CN105425076A - Method of carrying out transformer fault identification based on BP neural network algorithm - Google Patents
Method of carrying out transformer fault identification based on BP neural network algorithm Download PDFInfo
- Publication number
- CN105425076A CN105425076A CN201510918540.2A CN201510918540A CN105425076A CN 105425076 A CN105425076 A CN 105425076A CN 201510918540 A CN201510918540 A CN 201510918540A CN 105425076 A CN105425076 A CN 105425076A
- Authority
- CN
- China
- Prior art keywords
- discharge
- fault
- transformer
- carrying
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- 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
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
- G01R31/62—Testing of transformers
-
- 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
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention provides a method of carrying out transformer fault identification based on a BP neural network algorithm. The method comprises the following steps of step S1, through a partial discharge test system, collecting discharge pulse graphs of different transformer faults; step S2, carrying out power graph analysis on discharge pulses acquired through the step S1; step S3, extracting a training sample and a test sample from characteristic quantities acquired through the power graph analysis obtained from the step S2; step S4, constructing a BP network nerve; step S5, carrying out BP network nerve training; step S6, carrying out BP network nerve testing. By using the method of carrying out transformer fault identification based on the BP neural network algorithm, a fault type of a transformer can be accurately identified. And the method plays an important role in transformer fault diagnosis and a state assessment and the method is convenient.
Description
Technical field
The present invention relates to field of neural networks, relate more particularly to one and carry out transformer fault knowledge method for distinguishing based on BP neural network algorithm.
Background technology
Whether the running status of transformer concerns the conveying of whole electric energy and is reliably carried out, and because some accidental or non-accidental cause transformers there will be various fault in practical operation, and what wherein the most easily occur is exactly the insulation fault of inside transformer.The main cause of fault causes due to insulation ag(e)ing or reduction, and partial discharge phenomenon is the important symbol that transformer insulated level reduces, therefore, the result that shelf depreciation is tested is imported in analogue system electric discharge type in addition Classification and Identification, just can judge the insulation fault that inside transformer is potential accurately and rapidly, in time the fault being about to occur is got rid of, guarantee that transformer can continue reliably to run the normal operation ensureing the quality of power supply and whole electrical network with this.
In general, pass judgment on running state of transformer method to mainly contain: dissolved gas analysis method, thermometry, winding D.C. resistance mensuration, absorptance mensuration, dielectric loss measurement method and shelf depreciation etc.Have superiority, therefore usually used as one of the important detection method of power equipment in the state of insulation of partial discharge test in monitoring transformer.
Dissolved gas analysis method: transformer oil is inevitably exposed in the middle of air, the moisture that the existence in air is a large amount of and impurity will be partially dissolved in transformer oil.So transformer insulated fault just can by the content of analyzing each gas in oil with become to assign to carry out analysis and determine.This process need use gas chromatographicanalyzer device, just can determine whether transformer exists exception by observing the type of each gas and content, if having abnormal is belong to which kind of fault type, already present fault degree how, and future developing trend how.
Temperature monitoring method: on the one hand can mounting temperature sensor temperature sensor change on the wire near Transformer Winding, just can determine the hot spot of transformer by the changing value of observation and analysis temperature and process in time, but the method applies in reality exists cost intensive, the problem of technical sophistication.Can utilize the theory indirect inspection temperature conditions of heat trnasfer on the other hand, although this method does not have direct method definitely accurate, it needs the data of collection less, and process is simple, can widely use.
Winding D.C. resistance mensuration: the test of Transformer Winding direct current resistance can detect transformer and whether occurs winding interturn short-circuit, and whether shunting switch contacts well, and go between and whether rupture, tap changer is adjusted a wage scale the whether problem such as correct.The measurement of winding D.C. resistance has obviously advantage all the time on detection current return connectivity problem.
Absorptance mensuration: the method be based upon absorbing phenomenon basis on measure insulation resistance resistance over time with the insulation situation of this rapids judgement transformer.Insulation resistance R when usually measuring pressurization 60 seconds with megger
60resistance with measure 15 seconds time insulation resistance R
15the ratio of resistance, as absorptance, weighs humidified insulation situation, and it can reflect local defect, also can reflect general defect.
Dielectric loss angle: dielectric loss angle tangent (dielectric loss angle) is one of the insulation status judging transformer index clearly; The change of dielectric loss angle can humidified insulation, and the defects such as insulation ag(e)ing and insulating inner gas discharge, are particularly rich in advantage in humidified insulation, aging distributed defect.So measuring insulative dielectric loss angle concerning detection running state of transformer is a very important test item.
Shelf depreciation method: partial discharge of transformer test voltage is exactly the voltage applying to a certain degree certain hour on test specimen, make in test specimen insulation vulnerable area generation electric discharge phenomena, used partial discharge test can observe the parameters such as the starting potential of this process, extinction voltage and discharge capacity with the dielectric level of this rapids measurement test specimen; In test process on the basis avoiding core sataration, supply frequency should be reduced to reduce the capacity of compensating inductance as far as possible.
In recent years, neural network is widely used in every field such as productive lifes, and it is set up and logic supposition for realizing the complex data model be input between output as a kind of operational method.Artificial neural network has the essential characteristic of biological nervous system, has distributed treatment, large-scale parallel, self-organization, the features such as self study, can realize Nonlinear Mapping and be input to the well approximate of output; Wherein, BP (BackPropagation, multilayer feedforward) neural network has obvious advantage in pattern-recognition, therefore becomes one of the most widely used form in neural network.Nowadays, the research of neural network has reached certain degree of ripeness, based on the range of application of neural network also in continuous expansion, has obtained great achievement much obvious to all.Below some main application fields: (1) Pattern recognition and image processing: product hierarchy classification, fingerprint recognition, diseases analysis etc.; (2) control and optimize: the control of atmosphere quality detecting appraisal, semiconductor production process, the control of high voltage power transmission circulation etc.; (3) forecast and intelligent information management: electrical network short-term load forecasting, earthquake prediction, the management of intelligent grid household electrical appliances and traffic administration etc.Application of Neural Network will be possessed the more excellent effect of other Diagnosis Method of Transformer Faults in transformer fault diagnosis aspect.
Summary of the invention
The invention provides one and carry out transformer fault knowledge method for distinguishing based on BP neural network algorithm, it is characterized in that, comprise the steps:
Step S1: collected by the discharge pulse collection of illustrative plates of partial discharge test system to different transformer fault;
Step S2: power diagram analysis of spectrum is carried out to the discharge pulse obtained by step S1;
Step S3: extract training sample and test sample book from the characteristic quantity that the power diagram analysis of spectrum obtained by step S2 is obtained;
Step S4: build BP network neural;
Step S5: carry out the training of BP network neural;
Step S6: carry out the test of BP network neural.
Preferably, described partial discharge test system comprises: display instrument, ultra-high frequency antenna, electrode, ground wire, insulator sleeve pipe, high-tension insulator sleeve pipe, fuel tank, coupling condenser, protective resistance, transformer; Wherein, sparking voltage introduces one end of electrode through protective resistance and high-tension insulator sleeve pipe; again the other end of electrode is drawn connecting insulator sleeve pipe; insulator sleeve pipe also passes through ground connection; discharge signal introduces partial discharge test system through coupling capacitance, and from display instrument, a situation arises for the discharge pulse of observable Different electrodes shape.
Preferably, described transformer fault comprises internal fault and external fault.
Preferably, described internal fault comprises winding failure, iron core fault, major insulation fault.
Preferably, described external fault comprises shunting switch fault, sleeve pipe fault.
Preferably, described discharge pulse is the discharge pulse of creeping discharge type or the discharge pulse of corona discharge type.
Preferably, when carrying out step S1, carry out emptying of environmental variance.
Preferably, environmental variance comprises: floating potential discharge interference, Electromagnetic Interference, loose contact interference.
Preferably, power diagram analysis of spectrum is carried out by direct method or indirect method.
Preferably, get horizontal ordinate [65-75] interval as characteristic quantity, to get with imresize function the matrix that 1*250 dimension is expanded as in horizontal ordinate [65-75] interval, and the 1-50 group data will be set as test sample book, 51-250 group data will be set as training sample.
Preferably, step S4 comprises the selection of hidden layer node, the selection of activation function, the selection of learning rate.
Provided by the invention based on BP neural network algorithm carry out transformer fault know method for distinguishing, the fault type of transformer can be identified accurately, the vital role on transformer fault diagnosis and state estimation, and method is convenient.
Accompanying drawing explanation
Fig. 1 is the process flow diagram carrying out transformer fault knowledge method for distinguishing based on BP neural network algorithm provided by the invention;
Fig. 2 is the classification of common transformer fault type;
Fig. 3 is the schematic diagram of partial discharge test system;
Fig. 4 is the schematic diagram of corona discharge model;
Fig. 5 is the schematic diagram of creeping discharge model;
Fig. 6 is the schematic diagram of the electric discharge that floating potential causes;
Fig. 7 is Electromagnetic Interference collection of illustrative plates;
Fig. 8 is the power collection of illustrative plates of creeping discharge model;
Fig. 9 is the power collection of illustrative plates of the model of corona discharge;
Figure 10 is BP neural network test performance curve;
Figure 11 is the final training result display of BP neural network.
Embodiment
Fig. 1 is the process flow diagram carrying out transformer fault knowledge method for distinguishing based on BP neural network algorithm provided by the invention, as shown in Figure 1, the invention provides one and carry out transformer fault knowledge method for distinguishing based on BP neural network algorithm, it is characterized in that, comprise the steps:
Step S1: collected by the discharge pulse collection of illustrative plates of partial discharge test system to different transformer fault;
Step S2: power diagram analysis of spectrum is carried out to the discharge pulse obtained by step S1;
Step S3: extract training sample and test sample book from the characteristic quantity that the power diagram analysis of spectrum obtained by step S2 is obtained;
Step S4: build BP network neural;
Step S5: carry out the training of BP network neural;
Step S6: carry out the test of BP network neural.
Also by reference to the accompanying drawings each step is described below by way of specific embodiment.
Step S1: collected by the discharge pulse collection of illustrative plates of partial discharge test system to different transformer fault.
In day-to-day operation, owing to lacking many reasons such as good management and maintenance, insulation ag(e)ing, some failure conditions can be there are, affect the normal work of transformer, even cause major accident in transformer.To judge all kinds of fault, first will have the corresponding data of all kinds of fault, therefore just need to collect fault data, analyze, to form database.
As shown in Figure 2, be the classification of common transformer fault type, divide by structure and usually can be divided into internal fault and this two class of external fault.Internal fault comprises: winding failure, iron core fault, major insulation fault; External fault comprises: shunting switch fault, sleeve pipe fault.
Winding failure: due to long-term wearing and tearing, heating power, oxidation when transformer normally runs, the dielectric level of insulating material can reduce gradually.In this time, under the impact of surge power or other mechanical impetus, will physical damage be caused, the various problem such as short circuit, cause big current, superpotential thus make transformer-supplied interrupt even producing other more serious security incidents.
Iron core fault: multipoint earthing of iron core forms loop and causes local overheating to cause discharging fault, the insulation damages of iron core, scarce sheet, fastenings loosen or the surface distress of conductive material all can cause iron core to damage the generation causing transformer fault.
Major insulation fault: transformer is oil immersed type mostly, oil inevitably contacts with air and the moisture absorbed gradually in air and impurity; During operation, along with the continuous rising of temperature, insulating oil also constantly carries out oxidation and produces various acidic oxides by destroying the insulating effect of oil, also can accelerate the corrosion of other insulating material.
Shunting switch fault: the pressure of switch spring not, loose contact, tap joint position mistake, local overheating, the short circuit etc. that over-voltage breakdown causes causes shunting switch fault.
Sleeve pipe fault: sleeve pipe fastening force not, the supercycle is run, shelf depreciation causes insulation breakdown or cover fouling seriously all can cause sleeve pipe fault.
Many electric equipment products insulating inner all inevitably also exist air gap more or less, these air gaps are likely produced in the process of production, also be likely that the inside of product own also exists a little moisture, after energising, the cracking and produce the bubble that hydrogen formed under the effect of electric field of water energising electrolysis and the air gap that formed or oil.Specific inductive capacity due to air is less than insulating material specific inductive capacity, even if insulating inner field intensity is not very large, but insulating inner air gap is less than the cause of the dielectric constant of insulating material due to its dielectric constant, the field intensity of air gap inside will become very large, causes gas gap breakdown possibly.In addition, because insulating material inside has the electric field of possibility or its inside that there is impurity even not, at this moment also likely electric discharge can be produced.The reason of insulator arrangement local ageing is generally all caused by shelf depreciation, and its most serious impact shows to be accelerated insulation ag(e)ing rate or destroy the stable running status of device security.
The present invention collects the dissimilar discharge pulse of transformer by partial discharge test system, building database; Fig. 3 is the schematic diagram of partial discharge test system, comprising: display instrument 1, ultra-high frequency antenna 2, electrode 3, ground wire 4, insulator sleeve pipe 5, high-tension insulator sleeve pipe 6, fuel tank 7, coupling condenser 8, protective resistance 9, transformer 10.Sparking voltage introduces one end of electrode 3 through protective resistance 9 and high-tension insulator sleeve pipe 6, then the other end of electrode 3 is drawn connecting insulator sleeve pipe 5, and insulator sleeve pipe 5 is also by ground wire 4 ground connection.Discharge signal introduces partial discharge test system through coupling capacitance 8, and from display instrument 1, a situation arises for the discharge pulse of observable Different electrodes shape.
Partial discharge test system of the present invention can change electrode shape and simulate different transformer insulated fault, be illustrated with coronal model and creeping discharge model two type below, but the discharge pulse collection of partial discharge test system of the present invention to transformer insulated fault is not limited to this two type.
Fig. 4 is the schematic diagram of corona discharge model, corona discharge model comprise ground connection disc electrode 2, be arranged on the tip 1 that insulating oil 4 on disc electrode 2 and insulating oil 4 be provided separately.Fig. 5 is the schematic diagram of creeping discharge model, and creeping discharge model comprises disc electrode 2, be arranged on the plate electrode 3 that insulating oil 4 on disc electrode 2 and insulating oil 4 be provided separately.Table 1 be corona discharge model at different arcing distance and different voltage conditions, table 2 is creeping discharge model in the discharge pulse value of different arcing distance and different voltage.
Table 1: corona discharge model is in the discharge pulse value of different arcing distance and different voltage
Table 2: creeping discharge model is in the discharge pulse value of different arcing distance and different voltage
Be more than discharge pulse collection of illustrative plates respective under the different voltage of same arcing distance, can find out the continuous rising along with trial voltage, discharge pulse number of times and scope are all in continuous reinforcement.When impressed voltage constantly increases, the quantity of discharge pulse and scope all can be on the increase and expand.
In order to reduce partial discharge test system by environment impact, improve the correctness of test data, can environmental variance be emptied, environmental variance and to empty the method for environmental variance as follows:
Environmental variance (1): floating potential discharge disturbs
Fig. 6 is the schematic diagram of the electric discharge that floating potential causes, in partial discharge test system, insulator its vicinity is provided with other power equipments multiple, a certain position in the plurality of power equipment will have accumulated a large amount of electric charge owing to not having ground connection, define a potential difference (PD) between these electric charges and the earth to discharge, cause the deviation of test data.Settling mode is by each power equipment ground connection in partial discharge test system.
Environmental variance (2): Electromagnetic Interference
Shelf depreciation system does not have gauze screen, some high-frequency signals in space all can with the form of electromagnetic induction or electromagnetic radiation through stray electrical perhaps stray inductance be coupled to partial discharge test loop.Fig. 7 is Electromagnetic Interference collection of illustrative plates.Corresponding solution: about the interference reducing mains side and spatial electromagnetic ripple, Partial discharge detector can suppress part to be disturbed by the bound of adjustment measurement frequency band, but generally, interference still cannot thoroughly be got rid of, less interference of can only doing the best.Differentiate the type of interference according to some related datas, according to the time window in adjustment instrument for measuring partial discharge, utilize the undesired signal on time window rejecting fixed position, artificial cognition goes out the pulse signal of shelf depreciation.
Environmental variance (3): loose contact is disturbed
If the unreliable connection in the contact site of line also can produce interference to test loop in test process.When contacting complete conducting, interference can be eliminated voluntarily.Corresponding solution: each connecting portion will guarantee reliably to contact.
Step S2: power diagram analysis of spectrum is carried out to the discharge pulse obtained by step S1.
Carrying out power diagram analysis of spectrum to the discharge pulse collection of illustrative plates that local discharge test obtains, is to prepare for choosing of neural metwork training characteristic parameter:
Power spectrum refers to the situation that the power of the reflected signal of signal changes along with the change of frequency, in this test, the meaning of power spectrumanalysis is that namely the characteristic quantity extracting the test of BP neural network arranges weight, extracts suitable training sample and test sample book in other words.
The most important link of power diagram analysis of spectrum is rated output density function, mainly contain this two kinds of computing method: (1) direct method: refer to and directly the finite number certificate observed is regarded a discrete series, then its discrete Fourier transformation is calculated, obtain transformation results, get again the amplitude of transformation results square, finally divided by the number of this ordered series of numbers; The result obtained is namely as the power spectrumanalysis result to these group data.(2) indirect method: refer to and the data analysis observed is arranged the autocorrelation function being generated correspondence by program editing, then carry out Fourier transform to autocorrelation function, the result after conversion both can be used as the power spectrumanalysis result to reorganization data.Adopt indirect method to carry out power spectrumanalysis herein, be characterized in that resolution is high.
The present invention is by gathering 20 groups of corona discharge models, the creeping discharge model discharge pulse collection of illustrative plates of (interval is about 0.2s) not in the same time under same voltage effect respectively, according to above-mentioned power density functions production method, obtain power collection of illustrative plates, and the notable difference part analyzing two kinds of collection of illustrative plates carries out the characteristic quantity of neural network test as representative characteristic quantity.Preferably, for realizing obvious contrast, through considering to select change matrix dimension (parameter is adjusted to 1024 from 256) with amplification data display effect.
Fig. 8 is the power collection of illustrative plates of creeping discharge model, and Fig. 9 is the power collection of illustrative plates of the model of corona discharge, can analyze both difference near horizontal ordinate 70 the most obvious, therefore, preferably, choose the matrix parameter of horizontal ordinate [65-75] as characteristic quantity from figure.
Preferably, but for improving training accuracy, need this interval to expand with the coordinate figure extracting more available points, namely choose more characteristic quantity.With imresize function, this interval is expanded as the matrix of 1*250 dimension, the characteristic quantity obtained is as table 3.
Table 3: characteristic quantity
Step S3: extract training sample and test sample book from the characteristic quantity that the power diagram analysis of spectrum obtained by step S2 is obtained.In the present embodiment, from table 3, select the 1-50 group data as test sample book, select 51-250 group data as training sample.
Step S4: build BP network neural.The structure of neural network comprises: the selection of the selection of hidden layer node, the selection of activation function, learning rate.
(1) hidden layer node number: the selection of hidden layer node is not very large on the impact of final correct recognition rata in fact, but needs to select node number in right amount yet, and node can increase operand too much and reduce training speed; Node then can affect training precision very little.
(2) selection of activation function: the selection major effect discrimination of activation function or speed of convergence.Usually select sigmoid function, although its calculated amount is comparatively large, precision is very high;
(3) selection of learning rate: the setting of learning rate is the impact on network convergence speed, although learning rate setting is too little still can make network convergence, will make convergence time long like this; Learning rate arranges and then can not realize well too greatly restraining effect, finally affects correct recognition rata.
Step S5: carry out the training of BP network neural.
Target setting exports and is divided into other 0 and 1.0 represents creeping discharge, and 1 represents corona discharge.Through arriving emulation again to the training that is initialised to of BP neural network, finally characteristic parameter correctly being identified and achieving the target of automatic classification identification discharge mode.Figure 10 is BP neural network test performance curve, as shown in the figure, blue curve represents test performance curve, black curve represents target capabilities curve, can find out, performance curve converges on aim curve, illustrates that the training of BP neural network is successful, next just can utilize this network to test.
Step S6: carry out the test of BP network neural.Next just can utilize this network to test.Figure 10 is the final training result display of BP neural network, and sample recognition correct rate in the process of testing cassete training all reaches 100% as shown in Figure 11.Illustrate that this network process of establishing is very successful, it can identify creeping discharge model and corona discharge model well.
Although embodiments of the invention are described based on two kinds of electric discharge types, this does not limit the scope of the invention, and the method also can extend to and detect other faults of transformer.
Provided by the invention based on BP neural network algorithm carry out transformer fault know method for distinguishing, the fault type of transformer can be identified accurately, the vital role on transformer fault diagnosis and state estimation, and method is convenient.
Although content of the present invention is made complete with regard to embodiment disclosed in it and described clearly, it is not limited only to this.For the personnel of art, the present invention is made improvements and substitutes likely occur by the guidance of these statements, and these improve and to be alternatively included among the present invention.
Claims (10)
1. carry out transformer fault based on BP neural network algorithm and know a method for distinguishing, it is characterized in that, comprise the steps:
Step S1: collected by the discharge pulse collection of illustrative plates of partial discharge test system to different transformer fault;
Step S2: power diagram analysis of spectrum is carried out to the discharge pulse obtained by step S1;
Step S3: extract training sample and test sample book from the characteristic quantity that the power diagram analysis of spectrum obtained by step S2 is obtained;
Step S4: build BP network neural;
Step S5: carry out the training of BP network neural;
Step S6: carry out the test of BP network neural.
2. method of carrying out transformer discharge Fault Identification based on BP neural network algorithm as claimed in claim 1, it is characterized in that, described partial discharge test system comprises: display instrument, ultra-high frequency antenna, electrode, ground wire, insulator sleeve pipe, high-tension insulator sleeve pipe, fuel tank, coupling condenser, protective resistance, transformer; Wherein, sparking voltage introduces one end of electrode through protective resistance and high-tension insulator sleeve pipe; again the other end of electrode is drawn connecting insulator sleeve pipe; insulator sleeve pipe also passes through ground connection; discharge signal introduces partial discharge test system through coupling capacitance, and from display instrument, a situation arises for the discharge pulse of observable Different electrodes shape.
3. method of carrying out transformer discharge Fault Identification based on BP neural network algorithm as claimed in claim 1, it is characterized in that, described transformer fault comprises internal fault and external fault.
4. method of carrying out transformer discharge Fault Identification based on BP neural network algorithm as claimed in claim 3, it is characterized in that, described internal fault comprises winding failure, iron core fault, major insulation fault.
5. method of carrying out transformer discharge Fault Identification based on BP neural network algorithm as claimed in claim 3, it is characterized in that, described external fault comprises shunting switch fault, sleeve pipe fault.
6. method of carrying out transformer discharge Fault Identification based on BP neural network algorithm as claimed in claim 1, it is characterized in that, described discharge pulse is the discharge pulse of creeping discharge type or the discharge pulse of corona discharge type.
7. method of carrying out transformer discharge Fault Identification based on BP neural network algorithm as claimed in claim 1, is characterized in that, when carrying out step S1, carry out emptying of environmental variance.
8. method of carrying out transformer discharge Fault Identification based on BP neural network algorithm as claimed in claim 7, it is characterized in that, environmental variance comprises: floating potential discharge interference, Electromagnetic Interference, loose contact interference.
9. method of carrying out transformer discharge Fault Identification based on BP neural network algorithm as claimed in claim 1, is characterized in that, carry out power diagram analysis of spectrum by direct method or indirect method.
10. method of carrying out transformer discharge Fault Identification based on BP neural network algorithm as claimed in claim 1, it is characterized in that, get horizontal ordinate [65-75] interval as characteristic quantity, the matrix that 1*250 dimension is expanded as in horizontal ordinate [65-75] interval will be got with imresize function, and the 1-50 group data are set as test sample book, 51-250 group data are set as training sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510918540.2A CN105425076A (en) | 2015-12-11 | 2015-12-11 | Method of carrying out transformer fault identification based on BP neural network algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510918540.2A CN105425076A (en) | 2015-12-11 | 2015-12-11 | Method of carrying out transformer fault identification based on BP neural network algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105425076A true CN105425076A (en) | 2016-03-23 |
Family
ID=55503408
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510918540.2A Pending CN105425076A (en) | 2015-12-11 | 2015-12-11 | Method of carrying out transformer fault identification based on BP neural network algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105425076A (en) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106019098A (en) * | 2016-06-15 | 2016-10-12 | 国网北京市电力公司 | Partial discharge positioning method, device and system |
| CN106291275A (en) * | 2016-07-27 | 2017-01-04 | 西安西热节能技术有限公司 | A kind of local discharge superhigh frequency single waveform frequency domain character extracts and recognition methods |
| CN106503742A (en) * | 2016-11-01 | 2017-03-15 | 广东电网有限责任公司电力科学研究院 | A kind of visible images insulator recognition methods |
| CN106556781A (en) * | 2016-11-10 | 2017-04-05 | 华乘电气科技（上海）股份有限公司 | Shelf depreciation defect image diagnostic method and system based on deep learning |
| CN106771926A (en) * | 2017-01-04 | 2017-05-31 | 广西电网有限责任公司电力科学研究院 | One kind orientation introducing-type inside transformer electrical discharge generator |
| CN106841905A (en) * | 2017-04-14 | 2017-06-13 | 云南电网有限责任公司电力科学研究院 | A kind of recognition methods of transformer short circuit fault and device |
| CN106969836A (en) * | 2017-03-21 | 2017-07-21 | 华北电力大学（保定） | A kind of detection method of insulator heat generating spot |
| CN107037327A (en) * | 2016-10-09 | 2017-08-11 | 中国电力科学研究院 | Partial discharges fault judges feature extracting method and decision method |
| CN107122829A (en) * | 2017-06-16 | 2017-09-01 | 华北电力大学（保定） | A kind of method that utilization virtual sample trains Neural Network Diagnosis transformer fault |
| CN107364097A (en) * | 2016-05-12 | 2017-11-21 | 发那科株式会社 | The wear extent estimating device and abrasion method of estimating rate of the check-valves of injection machine |
| CN107505541A (en) * | 2017-06-26 | 2017-12-22 | 广西电网有限责任公司电力科学研究院 | A kind of recognition methods of partial discharge of transformer pattern |
| CN107561419A (en) * | 2017-08-24 | 2018-01-09 | 西南交通大学 | A kind of appraisal procedure of oil-immersed sleeve pipe insulation bubble effect risk |
| CN108621844A (en) * | 2018-05-10 | 2018-10-09 | 中南大学 | A kind of heavy rain road automatic driving vehicle power predicating method and early warning system |
| CN108896857A (en) * | 2018-07-06 | 2018-11-27 | 北京四方继保自动化股份有限公司 | A kind of transformer complex working condition recognition methods based on deep learning |
| CN108896948A (en) * | 2018-06-20 | 2018-11-27 | 成都理工大学 | A kind of no-load voltage ratio measuring device and measuring method of tandem test potential device |
| CN109978875A (en) * | 2019-04-03 | 2019-07-05 | 无锡立赫智能科技有限公司 | A kind of capacitor open defect recognition methods and identification device |
| CN110988617A (en) * | 2019-11-18 | 2020-04-10 | 国网河北省电力有限公司电力科学研究院 | Power transformer discharge fault diagnosis method and device |
| CN111157812A (en) * | 2019-12-28 | 2020-05-15 | 杭州拓深科技有限公司 | Power equipment fault monitoring method based on electromagnetic induction |
| RU2723155C1 (en) * | 2019-05-29 | 2020-06-09 | Акционерное общество "ПКК МИЛАНДР" | Method of electric discharge detection and classification on electric equipment |
| CN112597963A (en) * | 2020-12-31 | 2021-04-02 | 厦门天马微电子有限公司 | Display device, control method and circuit thereof |
| CN112763860A (en) * | 2020-12-11 | 2021-05-07 | 贵州电网有限责任公司 | Power transmission line fault and abnormal discharge monitoring system and method |
| CN112782614A (en) * | 2019-11-11 | 2021-05-11 | 株洲中车时代电气股份有限公司 | Fault early warning method and device of converter based on multi-information fusion |
| CN112926725A (en) * | 2020-12-01 | 2021-06-08 | 杭州电力设备制造有限公司 | Data analysis method and system based on thermal image |
| CN115542101A (en) * | 2022-11-30 | 2022-12-30 | 杭州兆华电子股份有限公司 | Voiceprint preprocessing method of transformer voiceprint detection system |
| CN117849498A (en) * | 2023-12-27 | 2024-04-09 | 慧之安信息技术股份有限公司 | Power grid transformer defect detection method and system based on multi-target identification |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5598509A (en) * | 1992-08-28 | 1997-01-28 | Hitachi, Ltd. | Method of configuring a neural network and a diagnosis/recognition system using the same |
| CN101408579A (en) * | 2008-11-21 | 2009-04-15 | 重庆大学 | Oil paper insulated electrothermic aging test and local discharge integrated detection device of transformer |
| CN102221651A (en) * | 2011-03-11 | 2011-10-19 | 太原理工大学 | Fault on-line diagnosis and early warning method of flameproof dry-type transformer for mine |
| CN102866377A (en) * | 2012-09-17 | 2013-01-09 | 中国电力科学研究院 | Detection and verification system and method of transformer partial discharge and positioning on-line monitoring assembly |
| CN103235206A (en) * | 2012-11-05 | 2013-08-07 | 王少夫 | Transformer fault diagnosis method |
| CN103576061A (en) * | 2013-10-17 | 2014-02-12 | 国家电网公司 | Method for discharge fault diagnosis of transformer |
| CN103576059A (en) * | 2013-10-10 | 2014-02-12 | 国家电网公司 | Integrated fault diagnosis method and system for turn-to-turn discharging of transformer |
| CN104020398A (en) * | 2014-06-03 | 2014-09-03 | 华北电力大学 | Method for extracting partial discharge waveform features of converter transformer |
-
2015
- 2015-12-11 CN CN201510918540.2A patent/CN105425076A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5598509A (en) * | 1992-08-28 | 1997-01-28 | Hitachi, Ltd. | Method of configuring a neural network and a diagnosis/recognition system using the same |
| CN101408579A (en) * | 2008-11-21 | 2009-04-15 | 重庆大学 | Oil paper insulated electrothermic aging test and local discharge integrated detection device of transformer |
| CN102221651A (en) * | 2011-03-11 | 2011-10-19 | 太原理工大学 | Fault on-line diagnosis and early warning method of flameproof dry-type transformer for mine |
| CN102866377A (en) * | 2012-09-17 | 2013-01-09 | 中国电力科学研究院 | Detection and verification system and method of transformer partial discharge and positioning on-line monitoring assembly |
| CN103235206A (en) * | 2012-11-05 | 2013-08-07 | 王少夫 | Transformer fault diagnosis method |
| CN103576059A (en) * | 2013-10-10 | 2014-02-12 | 国家电网公司 | Integrated fault diagnosis method and system for turn-to-turn discharging of transformer |
| CN103576061A (en) * | 2013-10-17 | 2014-02-12 | 国家电网公司 | Method for discharge fault diagnosis of transformer |
| CN104020398A (en) * | 2014-06-03 | 2014-09-03 | 华北电力大学 | Method for extracting partial discharge waveform features of converter transformer |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107364097A (en) * | 2016-05-12 | 2017-11-21 | 发那科株式会社 | The wear extent estimating device and abrasion method of estimating rate of the check-valves of injection machine |
| CN107364097B (en) * | 2016-05-12 | 2020-03-10 | 发那科株式会社 | Wear amount estimation device and wear amount estimation method for check valve of injection molding machine |
| CN106019098A (en) * | 2016-06-15 | 2016-10-12 | 国网北京市电力公司 | Partial discharge positioning method, device and system |
| CN106291275A (en) * | 2016-07-27 | 2017-01-04 | 西安西热节能技术有限公司 | A kind of local discharge superhigh frequency single waveform frequency domain character extracts and recognition methods |
| CN106291275B (en) * | 2016-07-27 | 2019-03-22 | 西安西热节能技术有限公司 | A kind of local discharge superhigh frequency single waveform frequency domain character extracts and recognition methods |
| CN107037327A (en) * | 2016-10-09 | 2017-08-11 | 中国电力科学研究院 | Partial discharges fault judges feature extracting method and decision method |
| CN106503742A (en) * | 2016-11-01 | 2017-03-15 | 广东电网有限责任公司电力科学研究院 | A kind of visible images insulator recognition methods |
| CN106503742B (en) * | 2016-11-01 | 2019-04-26 | 广东电网有限责任公司电力科学研究院 | A kind of visible images insulator recognition methods |
| CN106556781A (en) * | 2016-11-10 | 2017-04-05 | 华乘电气科技（上海）股份有限公司 | Shelf depreciation defect image diagnostic method and system based on deep learning |
| CN106771926A (en) * | 2017-01-04 | 2017-05-31 | 广西电网有限责任公司电力科学研究院 | One kind orientation introducing-type inside transformer electrical discharge generator |
| CN106969836A (en) * | 2017-03-21 | 2017-07-21 | 华北电力大学（保定） | A kind of detection method of insulator heat generating spot |
| CN106841905A (en) * | 2017-04-14 | 2017-06-13 | 云南电网有限责任公司电力科学研究院 | A kind of recognition methods of transformer short circuit fault and device |
| CN107122829A (en) * | 2017-06-16 | 2017-09-01 | 华北电力大学（保定） | A kind of method that utilization virtual sample trains Neural Network Diagnosis transformer fault |
| CN107505541A (en) * | 2017-06-26 | 2017-12-22 | 广西电网有限责任公司电力科学研究院 | A kind of recognition methods of partial discharge of transformer pattern |
| CN107505541B (en) * | 2017-06-26 | 2019-11-01 | 广西电网有限责任公司电力科学研究院 | A kind of mode identification method of partial discharge of transformer |
| CN107561419B (en) * | 2017-08-24 | 2019-02-22 | 西南交通大学 | A kind of appraisal procedure of oil-immersed sleeve pipe insulation bubble effect risk |
| CN107561419A (en) * | 2017-08-24 | 2018-01-09 | 西南交通大学 | A kind of appraisal procedure of oil-immersed sleeve pipe insulation bubble effect risk |
| CN108621844A (en) * | 2018-05-10 | 2018-10-09 | 中南大学 | A kind of heavy rain road automatic driving vehicle power predicating method and early warning system |
| CN108621844B (en) * | 2018-05-10 | 2020-01-03 | 中南大学 | Electric quantity prediction method and early warning system for unmanned vehicle on rainstorm road |
| CN108896948A (en) * | 2018-06-20 | 2018-11-27 | 成都理工大学 | A kind of no-load voltage ratio measuring device and measuring method of tandem test potential device |
| CN108896948B (en) * | 2018-06-20 | 2023-03-21 | 成都理工大学 | Transformation ratio measuring device and method for cascade test transformation equipment |
| CN108896857A (en) * | 2018-07-06 | 2018-11-27 | 北京四方继保自动化股份有限公司 | A kind of transformer complex working condition recognition methods based on deep learning |
| CN108896857B (en) * | 2018-07-06 | 2020-12-01 | 北京四方继保自动化股份有限公司 | Deep learning-based complex working condition identification method for transformer |
| CN109978875A (en) * | 2019-04-03 | 2019-07-05 | 无锡立赫智能科技有限公司 | A kind of capacitor open defect recognition methods and identification device |
| RU2723155C1 (en) * | 2019-05-29 | 2020-06-09 | Акционерное общество "ПКК МИЛАНДР" | Method of electric discharge detection and classification on electric equipment |
| CN112782614A (en) * | 2019-11-11 | 2021-05-11 | 株洲中车时代电气股份有限公司 | Fault early warning method and device of converter based on multi-information fusion |
| CN110988617A (en) * | 2019-11-18 | 2020-04-10 | 国网河北省电力有限公司电力科学研究院 | Power transformer discharge fault diagnosis method and device |
| CN111157812A (en) * | 2019-12-28 | 2020-05-15 | 杭州拓深科技有限公司 | Power equipment fault monitoring method based on electromagnetic induction |
| CN112926725A (en) * | 2020-12-01 | 2021-06-08 | 杭州电力设备制造有限公司 | Data analysis method and system based on thermal image |
| CN112926725B (en) * | 2020-12-01 | 2024-05-14 | 杭州电力设备制造有限公司 | Data analysis method and system based on thermal image |
| CN112763860A (en) * | 2020-12-11 | 2021-05-07 | 贵州电网有限责任公司 | Power transmission line fault and abnormal discharge monitoring system and method |
| CN112763860B (en) * | 2020-12-11 | 2022-05-24 | 贵州电网有限责任公司 | Power transmission line fault and abnormal discharge monitoring system and method |
| CN112597963A (en) * | 2020-12-31 | 2021-04-02 | 厦门天马微电子有限公司 | Display device, control method and circuit thereof |
| CN112597963B (en) * | 2020-12-31 | 2022-10-25 | 厦门天马微电子有限公司 | Display device, control method and circuit thereof |
| CN115542101A (en) * | 2022-11-30 | 2022-12-30 | 杭州兆华电子股份有限公司 | Voiceprint preprocessing method of transformer voiceprint detection system |
| CN117849498A (en) * | 2023-12-27 | 2024-04-09 | 慧之安信息技术股份有限公司 | Power grid transformer defect detection method and system based on multi-target identification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105425076A (en) | Method of carrying out transformer fault identification based on BP neural network algorithm | |
| CN104217110B (en) | GIS state evaluation method and device by using reference state analysis | |
| CN106199305A (en) | Underground coal mine electric power system dry-type transformer insulation health state evaluation method | |
| CN102221651A (en) | Fault on-line diagnosis and early warning method of flameproof dry-type transformer for mine | |
| Saleh et al. | Proposed diagnostic methodology using the cross‐correlation coefficient factor technique for power transformer fault identification | |
| CN109029959B (en) | Method for detecting mechanical state of transformer winding | |
| CN103760448A (en) | Comprehensive live test method of insulation defects of metal oxide arrester for electric power | |
| CN102081136A (en) | Method for on-site GIS (Gas-insulated metal-enclosed switchgear) partial discharge detection under impulse voltage | |
| Akhmetov et al. | A new diagnostic technique for reliable decision-making on transformer FRA data in interturn short-circuit condition | |
| CN107247204B (en) | State monitoring system and monitoring method for voltage limiters in ultra-high and extra-high voltage series compensation device | |
| Steennis et al. | Guarding MV cables on-line: With travelling wave based temperature monitoring, fault location, PD location and PD related remaining life aspects | |
| CN106770652A (en) | High-tension transformer health status monitoring device and monitoring method based on acoustic characteristic | |
| Cavallini et al. | Diagnosis of EHV and HV transformers through an innovative partial-discharge-based technique | |
| CN107525996A (en) | A kind of series compensation device voltage limiter On-line Monitoring of Leakage Current method and system | |
| CN110927620B (en) | Sleeve with winding deformation and overvoltage on-line monitoring sensor | |
| CN106980051B (en) | A kind of intermittence tandem type fault electric arc recognition methods | |
| CN115166450A (en) | Switch cabinet part partial discharge positioning method based on pulse current signal frequency spectrum | |
| CN105182116B (en) | A kind of transformer winding working state detecting method based on weighted gradient structural similarity | |
| CN112684311A (en) | Characteristic quantity extraction method for identifying oil paper insulation partial discharge type of transformer | |
| CN108181513B (en) | Inter-phase coupling capacitance testing method and system for lightning arrester based on pilot frequency method | |
| CN110645887A (en) | Winding deformation judgment method based on short-circuit reactance | |
| CN114113944B (en) | CVT capacitor unit breakdown defect test method and device | |
| CN107576259A (en) | Deformation of transformer winding online test method based on very fast transient overvoltage characteristic | |
| CN110794344B (en) | Frequency response test method for transformer winding deformation with influence of sleeve eliminated | |
| CN207380098U (en) | Countermeasure set is removed in a kind of bushing polarization depolarising on-the-spot test |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20160323 |
|
| WD01 | Invention patent application deemed withdrawn after publication |