CN112116160B - Important power transmission channel disaster monitoring method based on improved cellular automaton of optimized neural network - Google Patents
Important power transmission channel disaster monitoring method based on improved cellular automaton of optimized neural network Download PDFInfo
- Publication number
- CN112116160B CN112116160B CN202011021322.6A CN202011021322A CN112116160B CN 112116160 B CN112116160 B CN 112116160B CN 202011021322 A CN202011021322 A CN 202011021322A CN 112116160 B CN112116160 B CN 112116160B
- Authority
- CN
- China
- Prior art keywords
- cell
- monitoring
- disaster
- data
- neural network
- 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.)
- Active
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 85
- 230000005540 biological transmission Effects 0.000 title claims abstract description 60
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000001413 cellular effect Effects 0.000 title claims abstract description 47
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 46
- 238000012549 training Methods 0.000 claims abstract description 17
- 238000011156 evaluation Methods 0.000 claims abstract description 15
- 230000008569 process Effects 0.000 claims abstract description 15
- 238000003062 neural network model Methods 0.000 claims abstract description 8
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 238000007689 inspection Methods 0.000 claims abstract description 5
- 230000005856 abnormality Effects 0.000 claims abstract description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 41
- 239000002689 soil Substances 0.000 claims description 27
- 239000013598 vector Substances 0.000 claims description 18
- 238000006243 chemical reaction Methods 0.000 claims description 16
- 238000004458 analytical method Methods 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 10
- 230000010307 cell transformation Effects 0.000 claims description 9
- 230000002159 abnormal effect Effects 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 7
- 238000012502 risk assessment Methods 0.000 claims description 6
- 230000003068 static effect Effects 0.000 claims description 6
- 238000010586 diagram Methods 0.000 claims description 5
- 238000012876 topography Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000003247 decreasing effect Effects 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 238000007418 data mining Methods 0.000 claims description 3
- 229910052757 nitrogen Inorganic materials 0.000 claims description 3
- 230000007547 defect Effects 0.000 description 4
- 238000010606 normalization Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 210000000709 aorta Anatomy 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 229910052698 phosphorus Inorganic materials 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000006467 substitution reaction 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Game Theory and Decision Science (AREA)
- Mathematical Physics (AREA)
- Biomedical Technology (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Artificial Intelligence (AREA)
- Educational Administration (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an important power transmission channel disaster monitoring method based on an optimized neural network improved cellular automaton, which belongs to the field of important power transmission channel disaster monitoring and comprises the following steps: s1, data acquisition is carried out by monitoring resources, a meteorological department and a power grid meteorological site; s2, storing and preprocessing the data in the step S1; s3, determining a cell state based on an analytic hierarchy process; s4, acquiring an RBF neural network data center by adopting cosine similarity, optimizing weights from an implicit layer to an output layer by adopting an IWD algorithm, improving network accuracy, and training and learning by adopting an RBF neural network model; s5, predicting and evaluating the current monitoring data according to the rule to obtain a cellular state, namely a disaster prediction evaluation level, forming a current overall disaster state level distribution map of the monitoring area, and carrying out targeted inspection on the line passing through the disaster level abnormality and serious area by combining the trend and the topological structure of the important power transmission channel to eliminate the hidden danger of the disaster.
Description
Technical Field
The invention relates to the field of disaster monitoring of important power transmission channels, in particular to a disaster monitoring method of important power transmission channels based on an improved cellular automaton of an optimized neural network.
Background
With the rapid development of electric power construction, the scale of the power grid is continuously enlarged, the importance of an important power transmission channel mainly of 750kV in the power grid is more and more concentrated, and the important power transmission channel is an important energy aorta in the modern society. Therefore, when large-scale power grid accidents occur, huge economic losses can be caused, and great politics, social influence and personal casualties can be caused. Although the hazard of the important transmission line defect is smaller than the hazard degree of the fault, the occurrence frequency is far higher than that of the line fault, moreover, the defect can be changed from the variable quantity to the quality to the serious fault, and the factors causing the power grid fault of the important transmission channel not only comprise equipment fault, manual operation and possibly extreme natural disasters. Therefore, in order to ensure the normal operation of the important transmission line under the complex terrain condition, the monitoring work of the line body, the surrounding environment and the meteorological parameters must be enhanced, so that the defect is prevented.
At present, an important power transmission channel of a power grid is established with a power grid weather station, a power transmission monitoring device, a weather department station and other stand-up disaster monitoring and forecasting platforms, but the number of lines is numerous, monitoring equipment and data types are diversified, the prior art and means lack of effective integration and utilization of the existing monitoring data, the risk state of the important channel line cannot be effectively mastered in time, and the supporting force for operation, maintenance and overhaul work is limited. Meanwhile, because a certain barrier exists in the monitoring mode communication of different areas, a monitoring blind area exists on the basis of the existing monitoring equipment, and important channel power transmission line collapse, wire breakage and tripping phenomena occur.
Cellular automata has good advantages in terms of simulating complex space-time phenomena, behaviors and processes, and therefore, the cellular automata is widely used in the field of disaster monitoring. The neural network has important advantages in the aspect of mining the nonlinear relation among data, so that the neural network can be combined with a cellular automaton model to build an all-condition and all-dimensional disaster monitoring method of the power transmission line on important power transmission channels, and the environmental parameters and the running states of the important power transmission channels, especially disaster multiple areas, are subjected to centralized monitoring and risk assessment, so that the safety early warning and disaster prediction of the important power transmission lines of the regional power grid are realized, and the multi-parameter data fusion and intelligent analysis of the power transmission lines are realized.
Disclosure of Invention
1. Technical problem to be solved
Aiming at the problems existing in the prior art, the invention aims to provide an important power transmission channel disaster monitoring method based on an optimized neural network improved cellular automaton, which aims at the defects of the prior important power transmission channel disaster monitoring, and provides an important power transmission channel disaster monitoring method based on the optimized neural network improved cellular automaton, wherein data acquisition and pretreatment are carried out through a monitoring terminal, a power grid weather station and a weather station, then grid division with the same size is carried out on a disaster monitoring area, each grid is used as an independent cell, and then a hierarchical analysis method is carried out on the single cell according to the prior power transmission line monitoring history information to determine the cell state; determining the basis of the RBF neural network by combining cosine similarity with a mean value clustering algorithm; an improved intelligent water drop algorithm (IWD) is used for optimizing weights between hidden layers and output layers of a neural network, a cell conversion rule is obtained through training, disaster grade evaluation prediction is carried out on an existing important power transmission channel according to the cell conversion rule, and operation and maintenance personnel can adopt a targeted strategy on a fragile area carrier by combining a power transmission channel topological structure and a disaster evaluation result, so that real-time disaster monitoring and early warning evaluation of the important channel are realized, and the safety of the important power transmission channel is ensured.
2. Technical proposal
In order to solve the problems, the invention adopts the following technical scheme:
an important power transmission channel disaster monitoring method based on an optimized neural network improved cellular automaton comprises the following steps:
S1, data acquisition is carried out by monitoring resources, a meteorological department and a power grid meteorological site;
s2, storing and preprocessing the data in the step S1;
s3, determining a cell state based on an analytic hierarchy process;
S4, acquiring an RBF neural network data center by adopting cosine similarity, optimizing weights from an implicit layer to an output layer by adopting an IWD algorithm, improving network accuracy, training and learning by adopting an RBF neural network model, and obtaining a state conversion rule of a current cell at the next moment by adopting the states of neighbor cells;
s5, predicting and evaluating the current monitoring data according to the rule to obtain a cellular state, namely a disaster prediction evaluation level, forming a current overall disaster state level distribution map of the monitoring area, and carrying out targeted inspection on the line passing through the disaster level abnormality and serious area by combining the trend and the topological structure of the important power transmission channel to eliminate the hidden danger of the disaster.
In step S3, the monitoring area is divided into cells by combining with the geographic information system, the size of 3km is taken as one cell, and the single cell of the monitoring area is subjected to hierarchical analysis through historical data and expert experience to determine the cell state.
As a preferred embodiment of the present invention, in step S3, the method includes the steps of:
S301, preprocessing and normalizing data;
S302, forming a hierarchical structure diagram;
S303, acquiring a judgment matrix according to expert experience;
s304, consistency test;
S305, calculating a score according to the weight matrix.
As a preferred scheme of the present invention, in step S301, an important power transmission channel is divided into grids based on a geographic information system, each grid is used as a cell, data acquisition is performed according to an on-line monitoring terminal, a power grid weather site and a weather department site which are arranged in the range of the cell, and the data is normalized by adopting the following formula:
Wherein x i is the actual monitored quantity of certain monitored data, a and b are the threshold values of the evaluation indexes of the monitored quantity, and the larger the normalized data is, the more serious the monitored quantity deviates from the normal state, and the threshold value of each index is set according to the relevant expert reference rules and guidelines.
As a preferred embodiment of the present invention, in step S302, a hierarchical analysis model is constructed, including a target layer M, a criterion layer C, and a scheme layer P, in which: the target layer M is a disaster risk score; taking factors affecting the disaster model of the area into consideration, dividing the criterion layer C into 3 parts of climate factors, topography factors and tower factors; the scheme layer P is a specific influencing factor to be considered.
As a preferred embodiment of the present invention, in step S305, the result is evaluated by a hierarchical analysis method according to a hierarchical analysis model, and the finally obtained disaster risk score is classified into three classes, namely "normal", "abnormal" and "serious", according to historical experience.
As a preferred embodiment of the present invention, in step S4, a cellular automaton model is constructed; the cellular automaton model can be described by the following formula:
a= (C, S, N, R) formula (2)
Wherein C is a cell space, namely, a whole cell in a disaster monitoring area; s is a cell state set, and according to the analytic hierarchy process result, the cell states are divided into three types, namely, normal is 0, abnormal is 1, and serious is 2, so that the cell state set S= {0,1,2}; r represents a lattice bit of the cellular automaton, and S k (r, t) represents a kth state of a cell on the lattice bit r at a time t; n represents the neighbor of a cell taking r as the center, a Moore type cell neighbor model is adopted, namely 8 neighbors are contained around each cell, and N= { N 1,N2,…,Nq},Nq represents the position of the q-th neighbor of the cell r relative to r; r is a transformation rule of a cell, namely S (R, t) to S (R, t+1) R= (R 1,R2,…Rm),Rm is an mth transformation rule, which is only related to states of 8 neighbor cells around the cell at the time t, and a specific cell transformation function rule cannot be found due to complexity and uncertainty of disasters, so that data mining is performed through historical data, and the cell transformation rule is obtained by adopting an optimized RBF neural network model.
As a preferred embodiment of the present invention, in step S4, the method includes the steps of:
S401: uniformly selecting k data in cell historical data randomly as a sample center;
s402: the cosine similarity between each sample and the kth center is calculated, and the calculation formula is that
Where j represents an element in vectors i and k;
s403: calculating an average value rho of cosine similarity of all samples of the kth center point; the formula is
S404: judging whether cos (i, k) of the ith sample and the kth center is smaller than rho k;
s405: if S404 is yes, the ith sample belongs to the kth class and the traversal is continued;
S406: if S404 is negative, the ith vector is added as the center of the (k+1) th sample, and the traversal is continued until the whole data set is obtained.
As a preferable scheme of the invention, the method further comprises the step of optimizing the weight from the hidden layer to the output layer of the RBF neural network by adopting an intelligent water drop algorithm, and comprises the following steps:
s407: obtaining a central vector of an hidden layer and the number of the hidden layers according to the S406 result, setting RBF network parameters, and generating l k-dimensional random number vectors based on levy distribution as initial weights;
s408: initializing IWD algorithm parameters;
s409: each water droplet will flow with a greater probability to a path of lesser soil volume, assuming that the water droplet is at node a, the probability to flow to the next node b is:
wherein f (soil (a, b)) is the amount of soil between nodes a to b; in order to solve the possible local convergence, introducing a linear decreasing function, wherein k iter is the current iteration number, T max is the maximum iteration number, searching in the global with a larger probability at the initial stage of iteration, and searching in the local with a smaller probability at the later stage of iteration; p_is the roulette option;
s4010: the water droplet velocity, the amount of soil in the water droplet, and the amount of soil in the path are updated.
The water drop speed updating formula is as follows:
Wherein, vel is the water drop junction speed;
The updated formula of the soil amount in the water drops is:
the soil quantity update formula in the path is:
Wherein alpha n and sp, sq, sr, vp, vq, vr are static parameters and are set during initialization
S4011: traversing all nodes continuously;
S4012: selecting paths to calculate network output according to all water drops obtained at present, calculating an error E according to an error function, selecting a path with the smallest error to be reserved as the optimal path, and updating the soil quantity in the path according to whether the water drops of the optimal path pass through or not, wherein an updating formula is as follows:
wherein lambda is a static parameter, mu is the number of optimal water drops
S4013: judging whether the iteration times are reached;
S4014: and outputting an optimal path, namely a neural network weight, completing network training, obtaining a cell conversion rule, and carrying out risk assessment prediction.
In step S5, the optimized RBF neural network is adopted to obtain a cell transformation rule, and the trained RBF network model is used to predict the next state of the r-th cell, so as to extend the next state to the whole cell space, and a disaster risk level prediction distribution map of the power transmission channel is obtained.
3. Advantageous effects
Compared with the prior art, the invention has the advantages that:
(1) The first innovation point of the invention is that the monitoring area is subjected to grid division, disaster monitoring is performed by expanding points to the surface, comprehensive evaluation analysis is performed on the current disaster monitoring area by combining weather factors, topography factors and tower self factors of an important power transmission channel through an analytic hierarchy process, an evaluation result is directly converted into a cell state, and no application case for directly determining the cell state by applying analytic hierarchy process on the aspect of disaster monitoring of the power transmission channel exists at present.
(2) At present, in the aspect of disaster monitoring of important transmission channels, no case of combining a neural network with a cellular automaton exists. In other fields, the method is different from the combination method of the neural network and the cellular automaton in other fields, and the method is used for determining the center of the RBF network by combining cosine similarity with a mean value clustering algorithm aiming at the process of calculating the RBF neural network, so that the difference between individuals can be better distinguished compared with Euclidean distance; in addition, levy flight and linear decreasing weights are introduced into the IED algorithm, so that the probability of converging on local optimum is effectively reduced, the weight of the RBF neural network is optimized through the improved IWD algorithm, the optimal weight is found, and the training efficiency of the network is improved.
(3) The cell state conversion rule is also a classification problem in essence, and the RBF neural network classification capability is obviously superior to other networks, and the training speed is high, so that the input training is directly carried out through the optimized RBF neural network, the conversion rule of the current cell is deeply mined, the algorithm precision is improved, and the obtained cell conversion rule is more approximate to a true value, so that the method is an improvement and expansion of a cell automaton model. Meanwhile, in the next process of accumulating new data, the model can be continuously corrected, and the precision is improved.
Drawings
FIG. 1 is a general flow chart of an important power transmission channel disaster monitoring method based on an optimized neural network improved cellular automaton;
FIG. 2 is a flow chart of an analytic hierarchy process in the method for monitoring disaster of important transmission channels based on an improved cellular automaton of an optimized neural network;
FIG. 3 is a hierarchical structure diagram of an analytic hierarchy process in the method for monitoring disaster of important transmission channels based on an improved cellular automaton of the optimized neural network;
FIG. 4 is a flowchart of RBF neural network in an important transmission channel disaster monitoring method based on an improved cellular automaton of an optimized neural network;
Fig. 5 is a diagram of RBF neural network in an important transmission channel disaster monitoring method based on an improved cellular automaton of an optimized neural network.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. It is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments, and that all other embodiments obtained by persons of ordinary skill in the art without making creative efforts based on the embodiments in the present invention are within the protection scope of the present invention.
Examples:
Referring to fig. 1-5, an important power transmission channel disaster monitoring method based on an optimized neural network for improving cellular automata includes the following steps:
S1, data acquisition is carried out by monitoring resources, meteorological departments and power grid meteorological sites: collecting field data, including monitoring terminals, weather department stations and power grid weather stations which are arranged on a power transmission line and a tower; the monitoring terminal comprises image monitoring, wire temperature monitoring, microclimate monitoring, icing monitoring, windage yaw monitoring, breeze vibration monitoring, tower inclination monitoring, wire galloping monitoring and on-site pollution degree monitoring; meanwhile, weather department data and power grid weather site data are respectively acquired through different data communication modes to acquire information including temperature, humidity, wind speed, wind direction, rainfall and the like and omnibearing multidimensional information including geographic information;
s2, storing and preprocessing the data in the step S1: the multi-dimensional monitoring information collected from different data sources is stored in a special Oracle database in the power system in a safe transmission mode and an encryption mode, and is stored in a grading mode according to different types of data information, so that data information of a single tower node and integral channel weather information are formed, and the monitoring center system can call the data information at any time through a special interface; preprocessing data, including data cleaning, unstructured conversion, data normalization and the like;
S3, determining a cell state based on an analytic hierarchy process: dividing grids based on an important transmission channel GIS system, wherein the dividing size is 3km by 3km, each grid is used as a cell, normalizing the data acquired in the step S4, then constructing a hierarchical analysis model, wherein a target layer M is a disaster risk score, a criterion layer C is divided into 3 parts of climate factors, topography factors and tower factors, a scheme layer P is a specific factor to be considered, the hierarchical model is shown in fig. 3, finally, a disaster evaluation grade of each cell is obtained, and the grade is converted into a cell state, so that the cell state of the whole cell space at a certain moment is obtained; the specific flow is shown in figure 2, and the specific steps are as follows:
S301, data preprocessing and normalization: dividing grids of an important power transmission channel based on a geographic information system, taking each grid as a cell, acquiring data according to an on-line monitoring terminal, a power grid weather site and a weather department site which are arranged in the range of the cell, carrying out normalization processing on the data, and carrying out normalization by adopting the following formula:
Wherein x i is the actual monitored quantity of certain monitored data, a and b are thresholds of evaluation indexes of the monitored quantity, and the larger the normalized data is, the more serious the monitored quantity deviates from a normal state, and the threshold of each index is set according to relevant expert reference rules and guidelines;
S302, forming a hierarchy structure diagram: carrying out hierarchical structure division according to factors influencing disaster evaluation of an important power transmission channel, and constructing a hierarchical analysis model, wherein the hierarchical analysis model comprises a target layer M, a criterion layer C and a scheme layer P, and the method comprises the following steps of: the target layer M is a disaster risk score; taking factors affecting the disaster model of the area into consideration, dividing the criterion layer C into 3 parts of climate factors, topography factors and tower factors; the scheme layer P is a specific influence factor to be considered;
S303, acquiring a judgment matrix according to expert experience: constructing a judgment matrix according to the pairwise comparison importance degree of the criterion layer C and the scheme layer P data according to expert experience;
s304, consistency test: the matrix eigenvalue lambda max is calculated according to the formula Consistency test is carried out to obtain a weight matrix;
S305, calculating a score according to the weight matrix: according to the analytic hierarchy process, the final obtained disaster risk score is divided into three grades according to historical experience, namely normal grade, abnormal grade and serious grade, the weight matrix is multiplied with the current normalized monitoring data to obtain a disaster evaluation score grade, the state of each cell in the cell space is obtained according to the following table, and the cell state is vectorized and used for output calculation of the RBF network.
S4, acquiring an RBF neural network data center through adopting cosine similarity, optimizing weights from an implicit layer to an output layer through adopting an IWD algorithm, improving network accuracy, training and learning through adopting an RBF neural network model, and obtaining a state conversion rule of a current cell at the next moment through the states of neighbor cells: optimizing an RBF neural network, training by taking cell states of a historical cell space as training data, obtaining a cell conversion rule, and constructing a cell automaton model; the cellular automaton model can be described by the following formula:
a= (C, S, N, R) formula (2)
Wherein C is a cell space, namely, a whole cell in a disaster monitoring area; s is a cell state set, and according to the analytic hierarchy process result, the cell states are divided into three types, namely, normal is 0, abnormal is 1, and serious is 2, so that the cell state set S= {0,1,2}; r represents a lattice bit of the cellular automaton, and S k (r, t) represents a kth state of a cell on the lattice bit r at a time t; n represents the neighbor of a cell taking r as the center, a Moore type cell neighbor model is adopted, namely 8 neighbors are contained around each cell, and N= { N 1,N2,…,Nq},Nq represents the position of the q-th neighbor of the cell r relative to r; r is a transformation rule of a cell, namely S (R, t) to S (R, t+1) R= (R 1,R2,…Rm),Rm is an mth transformation rule, which is only related to states of 8 neighbor cells around the cell at the time t, and a specific cell transformation function rule cannot be found due to complexity and uncertainty of disasters, so that data mining is performed through historical data, and the cell transformation rule is obtained by adopting an optimized RBF neural network model, wherein a specific flow chart is shown in fig. 4, and the specific steps are as follows:
S401: uniformly selecting k data in cell historical data randomly as a sample center;
s402: the cosine similarity between each sample and the kth center is calculated, and the calculation formula is that
Where j represents an element in vectors i and k;
s403: calculating an average value rho of cosine similarity of all samples of the kth center point; the formula is
S404: judging whether cos (i, k) of the ith sample and the kth center is smaller than rho k;
s405: if S404 is yes, the ith sample belongs to the kth class and the traversal is continued;
S406: if S404 is negative, the ith vector is added as the center of the (k+1) th sample, and the traversal is continued until the whole data set is obtained.
S407: obtaining a central vector of an hidden layer and the number of the hidden layers according to the S406 result, setting RBF network parameters, and generating l k-dimensional random number vectors based on levy distribution as initial weights;
s408: initializing IWD algorithm parameters;
s409: each water droplet will flow with a greater probability to a path of lesser soil volume, assuming that the water droplet is at node a, the probability to flow to the next node b is:
wherein f (soil (a, b)) is the amount of soil between nodes a to b; in order to solve the possible local convergence, introducing a linear decreasing function, wherein k iter is the current iteration number, T max is the maximum iteration number, searching in the global with a larger probability at the initial stage of iteration, and searching in the local with a smaller probability at the later stage of iteration; p_is the roulette option;
S4010: the water droplet velocity, the amount of soil in the water droplet and the amount of soil in the path are updated.
The water drop speed updating formula is as follows:
Wherein, vel is the water drop junction speed;
The updated formula of the soil amount in the water drops is:
the soil quantity update formula in the path is:
Wherein alpha n and sp, sq, sr, vp, vq, vr are static parameters and are set during initialization
S4011: traversing all nodes continuously;
S4012: selecting paths to calculate network output according to all water drops obtained at present, calculating an error E according to an error function, selecting a path with the smallest error to be reserved as the optimal path, and updating the soil quantity in the path according to whether the water drops of the optimal path pass through or not, wherein an updating formula is as follows:
wherein lambda is a static parameter, mu is the number of optimal water drops
S4013: judging whether the iteration times are reached;
S4014: outputting an optimal path, namely a neural network weight, completing network training, obtaining a cell conversion rule, carrying out risk assessment prediction, and recording 1 when elements in a predicted output vector are more than or equal to 0.5, otherwise recording 0;
S5, predicting and evaluating the current monitoring data according to the acquired cellular automaton conversion rule to obtain a cellular state, namely a disaster prediction evaluation level, and forming a current overall disaster state level distribution map of the monitoring area: the optimized RBF neural network is adopted to obtain a cell transformation rule, the RBF network model trained in the step S4 is used for predicting the next state of the r-th cell, so that the next state is expanded to the whole cell space, a power transmission channel disaster risk level prediction distribution map is obtained, and the important power transmission channel trend and the topological structure are combined to conduct targeted inspection on the lines passing through disaster level abnormal and serious areas, so that disaster hidden danger is eliminated. For example, as shown in fig. 5, the input vector is x= [1,1,2,0,0,0,1,2], the output vector o= [1, 0], and the disaster risk assessment is performed according to the disaster grade assessment result obtained by combining the trained RBF neural network with the present monitoring data, the risk map is drawn for the whole cellular space, and if the important power transmission channel passes through the disaster anomaly or serious area, the area is targeted for inspection.
The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art, who is within the scope of the present invention, should make equivalent substitutions or modifications according to the technical solution and the modified concept thereof, within the scope of the present invention.
Claims (9)
1. An important power transmission channel disaster monitoring method based on an optimized neural network improved cellular automaton is characterized by comprising the following steps:
S1, data acquisition is carried out by monitoring resources, a meteorological department and a power grid meteorological site;
s2, storing and preprocessing the data in the step S1;
S3, expressing the divided grids by using cells, expressing the state of each divided grid at the current moment by using the cell state, and determining the cell state based on an analytic hierarchy process, wherein the method specifically comprises the following steps of:
S301, preprocessing and normalizing data;
S302, forming a hierarchical structure diagram;
S303, acquiring a judgment matrix according to expert experience;
s304, consistency test;
s305, calculating a score according to the weight matrix;
S4, constructing a cellular automaton model, acquiring a cellular conversion rule by improving an intelligent water drop algorithm, acquiring an RBF neural network data center by adopting cosine similarity, optimizing weights from an hidden layer to an output layer by adopting the intelligent water drop algorithm, improving network accuracy, training and learning by adopting the RBF neural network model, and acquiring a state conversion rule of a current cellular at the next moment by adopting the state of a neighbor cellular;
the method for acquiring the cellular transformation rule by constructing the cellular automaton model specifically comprises the following steps:
s401, randomly and uniformly selecting k data in cell historical data as a sample center;
S402: calculating cosine similarity between each sample and the kth center;
S403: calculating an average value rho k of cosine similarity of all samples of the kth center point;
S404: judging whether the cosine similarity between the ith sample and the kth center is smaller than rho k or not;
s405: if S404 is yes, the ith sample belongs to the kth class and the traversal is continued;
S406: if S404 is not, adding the ith vector as the (k+1) th sample center, and continuing to traverse until the whole data set is obtained;
for adopting intelligent water drop algorithm to optimize RBF neural network hidden layer to output layer weight and adopting RBF neural network model to make training study, specifically includes:
s407: obtaining a central vector of an hidden layer and the number of the hidden layers according to the S406 result, setting RBF network parameters, and generating l k-dimensional random number vectors based on levy distribution as initial weights;
s408: initializing IWD algorithm parameters;
S409: performing local search based on the probability that each water drop flows to the next node in the node a;
s4010: updating the water drop speed, the soil quantity in the water drops and the soil quantity in the paths;
S4011: traversing all nodes continuously;
S4012: selecting paths to calculate network output according to all water drops obtained at present, calculating an error E according to an error function, selecting a path with the smallest error to be reserved as the optimal path, and updating the soil quantity in the path according to whether the water drops of the optimal path pass through or not;
s4013: judging whether the iteration times are reached;
S4014: outputting an optimal path, namely a neural network weight, completing network training, obtaining a cell conversion rule, and carrying out risk assessment prediction;
s5, predicting and evaluating the current monitoring data according to the rule to obtain a cellular state, namely a disaster prediction evaluation level, forming a current overall disaster state level distribution map of the monitoring area, and carrying out targeted inspection on the line passing through the disaster level abnormality and serious area by combining the trend and the topological structure of the important power transmission channel to eliminate the hidden danger of the disaster.
2. The method for monitoring the disaster of the important power transmission channel based on the improved cellular automaton of the optimized neural network according to claim 1, wherein in the step S3, the monitoring area is divided into cells by combining a geographic information system, the size of 3km is taken as one cell, and the single cell of the monitoring area is subjected to hierarchical analysis through historical data and expert experience to determine the cell state.
3. The method for monitoring disaster in an important power transmission channel based on an improved cellular automaton of claim 1, wherein in step S301, the important power transmission channel is divided into grids based on a geographic information system, each grid is used as a cell, data acquisition is performed according to an on-line monitoring terminal, a power grid weather site and a weather department site which are arranged in the range of the cell, and the data is normalized by adopting the following formula:
wherein x im represents normalized data of the monitoring data, x i represents actual monitoring quantity of a certain monitoring data, a and b are thresholds of evaluation indexes of the monitoring quantity, and the larger the normalized data is, the more serious the monitoring quantity deviates from a normal state, and the threshold of each index is set according to a relevant expert reference rule and a guide rule.
4. The method for monitoring important power transmission channel disasters based on the improved cellular automaton of claim 1, wherein in step S302, a hierarchical analysis model is constructed, including a target layer M, a criterion layer C and a scheme layer P, wherein: the target layer M is a disaster risk score; taking factors affecting the disaster model of the area into consideration, dividing the criterion layer C into 3 parts of climate factors, topography factors and tower factors; the scheme layer P is a specific influencing factor to be considered.
5. The method for monitoring disaster in an important power transmission channel based on an improved cellular automaton of claim 1, wherein in step S305, the result is evaluated by using a hierarchical analysis method according to a hierarchical analysis model, and the finally obtained disaster risk score is classified into three classes, namely "normal", "abnormal" and "serious", according to historical experience.
6. The method for monitoring important power transmission channel disasters based on the improved cellular automaton of the optimized neural network according to claim 5, wherein in step S4, a cellular automaton model is constructed; the cellular automaton model can be described by the following formula:
a= (C, S, N, R) formula (2)
Wherein C is a cell space, namely, a whole cell in a disaster monitoring area; s is a cell state set, and according to the analytic hierarchy process result, the cell states are divided into three types, namely, normal is 0, abnormal is 1, and serious is 2, so that the cell state set S= {0,1,2}; r represents a lattice bit of the cellular automaton, and S k (r, t) represents a kth state of a cell on the lattice bit r at a time t; n represents the neighbor of a cell taking r as the center, a Moore type cell neighbor model is adopted, namely 8 neighbors are contained around each cell, and N= { N 1,N2,…,Nq},Nq represents the position of the q-th neighbor of the cell r relative to r; r is a transformation rule of a cell, namely S (R, t) to S (R, t+1) R= (R 1,R2,…Rm),Rm is an mth transformation rule, which is only related to states of 8 neighbor cells around the cell at the time t, and a specific cell transformation function rule cannot be found due to complexity and uncertainty of disasters, so that data mining is performed through historical data, and the cell transformation rule is obtained by adopting an optimized RBF neural network model.
7. The method for monitoring the disaster of the important transmission channel based on the improved cellular automaton of the optimized neural network according to claim 6, wherein in step S4, the method comprises the steps of:
S401: uniformly selecting k data in cell historical data randomly as a sample center;
s402: the cosine similarity between each sample and the kth center is calculated, and the calculation formula is that
Where j represents an element in vectors i and k;
s403: calculating an average value rho of cosine similarity of all samples of the kth center point; the formula is
S404: judging whether cos (i, k) of the ith sample and the kth center is smaller than rho k;
s405: if S404 is yes, the ith sample belongs to the kth class and the traversal is continued;
S406: if S404 is negative, the ith vector is added as the center of the (k+1) th sample, and the traversal is continued until the whole data set is obtained.
8. The method for monitoring the disaster of the important power transmission channel based on the improved cellular automaton of the optimized neural network according to claim 7, further comprising the step of optimizing weights from an hidden layer to an output layer of the RBF neural network by adopting an intelligent water drop algorithm, and comprising the following steps:
s407: obtaining a central vector of an hidden layer and the number of the hidden layers according to the S406 result, setting RBF network parameters, and generating l k-dimensional random number vectors based on levy distribution as initial weights;
s408: initializing IWD algorithm parameters;
s409: each water droplet will flow with a greater probability to a path of lesser soil volume, assuming that the water droplet is at node a, the probability to flow to the next node b is:
Wherein f (soil (a, b)) is the amount of soil between nodes a to b; in order to solve the possible local convergence, introducing a linear decreasing function, wherein k iter is the current iteration number, T max is the maximum iteration number, searching in the global with a larger probability at the initial stage of iteration, and searching in the local with a smaller probability at the later stage of iteration; selecting a mode for roulette wheel;
s4010: updating the water drop speed, the soil quantity in the water drops and the soil quantity in the paths;
the water drop speed updating formula is as follows:
Wherein, vel is the water drop junction speed;
The updated formula of the soil amount in the water drops is:
the soil quantity update formula in the path is:
Wherein alpha n and sp, sq, sr, vp, vq, vr are static parameters and are set during initialization
S4011: traversing all nodes continuously;
S4012: selecting paths to calculate network output according to all water drops obtained at present, calculating an error E according to an error function, selecting a path with the smallest error to be reserved as the optimal path, and updating the soil quantity in the path according to whether the water drops of the optimal path pass through or not, wherein an updating formula is as follows:
wherein lambda is a static parameter, mu is the number of optimal water drops
S4013: judging whether the iteration times are reached;
S4014: and outputting an optimal path, namely a neural network weight, completing network training, obtaining a cell conversion rule, and carrying out risk assessment prediction.
9. The method for monitoring the disaster of the important transmission channel based on the improved cellular automaton of claim 8, wherein in the step S4, the cell transformation rule is obtained by adopting the optimized RBF neural network, in the learning stage, the cell space history data is selected as training data, the states of 8 cells around the t moment are taken as input, and the state of the r cell at the t+1 moment is taken as output for training; in step S5, predicting the next state of the r-th cell by using the trained RBF network model, so as to extend the next state to the whole cell space, and obtaining a disaster risk level prediction distribution map of the power transmission channel.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011021322.6A CN112116160B (en) | 2020-09-25 | 2020-09-25 | Important power transmission channel disaster monitoring method based on improved cellular automaton of optimized neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011021322.6A CN112116160B (en) | 2020-09-25 | 2020-09-25 | Important power transmission channel disaster monitoring method based on improved cellular automaton of optimized neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112116160A CN112116160A (en) | 2020-12-22 |
| CN112116160B true CN112116160B (en) | 2024-07-02 |
Family
ID=73796767
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011021322.6A Active CN112116160B (en) | 2020-09-25 | 2020-09-25 | Important power transmission channel disaster monitoring method based on improved cellular automaton of optimized neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112116160B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113077159B (en) * | 2021-04-13 | 2023-12-15 | 中能融合智慧科技有限公司 | Data processing method and data processing device |
| CN113379942B (en) * | 2021-06-10 | 2023-11-24 | 平安国际智慧城市科技股份有限公司 | Construction site facility inspection method and device, computer equipment and storage medium |
| CN113537528B (en) * | 2021-07-28 | 2022-05-03 | 贵州电网有限责任公司 | Preprocessing method and system for state monitoring data of power transmission and transformation equipment |
| CN114089091A (en) * | 2021-09-27 | 2022-02-25 | 杭州柯林电气股份有限公司 | Power transmission line on-line monitoring method and system based on non-contact sensor |
| CN114841084B (en) * | 2022-07-04 | 2022-09-09 | 西南交通大学 | Aerial image-based personnel evacuation simulation method, system and equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104361393A (en) * | 2014-09-06 | 2015-02-18 | 华北电力大学 | Method for using improved neural network model based on particle swarm optimization for data prediction |
| CN105956934A (en) * | 2016-05-05 | 2016-09-21 | 国网湖南省电力公司防灾减灾中心 | Power grid forest fire and icing disaster safety evaluation method based on fuzzy comprehensive evaluation approach |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103095494B (en) * | 2012-12-31 | 2016-06-29 | 北京邮电大学 | A kind of electric power telecommunication network risk assessment method |
| WO2014203039A1 (en) * | 2013-06-19 | 2014-12-24 | Aselsan Elektronik Sanayi Ve Ticaret Anonim Sirketi | System and method for implementing reservoir computing using cellular automata |
| CN106021666B (en) * | 2016-05-10 | 2019-03-12 | 四川大学 | A kind of mountain fire disaster alarm method of overhead transmission line |
| TR201619702A1 (en) * | 2016-12-27 | 2018-07-23 | Univ Yeditepe | A METHOD FOR THERMAL DISPOSAL BASED SOLUTION OF CLUSTERING PROBLEM WITH CELLULAR AUTOMATION |
| US10832180B2 (en) * | 2017-10-30 | 2020-11-10 | The Aerospace Corporation | Artificial intelligence system that employs windowed cellular automata to create plausible alternatives |
| JP6596543B2 (en) * | 2017-11-28 | 2019-10-23 | 学校法人梅花学園 | Data conversion method, data conversion device, program, and cellular automaton configuration method |
| CN108182484A (en) * | 2017-12-04 | 2018-06-19 | 国网吉林省电力有限公司长春供电公司 | Spatial Load Forecasting method based on gridding technology and BP neural network |
| CN109508360B (en) * | 2018-11-07 | 2021-03-16 | 武汉大学 | Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton |
| CN109839824A (en) * | 2019-01-24 | 2019-06-04 | 青岛理工大学 | Network control system delay compensation method based on predictive control |
| CN111144611A (en) * | 2019-11-22 | 2020-05-12 | 国网辽宁省电力有限公司经济技术研究院 | Spatial load prediction method based on clustering and nonlinear autoregression |
-
2020
- 2020-09-25 CN CN202011021322.6A patent/CN112116160B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104361393A (en) * | 2014-09-06 | 2015-02-18 | 华北电力大学 | Method for using improved neural network model based on particle swarm optimization for data prediction |
| CN105956934A (en) * | 2016-05-05 | 2016-09-21 | 国网湖南省电力公司防灾减灾中心 | Power grid forest fire and icing disaster safety evaluation method based on fuzzy comprehensive evaluation approach |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112116160A (en) | 2020-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112116160B (en) | Important power transmission channel disaster monitoring method based on improved cellular automaton of optimized neural network | |
| CN108280553A (en) | Regional Torrent Risk Zonation based on GIS- Artificial neural network ensembles and prediction technique | |
| US8606418B1 (en) | Wind prediction for wind farms through the use of weather radar | |
| CN102411729B (en) | Wind power prediction method based on adaptive linear logic network | |
| CN106897826A (en) | A kind of street accidents risks appraisal procedure and system | |
| CN109522599A (en) | Transmission line of electricity catastrophic failure method for early warning caused by a kind of typhoon | |
| CN111428942B (en) | Line icing thickness prediction method for extracting micro-terrain factors based on variable grid technology | |
| CN113988273A (en) | Ice disaster environment active power distribution network situation early warning and evaluation method based on deep learning | |
| CN113344408B (en) | Processing method for multi-scale situation awareness process of civil aviation traffic control operation | |
| CN107480781A (en) | The nuclear accident Source Term Inversion method of neutral net adaptive Kalman filter | |
| CN115526382B (en) | Road network level traffic flow prediction model interpretability analysis method | |
| CN111126672A (en) | High-voltage overhead transmission line typhoon disaster prediction method based on classification decision tree | |
| CN109376863A (en) | Photovoltaic power based on MIV-BP neural network is classified prediction technique in short term | |
| CN106919645A (en) | The sight spot meteorological element Intelligent fine Forecasting Methodology at the big scenic spot of complex landform | |
| CN116205265A (en) | Power grid fault diagnosis method and device based on deep neural network | |
| CN114357670A (en) | Power distribution network power consumption data abnormity early warning method based on BLS and self-encoder | |
| CN115688032A (en) | Power grid risk early warning method and system under typhoon disaster based on multi-source data fusion | |
| CN109948839A (en) | The prediction of overhead transmission line galloping risk and method for early warning and system | |
| CN113610309A (en) | Fire station site selection method and device based on big data and artificial intelligence | |
| CN112434887B (en) | Water supply network risk prediction method combining network kernel density estimation and SVM | |
| CN114417732A (en) | Self-adaptive identification method and system for multi-source load damage of power distribution network under strong typhoon | |
| US12007075B2 (en) | Methods and internet of things (IoT) systems for managing smart gas safety hazard items based on geographic information systems | |
| CN111912393A (en) | Hydrological environment monitoring method based on water surface mobile robot | |
| CN112257329A (en) | Method for judging influence of typhoon on line | |
| CN112819208A (en) | Spatial similarity geological disaster prediction method based on feature subset coupling model |
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 |