CN114662976A - Power distribution network overhead line state evaluation method based on convolutional neural network - Google Patents

Power distribution network overhead line state evaluation method based on convolutional neural network Download PDF

Info

Publication number
CN114662976A
CN114662976A CN202210376163.4A CN202210376163A CN114662976A CN 114662976 A CN114662976 A CN 114662976A CN 202210376163 A CN202210376163 A CN 202210376163A CN 114662976 A CN114662976 A CN 114662976A
Authority
CN
China
Prior art keywords
overhead line
convolutional neural
neural network
image
power distribution
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.)
Withdrawn
Application number
CN202210376163.4A
Other languages
Chinese (zh)
Inventor
刘志刚
马宏光
张明明
罗进
商经锐
林阳坡
陈耀高
唐超
胡东
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Yunnan Power Grid Co ltd Dehong Power Supply Bureau
Original Assignee
Yunnan Power Grid Co ltd Dehong Power Supply Bureau
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Yunnan Power Grid Co ltd Dehong Power Supply Bureau filed Critical Yunnan Power Grid Co ltd Dehong Power Supply Bureau
Priority to CN202210376163.4A priority Critical patent/CN114662976A/en
Publication of CN114662976A publication Critical patent/CN114662976A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • 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
    • 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)
  • Health & Medical Sciences (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Strategic Management (AREA)
  • Software Systems (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • General Business, Economics & Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Game Theory and Decision Science (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • Primary Health Care (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a power distribution network overhead line state evaluation method based on a convolutional neural network, which comprises the following steps of: collecting a real-time image of an overhead line; setting m different image types according to the conventional component composition of the overhead line of the power distribution network, and inputting the real-time image into the trained convolutional neural network model to identify the image type corresponding to the real-time image; and taking the health degree of the overhead line as a target layer, the image type of the overhead line as a criterion layer and the fault type of the overhead line as an index layer, thereby constructing an overhead line hierarchical structure model, and judging the result output by the convolutional neural network model by adopting an analytic hierarchy process, thereby deciding the fault state of the overhead line. The invention develops a power distribution network overhead line state evaluation method by combining a convolutional neural network and a chromatographic analysis method, overcomes the defects of manual line patrol, and has the advantages of simple algorithm, strong universality, high practicability and the like.

Description

Power distribution network overhead line state evaluation method based on convolutional neural network
Technical Field
The invention relates to the technical field of power industry, in particular to a power distribution network overhead line state evaluation method based on a convolutional neural network.
Background
The operation environment of the overhead line of the power distribution network is complex and changeable, the overhead line is often in the environments of bird nest movement, ice disasters, fire disasters, vegetation growth, hanging objects and the like, if the potential fault hazard occurs, the safety of the power distribution network can be threatened, and loss which is difficult to estimate can be caused. Therefore, the evaluation of the running state of the overhead line is an important key link affecting the reliability of the power system. However, the state of the overhead line is evaluated by adopting a manual line patrol method in the traditional method, a large amount of manpower, material resources and financial resources are consumed, the problems of low evaluation efficiency, low accuracy and the like exist, and the method is influenced by various external environments.
Based on the condition, the power distribution network overhead line state evaluation method based on the convolutional neural network is provided.
Disclosure of Invention
The invention provides a power distribution network overhead line state evaluation method based on a convolutional neural network, and mainly aims to solve the problems in the prior art.
The invention adopts the following technical scheme:
a power distribution network overhead line state evaluation method based on a convolutional neural network is characterized by comprising the following steps: the method comprises the following steps:
s1, acquiring a real-time image of the overhead line through an image acquisition device;
s2, setting m different image types according to the conventional component composition of the overhead line of the power distribution network, and inputting the real-time image into the trained convolutional neural network model to identify the image type corresponding to the real-time image;
s3, taking the health degree of the overhead line as a target layer, the image type of the overhead line as a standard layer and the fault type of the overhead line as an index layer, so as to construct an overhead line hierarchical structure model, and judging the result output by the convolutional neural network model by adopting an analytic hierarchy process, thereby deciding the fault state of the overhead line.
Further, in step S2, the image types include 9 types, including foundation and protection facilities, towers, wires, insulator strings, disconnecting switches, hardware fittings, lightning protection facilities and grounding devices, auxiliary facilities, and line protection areas.
Further, in step S2, the convolutional neural network model is an improved SSD convolutional neural network model, which includes 4 convolutional layers and 3 fully-connected layers, wherein the first two convolutional layers are connected with the pooling layer, and the last fully-connected layer uses the Softmax function for classification.
Further, before training the convolutional neural network model, a convolutional neural network platform tensoflow is adopted to label the image data set of the overhead line, and the labeling processing method comprises the following steps:
s21, respectively acquiring images of the overhead line in different environments through an image acquisition device to construct a training set and a verification set, and converting all image sizes into set sizes by adopting a bilinear interpolation method;
and S22, classifying the images in the training set and the verification set according to the set image types, and converting the classified training set and the classified verification set into two independent binary files respectively, so as to obtain an image data set of the overhead line.
Further, in step S3, the overhead line has 7 types of faults, including corrosion, shedding, foreign matter, dirt, discoloration, hidden danger, and loss.
Further, in step S3, the process of making a decision for the chromatography includes the following steps:
s31, according to the constructed hierarchical index system, enabling experts to compare the importance degrees of all indexes of the same level pairwise, and establishing a judgment matrix for the indexes of each level;
s32, solving the eigenvectors and the maximum eigenvalues of the judgment matrixes;
s33 maximum characteristic root using judgment matrix
Figure 104267DEST_PATH_IMAGE001
Carrying out consistency check;
s34, normalizing the feature vectors passing the test to obtain the weight value of the level;
and S35, comprehensively evaluating the state and health degree of the overhead line according to the calculated weight value and by combining with the pre-specified overhead line evaluation grade standard.
Further, the judgment matrix equation of the analytic hierarchy process is as follows:
Figure 370163DEST_PATH_IMAGE002
in the formula: m is the number of image type types, i is the serial number of the image type corresponding to the real-time image; n is the number of the fault types of the overhead lines, and j is the fault type serial number of the overhead lines corresponding to the real-time images.
Compared with the prior art, the invention has the beneficial effects that:
the invention develops a power distribution network overhead line state evaluation method by combining a convolutional neural network and a chromatographic analysis method, can accurately judge the fault hidden danger of the overhead line in time through a visual image, and accurately evaluate the overall state of the power distribution network overhead line, thereby overcoming the defects of manual line inspection, and having the advantages of simple algorithm, strong universality, high practicability and the like.
Drawings
FIG. 1 is a schematic diagram of the algorithm flow of the present invention.
FIG. 2 is a diagram of a convolutional neural network architecture of the present invention.
FIG. 3 is a framework diagram of a chromatographic assay of the invention.
FIG. 4 is a diagram of a hierarchical analysis structure according to the present invention.
Detailed Description
The following describes embodiments of the present invention with reference to the drawings.
Referring to fig. 1 to 4, the invention provides a power distribution network overhead line state evaluation method based on a convolutional neural network, which comprises the following steps:
and S1, acquiring the real-time image of the overhead line through an image acquisition device. In practical applications, a device such as a mobile phone, a camera, or a drone may be used to capture the image, but the resolution of the real-time image is required to be 1920 × 1440.
S2, setting m different image types according to the conventional component composition of the overhead line of the power distribution network, and inputting the real-time image into the trained convolutional neural network model to identify the image type corresponding to the real-time image.
Preferably, the image types are 9, including foundation and protection facilities, towers, wires, insulator strings, isolating switches, hardware fittings, lightning protection facilities and grounding devices, auxiliary facilities and line protection areas.
Before training a convolutional neural network model, a convolutional neural network platform Tensorflow is adopted to label an image data set of an overhead line, and the processing process comprises the following steps:
and S21, respectively acquiring images of the overhead line in different environments through an image acquisition device to construct a training set and a verification set, and converting all image sizes into set sizes by adopting a bilinear interpolation method. Preferably, in the embodiment, the overhead line images are respectively collected in different environments such as bird nest movement, ice disasters, fires, vegetation growth, hanging objects and the like, and an overhead line image data set with the capacity of 5 ten thousand is constructed. Wherein 4 ten thousand images are used for constructing a training set, 1 ten thousand images are used for constructing a verification set, and the two images are independent and have no overlap. The number of overhead line images in each state is set according to the proportion of the collected original images, and meanwhile, in order to meet the network input requirement, a bilinear interpolation method is adopted to convert all the image sizes into sizes of 64 multiplied by 64.
And S22, classifying the images in the training set and the verification set according to the set image types, and converting the classified training set and the classified verification set into two independent binary files respectively, thereby obtaining an image data set of the overhead line. The convolutional neural network algorithm belongs to supervised classification and needs to label a mass data set. In the research, a convolutional neural network platform Tensorflow of Google company is adopted to label an image data set of the overhead line. Firstly, storing the arranged training set and verification set under two folders, wherein each folder comprises the 9 image types, namely, a foundation and protection facility, a tower, a lead, an insulator string, an isolating switch, hardware fittings, a lightning protection facility and grounding device, an auxiliary facility and a line protection area, and respectively establishing corresponding folder labels. Then, each picture is converted into binary data of a fixed length using a built-in function of Tensorflow, where the first byte is an image tag and the remaining 64 × 64 × 3 bytes are image information. And finally, respectively converting the training set and the verification set into two independent binary files to obtain an image data set of the overhead line after the labeling processing is finished.
Referring to fig. 2, preferably, the convolutional neural network model in this embodiment is an improved SSD convolutional neural network model, which includes 4 convolutional layers and 3 fully-connected layers, wherein the first two convolutional layers are connected with the pooling layer, and the last fully-connected layer uses the Softmax function for classification. Specifically, the method comprises the following steps:
if it is
Figure 447841DEST_PATH_IMAGE003
I-th layer feature map representing convolutional neural network, network input
Figure 355754DEST_PATH_IMAGE004
A 64 × 64 × 3 original image, a convolution layer
Figure 835277DEST_PATH_IMAGE003
The calculation process of (a) can be described as formula (1):
Figure 22807DEST_PATH_IMAGE005
wherein,
Figure 853360DEST_PATH_IMAGE006
representing the weight of the i-th layer convolution kernel; symbol
Figure 768226DEST_PATH_IMAGE007
Representing the convolution operation of the convolution kernel with the i-th layer input vector, the value of convolution and the offset value
Figure 102255DEST_PATH_IMAGE008
Adding; then through a non-linear excitation functionf(x)And obtaining a characteristic diagram of the i-th layer, wherein a Relu function is selected as a nonlinear excitation function in the network, and the function has low computational complexity in the back propagation process and can enable the model to be rapidly converged.
The loss function loss is calculated as:
Figure 959222DEST_PATH_IMAGE009
Figure 11491DEST_PATH_IMAGE010
the actual label is represented by a representation of,
Figure 730049DEST_PATH_IMAGE011
the output of the prediction is represented by,rrepresenting the total number of samples.
Common pooling methods are average pooling and maximum pooling,the present example employs a maximum pooling method. Maximum pooling layer
Figure 449743DEST_PATH_IMAGE012
The calculation process of (c) can be described as formula (3):
Figure 710567DEST_PATH_IMAGE013
fig. 2 contains detailed network design parameters of the convolutional neural network model in this embodiment. Specifically, the C1 layer uses 32 convolution kernels of 5 × 5 × 3 with a step size of 1, 32 features are extracted after the input image is convolved, 32 feature maps of 60 × 60 are obtained, and 32 feature maps of 29 × 29 are obtained through the S1 layer. The C2 layer has 64 convolution kernels of 5 × 5 × 32, the output feature map of the S1 layer is convoluted to obtain 64 feature maps of 29 × 29, and 64 feature maps of 14 × 14 are obtained through the S2 layer. The C3 layer has 128 convolution kernels of 5 multiplied by 64, and the output characteristic diagram of the S2 layer is convoluted to obtain 128 characteristic diagrams of 14 multiplied by 14; the C4 layer uses 256 convolution kernels of 5 × 5 × 128 to convolve the output feature maps of the C3 layer, so as to obtain 256 feature maps of 14 × 14, and the feature maps are output to the F5 layer. The F5 layer adopts 384 neurons to carry out full connection processing on 256 14 multiplied by 14 feature maps; the F6 layer adopts 192 neurons to carry out full connection processing on 256 neurons; the F7 layer uses the Softmax function to classify the feature vector processing results into 9 image types.
S3, taking the health degree of the overhead line as a target layer, the image type of the overhead line as a criterion layer, and the fault type of the overhead line as an index layer, so as to construct an overhead line hierarchical structure model, and adopting an analytic hierarchy process to decide the fault state of the overhead line.
Referring to fig. 3 and 4, the analytic hierarchy process is a quantitative method, and comprises layering a complex problem, decomposing the problem into a target layer, a criterion layer and an index layer according to the property and the total target of the problem, comparing the quantities of the same layer in pairs, quantizing the importance of the quantities in the layer by using relative scales, establishing a judgment matrix, solving the eigenvector of the matrix to obtain the weight distribution of each layer, and finally calculating the relative weight of the evaluation index layer to the target layer from bottom to top. Specifically, the chromatography comprises the following steps:
s31, according to the constructed hierarchical index system, enabling experts to compare the importance degrees of each index of the same level pairwise, and establishing a judgment matrix for each level of index. In the process of pairwise comparison, scales of 1-9 are introduced to obtain a quantized judgment matrix. Wherein 1, 3, 5, 7, 9 respectively represent the same, slightly, more, very, and absolutely important.
TABLE 1 quantization scale & degree LUT
Figure 250132DEST_PATH_IMAGE014
The formula of the judgment matrix is:
Figure 37960DEST_PATH_IMAGE015
in the formula: m is the number of image type types, i is the serial number of the image type corresponding to the real-time image; n is the number of the fault types of the overhead lines, and j is the fault type serial number of the overhead lines corresponding to the real-time images.
And S32, solving the eigenvector and the maximum eigenvalue of each judgment matrix.
The purpose of solving the judgment matrix is to obtain the eigenvector and the maximum eigenvalue of the judgment matrix, solve the judgment matrix by a sum-product method, and add each column of the matrix according to rows to obtain a formula (5):
Figure 346581DEST_PATH_IMAGE016
the calculation of the maximum eigenvalue is expressed as formula (6):
Figure 561662DEST_PATH_IMAGE017
s33 maximum characteristic root using judgment matrix
Figure 572212DEST_PATH_IMAGE018
And (5) carrying out consistency check.
And introducing a deviation consistency index CI in the analytic hierarchy process to check the consistency of the judgment matrix. The expression for the consistency test is shown in equation (7):
Figure 898151DEST_PATH_IMAGE019
in the formula, CR is a consistency ratio, RI is an average random consistency index, and the values thereof are shown in table 2. When CR is <0.1, the decision matrix can be considered to have satisfactory consistency, otherwise the decision matrix is adjusted.
TABLE 2 average random consensus index (RI)
Figure 326859DEST_PATH_IMAGE021
If the consistency check is not passed, the judgment matrix is subjected to normalization correction, and then steps S32 and S33 are repeated until the check is passed.
And S34, normalizing the feature vectors passing the test by adopting a formula (8) to obtain the weight value of the level.
Figure 447261DEST_PATH_IMAGE022
In the formula: m represents the number of image type categories;
Figure 712152DEST_PATH_IMAGE023
representing state i versus some property of the k +1 layer
Figure 841782DEST_PATH_IMAGE024
The relative weight of (2); n represents
Figure 859416DEST_PATH_IMAGE024
The number of sub-attributes of (2);
Figure 416300DEST_PATH_IMAGE025
to represent
Figure 417754DEST_PATH_IMAGE024
The relative weight of the sub-attribute j at layer k,
Figure 334763DEST_PATH_IMAGE026
representing the relative weight of state j with respect to sub-attribute j.
And S35, comprehensively evaluating the state and health degree of the overhead line according to the calculated weight value and by combining with the pre-specified overhead line evaluation grade standard. The standard of the overhead line evaluation grade is established by experts according to experience and relevant requirements, and the table 3 can be specifically referred.
TABLE 3 overhead line evaluation grade Standard
Figure 206904DEST_PATH_IMAGE028
The above description is only an embodiment of the present invention, but the design concept of the present invention is not limited thereto, and any insubstantial modifications made by using the design concept should fall within the scope of infringing the present invention.

Claims (7)

1. A power distribution network overhead line state evaluation method based on a convolutional neural network is characterized by comprising the following steps: the method comprises the following steps:
s1, acquiring a real-time image of the overhead line through an image acquisition device;
s2, setting m different image types according to the conventional component composition of the overhead line of the power distribution network, and inputting the real-time image into the trained convolutional neural network model to identify the image type corresponding to the real-time image;
and S3, taking the health degree of the overhead line as a target layer, taking the image type of the overhead line as a standard layer, taking the fault type of the overhead line as an index layer, thereby constructing an overhead line hierarchical structure model, and judging the result output by the convolutional neural network model by adopting an analytic hierarchy process, thereby deciding the fault state of the overhead line.
2. The power distribution network overhead line state evaluation method based on the convolutional neural network as claimed in claim 1, characterized in that: in step S2, the image types include 9 types, including foundation and protection facilities, towers, wires, insulator strings, disconnecting switches, hardware fittings, lightning protection facilities and grounding devices, auxiliary facilities, and line protection areas.
3. The power distribution network overhead line state evaluation method based on the convolutional neural network as claimed in claim 1, characterized in that: in step S2, the convolutional neural network model is an improved SSD convolutional neural network model, which includes 4 convolutional layers and 3 fully-connected layers, wherein the first two convolutional layers are connected to the pooling layer, and the last fully-connected layer uses the Softmax function for classification.
4. The power distribution network overhead line state evaluation method based on the convolutional neural network as claimed in claim 1, characterized in that: before training a convolutional neural network model, a convolutional neural network platform Tensorflow is adopted to label an image data set of an overhead line, and the labeling processing method comprises the following steps:
s21, respectively acquiring images of the overhead line in different environments through an image acquisition device to construct a training set and a verification set, and converting all image sizes into set sizes by adopting a bilinear interpolation method;
and S22, classifying the images in the training set and the verification set according to the set image types, and converting the classified training set and the classified verification set into two independent binary files respectively, so as to obtain an image data set of the overhead line.
5. The power distribution network overhead line state evaluation method based on the convolutional neural network as claimed in claim 1, characterized in that: in step S3, the overhead line has 7 types of faults, including corrosion, shedding, foreign matter, dirt, discoloration, hidden danger, and loss.
6. The power distribution network overhead line state evaluation method based on the convolutional neural network as claimed in claim 1, characterized in that: in step S3, the process of making a chromatographic analysis decision includes the following steps:
s31, according to the constructed hierarchical index system, enabling experts to compare the importance degree of each index of the same layer pairwise, and establishing a judgment matrix for each layer of index;
s32, solving the eigenvectors and the maximum eigenvalues of the judgment matrixes;
s33 maximum characteristic root using judgment matrix
Figure 879421DEST_PATH_IMAGE001
Carrying out consistency check;
s34, normalizing the feature vectors passing the test to obtain the weight value of the level;
and S35, comprehensively evaluating the state and health degree of the overhead line according to the calculated weight value and by combining with the pre-specified overhead line evaluation grade standard.
7. The power distribution network overhead line state evaluation method based on the convolutional neural network as claimed in claim 6, characterized in that: the judgment matrix equation of the analytic hierarchy process is as follows:
Figure 102592DEST_PATH_IMAGE002
in the formula: m is the number of image type types, i is the serial number of the image type corresponding to the real-time image; n is the number of the fault types of the overhead lines, and j is the fault type serial number of the overhead lines corresponding to the real-time images.
CN202210376163.4A 2022-04-11 2022-04-11 Power distribution network overhead line state evaluation method based on convolutional neural network Withdrawn CN114662976A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210376163.4A CN114662976A (en) 2022-04-11 2022-04-11 Power distribution network overhead line state evaluation method based on convolutional neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210376163.4A CN114662976A (en) 2022-04-11 2022-04-11 Power distribution network overhead line state evaluation method based on convolutional neural network

Publications (1)

Publication Number Publication Date
CN114662976A true CN114662976A (en) 2022-06-24

Family

ID=82035130

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210376163.4A Withdrawn CN114662976A (en) 2022-04-11 2022-04-11 Power distribution network overhead line state evaluation method based on convolutional neural network

Country Status (1)

Country Link
CN (1) CN114662976A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117421998A (en) * 2023-12-18 2024-01-19 国网湖北省电力有限公司经济技术研究院 Multi-mode data-based power transmission overhead line health state evaluation system
CN117783769A (en) * 2024-02-28 2024-03-29 国网山西省电力公司太原供电公司 Power distribution network fault positioning method, system, equipment and storage medium based on visual platform

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117421998A (en) * 2023-12-18 2024-01-19 国网湖北省电力有限公司经济技术研究院 Multi-mode data-based power transmission overhead line health state evaluation system
CN117421998B (en) * 2023-12-18 2024-03-12 国网湖北省电力有限公司经济技术研究院 Multi-mode data-based power transmission overhead line health state evaluation system
CN117783769A (en) * 2024-02-28 2024-03-29 国网山西省电力公司太原供电公司 Power distribution network fault positioning method, system, equipment and storage medium based on visual platform
CN117783769B (en) * 2024-02-28 2024-05-10 国网山西省电力公司太原供电公司 Power distribution network fault positioning method, system, equipment and storage medium based on visual platform

Similar Documents

Publication Publication Date Title
CN106595551B (en) Ice covering thickness detection method in powerline ice-covering image based on deep learning
CN114662976A (en) Power distribution network overhead line state evaluation method based on convolutional neural network
CN107728234B (en) Thunder and lightning strength value prediction method based on atmospheric electric field data
CN113723669B (en) Power transmission line icing prediction method based on Informmer model
CN111008641B (en) Power transmission line tower external force damage detection method based on convolutional neural network
CN112288711A (en) Unmanned aerial vehicle inspection image defect image identification method, device, equipment and medium
CN111178585A (en) Fault reporting amount prediction method based on multi-algorithm model fusion
CN114821155A (en) Multi-label classification method and system based on deformable NTS-NET neural network
CN112836937A (en) Flood disaster loss evaluation method based on entropy weight and BP neural network technology
CN114881286A (en) Short-time rainfall prediction method based on deep learning
CN110503135A (en) Deep learning model compression method and system for the identification of power equipment edge side
Song et al. Deformable YOLOX: Detection and rust warning method of transmission line connection fittings based on image processing technology
CN111143835B (en) Non-invasive protection method for business logic of electric power metering system based on machine learning
CN114444757A (en) Combined prediction method for plateau mountain multi-model multi-scale new energy power station output
CN116206158A (en) Scene image classification method and system based on double hypergraph neural network
CN115641498A (en) Medium-term rainfall forecast post-processing correction method based on space multi-scale convolutional neural network
CN116070902A (en) Power transmission line state evaluation method, device, computer equipment, medium and product
CN111507393A (en) Display card interface machine testing method based on Laplace feature mapping learning
CN115017828A (en) Power cable fault identification method and system based on bidirectional long-short-time memory network
CN115236451A (en) Power distribution network overhead line fault monitoring device and method
CN118298149A (en) Target detection method for parts on power transmission line
CN116485802B (en) Insulator flashover defect detection method, device, equipment and storage medium
CN117113243B (en) Photovoltaic equipment abnormality detection method
CN116738355A (en) Electric power multi-mode feature level fusion method considering data difference
CN117592619B (en) GNN-LSTM-based county winter wheat estimated yield analysis method and system

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
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20220624