CN106599777A - Cable partial discharge signal identification method based on energy percentage - Google Patents

Cable partial discharge signal identification method based on energy percentage Download PDF

Info

Publication number
CN106599777A
CN106599777A CN201610943645.8A CN201610943645A CN106599777A CN 106599777 A CN106599777 A CN 106599777A CN 201610943645 A CN201610943645 A CN 201610943645A CN 106599777 A CN106599777 A CN 106599777A
Authority
CN
China
Prior art keywords
discharge signal
local discharge
decomposition
neural network
detail coefficients
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201610943645.8A
Other languages
Chinese (zh)
Inventor
许佳
牛海清
郑文坚
吴炬卓
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.)
South China University of Technology SCUT
Original Assignee
South China University of Technology SCUT
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 South China University of Technology SCUT filed Critical South China University of Technology SCUT
Priority to CN201610943645.8A priority Critical patent/CN106599777A/en
Publication of CN106599777A publication Critical patent/CN106599777A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • G06F2218/06—Denoising by applying a scale-space analysis, e.g. using wavelet analysis
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
    • G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
    • G01R31/1263—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation
    • G01R31/1272—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation of cable, line or wire insulation, e.g. using partial discharge measurements
    • 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
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Data Mining & Analysis (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Signal Processing (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Testing Relating To Insulation (AREA)

Abstract

The present invention discloses a cable partial discharge signal identification method based on an energy percentage. The method comprises the following steps: obtaining a plurality of partial discharge signals with known sources; performing wavelet decomposition of each partial discharge signals with the known sources, obtaining the approximation coefficients of the highest decomposition scale and the detail coefficients of each decomposition scale, calculating the energy percentage of the approximation coefficients of the highest decomposition scale and the detail coefficients of each decomposition scale, and establishing a partial discharge signal feature sample database; taking the energy percentage of the approximation coefficients of the highest decomposition scale and the detail coefficients of each decomposition scale as the input, and constructing a BP neural network with the preset number of layers; taking the partial discharge signals with the known sources as samples, performing training of the BP neural network, and obtaining a trained BP neural network; and inputting the partial discharge signals to be identified into the trained BP neural network to rapidly detect the sources. The detection steps are simple, the detection speed is fast, and the detection precision is high.

Description

A kind of cable local discharge signal recognition method based on energy percentage
Technical field
The present invention relates to cable local discharge on-line monitoring technique field, and in particular to a kind of electricity based on energy percentage Cable local discharge signal recognition methods.
Background technology
In cable local discharge on-line monitoring, the local discharge signal for detecting may be from cable body and cable termination Head, it is also possible to from coupled switch cubicle.Because the shelf depreciation of separate sources is to equipment harm difference, criterion Difference, so being identified important realistic meaning to local discharge signal source.
In terms of local discharge signal identification, signal characteristic abstraction and grader select to be most critical part.Feature extraction It is the local discharge signal identification first step, the quality of feature extraction directly influences the effect of identification.At present, local discharge signal Feature extracting method mainly has statistical nature method and the big class of temporal analysis two.Wherein statistical nature method is directed to shelf depreciation The phase place of signal, and distribution cable is generally three-core cable and totally one ground wire, when there is shelf depreciation in two-phase or three-phase, detection The phase property of local discharge signal becomes hardly possible.Temporal analysis be for high speed acquisition once discharge generation when Wave character or corresponding transformation results obtained by the pulse of domain carries out pattern-recognition, mainly including Fourier analysis method, small echo Analytic approach and waveform parameter direct extraction method etc..Pattern recognition classifier device mainly has neural network classifier, minimum distance classification Device and fuzzy diagnosis grader.Wherein neural network classifier is mutual by the fairly simple neuron of substantial amounts of function and form The complex networks system for connecting and constituting, network can be regarded as from a Nonlinear Mapping for being input to output.It is used as one Plant successful mode identification technology and apply to many fields.
The content of the invention
The invention aims to solve drawbacks described above of the prior art, disclose a kind of based on energy percentage BP neural network distribution cable local discharge signal recognition methods, the method recognition speed is fast, and accuracy of identification is high.
The purpose of the present invention can be reached by adopting the following technical scheme that：
A kind of cable local discharge signal recognition method based on energy percentage, the recognition methods includes following step Suddenly：
Obtain the several local discharge signal in known source；
Wavelet decomposition is carried out to each local discharge signal in known source, obtain highest decomposition yardstick approximation coefficient and The detail coefficients of each decomposition scale, and calculate the approximation coefficient of highest decomposition yardstick and the detail coefficients of each decomposition scale Energy percentage, sets up local discharge signal feature samples storehouse；
With the energy percentage of the approximation coefficient of highest decomposition yardstick and the detail coefficients of each decomposition scale as input, structure Build the BP neural network of the default number of plies；
Using the local discharge signal in known source as sample, BP neural network is trained, obtains the BP for training Neutral net；
Local discharge signal to be identified is input to the BP neural network for training, shelf depreciation letter to be identified is obtained Number source.
Further, it is described that wavelet decomposition is carried out to each local discharge signal, obtain the approximate system of highest decomposition yardstick The detail coefficients of number and each decomposition scale, and calculate the approximation coefficient of highest decomposition yardstick and the details system of each decomposition scale Several energy percentages, specifically includes the step of set up local discharge signal feature samples storehouse：
Wavelet decomposition is carried out to each local discharge signal, Decomposition order is 4, approximate on highest decomposition yardstick 4 Detail coefficients on coefficient and decomposition scale 1,2,3,4；
Calculate the energy percentage E of the approximation coefficient on highest decomposition yardstick 4aFor：
Wherein, a4,kFor the approximation coefficient of highest decomposition yardstick, dj,kFor the detail coefficients of jth decomposition scale, j is little wavelength-division The yardstick (j=1 ... 4) of solution, k is the approximation coefficient of jth decomposition scale or the length of detail coefficients；
Calculate the energy percentage E of the detail coefficients of each decomposition scalejFor：
Wherein, a4,kFor the approximation coefficient of highest decomposition yardstick, dj,kFor the detail coefficients of jth decomposition scale, j is to decompose chi Degree, k is the approximation coefficient of jth decomposition scale or the length of detail coefficients；
The energy percentage for extracting the approximation coefficient on highest decomposition yardstick 4 and the detail coefficients of each decomposition scale is constituted Local discharge signal characteristic vector λ：
λ=[Ea,E1,E2,E3,E4]；
When local discharge signal feature samples storehouse is set up, using two bits the difference of local discharge signal is marked Source.
Further, the local discharge signal in the known source includes：Cable body local discharge signal, cable termination The surface-discharge signal of head local discharge signal, the corona discharge signal of switch cubicle and switch cubicle.
Further, it is described to adopt the separate sources of two bits mark local discharge signal to be specially：
Corona in the cable local discharge signal, the cable terminal local discharge signal, the switch cubicle is put Surface-discharge signal in electric signal and the switch cubicle corresponds to respectively 00,01,10,11.
Further, the energy hundred of the detail coefficients of the approximation coefficient with highest decomposition yardstick and each decomposition scale The step of dividing than to be input into, building the BP neural network for presetting the number of plies is specially：
Four layers of BP neural network are set, and the neutral net includes an input layer, two hidden layers and an output layer, Wherein X for network input vector, Y for network output vector, W1For input layer and the weight matrix of first hidden layer, W2 For first hidden layer and the weight matrix of second hidden layer, W3For second hidden layer and the weight matrix of output layer, b1For The threshold vector of first hidden layer, b2For the threshold vector of second hidden layer, b3For the threshold vector of output layer, hidden layer Activation primitive adopt sigmoid functions f (x), the activation primitive of output layer to adopt purelin linear functions g (x), network Output vector Y be：
Further, the use sample is trained to BP neural network, obtains the step of BP neural network for training It is rapid to be specially：
Each certain amount of selecting at random is input to and sets as training sample from the several discharge signal sample Neutral net in, carry out network training using BP algorithm.
Further, it is described that local discharge signal to be identified is input to the BP neural network for training, obtain waiting to know The step of source of other local discharge signal, is specially：
According to the output valve of the BP neural network, contrasted with the category label, obtained shelf depreciation to be identified The source of signal：
When the output valve is between 0 to 1, with the category label pair after the output valve is rounded up Than.
The present invention has the following advantages and effect relative to prior art：
A kind of BP neural network distribution cable local discharge signal identification side based on energy percentage disclosed by the invention Method, wavelet decomposition is carried out to local discharge signal, extracts the energy percentage feature of waveform, and sets up Sample Storehouse；Using training Sample is trained to neutral net, obtains the neutral net for training；It is input into after the local discharge signal in source to be identified fast Speed detects its source, and simple with detecting step, detection speed is fast, the features such as accuracy of detection is high.
Description of the drawings
Fig. 1 is that a kind of BP neural network distribution cable local discharge signal based on energy percentage disclosed by the invention is known The schematic flow sheet of other method；
Fig. 2 (a) is cable body local discharge signal oscillogram；
Fig. 2 (b) is cable termination head discharge signal oscillogram；
Fig. 2 (c) is corona local discharge signal oscillogram；
Fig. 2 (d) is cable body local discharge signal oscillogram；
Fig. 3 is the four layers of BP neural network topology diagram set up.
Specific embodiment
To make purpose, technical scheme and the advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is The a part of embodiment of the present invention, rather than the embodiment of whole.Based on the embodiment in the present invention, those of ordinary skill in the art The every other embodiment obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.
As shown in figure 1, being that a kind of BP neural network distribution cable local based on energy percentage that the present invention is provided is put The schematic flow sheet of electric signal recognition methods, comprises the following steps：
S1, the local discharge signal for obtaining known source；
In this example, it is known that the local discharge signal in source includes cable body local discharge signal, cable end The surface-discharge signal in corona discharge signal and switch cubicle in termination local discharge signal, switch cubicle.
Oscillogram as shown in Fig. 2 (a) to Fig. 2 (d), the sample frequency of waveform is 100MHz, and the time domain of each waveform is long Spend for 1500 sampled points.
S12, wavelet decomposition is carried out to each local discharge signal, obtain the approximation coefficient and each point of highest decomposition yardstick The detail coefficients of solution yardstick, and calculate the energy hundred of the approximation coefficient of highest decomposition yardstick and the detail coefficients of each decomposition scale Divide ratio, set up local discharge signal feature samples storehouse；
Specifically, it may include following steps：
Wavelet decomposition is carried out to each local discharge signal, Decomposition order is 4, obtain approximate on highest decomposition yardstick 4 Coefficient and decomposition scale 1, the detail coefficients on 2,3,4.
Calculate the energy percentage E of the approximation coefficient on highest decomposition yardstick 4aFormula is：
Wherein, aJ,kFor the approximation coefficient of highest decomposition yardstick, dj,kFor the detail coefficients of jth decomposition scale, k is that jth is divided The approximation coefficient of solution yardstick or the length of detail coefficients.
Calculate the energy percentage E of the detail coefficients on decomposition scale 1,2,3,4jFormula is：
Wherein, a4,kFor the approximation coefficient of highest decomposition yardstick, dj,kFor the detail coefficients of jth decomposition scale, k is that jth is divided The approximation coefficient of solution yardstick or the length of detail coefficients.
The energy percentage for extracting the approximation coefficient on highest decomposition yardstick 4 and the detail coefficients of each decomposition scale is constituted Local discharge signal characteristic vector λ is：
λ=[Ea,E1,E2,E3,E4]。
In a preferred embodiment, also including step：The separate sources of local discharge signal is marked using binary number；Such as The table in corona discharge signal and switch cubicle in cable local discharge signal, cable terminal local discharge signal, switch cubicle Face discharge signal can respectively correspond to 00,01,10,11.
S13, with the energy percentage of the approximation coefficient of highest decomposition yardstick and the detail coefficients of each decomposition scale as defeated Enter, build the BP neural network of the default number of plies；
The BP neural network that this example is used is 4 layer models, as shown in Figure 3.
In a preferred embodiment, the step of BP neural network of the structure default number of plies may include：
4 layers of BP neural network are set, and the neutral net includes an input layer, two hidden layers and an output layer； Wherein X for network input vector, Y for network output vector, W1For input layer and the weight matrix of hidden layer, W2For hidden layer and The weight matrix of hidden layer, W3For hidden layer and the weight matrix of output layer, b1For the threshold vector of first hidden layer, b2For second The threshold vector of individual hidden layer, b3For the threshold vector of output layer, the activation primitive of hidden layer adopts sigmoid functions f (x), The activation primitive of output layer adopts purelin linear functions g (x), network to export Y and be：
S14, BP neural network is trained using sample, obtains the BP neural network for training；
In this example, from the several local discharge signal sample it is each select at random certain amount as training sample This, in being input to the neutral net for setting, using BP algorithm network training is carried out, and obtains the neutral net for training.
S15, local discharge signal to be identified is input to the BP neural network for training, obtains local to be identified and put The source of electric signal.
In a preferred embodiment, local discharge signal to be identified is input to the BP neural network for training, is obtained The step of source of local discharge signal to be identified is
According to described BP neural network output valve, contrasted with the category label, obtained shelf depreciation to be identified The source of signal：
When the output valve is between 0 to 1, with the category label pair after the output valve is rounded up Than.
In this example, BP neural network is trained using training sample, error precision is set to 0.001, learning efficiency 0.1 is set to, iterations is set to 1000, and recognition effect is as shown in table 1,.
The neural network recognization effect of table 1.
A kind of BP neural network distribution cable local discharge signal identification side based on energy percentage that the present invention is provided Method, wavelet decomposition is carried out to local discharge signal, extracts the energy percentage feature of waveform, and sets up Sample Storehouse；Using training Sample is trained to neutral net, obtains the neutral net for training；It is input into after the local discharge signal in source to be identified fast Speed detects its source, and simple with detecting step, detection speed is fast, the features such as accuracy of detection is high.
Above-described embodiment is the present invention preferably embodiment, but embodiments of the present invention not by above-described embodiment Limit, other any Spirit Essences without departing from the present invention and the change, modification, replacement made under principle, combine, simplification, Equivalent substitute mode is should be, is included within protection scope of the present invention.

Claims (7)

1. a kind of cable local discharge signal recognition method based on energy percentage, it is characterised in that the recognition methods bag Include following steps：
Obtain the several local discharge signal in known source；
Wavelet decomposition is carried out to each local discharge signal in known source, obtain highest decomposition yardstick approximation coefficient and each The detail coefficients of decomposition scale, and calculate the energy of the approximation coefficient of highest decomposition yardstick and the detail coefficients of each decomposition scale Percentage, sets up local discharge signal feature samples storehouse；
With the energy percentage of the approximation coefficient of highest decomposition yardstick and the detail coefficients of each decomposition scale as input, build pre- If the BP neural network of the number of plies；
Using the local discharge signal in known source as sample, BP neural network is trained, the BP for obtaining training is neural Network；
Local discharge signal to be identified is input to the BP neural network for training, local discharge signal to be identified is obtained Source.
2. a kind of cable local discharge recognition methods based on energy percentage according to claim 1, it is characterised in that It is described that wavelet decomposition is carried out to each local discharge signal, obtain the approximation coefficient and each decomposition scale of highest decomposition yardstick Detail coefficients, and the energy percentage of the approximation coefficient of highest decomposition yardstick and the detail coefficients of each decomposition scale is calculated, build The step of vertical local discharge signal feature samples storehouse, specifically includes：
Wavelet decomposition is carried out to each local discharge signal, Decomposition order is 4, obtains the approximation coefficient on highest decomposition yardstick 4 With the detail coefficients on decomposition scale 1,2,3,4；
Calculate the energy percentage E of the approximation coefficient on highest decomposition yardstick 4aFor：
E a = Σ k a 4 , k 2 Σ k a 4 , k 2 + Σ j Σ k d j , k 2 , j = 1 , 2 , ... , 4
Wherein, a4,kFor the approximation coefficient of highest decomposition yardstick, dj,kFor the detail coefficients of jth decomposition scale, j is wavelet decomposition Yardstick (j=1 ... 4), k is the approximation coefficient of jth decomposition scale or the length of detail coefficients；
Calculate the energy percentage E of the detail coefficients of each decomposition scalejFor：
E j = Σ k d j , k 2 Σ k a 4 , k 2 + Σ j Σ k d j , k 2 , j = 1 , 2 , ... , 4
Wherein, a4,kFor the approximation coefficient of highest decomposition yardstick, dj,kFor the detail coefficients of jth decomposition scale, j is decomposition scale, k The length of approximation coefficient or detail coefficients for jth decomposition scale；
The energy percentage for extracting the approximation coefficient on highest decomposition yardstick 4 and the detail coefficients of each decomposition scale constitutes local Discharge signal characteristic vector λ：
λ=[Ea,E1,E2,E3,E4]；
When local discharge signal feature samples storehouse is set up, using two bits mark local discharge signal difference come Source.
3. a kind of cable local discharge recognition methods based on energy percentage according to claim 1 and 2, its feature exists In the several local discharge signal in the known source includes：Cable body local discharge signal, cable terminal local are put The surface-discharge signal of electric signal, the corona discharge signal of switch cubicle and switch cubicle.
4. a kind of cable local discharge recognition methods based on energy percentage according to claim 3, it is characterised in that
It is described to adopt the separate sources of two bits mark local discharge signal to be specially：
Corona discharge letter in the cable local discharge signal, the cable terminal local discharge signal, the switch cubicle Number and the switch cubicle in surface-discharge signal respectively correspond to 00,01,10,11.
5. a kind of cable local discharge recognition methods based on energy percentage according to claim 1, it is characterised in that The energy percentage of the detail coefficients of the approximation coefficient with highest decomposition yardstick and each decomposition scale is input, builds pre- If the step of BP neural network of the number of plies, is specially：
Four layers of BP neural network are set, and the neutral net includes an input layer, two hidden layers and an output layer, wherein X for network input vector, Y for network output vector, W1For input layer and the weight matrix of first hidden layer, W2For The weight matrix of one hidden layer and second hidden layer, W3For second hidden layer and the weight matrix of output layer, b1For first The threshold vector of individual hidden layer, b2For the threshold vector of second hidden layer, b3For the threshold vector of output layer, hidden layer swashs Function living adopts sigmoid functions f (x), and the activation primitive of output layer adopts purelin linear functions g (x), network it is defeated Outgoing vector Y is：
Y=g (W3 Tf(W2 Tf(W1 T+b1)+b2)+b3)。
6. a kind of cable local discharge recognition methods based on energy percentage according to claim 1, it is characterised in that The use sample is trained to BP neural network, is specially the step of obtain the BP neural network for training：
Each certain amount of selecting at random is input to the god for setting as training sample from the several discharge signal sample In Jing networks, using BP algorithm network training is carried out.
7. a kind of cable local discharge recognition methods based on energy percentage according to claim 1, it is characterised in that It is described that local discharge signal to be identified is input to the BP neural network for training, obtain local discharge signal to be identified The step of source, is specially：
According to the output valve of the BP neural network, contrasted with the category label, obtained local discharge signal to be identified Source：
When the output valve is between 0 to 1, contrast with the category label after the output valve is rounded up.
CN201610943645.8A 2016-11-02 2016-11-02 Cable partial discharge signal identification method based on energy percentage Pending CN106599777A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610943645.8A CN106599777A (en) 2016-11-02 2016-11-02 Cable partial discharge signal identification method based on energy percentage

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610943645.8A CN106599777A (en) 2016-11-02 2016-11-02 Cable partial discharge signal identification method based on energy percentage

Publications (1)

Publication Number Publication Date
CN106599777A true CN106599777A (en) 2017-04-26

Family

ID=58590352

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610943645.8A Pending CN106599777A (en) 2016-11-02 2016-11-02 Cable partial discharge signal identification method based on energy percentage

Country Status (1)

Country Link
CN (1) CN106599777A (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107271868A (en) * 2017-06-29 2017-10-20 国家电网公司 A kind of shelf depreciation time-delay calculation error compensating method based on multiple neural network
CN107703418A (en) * 2017-08-30 2018-02-16 上海交通大学 Shelf depreciation location error compensation method based on more radial base neural nets
CN107862320A (en) * 2017-11-28 2018-03-30 广东电网有限责任公司珠海供电局 A kind of porcelain shell for cable terminal Infrared Image Features vector extracting method
CN109324274A (en) * 2018-11-29 2019-02-12 广东电网有限责任公司 A kind of local discharge signal wavelet decomposition optimal base wavelet choosing method
CN109387757A (en) * 2018-12-14 2019-02-26 广东电网有限责任公司 A kind of local discharge signal characteristic vector pickup method
CN109870654A (en) * 2019-02-02 2019-06-11 福州大学 The online method for dynamic estimation of accumulator capacity based on impact load response characteristic
CN110646708A (en) * 2019-09-27 2020-01-03 中国矿业大学 10kV single-core cable early state identification method based on double-layer long-and-short-term memory network
CN112001246A (en) * 2020-07-20 2020-11-27 中国南方电网有限责任公司超高压输电公司广州局 Partial discharge type identification method and device based on singular value decomposition

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
吴炬卓: ""中压电缆局部放电带电检测的白噪声抑制和放电类型识别方法研究"", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 *
吴炬卓等: ""基于逐层最优基小波和贝叶斯估计的电缆瓷套终端红外图像自适应去噪方法"", 《电测与仪表》 *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107271868A (en) * 2017-06-29 2017-10-20 国家电网公司 A kind of shelf depreciation time-delay calculation error compensating method based on multiple neural network
CN107703418A (en) * 2017-08-30 2018-02-16 上海交通大学 Shelf depreciation location error compensation method based on more radial base neural nets
CN107703418B (en) * 2017-08-30 2019-10-18 上海交通大学 Shelf depreciation location error compensation method based on more radial base neural nets
CN107862320A (en) * 2017-11-28 2018-03-30 广东电网有限责任公司珠海供电局 A kind of porcelain shell for cable terminal Infrared Image Features vector extracting method
CN109324274A (en) * 2018-11-29 2019-02-12 广东电网有限责任公司 A kind of local discharge signal wavelet decomposition optimal base wavelet choosing method
CN109387757A (en) * 2018-12-14 2019-02-26 广东电网有限责任公司 A kind of local discharge signal characteristic vector pickup method
CN109387757B (en) * 2018-12-14 2020-12-04 广东电网有限责任公司 Partial discharge signal feature vector extraction method
CN109870654A (en) * 2019-02-02 2019-06-11 福州大学 The online method for dynamic estimation of accumulator capacity based on impact load response characteristic
CN110646708A (en) * 2019-09-27 2020-01-03 中国矿业大学 10kV single-core cable early state identification method based on double-layer long-and-short-term memory network
CN110646708B (en) * 2019-09-27 2020-07-17 中国矿业大学 10kV single-core cable early state identification method based on double-layer long-and-short-term memory network
CN112001246A (en) * 2020-07-20 2020-11-27 中国南方电网有限责任公司超高压输电公司广州局 Partial discharge type identification method and device based on singular value decomposition

Similar Documents

Publication Publication Date Title
CN106599777A (en) Cable partial discharge signal identification method based on energy percentage
Ekici et al. Energy and entropy-based feature extraction for locating fault on transmission lines by using neural network and wavelet packet decomposition
CN103728535B (en) A kind of extra-high-voltage direct-current transmission line fault location based on wavelet transformation transient state energy spectrum
CN103675616B (en) The online partial discharge detection signal recognition method of cable
CN106680576B (en) Power distribution network internal overvoltage recognition methods based on piecemeal time-frequency spectrum and deep learning algorithm
CN106405339A (en) Power transmission line fault reason identification method based on high and low frequency wavelet feature association
CN106203531A (en) A kind of transmission line fault sorting technique based on the sparse autocoder of convolution
CN102279358B (en) MCSKPCA based neural network fault diagnosis method for analog circuits
CN104808107A (en) XLPE cable partial discharge defect type identification method
CN105223466B (en) It is a kind of using modulus maximum than extra high voltage direct current transmission line method of single end distance measurement
CN111239549A (en) Power distribution fault rapid positioning method based on discrete wavelet transform
CN103995237A (en) Satellite power supply system online fault diagnosis method
CN103439647A (en) Analog circuit fault diagnosis method for dynamic supply current monitoring
CN111308260B (en) Electric energy quality monitoring and electric appliance fault analysis system based on wavelet neural network and working method thereof
CN103064008A (en) Nolinear analog circuit soft fault diagnostic method based on Hilbert-huang transform
CN109783919A (en) Frequency domain multi-conductor transmission lines electromagnetic pulse response quickly modeling method based on waveform relaxation iteration
CN113985194A (en) Power distribution network fault positioning method based on stack self-encoder
CN106443380B (en) A kind of distribution cable local discharge signal recognition methods and device
CN106548013A (en) Using the voltage sag source identification method for improving incomplete S-transformation
CN106124830A (en) The method obtaining lightning current based on artificial neural network
CN107832729A (en) A kind of bearing rust intelligent diagnosing method
CN114545292A (en) Cable fault type identification method based on deep learning
CN113514743A (en) Construction method of GIS partial discharge pattern recognition system based on multi-dimensional features
CN112381667A (en) Distribution network electrical topology identification method based on deep learning
CN112946425A (en) Fault positioning method for mining travelling wave time-frequency domain characteristics by utilizing deep learning

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
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20170426

WD01 Invention patent application deemed withdrawn after publication