CN110533331B - Fault early warning method and system based on transmission line data mining - Google Patents
Fault early warning method and system based on transmission line data mining Download PDFInfo
- Publication number
- CN110533331B CN110533331B CN201910818412.9A CN201910818412A CN110533331B CN 110533331 B CN110533331 B CN 110533331B CN 201910818412 A CN201910818412 A CN 201910818412A CN 110533331 B CN110533331 B CN 110533331B
- Authority
- CN
- China
- Prior art keywords
- fault
- transmission line
- data
- characteristic data
- power transmission
- 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
- 230000005540 biological transmission Effects 0.000 title claims abstract description 151
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000007418 data mining Methods 0.000 title claims abstract description 22
- 230000007613 environmental effect Effects 0.000 claims abstract description 30
- 238000012706 support-vector machine Methods 0.000 claims abstract description 16
- 238000009825 accumulation Methods 0.000 claims abstract description 14
- 238000012544 monitoring process Methods 0.000 claims description 28
- 238000012545 processing Methods 0.000 claims description 27
- 238000011176 pooling Methods 0.000 claims description 22
- 238000013527 convolutional neural network Methods 0.000 claims description 21
- 238000010606 normalization Methods 0.000 claims description 21
- 230000004913 activation Effects 0.000 claims description 17
- 238000004140 cleaning Methods 0.000 claims description 16
- 238000004422 calculation algorithm Methods 0.000 claims description 13
- 238000007726 management method Methods 0.000 claims description 12
- 238000009826 distribution Methods 0.000 claims description 11
- 238000004519 manufacturing process Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 230000010354 integration Effects 0.000 claims description 8
- 238000003064 k means clustering Methods 0.000 claims description 7
- 208000025274 Lightning injury Diseases 0.000 claims description 6
- 238000013501 data transformation Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 4
- 238000012549 training Methods 0.000 description 18
- 230000006870 function Effects 0.000 description 14
- 238000013145 classification model Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 239000004020 conductor Substances 0.000 description 3
- 238000012217 deletion Methods 0.000 description 3
- 230000037430 deletion Effects 0.000 description 3
- 230000005684 electric field Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 239000012212 insulator Substances 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 230000002093 peripheral effect Effects 0.000 description 3
- 238000003915 air pollution Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000013139 quantization Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003213 activating effect Effects 0.000 description 1
- 230000003373 anti-fouling effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000003631 expected effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000011155 quantitative monitoring Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 239000000725 suspension Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Bioinformatics & Computational Biology (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Operations Research (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Public Health (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- Remote Monitoring And Control Of Power-Distribution Networks (AREA)
Abstract
The invention provides a fault early warning method based on transmission line data mining, which comprises the following steps: collecting transmission line fault characteristic data, wherein the transmission line fault characteristic data comprises electric characteristic data, mechanical characteristic data and environmental characteristic data; acquiring the fault type of a power transmission line; according to different power transmission line fault types, obtaining fault accumulation values of the different power transmission line fault types based on the electrical characteristic data, the mechanical characteristic data and the environmental characteristic data; and comparing the fault accumulation values of the fault types of different power transmission lines, judging the actual fault type of the power transmission lines, and evaluating the risk level by using a support vector machine to ensure the stable operation of the power transmission lines.
Description
Technical Field
The invention belongs to the technical field of power transmission, and particularly relates to a fault early warning method and system based on power transmission line data mining.
Background
There is no mature power transmission line state on-line monitoring and fault early warning scheme in China. Researches for realizing identification of different fault reasons of the power transmission line by utilizing a direct monitoring technology are still in a starting stage, and a fault state quantitative analysis and monitoring technology based on broadband voltage and current measurement is not realized yet.
For various reasons such as geographic conditions and histories, unbalance exists between the natural energy distribution and regional economic development degree of China, and the situation is particularly reflected in the power industry, in order to eliminate the situation, the China actively develops western hydropower and thermal power resources at the beginning of the 21 st century, builds a large-scale south-north mutual supply and western electric power transmission channel, builds a nationwide electric power networking system, balances the supply and demand relationship between energy production places and load production places, realizes the optimal supply and configuration of resources and energy among nationwide regions, particularly the eastern-west regions, is connected with each other through alternating current or direct current transmission lines, and the regional local power grids are connected to gradually form a power supply network which traverses things and is a large amount of alternating current-direct current series-parallel connection across the north and south.
In order to reduce the use of non-renewable resources such as coal, renewable resources such as wind energy and solar energy are widely used nowadays, however, both wind energy and solar energy are unstable, and larger harmonic waves are introduced to a power system.
Further complicating matters, an energy internet system has now evolved globally. The power system becomes more complex and fragile, any one of equipment faults, line faults and human errors can cause large faults of the power system in a plurality of areas, and huge influences are brought to economy, production and life, so that a method capable of early warning the faults of the power transmission line is needed.
The state maintenance work on the premise of on-line monitoring is firstly carried out in the United states; state maintenance work based on state analysis and on-line monitoring is carried out in japan from the last 80 th century; state analysis and on-line monitoring technologies are adopted in many countries in europe to improve the maintenance efficiency of electrical equipment; the JPS company in japan has studied on monitoring and controlling of a power transmission line, including line fault location monitoring, weather environment monitoring, line temperature reporting, ground resistance measurement, wire tension, audible noise, and the like.
Disclosure of Invention
Aiming at the problems, the invention provides a fault early warning method and system based on transmission line data mining.
The invention provides a fault early warning method based on transmission line data mining, which comprises the following steps:
collecting transmission line fault characteristic data, wherein the transmission line fault characteristic data comprises electric characteristic data, mechanical characteristic data and environmental characteristic data;
acquiring the fault type of a power transmission line;
according to different power transmission line fault types, obtaining fault accumulation values of the different power transmission line fault types based on the electrical characteristic data, the mechanical characteristic data and the environmental characteristic data;
and comparing the fault accumulation values of the fault types of the different power transmission lines, and judging the actual fault type of the power transmission line.
Preferably, the failure accumulation value is obtained by the following formula:
wherein the method comprises the steps of
In the method, in the process of the invention,
y is a fault accumulation value of the power transmission line fault;
w ij weight, w, representing ith electrical or mechanical signature data for j fault types ij ∈[1，10]，x i Represents the ith electrical characteristic data or mechanical characteristic data normalization factor, x i E (0, 1), H represents the correction parameter, w jk Weights representing kth environmental feature data under j fault types, w jk ∈[1，10]。
Preferably, the power transmission line fault characteristic data is adopted to train a support vector machine, the fault characteristic data associated with the fault type is based on the fault characteristic data, and the risk level of the fault type is estimated based on the support vector machine.
Preferably, the transmission line fault characteristic data is obtained from an enterprise resource planning system, a transmission automation system, a transmission line on-line monitoring system, an information acquisition system, a production management system, a meteorological information system, a transmission geographic information system and an intelligent shared distribution transformer monitoring system.
Preferably, preprocessing is carried out on the collected transmission line fault characteristic data, including data cleaning, data transformation, data integration and outlier sample removal, wherein a k-means clustering algorithm is adopted to carry out data cleaning on the transmission line fault characteristic data.
Preferably, the convolutional neural network comprises an input layer, a convolutional layer, an activation layer, a pooling layer and a full connection layer;
the input layer is used for receiving the preprocessed transmission line fault characteristic data and carrying out normalization processing on the transmission line fault characteristic data;
the convolution layer is used for identifying the transmission line fault characteristic data after normalization processing;
the activation layer is used for carrying out nonlinear processing on the identified transmission line fault characteristic data;
the pooling layer is used for screening the data after nonlinear processing and extracting the most representative data;
the full connection layer is used for summarizing and outputting the data extracted by the pooling layer.
The invention also provides a fault early warning system based on the data mining of the transmission line, which comprises an acquisition module, a fault identification module and a risk early warning module;
the acquisition module is used for acquiring transmission line fault characteristic data, wherein the transmission line fault characteristic data comprises electric characteristic data, mechanical characteristic data and environmental characteristic data;
the acquisition module is also used for acquiring fault types of the power transmission line, wherein the fault types comprise lightning stroke, icing, bird damage, pollution flashover and forest fire;
the fault identification module is used for carrying out fault type identification by adopting a convolutional neural network algorithm according to the transmission line fault characteristic data;
the risk early warning module trains a support vector machine by adopting the transmission line fault characteristic data, the support vector machine builds a model by adopting a radial basis function, and evaluates the risk level of the fault type based on the fault characteristic data associated with the fault type.
Preferably, the acquisition module is specifically configured to acquire the transmission line fault characteristic data from an enterprise resource planning system, a transmission automation system, a transmission line online monitoring system, an information acquisition system, a production management system, a weather information system, a transmission geographic information system and an intelligent shared distribution transformer monitoring system.
Preferably, the acquisition module comprises a preprocessing unit, the preprocessing unit comprises a data cleaning unit, a data transformation unit, a data integration unit and an outlier sample rejection unit, wherein the data cleaning unit is used for cleaning the data of the transmission line fault characteristic data by adopting a k-means clustering algorithm.
Preferably, the fault identification module comprises a normalization unit, an activation unit, a pooling unit and a full connection unit;
the normalization unit is used for performing normalization processing on the preprocessed transmission line fault characteristic data;
the activation unit is used for carrying out nonlinear processing on the normalized transmission line fault characteristic data;
the pooling unit is used for screening the data after nonlinear processing and extracting the most representative data;
the full connection unit is used for summarizing and outputting the data extracted by the pooling layer.
According to the fault early warning method based on the power transmission line data mining, the power transmission line fault characteristic data are collected, the power transmission line fault data come from an enterprise resource planning system, a power transmission automation system, a power transmission line on-line monitoring system, an information acquisition system, a production management system and the like, a convolutional neural network algorithm is applied to identify faults, and a support vector machine is applied to evaluate risk grades, so that stable operation of the power transmission line is ensured.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 shows an overall block diagram of a project of a fault early warning system based on transmission line data mining;
fig. 2 shows a schematic structural diagram of a fault pre-warning method based on transmission line data mining according to an embodiment of the present invention;
fig. 3 shows a schematic structural diagram of a convolutional neural network according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
FIG. 1 shows an overall block diagram of a project of a fault early warning system based on transmission line data mining, relying on high-precision sensors to obtain information of electrical, mechanical and environmental characteristic data of a transmission line; the data is transmitted to the background through the on-line monitoring system and the information transmission system, and effective data is obtained through data cleaning, data transformation, data integration and outlier sample removal. And (3) applying the convolutional neural network algorithm to the data identification and classification process, identifying the state of the transmission line, taking the limit value obtained according to the historical database as a reference, alarming if the limit value exceeds the reference value, and giving out the fault type.
The fault types of the power transmission line mainly comprise lightning stroke, icing, bird damage, pollution flashover, mountain fire and the like, the fault early warning method can accurately monitor the fault types of the power transmission line, and in order to accurately monitor the fault types, the characteristic data required to be collected are different, and the fault characteristics of the power transmission line are shown in a table 1;
TABLE 1
As can be seen from table 1, different data needs to be collected for early warning of different faults, and in this embodiment, electrical feature data, mechanical feature data and environmental feature data are mainly collected:
the electrical characteristic data comprise transmission line attributes, lightning conductor data, transmission line insulativity, insulator attributes, suspension modes, tower types, insulator attributes, line-to-ground heights, antifouling equipment installation data and fireproof equipment information;
the mechanical characteristic data comprise wire tension, wire galloping amplitude, sag deviation, tower inclination rate and vertical load;
the environmental characteristic data comprises meteorological data, geographic information data, peripheral magnetic fields and electric fields, environmental resistivity, lightning stroke records, meteorological data, icing thickness, environmental wind speed, icing fault records, regional characteristics, vegetation coverage, seasons and time, bird-preventing equipment installation data, bird activity habits, line bird flash records, industrial environment distribution, air pollution data, pollution component analysis, line pollution collection rate, line pollution flash records, vegetation conditions, line network mountain distribution, forest fire grades and mountain fire fault records.
The above electrical characteristic data, mechanical characteristic data and environmental characteristic data are collected in an existing power transmission line management system, where the power transmission line management system includes an Enterprise Resource Planning (ERP) system, a power transmission automation system, a power transmission line on-line monitoring system, an information collection system, a weather information system, a production management system, a power transmission geographic information system and an intelligent common distribution transformer monitoring system, and table 2 shows data that can be collected by part of the power transmission line management system:
TABLE 2
Because the electrical characteristic data, the mechanical characteristic data and the environmental characteristic data are obtained from different systems, the data structures are different, and certain repetition and crossover exist in partial data, the data needs to be preprocessed, and the preprocessing specifically comprises data cleaning, data conversion, data integration and outlier sample rejection.
Data cleaning: including data null value processing, data outlier processing, and data duplicate value processing. The data vacancy value processing mainly comprises the steps of eliminating or supplementing record deletion in original data and deletion of a certain deletion field in the record; the data outlier processing is to make corresponding rules to reject or replace the data with overlarge deviation according to the characteristics of the original data; the data repetition value processing is to reject repeated data according to the characteristics of the data.
For example, a voltage set u= { U of monitoring points is established 1 ，u 2 ，u 3 …u k …u m M is the total number of target monitoring network monitoring points in the transmission line, u k For the voltage of the kth monitoring point of the target monitoring network in the power transmission line, putting the collected voltage values of the monitoring points into a set, if Or u k Is null or has two identical u k Then reject u k 。
And (3) data transformation: the original data is converted into a form which is easy to analyze and apply, and the main content comprises characteristic construction, data grading, data quantization and the like, such as quantization position information, operation time construction characteristic attribute, grading analysis of weather data and the like. Taking the month average temperature as an example, the months with the month average temperature higher than 30 ℃ in the ground city are mainly concentrated in 6-9 months, and the months with the month average temperature between 20 ℃ and 30 ℃ are mainly concentrated in 3-5 months and 10-11 months; the months with the average temperature of 10-20 ℃ are mainly concentrated in 1-2 months and 12 months, so that the average temperature of the months can be divided into 3 grades; from the results of the data analysis, the fault condition of the transmission line is closely related to the ambient temperature and varies with time throughout the year.
Data integration: and carrying out data statistics, combining the data into a unified database, wherein the data required by the power transmission line fault risk early warning come from different power transmission line management systems, so that the original data is required to be subjected to statistical analysis and combination.
Outlier sample rejection: the preprocessed original data may also contain abnormal samples, which are very different from most of the data in the same data set, and the data are called outlier samples, and can be identified and removed by adopting a method based on statistics, based on neighboring values or based on clustering, so that effective data are finally obtained.
In the embodiment, the collected data is cleaned by adopting a K-means clustering algorithm, and the K-means clustering algorithm is a simple and efficient clustering algorithm. The method is characterized in that data sets are divided into groups according to the similarity of the data, objects in the same cluster are similar to each other, and objects in different clusters are different. Using cluster c i The centroid (the mean of the objects assigned to the cluster) represents the cluster, cluster c i The quality of the image can be improved by intra-cluster degradation (cluster c i Sum of squares of errors between all objects and centroid ci), defined as
Wherein E represents the sum of squares of errors of all data records, p represents the data records, dist (p, c i ) Representing object p E c i And the cluster represents c i And (3) a difference. Assuming that the nearest center to object p is c p ，c p And assigned to c p Average distance lc between objects p The definition ratio judges the outlier sample according to the ratio R and rejects.
Setting 5 fault types, namely lightning stroke, icing, bird damage, pollution flashover and forest fire.
Referring to fig. 3, the convolutional neural network is composed of an input layer, a convolutional layer, an active layer, a pooling layer, and a full-connection layer.
The convolutional neural network needs to learn and train, so that a sample set for training the convolutional neural network is firstly established, training data is obtained, the used training data is normal operation data of electric characteristic data, mechanical characteristic data and environmental characteristic data of a power transmission line, historical data of a certain period of time are read from a plurality of power transmission line management systems, and healthy data with good operation states are selected from all the historical data to serve as the training data for constructing the convolutional neural network; and different sample target values are made for different fault types.
Building a framework of a Convolutional Neural Network (CNN), importing training data into a deep learning program for training, and completing a training process by adjusting training parameters.
Specifically, the convolutional neural network training process firstly carries out forward transmission and then reverse transmission of errors, and updates weight parameters so that training errors are minimized; the forward transmission is to input training data into an input layer, process the training data through the input layer, process the training data through a convolution layer, an activation layer, a pooling layer and a full-connection layer in sequence, and finally output results, namely connection weights for connecting the neurons are obtained through calculation, the reverse transmission errors are calculated according to the forward transmission output results, the output errors are reversely transmitted according to the output errors, the errors are distributed to all units of all the layers, and accordingly error signals of all the units of all the layers are obtained, and the weights of all the units are corrected.
And carrying out fault identification on the power transmission line by the trained convolutional neural network, giving out an identification result, and obtaining a fault type.
The layers of the convolutional neural network are described below.
Input layer
The Input Layer (Input Layer) receives the preprocessed data and performs normalization processing on the data, and since the influence degree of each data has close relation with the value range, all the data are subjected to normalization processing according to the following formula,
for normalized data, x is the collected data, x max And x min Respectively, the maximum and minimum of the data.
Convolutional layer
The convolution network transmits the data of the input layer to a series of convolution operations, the convolution operations are similar to the filtering process, namely, a predefined convolution kernel is adopted to slide on the data, the slid part of the convolution kernel is multiplied with the original data, the whole is added, and a new matrix is formed by the added result, so that feature extraction is realized, and the calculation formula of the convolution layer is as follows:
where l represents the number of levels of the layer network,represents the j-th feature map of the first layer, m is a set of input feature maps, b j For each feature in the convolutional layer, an offset term, x j-1 Representing the output of the previous layer, x j Representing the output of the current layer.
An activation layer
The Activation Layer (Activation Layer) is responsible for activating the features extracted by the convolution Layer, and because the convolution operation is a linear change relation of phase difference between an input matrix and a convolution kernel matrix, the Activation Layer is required to perform nonlinear mapping on the features. The activation layer mainly comprises an activation function, namely a nonlinear function is nested on the basis of the output result of the convolution layer, so that the output characteristic diagram has a nonlinear relation, the nonlinear transformation function is usually a sigmoid function, and the function expression is as follows:
pooling layer
The pooling layer is also called a downsampling layer (Downsampling Layer), and has the function of screening the characteristics in the experience area, extracting the most representative characteristics in the area, and effectively reducing the output characteristic scale so as to reduce the parameter quantity required by the model; in this embodiment, mean pooling is used, and the dimensions of the input, output and pooling matrices satisfy m=n/k. Feature extraction can be performed again, and the neuron calculation method is as follows:
wherein the method comprises the steps of
Where k is the dimension of the pooling matrix.
Full connection layer
The full-connection layer is responsible for summarizing and outputting results of features extracted by convolutional neural network learning, mapping multidimensional feature input into two-dimensional feature output, wherein high dimensions represent sample batches, low dimensions often correspond to task targets, and the full-connection layer output function is as follows:
wherein the method comprises the steps of
Wherein y is a fault accumulation value of transmission line faults, and w ij Weight, w, representing ith electrical or mechanical signature data for j fault types ij ∈[1，10]，x i Represents the ith electrical characteristic data or mechanical characteristic data normalization factor, x i E (0, 1), H represents the correction parameter, w jk Weights representing kth environmental feature data under j fault types, w jk ∈[1，10]. x represents the argument of the f () function, exemplary as:
based on the collected electrical characteristic data, mechanical characteristic data and environmental characteristic data, all fault characteristic data related under different fault types are put into the formula to be calculated, the y value of each fault type is obtained, and the maximum y value is obtained, wherein the fault type corresponding to the maximum y value is the fault actually happened to the power transmission line.
Illustratively, the fault type is calculated as a y value of a lightning strike, j=1, the collected electrical signature data comprises lightning conductor data, i=1, the mechanical signature data comprises wire tension, i=2, the environmental signature data comprises environmental resistivity, k=1, where y=f [ Hw 11 f(w 11 x 1 +w 21 x 2 )]Let the calculated y value be y1;
calculating a y value of ice coating with a fault type, wherein j=2, the collected electrical characteristic data comprise the insulation property of the power transmission line, i=1, the mechanical characteristic data comprise sag deviation, i=2, the environmental characteristic data comprise the environmental wind speed, and k=1, and y=f [ Hw ] at the moment 21 f(w 12 x 1 +w 22 x 2 )]Let the calculated y value be y2;
calculating a y value of the fault type of bird damage, wherein j=3, the collected electrical characteristic data comprise peripheral electric field data, i=1, the environmental characteristic data comprise bird activity habit, and k=1, and y=f [ Hw 31 f(w 13 x 1 )]Let the calculated y value be y3;
calculating a y value of the fault type of bird damage, j=4, wherein the collected electrical characteristic data comprise insulator attributes, i=1, the environmental characteristic data comprise industrial environmental distribution, k=1, the air pollution data, and k=2, and y=f [ (Hw) at the moment 41 +Hw 42 )f(w 14 x 1 )]Let the calculated y value be y4;
calculating a y value of a fault type of mountain fire, wherein j=5, the collected electrical characteristic data comprise line-to-ground heights, i=1, and the environmental characteristic data comprise vegetation conditionsK=1, the line network mountain distribution, k=2, where y=f [ (Hw) 51 +Hw 52 )f(w 15 x 1 )]Let the calculated y value be y5;
and comparing y1, y2, y3, y4 and y5 to obtain that y1 is the largest, and determining that the fault type of the power transmission line is lightning stroke.
After the convolutional neural network identifies the fault of the power transmission line, the risk level of the fault needs to be calculated, an early warning result is given, in the embodiment, the power transmission line risk early warning is carried out based on the RBF-SVM, the relevant data of the calculated fault is specifically extracted and input into the RBF-SVM model, and whether the early warning is sent out is determined according to the output condition of the RBF-SVM.
Support Vector Machines (SVMs) are techniques used for classification recognition in data mining, one of which constructs a hyperplane or high-dimensional space for classification, regression, or other tasks. One good separation result is achieved by the hyperplane of the greatest distance nearest to any class of training data points (i.e., functional margin) because the larger the margin, the smaller the classifier generalization error.
The embodiment adopts a support vector machine based on Radial Basis Function (RBF), the radial basis function is a multi-layer feedforward network, an effective means is provided for machine learning, and the method has the advantages of better generalization capability, small calculation error, rapid convergence, no local minimum point and the like
Radial basis function
The regression function constructed based on the radial basis function is:
in the embodiment, the risk grades of the power transmission line are divided into three grades of normal risk and serious risk, and the risk grade of the power transmission line fault is predicted according to the trained support vector machine.
Firstly, a sample set for training a support vector machine is established, and feature quantity extraction is carried out aiming at three levels of normal risk and general risk and serious risk; the characteristic quantity comprises transmission line attribute, meteorological data, geographic information data, environment resistivity, peripheral magnetic field and electric field, lightning conductor data, geographic coordinates and the like.
And randomly selecting part of sample data as training data, and taking the rest sample data as discrimination sample data.
And carrying out normalization processing on the extracted characteristic quantity, and constructing an input space vector by the characteristic quantity after normalization processing.
And taking the input space vector as input and taking the results of three risk levels corresponding to the input space vector as output, establishing an RBF-SVM classification model and training the RBF-SVM classification model according to training sample data.
And evaluating and verifying the trained RBF-SVM classification model according to the discrimination sample data, if the evaluation and verification result has high precision and the expected effect is obtained, the model can be used as a fault risk level classification model, and if the evaluation and verification model is unsatisfactory, the sample can be reselected or the number of the sample is increased, the relevant parameters are adjusted and the like to reconstruct and train the model.
And finally, after the trained RBF-SVM classification model is obtained, inputting the relevant feature quantity of the fault identified by the convolutional neural network into the trained RBF-SVM classification model to obtain the risk grade.
The fault early warning system based on the data mining of the power transmission line comprises an acquisition module, a fault identification module and a risk early warning module;
the acquisition module is used for acquiring the data of the transmission line fault influence factors;
the fault identification module is used for carrying out fault type identification by adopting a convolutional neural network algorithm according to the transmission line fault influence factor data;
and the risk early warning module trains a support vector machine by adopting the power transmission line fault influence factor data, and evaluates the risk level of the fault type based on the fault influence factor data related to the fault type.
The acquisition module is specifically used for acquiring the power transmission line fault influence factor data from an enterprise resource planning system, a power transmission automation system, a power transmission line on-line monitoring system, an information acquisition system, a production management system, a meteorological information system, a power transmission geographic information system and an intelligent shared distribution transformer monitoring system.
The acquisition module comprises a preprocessing unit, wherein the preprocessing unit comprises a data cleaning unit, a data conversion unit, a data integration unit and an outlier sample removing unit, and the data cleaning unit is used for cleaning data of the transmission line fault influence factor data by adopting a k-means clustering algorithm.
The fault identification module comprises a normalization unit, an activation unit, a pooling unit and a full connection unit;
the normalization unit is used for performing normalization processing on the preprocessed transmission line fault influence factor data;
the activation unit is used for carrying out nonlinear processing on the normalized transmission line fault influence factor data;
the pooling unit is used for screening the data after nonlinear processing and extracting the most representative data;
the full connection unit is used for summarizing and outputting the data extracted by the pooling layer.
Although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (9)
1. The fault early warning method based on the data mining of the power transmission line is characterized by comprising the following steps of:
collecting transmission line fault characteristic data, wherein the transmission line fault characteristic data comprises electric characteristic data, mechanical characteristic data and environmental characteristic data;
acquiring the fault type of a power transmission line;
according to different power transmission line fault types, obtaining fault accumulation values of the different power transmission line fault types based on the electrical characteristic data, the mechanical characteristic data and the environmental characteristic data;
comparing the fault accumulation values of different power transmission line fault types through a convolutional neural network, and judging the actual fault type of the power transmission line;
the failure accumulation value is obtained by the following formula:
wherein the method comprises the steps of
In the method, in the process of the invention,
y is a fault accumulation value of the power transmission line fault;
wij represents the weight of the ith electrical characteristic data or mechanical characteristic data under the j fault type, and is E [1, 10];
xi represents the ith electrical characteristic data or mechanical characteristic data normalization factor, xi e (0, 1);
h represents a correction parameter;
wjk represents the weight of the kth environmental feature data under the j fault type, wjk e 1, 10.
2. The transmission line data mining-based fault early warning method according to claim 1, wherein a support vector machine is trained by adopting the transmission line fault characteristic data, and the risk level of the fault type is evaluated based on the support vector machine based on the fault characteristic data associated with the fault type.
3. The transmission line data mining-based fault pre-warning method according to claim 1, wherein the transmission line fault characteristic data is taken from an enterprise resource planning system, a transmission automation system, a transmission line on-line monitoring system, an information acquisition system, a production management system, a weather information system, a transmission geographic information system and an intelligent shared distribution transformer monitoring system.
4. The power transmission line data mining-based fault early warning method according to claim 1, wherein preprocessing is performed on collected power transmission line fault characteristic data, including data cleaning, data transformation, data integration and outlier sample rejection, and wherein a k-means clustering algorithm is adopted to perform data cleaning on the power transmission line fault characteristic data.
5. The transmission line data mining-based fault early warning method according to claim 4, wherein the convolutional neural network comprises an input layer, a convolutional layer, an activation layer, a pooling layer and a full connection layer;
the input layer is used for receiving the preprocessed transmission line fault characteristic data and carrying out normalization processing on the transmission line fault characteristic data;
the convolution layer is used for identifying the transmission line fault characteristic data after normalization processing;
the activation layer is used for carrying out nonlinear processing on the identified transmission line fault characteristic data;
the pooling layer is used for screening the data after nonlinear processing and extracting the most representative data;
the full connection layer is used for summarizing and outputting the data extracted by the pooling layer.
6. The fault early warning system based on the data mining of the power transmission line is characterized by comprising an acquisition module, a fault identification module and a risk early warning module;
the acquisition module is used for acquiring transmission line fault characteristic data, wherein the transmission line fault characteristic data comprises electric characteristic data, mechanical characteristic data and environmental characteristic data;
the acquisition module is also used for acquiring fault types of the power transmission line, wherein the fault types comprise lightning stroke, icing, bird damage, pollution flashover and forest fire;
the fault identification module is used for carrying out fault type identification by adopting a convolutional neural network algorithm according to the transmission line fault characteristic data;
based on the electrical characteristic data, the mechanical characteristic data and the environmental characteristic data, obtaining fault accumulation values of different power transmission line fault types;
the failure accumulation value is obtained by the following formula:
wherein the method comprises the steps of
In the method, in the process of the invention,
y is a fault accumulation value of the power transmission line fault;
wij represents the weight of the ith electrical characteristic data or mechanical characteristic data under the j fault type, and is E [1, 10];
xi represents the ith electrical characteristic data or mechanical characteristic data normalization factor, xi e (0, 1);
h represents a correction parameter;
wjk represents the weight of the kth environmental feature data under the j fault type, wjk epsilon [1, 10];
the risk early warning module trains a support vector machine by adopting the transmission line fault characteristic data, the support vector machine builds a model by adopting a radial basis function, and evaluates the risk level of the fault type based on the fault characteristic data associated with the fault type.
7. The transmission line data mining-based fault pre-warning system of claim 6, wherein the collection module is specifically configured to collect the transmission line fault signature data from an enterprise resource planning system, a transmission automation system, a transmission line on-line monitoring system, an information collection system, a production management system, a weather information system, a transmission geographic information system, and an intelligent common distribution transformer monitoring system.
8. The power transmission line data mining-based fault early warning system according to claim 6, wherein the acquisition module comprises a preprocessing unit, the preprocessing unit comprises a data cleaning unit, a data transformation unit, a data integration unit and an outlier sample rejection unit, and the data cleaning unit is used for performing data cleaning on the power transmission line fault characteristic data by adopting a k-means clustering algorithm.
9. The transmission line data mining-based fault early warning system of claim 8, wherein the fault identification module comprises a normalization unit, an activation unit, a pooling unit, and a fully connected unit;
the normalization unit is used for performing normalization processing on the preprocessed transmission line fault characteristic data;
the activation unit is used for carrying out nonlinear processing on the normalized transmission line fault characteristic data;
the pooling unit is used for screening the data after nonlinear processing and extracting the most representative data;
the full connection unit is used for summarizing and outputting the data extracted by the pooling layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910818412.9A CN110533331B (en) | 2019-08-30 | 2019-08-30 | Fault early warning method and system based on transmission line data mining |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910818412.9A CN110533331B (en) | 2019-08-30 | 2019-08-30 | Fault early warning method and system based on transmission line data mining |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110533331A CN110533331A (en) | 2019-12-03 |
| CN110533331B true CN110533331B (en) | 2024-03-15 |
Family
ID=68666025
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910818412.9A Active CN110533331B (en) | 2019-08-30 | 2019-08-30 | Fault early warning method and system based on transmission line data mining |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110533331B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111144447B (en) * | 2019-12-09 | 2022-05-31 | 国网新疆电力有限公司电力科学研究院 | Power grid peak-valley time interval division method for preventing peak regulation risk caused by new energy output |
| CN112232370A (en) * | 2020-09-16 | 2021-01-15 | 中国电子产品可靠性与环境试验研究所（（工业和信息化部电子第五研究所）（中国赛宝实验室）） | Fault analysis and prediction method for engine |
| CN112417763A (en) * | 2020-11-25 | 2021-02-26 | 杭州凯达电力建设有限公司 | Defect diagnosis method, device and equipment for power transmission line and storage medium |
| CN112508429A (en) * | 2020-12-16 | 2021-03-16 | 常州大学 | Fault diagnosis method for cathode protection system of buried pipeline based on convolutional neural network |
| CN113033889B (en) * | 2021-03-19 | 2022-12-20 | 国网河北省电力有限公司邢台供电分公司 | High-voltage transmission line fault prediction method and device and terminal equipment |
| CN113205487B (en) * | 2021-04-12 | 2022-09-13 | 武汉大学 | Cable state detection method based on residual error network fusion heterogeneous data |
| CN113538371B (en) * | 2021-07-14 | 2022-05-17 | 贵州电网有限责任公司 | Power distribution network overhead line icing thickness monitoring method for improving K-means clustering |
| CN115061009B (en) * | 2022-06-01 | 2024-05-31 | 国网浙江省电力有限公司嘉兴供电公司 | Power transmission line fault analysis and early warning method |
| CN117633479B (en) * | 2024-01-26 | 2024-04-09 | 国网湖北省电力有限公司 | Method and system for analyzing and processing faults of charging piles |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105974270A (en) * | 2016-07-22 | 2016-09-28 | 广东电网有限责任公司电力科学研究院 | Power transmission line fault type diagnosis method and system |
| CN106779505A (en) * | 2017-02-28 | 2017-05-31 | 中国南方电网有限责任公司 | A kind of transmission line malfunction method for early warning driven based on big data and system |
| CN107609574A (en) * | 2017-08-18 | 2018-01-19 | 上海电力学院 | Wind turbines fault early warning method based on data mining |
| CN108133225A (en) * | 2017-10-23 | 2018-06-08 | 南京南瑞集团公司 | A kind of icing flashover fault early warning method based on support vector machines |
| CN108959732A (en) * | 2018-06-15 | 2018-12-07 | 西安科技大学 | A kind of transmission line malfunction kind identification method based on convolutional neural networks |
| CN109188198A (en) * | 2018-08-30 | 2019-01-11 | 武汉大学 | A kind of fault data matching process based on convolutional neural networks |
| CN109270407A (en) * | 2018-11-16 | 2019-01-25 | 国网山东省电力公司电力科学研究院 | Extra high voltage direct current transmission line failure cause discrimination method based on Multi-source Information Fusion |
| CN109739209A (en) * | 2018-12-11 | 2019-05-10 | 深圳供电局有限公司 | A kind of electric network failure diagnosis method based on Classification Data Mining |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3460496B1 (en) * | 2017-09-26 | 2022-06-08 | Siemens Aktiengesellschaft | A method and apparatus for automatic localization of a fault |
-
2019
- 2019-08-30 CN CN201910818412.9A patent/CN110533331B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105974270A (en) * | 2016-07-22 | 2016-09-28 | 广东电网有限责任公司电力科学研究院 | Power transmission line fault type diagnosis method and system |
| CN106779505A (en) * | 2017-02-28 | 2017-05-31 | 中国南方电网有限责任公司 | A kind of transmission line malfunction method for early warning driven based on big data and system |
| CN107609574A (en) * | 2017-08-18 | 2018-01-19 | 上海电力学院 | Wind turbines fault early warning method based on data mining |
| CN108133225A (en) * | 2017-10-23 | 2018-06-08 | 南京南瑞集团公司 | A kind of icing flashover fault early warning method based on support vector machines |
| CN108959732A (en) * | 2018-06-15 | 2018-12-07 | 西安科技大学 | A kind of transmission line malfunction kind identification method based on convolutional neural networks |
| CN109188198A (en) * | 2018-08-30 | 2019-01-11 | 武汉大学 | A kind of fault data matching process based on convolutional neural networks |
| CN109270407A (en) * | 2018-11-16 | 2019-01-25 | 国网山东省电力公司电力科学研究院 | Extra high voltage direct current transmission line failure cause discrimination method based on Multi-source Information Fusion |
| CN109739209A (en) * | 2018-12-11 | 2019-05-10 | 深圳供电局有限公司 | A kind of electric network failure diagnosis method based on Classification Data Mining |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110533331A (en) | 2019-12-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110533331B (en) | Fault early warning method and system based on transmission line data mining | |
| CN112749904B (en) | Power distribution network fault risk early warning method and system based on deep learning | |
| CN109871976A (en) | A kind of prediction technique of power quality containing distributed power distribution network based on cluster and neural network | |
| CN109002915B (en) | Photovoltaic power station short-term power prediction method based on Kmeans-GRA-Elman model | |
| CN108388962B (en) | Wind power prediction system and method | |
| CN111753893A (en) | Wind turbine generator power cluster prediction method based on clustering and deep learning | |
| CN111488896B (en) | Distribution line time-varying fault probability calculation method based on multi-source data mining | |
| Kolhe et al. | GA-ANN for short-term wind energy prediction | |
| CN110766200A (en) | Method for predicting generating power of wind turbine generator based on K-means mean clustering | |
| CN110378549B (en) | Transmission tower bird damage grade assessment method based on FAHP-entropy weight method | |
| Eseye et al. | Short-term wind power forecasting using a double-stage hierarchical hybrid GA-ANN approach | |
| Chen et al. | Research on wind power prediction method based on convolutional neural network and genetic algorithm | |
| CN115796434A (en) | Management and control method and device for power distribution network, electronic equipment and storage medium | |
| CN114399081A (en) | Photovoltaic power generation power prediction method based on weather classification | |
| CN111461921A (en) | Load modeling typical user database updating method based on machine learning | |
| CN112241606A (en) | Cooperative decision-making method for operation and maintenance of ship intelligent equipment based on CPS decision-making module | |
| CN113987912A (en) | Pollutant on-line monitoring system based on geographic information | |
| CN112288157A (en) | Wind power plant power prediction method based on fuzzy clustering and deep reinforcement learning | |
| CN116467648A (en) | Early monitoring method for nonlinear platform power failure based on Internet of things table | |
| CN115829145A (en) | Photovoltaic power generation capacity prediction system and method | |
| CN112085350B (en) | Evaluation method for photovoltaic array state in large photovoltaic power station | |
| KR20190071174A (en) | Method and system for short-term wind speed prediction based on pressure data | |
| CN111612227A (en) | Load prediction method based on K-means clustering and bat optimization neural network | |
| CN108345996B (en) | System and method for reducing wind power assessment electric quantity | |
| CN115907228A (en) | Short-term power load prediction analysis method based on PSO-LSSVM |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |