obtaining a slope change rate data I of the original DEM data by using a slope tool to extract a slope from the slope data II, adding the slope change rate data I of the original DEM data and the slope change rate data II of the opposite topography by using a grid calculator again, and subtracting one half of the absolute value of the difference between the slope change rate I of the original DEM data and the slope change rate data II of the opposite topography to obtain an error-free DEM slope change rate;
terrain relief vector: using a focus statistics tool, respectively selecting a maximum elevation value of the elevation data of the DEM and a minimum elevation value of the elevation data of the DEM according to the statistics type, obtaining a data layer A and a data layer B respectively without changing other parameters, and subtracting the data layer B from the data layer A to obtain the extremely poor in the same neighborhood range in the image layer;
Surface coverage type: splicing the images after loading the original data, embedding the images into a new grid by using tools in a tool box according to the sequence of a data management tool, the grid and a grid data set, extracting and analyzing by using a space analysis tool, and finally extracting according to a mask to obtain tower ground surface coverage classification data;
step 1.2, calculating the thunder density: converting the lightning current latitude data counted in the step 1 into an Excel data format file with the extension name of xls, defining X-axis and Y-axis data by using Arcgis, selecting map projection, importing the selected lightning current latitude data, dividing the selected lightning current latitude data into grids consistent with the step 2.1, counting the grid area and the number of lightning drops in the grids, and calculating the lightning drop density by using a field calculator through a density function.
3. The grading evaluation method for lightning protection feature parameters of a distribution line tower according to claim 1, wherein the step 2 comprises the following steps:
step 2.1, converting the lightning current latitude data counted in the step 1 into an Excel data format file, wherein the expansion name is xls, defining X-axis and Y-axis data by using Arcgis, selecting map projection, importing the selected lightning current latitude data, equally determining the size of each grid layer, dividing the grid layer into grids, counting the number of lightning drops in the grids, and obtaining a lightning drop density grid vector diagram through density function calculation processing by using a field calculator; extracting the magnitude of lightning current of each lightning point by adopting a multi-value extraction to point tool;
Step 2.2, analyzing lightning current amplitude rules under different distribution line scene factor conditions: the lightning current on the typical scene factor condition is selected for analysis, the corresponding lightning current amplitude value is extracted, and parameter fitting is carried out on the lightning current amplitude value;
step 2.3, carrying out lightning-proof level calculation on a typical overhead distribution line: the direct impact Lei Nailei level comprises a lightning stroke wire lightning resistance level and a lightning stroke tower lightning resistance level, and is calculated by using a lightning stroke wire lightning resistance level formula (3) and a lightning stroke tower lightning resistance level formula (4):
wherein I is G The lightning-proof level of the lightning strike wire is; i D The lightning resistance level of the lightning stroke tower is; u (U) 50％ 50% flashover voltage for the insulator; r is R ch The grounding resistor of the tower; l (L) gt Equivalent inductance of the tower; h is a d Is the average height of the line;
the induction Lei Nai is calculated by using the formula (5) from the magnitude of the current and the vertical distance between the ground flash point and the protected line to obtain the induction Lei Nailei level:
wherein I is g Sensing Lei Nailei level for distribution lines; s is the lightning strike point toThe closest point distance of the wire.
4. The grading evaluation method for lightning damage characteristic parameters of power distribution line pole and tower according to claim 1, wherein the step 3 comprises the following procedures:
step 3.1: calculating the tripping condition of a typical distribution line to obtain the level of direct impact Lei Nailei, directly tripping the line, and adjusting the risk grade of a tower with a statistical result of direct impact on the basis of the evaluation of the model by one level; analyzing and comparing the induction Lei Nailei level calculated in the step 2 with the median number of the lightning current amplitude values fitted by the lightning current amplitude values in the step 3 and the speed of the change of the lightning current amplitude values, so as to determine the lightning current amplitude coefficient;
Step 3.2: multiplying the lightning current amplitude influence coefficient obtained in the step 3.1 with the lightning intensity to obtain lightning hazard risk intensity indexes, and classifying the risks of a typical line scene according to the lightning hazard risk intensity indexes.
5. The grading evaluation method for lightning damage characteristic parameters of power distribution line pole and tower according to claim 1, wherein the specific process of the step 4 is as follows:
step 4.1: in Arcgis, generating a shapefile by using the risk level obtained in the step 3.2 and the corresponding scene factor in a corresponding way, defining the element type as a line type, opening an advanced editor, and performing surface construction on the line by using a construction surface tool;
step 4.2: in matlab, reading the surface data in Arcgis into matlab by shape, judging data attribution by using an index command in matlab, and correspondingly generating a data set Q.
6. The grading evaluation method for lightning damage characteristic parameters of a distribution line pole tower according to claim 1, wherein the specific process of the step 5 is as follows:
step 5.1, improving a sparrow search algorithm: (1) reverse learning strategy initializing population: first randomly initializing the positions x of N sparrow individuals in a search space i Wherein i is the number of sparrow individuals as the initial population RP; then constructing a reverse population OP, wherein the OP is formed by each sparrow individual x in the initial population RP i Is the inverse of individual x i ' composition; finally merging the population RP and OP, arranging the adaptive values of 2N sparrow individuals according to ascending order, and selecting the first N sparrow individuals with the adaptive values as an initial population; (2) sine and cosine algorithm ideas improve finder position: adding a nonlinear sine learning factor, wherein the nonlinear sine learning factor and the improved finder position formula are as follows:
w＝w min +(w max -w min )·sin(tπ/iter max ) (7)
X(t+1)＝X(t)+r 1 ×sinr 2 |r 3 ×P t -X(t)|，r 4 <0.5 (8)
X(t+1)＝X(t)+r 1 ×cosr 2 |r 3 ×P t -X(t)| r 4 ≥0.5 (9)
wherein w is a nonlinear sine learning factor, w min Is nonlinear sine learning factor minimum value, w max Is the maximum value of nonlinear sine learning factors, item is a degree of freedom parameter, t is the current iteration number, and P t R is the current optimal solution position 2 、r 3 、r 4 Is a random parameter, X is a randomly generated solution, t represents the iteration number, and the parameter r 1 Is responsible for balancing the whole algorithm detection and exploitation process, and has the expression r 1 =h-h (u/v), h being a constant, u, v being the current iteration number and the maximum iteration number, respectively; (3) initializing a population by a reverse learning strategy: the levy flight strategy improves follower position: the follower position is updated according to the following formula:
wherein,,for the previous finder's occupied position, < >>For the location occupied by the current finder, +. >For the worst position occupied by the previous finder, n is the maximum of the expandable spatial dimension, +.>Is the best place for the current finder to occupy, and the Levy flight mechanism is as follows:
wherein epsilon is a flight factor, sigma is a flight smoothness parameter, epsilon takes 1.5, and sigma is calculated as follows:
(4) Adaptive t distribution variation strategy: updating formulas (8), (9) and (10) for target positions, improving global exploration performance by using disturbance capability of t distribution mutation, and updating sparrow positions by adopting a self-adaptive t distribution mutation strategy, wherein the method specifically comprises the following steps:
in the formula (I)Indicating the position of sparrow after mutation, x i For the position of sparrow individual, t (ite) represents the t distribution regarding the iteration number as the parameter degree of freedom;
step 5.2, constructing an improved sparrow search algorithm-radial basis function neural network classifier: initializing parameters for improving sparrow search algorithm, and carrying out central vector C on jth node of hidden layer of radial basis function neural network algorithm j Width vector D of the j-th node of the hidden layer j EncodingGenerating an initialized sparrow population, receiving information transmitted by the sparrow population by a radial basis function neural network algorithm, and decoding the sparrow population to obtain a corresponding optimized center vector C' j And width vector D' j Calculating the fitness value of each sparrow, finding out the current optimal value, the worst value and the corresponding position, selecting part from sparrows with better fitness value as a finder, updating the positions according to the formulas (8) and (9), taking the rest sparrows as followers, and introducingA flight strategy, updating the position according to formula (10); and (3) carrying out self-adaptive t distribution variation according to a formula (13), perturbing the current position, generating a new solution, calculating an fitness value and updating the individual position of the sparrow, and constructing an improved sparrow search algorithm-radial basis function neural network classifier.
7. The grading evaluation method for lightning damage characteristic parameters of a distribution line pole tower according to claim 1, wherein the specific process of the step 6 is as follows:
step 6.1: according to the data set Q obtained in the step 5.2, establishing the data set Q as a sequence sample, and obtaining a risk classification evaluation model by the improved sparrow search algorithm-radial basis function network classifier through collecting the sequence sample, preprocessing data, determining input and output neurons, constructing the improved sparrow search algorithm-radial basis function network, training the improved sparrow search algorithm-radial basis function network and testing the improved sparrow search algorithm-radial basis function network;
Step 6.2: carrying out lightning hazard risk classification on the towers: and selecting an actual power distribution network line of the lightning accident, and classifying and identifying scenes of each base tower by using a risk classification evaluation model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310368062.7A CN116523299A (en) | 2023-04-08 | 2023-04-08 | Grading evaluation method for lightning damage characteristic parameters of distribution line pole tower |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310368062.7A CN116523299A (en) | 2023-04-08 | 2023-04-08 | Grading evaluation method for lightning damage characteristic parameters of distribution line pole tower |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116523299A true CN116523299A (en) | 2023-08-01 |
Family
ID=87405559
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310368062.7A Pending CN116523299A (en) | 2023-04-08 | 2023-04-08 | Grading evaluation method for lightning damage characteristic parameters of distribution line pole tower |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116523299A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117349692A (en) * | 2023-12-04 | 2024-01-05 | 国网江西省电力有限公司南昌供电分公司 | Distribution line lightning early warning method integrating multiple lightning early warning factors |
-
2023
- 2023-04-08 CN CN202310368062.7A patent/CN116523299A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117349692A (en) * | 2023-12-04 | 2024-01-05 | 国网江西省电力有限公司南昌供电分公司 | Distribution line lightning early warning method integrating multiple lightning early warning factors |
| CN117349692B (en) * | 2023-12-04 | 2024-05-07 | 国网江西省电力有限公司南昌供电分公司 | Distribution line lightning early warning method integrating multiple lightning early warning factors |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Dodangeh et al. | Novel hybrid intelligence models for flood-susceptibility prediction: Meta optimization of the GMDH and SVR models with the genetic algorithm and harmony search | |
| CN108831115B (en) | Adaboost-based power transmission line rainstorm disaster risk early warning method | |
| CN104463700B (en) | A kind of electric power line pole tower Lightning Strike Risk Evaluation method based on data mining technology | |
| CN108133225A (en) | A kind of icing flashover fault early warning method based on support vector machines | |
| CN116523299A (en) | Grading evaluation method for lightning damage characteristic parameters of distribution line pole tower | |
| CN111860947A (en) | Transmission tower bird damage occurrence probability prediction method based on BP neural network | |
| CN114966233A (en) | Lightning forecasting system and method based on deep neural network | |
| Bao et al. | Lightning performance evaluation of transmission line based on data-driven lightning identification, tracking, and analysis | |
| CN111144637A (en) | Regional power grid geological disaster forecasting model construction method based on machine learning | |
| JP2023535438A (en) | Spatial Power Outage Duration Estimation for Natural Disasters Using Optimal Composite Power Networks | |
| CN116109133A (en) | Risk early warning method for thunderstorm weather power transmission line | |
| CN110929952B (en) | Optical cable fault probability prediction method based on circuit surrounding environment and laying type | |
| CN110197048B (en) | Distribution network line lightning protection measure configuration method based on electric field and lightning damage analysis | |
| CN103164619B (en) | Grounding net of transformer substation corrosion rate pattern recognition method based on best discriminant technique criterion | |
| Ilmy et al. | Application of machine learning on google earth engine to produce landslide susceptibility mapping (case study: Pacitan) | |
| CN115115163A (en) | Method for evaluating fault risk degree of transmission line tower under typhoon crossing and storage medium | |
| CN112508060B (en) | Landslide body state judging method and system based on graph convolution neural network | |
| Francipane et al. | A comparison between heuristic, statistical, and data-driven methods in landslide susceptibility assessment: an application to the briga and giampilieri catchments | |
| CN110866693B (en) | Overhead transmission line icing risk assessment method based on GIS model | |
| CN115018014B (en) | Machine learning-assisted communication scene classification method based on multi-source information | |
| Pasang et al. | Information value model based landslide susceptibility mapping at Phuentsholing, Bhutan | |
| CN114611777A (en) | Method for predicting reservoir warehousing flow of flood season based on stacked long and short memory networks | |
| Harsono et al. | An overview of lightning data evaluation in Central Java Province for transmission line management | |
| Zhang et al. | Assessment of Geological Hazards in Ningde Based on Hybrid Intelligent Algorithm. | |
| CN117349692B (en) | Distribution line lightning early warning method integrating multiple lightning early warning factors |
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 |