CN106771938A - A kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method and device - Google Patents
A kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method and device Download PDFInfo
- Publication number
- CN106771938A CN106771938A CN201710174238.XA CN201710174238A CN106771938A CN 106771938 A CN106771938 A CN 106771938A CN 201710174238 A CN201710174238 A CN 201710174238A CN 106771938 A CN106771938 A CN 106771938A
- Authority
- CN
- China
- Prior art keywords
- suprasphere
- partial discharge
- optimal
- characteristic parameter
- radius
- 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.)
- Granted
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
- G01R31/12—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
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention discloses a kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method and its device, including the training sample that the characteristic parameter of multiple Partial discharge signals is set；Determine the separation suprasphere per class discharge mode according to training sample and support vector description algorithm, each separates the interior only whole training samples comprising a class discharge mode of suprasphere；The current radius of each separation suprasphere are carried out into self-adaptive processing calculating as initial value respectively, each optimal radius for separating suprasphere is obtained, so as to obtain the corresponding optimal suprasphere of every class discharge mode；The actual Partial discharge signal of collection, obtains the characteristic parameter in actual Partial discharge signal；It is determined that the optimal suprasphere residing for characteristic parameter in actual Partial discharge signal, the corresponding discharge mode of the optimal suprasphere is the discharge mode of actual Partial discharge signal.The present invention is the discharge mode that can determine whether actual Partial discharge signal without the concern for the regularity of distribution of the characteristic parameter of Partial discharge signal, can process big-sample data, and identification accuracy is high.
Description
Technical field
The present invention relates to field, more particularly to a kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method and its dress
Put.
Background technology
Solid insulation ring main unit is exhausted by itself during long-term use as the key equipment in intelligent distribution network
Edge is aging and outside environmental elements are disturbed, and easily causes that ring main unit inside produces shelf depreciation (hereinafter referred to as " partial discharge ") phenomenon,
From in terms of ring main unit service condition in recent years, the accident number caused by shelf depreciation account for more than half of whole distribution accident.
Accordingly, it would be desirable to the shelf depreciation for being directed to switchgear carries out detection monitoring, shelf depreciation produces difference because different factors influence
Shelf depreciation pattern, shelf depreciation pattern have 4 kinds be respectively metallic projections electric discharge, insulator surface metal discharge, air gaps
Electric discharge, the electric discharge of free metal particulate, the harm that different discharge patterns are caused also are not quite similar, and effective mould is carried out to Partial discharge signal
Formula identification, further can exactly understand and grasp insulation defect type and feature that switchgear inside occurs, to enter
One step judges its insulating reliability, and highly important engineering significance is suffered from for service work.
At present, it is commonly used in Partial Discharge Pattern Recognition Method and initially sets up typical partial discharge model, by applied voltage test
Obtain the local discharge signal of each model；Then according to the local discharge signal that obtains draw shelf depreciation the q of middle φ mono- figures or
φ-n scheme, and the statistical nature parameter of each quasi-representative shelf depreciation is extracted from figure；Again with the statistics of each quasi-representative shelf depreciation
Characteristic parameter is trained as initial training sample to support vector machine classifier；Finally with the SVMs point for training
Class device is identified to partial discharge of switchgear pattern.
But, because the distribution of the characteristic parameter of local discharge signal is overlap, nonlinear and complicated, standard branch
The method for holding vector machine cannot process nonlinear problem, can only process Small Sample Database, cause to be processed when sample size is more
Accuracy is low or even cannot process.
Therefore, how to provide a kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method for being capable of big-sample data and
Its device is the problem that those skilled in the art need to solve at present.
The content of the invention
It is an object of the invention to provide a kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method and its device, it is not required to
Consider the regularity of distribution of the characteristic parameter of Partial discharge signal, you can judge the discharge mode of actual Partial discharge signal, can be to full-page proof
Notebook data is processed, and identification accuracy is high.
In order to solve the above technical problems, the invention provides a kind of solid insulation ring main unit PD Pattern Recognition side
Method, including：
The training sample of the characteristic parameter of multiple Partial discharge signals is set；
Determine the separation suprasphere per class discharge mode according to the training sample and support vector description algorithm, its
In, whole training samples of itself corresponding class discharge mode are only included in each described separation suprasphere；
Each described current radius for separating suprasphere is carried out into self-adaptive processing calculating as initial value respectively, obtains each
The individual optimal radius for separating suprasphere, according to the optimal radius, adjust each described separation suprasphere and obtain every class and put
The corresponding optimal suprasphere of power mode；
The actual Partial discharge signal of collection, obtains the characteristic parameter in the actual Partial discharge signal；
Determine the optimal suprasphere residing for the characteristic parameter in the actual Partial discharge signal, optimal suprasphere is corresponding puts for this
Power mode is the discharge mode of the actual Partial discharge signal.
Preferably, the characteristic parameter is Fn、m、g、p={ Δ Q, Δ T, Δ U, Au,Ku,mcc,Sk, Γ }, wherein, Δ Q is to put
The electric quantity of electric charge, Δ T are neighbouring discharge time interval, Δ U is adjacent partial discharge pulse amplitude different information, AuIt is maximum discharge pulse
Amplitude ratio, SkIt is degree of skewness, KuIt is steepness, mccFor amendment coefficient correlation, Γ are relative magnitude dispersion.
Preferably, also include：
If the characteristic parameter in the actual Partial discharge signal is in outside all optimal supraspheres, the actual partial discharge letter
Number be interference signal.
Preferably, the process that the self-adaptive processing is calculated is specially：
Step s201：Using the current radius of the separation suprasphere as initial value input adaptive processor, the is obtained
One output result y (n), the relational expression of output radius that self-adaptive processing is calculated is：
R (n)=d (n)-y (n)
Minimal error is ε (n)；
Wherein, λ is exponential weighting factor；D (n) is Expected Response；Obtained according to adaptive algorithm minimum mean square error criterion
To standard relationship：
Step s202：The output radius that will be obtained are computed repeatedly as initial value, return to step s201, until obtain
Error meets error precision, and the output radius for now obtaining are the optimal radius.
In order to solve the above technical problems, present invention also offers a kind of solid insulation ring main unit PD Pattern Recognition dress
Put, including：
Sample setup module, the training sample of the characteristic parameter for setting multiple Partial discharge signals；
The originally determined module of grader, for determining to be put per class according to the training sample and support vector description algorithm
The separation suprasphere of power mode, wherein, each described separation only includes the complete of itself corresponding class discharge mode in suprasphere
Portion's training sample；
Optimal correction module, is carried out each described current radius for separating suprasphere as initial value for respectively adaptive
Calculating should be processed, each optimal radius for separating suprasphere is obtained, according to the optimal radius, each separation is adjusted
Suprasphere obtains the corresponding optimal suprasphere of every class discharge mode；
Actual characteristic identification module, for gathering actual Partial discharge signal, obtains the feature ginseng in the actual Partial discharge signal
Number；
Classification judge module, for determining the optimal suprasphere residing for the characteristic parameter in the actual Partial discharge signal, should
The corresponding discharge mode of optimal suprasphere is the discharge mode of the actual Partial discharge signal.
Preferably, the classification judge module also includes：
Interference judging unit, for determining the characteristic parameter in the actual Partial discharge signal whether in all optimal hyperspheres
Outside body, if so, then judging that the actual Partial discharge signal is interference signal.
The invention provides a kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method and its device, multiple offices are obtained
After the training sample of the characteristic parameter of discharge signal, every class discharge mode is obtained according to training sample and support vector description algorithm
Separation suprasphere, now each separate only comprising the whole training samples under its corresponding class discharge mode in suprasphere,
The radius for separating suprasphere is optimized according to adaptive algorithm again afterwards, obtains optimal suprasphere, the optimal hypersphere physical efficiency
Enough whole actual samples included as far as possible under its corresponding class discharge mode.It can be seen that, the present invention is without the concern for partial discharge letter
Number characteristic parameter the regularity of distribution, it is only necessary to judge which optimal suprasphere it is in, you can judge actual Partial discharge signal
Discharge mode, therefore the present invention can be processed big-sample data, and identification accuracy is higher.
Brief description of the drawings
Technical scheme in order to illustrate more clearly the embodiments of the present invention, below will be to institute in prior art and embodiment
The accompanying drawing for needing to use is briefly described, it should be apparent that, drawings in the following description are only some implementations of the invention
Example, for those of ordinary skill in the art, on the premise of not paying creative work, can also obtain according to these accompanying drawings
Obtain other accompanying drawings.
A kind of flow of the process of solid insulation ring main unit Partial Discharge Pattern Recognition Method that Fig. 1 is provided for the present invention
Figure；
A kind of structural representation of solid insulation ring main unit PD Pattern Recognition device that Fig. 2 is provided for the present invention.
Specific embodiment
Core of the invention is to provide a kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method and its device, is not required to
Consider the regularity of distribution of the characteristic parameter of Partial discharge signal, you can judge the discharge mode of actual Partial discharge signal, can be to full-page proof
Notebook data is processed, and identification accuracy is high.
To make the purpose, technical scheme and 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
A part of embodiment of the present invention, rather than whole embodiments.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.
Shown in Figure 1 the invention provides a kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method, Fig. 1 is
A kind of flow chart of the process of solid insulation ring main unit Partial Discharge Pattern Recognition Method that the present invention is provided；The method includes：
Step s101：The training sample of the characteristic parameter of multiple Partial discharge signals is set；
Step s102：The separation hypersphere per class discharge mode is determined according to training sample and support vector description algorithm
Body, wherein, each separates whole training samples that itself corresponding class discharge mode is only included in suprasphere；
Step s103：The current radius of each separation suprasphere are carried out into self-adaptive processing calculating as initial value respectively,
Each optimal radius for separating suprasphere is obtained, according to optimal radius, each separation suprasphere is adjusted and is obtained every class discharge mode
Corresponding optimal suprasphere；
Step s104：The actual Partial discharge signal of collection, obtains the characteristic parameter in actual Partial discharge signal；
Step s105：It is determined that the optimal suprasphere residing for characteristic parameter in actual Partial discharge signal, the optimal suprasphere pair
The discharge mode answered is the discharge mode of actual Partial discharge signal.
Specifically, characteristic parameter here is Fn、m、g、p={ Δ Q, Δ T, Δ U, Au,Ku,mcc,Sk, Γ }, wherein, Δ Q is
Discharge charge amount, Δ T are neighbouring discharge time interval, Δ U is adjacent partial discharge pulse amplitude different information, AuIt is maximum electric discharge arteries and veins
Rush amplitude ratio, SkIt is degree of skewness, KuIt is steepness, mccFor amendment coefficient correlation, Γ are relative magnitude dispersion.
Wherein, above-mentioned 8 information：ΔQ、ΔT、ΔU、Au、Sk、Ku、mcc, Γ be to gather the feature that obtains after Partial discharge signal
Information, is characterized parameter as F by its bout afterwardsn、m、g、p.In addition, preceding four Δ Q, Δ T, Δ U, AuIt is different discharge patterns
Under comprehensive characteristics information, for recognizing whether the signal for collecting really be from shelf depreciation, the shadow for excluding the interference signal
Ring；Four statistical parameter S afterwardsk、Ku、mcc, Γ be then the parameter areas different for limiting different discharge patterns so that realization office
Mode playback is accurately identified.
It is further known that：
First：Discharge charge amount Δ Q can characterize shelf depreciation information from macroscopic perspective, wherein, the width of Partial discharge signal
It is proportional to there is quadratic integral in value and the discharge charge amount of Partial discharge signal, therefore：
Wherein, N is discharge time in a power frequency period；uiRepresent i-th discharge pulse amplitude.
Second：Neighbouring discharge time interval Δ T refers to the difference of the dense degree of positive-negative half-cycle discharge pulse, i.e., positive and negative half
The difference of the time interval of Zhou Xianglin discharge pulses.I.e.：
ΔTmax=max (Δ T1,ΔT2,...,ΔTN)
Wherein, N+, N-A power frequency period positive-negative half-cycle discharge time is represented respectively；Δti +(i=1,2 ... N+), Δ ti -
(i=1,2 ... N-) time interval for power frequency period positive-negative half-cycle neighbouring discharge pulse is represented respectively；ΔTi=Δ T++Δ
T-(i=1,2 ... N-1) represent adjacent the interspike interval twice in a power frequency period；ΔTmaxRepresent whole power frequency
Adjacent discharge pulse maximum time interval twice in cycle.Very it is difficult to rapidly catch accurate due to discharge time, therefore with adjacent
The interval of discharge time takes average as the interspike interval twice.
3rd：The relational expression of the adjacent partial discharge pulse amplitude different informations of Δ U is：
ΔUn=Un+1-Un
By being analyzed discovery to positive-negative half-cycle neighbouring discharge pulse amplitude feature, Δ U as shelf depreciation sign
Information, the adjacent two partial discharges pulse signal amplitude difference of the smaller explanation of value of Δ U is smaller, and strength of discharge is stronger.Conversely, strength of discharge
It is smaller.
4th, because partial discharge pulse signal amplitude is change, therefore use maximum discharge pulse amplitude ratio AuTo characterize just
Negative half period electric discharge most serious degree difference.AuRelationship between expression formula be：
5th, degree of skewness SkDefinition be：
Wherein, degree of skewness SkReflect left and right deflection situation of the spectral shape relative to normal distribution, Sk=0 explanation spectrogram
It is symmetrical；Sk>0 illustrates spectrogram toward left avertence；Sk<0 illustrates spectrogram toward right avertence.xiIt is i-th phase of phase window；μ is that signal is equal
Value,It is the i average of phase, σ is variance, and W is the phase window sum in the half period.
6th：Steepness KuFor describing certain partial discharge model shape in contrast to its definition of the projection degree of normal distribution
For：
Wherein,xiIt is i-th phase of phase window, Δ x is the width of phase window, and W is the phase window in the half period
Number, Ku>0 illustrates that the collection of illustrative plates under the pattern is sharply more precipitous than normal distribution profile, conversely, more flat than normal distribution profile.
7th：Amendment coefficient correlation mccIt is to propose to be put in spectrogram positive-negative half-cycle for evaluating under original cross-correlation coefficient cc
The difference of power mode, its relational expression is：
Wherein, Q is discharge capacity,It is phase, qiIt is the mean discharge magnitude in phase window i, subscript "+" "-" corresponds to spectrogram
Positive-negative half-cycle, cross-correlation coefficient cc reflects shape similarity degree of the spectrogram in positive-negative half-cycle, cross-correlation coefficient cc close to 1,
Mean that positive-negative half-cycle profile is quite similar, conversely, positive-negative half-cycle shape difference is big.
8th：Relative magnitude dispersion Γ characterizes the dense degree of positive-negative half-cycle magnitude parameters under different partial discharge patterns, its
Relational expression is as follows：
Γ=Γ+/Γ-
Wherein, Γ+And Γ-Positive half cycle and negative half period dispersion degree are represented respectively.
In addition, the process of step s102 is specially：
Appoint first and take a kind of grader, the training sample F that n metalloids protrusion is dischargednUsed as target sample, remaining is put
The training sample of electric model carries out SA-SVDD study as non-targeted samples, and obtaining can be by n metalloid protrusion discharging model
The SA-SVDDn that with the training sample of other discharging models separate of training sample；Similarly, according to characteristic parameter FmCan be used for
Separate m class insulator surface metal discharge models SA-SVDDm, according to characteristic parameter FgCan be used to separate g class bubble-discharges
Model SA-SVDDg divides and according to characteristic parameter FpThe SA- of p type free metal particle electric discharge types can be used to separate
SVDDp.Wherein, SA-SVDDn, SA-SVDDm, SA-SVDDg, SA-SVDDp are separation suprasphere.
Afterwards, will be respectively using all kinds of discharging models as to calculate its corresponding most in target sample feeding adaptive algorithm
Excellent radius value, constitutes the optimal suprasphere Θ 0 that can be separated all kinds of discharging models and other three classes discharging models.It is used to distinguish
The optimal radius R expression formulas of other three classes areWherein, N (n) is primary noise letter
Number；It is noise signal that device is produced in itself；It is the best estimate of output y (n).
Wherein, the process that above-mentioned self-adaptive processing is calculated is specially：
Step s201：The current radius of suprasphere will be separated as initial value input adaptive processor, first is obtained defeated
Go out result y (n), the relational expression of the output radius that self-adaptive processing is calculated is：
R (n)=d (n)-y (n)
Minimal error is ε (n), and its relational expression is：
Wherein, λ is exponential weighting factor；D (n) is Expected Response；Obtained according to adaptive algorithm minimum mean square error criterion
To standard relationship：
It is understood that above-mentioned relation formula is in order to function expression to be write as the form of the Wiener equation of standard, side
Face obtains optimal radius value using the error amount that iterative formula solves minimum after an action of the bowels.This part is defined as Q in the bracket of the left side
N (), this part of the right is defined as P (n), then above formula is reduced to Q (n) w=P (n) and show that wiener solution is w=Q-1(n) P (n), so
Afterwards, launched using iteration form, be by above formula abbreviation：
W (n)=w (n-1)+Q-1(n)x(n)R(n|n-1)
Step s202：The output radius that will be obtained are computed repeatedly as initial value, return to step s201, until obtain
Error meets error precision, and the output radius for now obtaining are optimal radius.
It is understood that adaptive algorithm is actually a kind of recursive algorithm, with least squared criterion as foundation.Calculate
The key of method is, with the time averaging minimum criterion of square, to replace lowest mean square criterion, and well is temporally iterated meter
Calculate.Specifically, be will square carrying out mean deviation and minimize it to initial time to all errors of moment at that time, according still further to
This criterion determines weight coefficient vector.
Wherein, the minimal error of self-adaptive processing calculating is：
For non-stationary input signal, in order to be able to be tracked well, often an exponential weighting factor is introduced (namely
Forgetting factor λ) formula is modified, obtain：
According to adaptive algorithm minimum mean square error criterion judge the relational expression of weight vector as：
Above-mentioned standard relational expression is obtained after arrangement：
The calculating that iterates is carried out to above-mentioned standard relational expression, is obtained by meeting on the premise of error precision and is put per class
The optimal R (n) of electric model.
Preferably, the method also includes：
If the characteristic parameter in actual Partial discharge signal is in outside all optimal supraspheres, actual Partial discharge signal is interference
Signal.
It is understood that
The invention provides a kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method, multiple Partial discharge signals are obtained
After the training sample of characteristic parameter, the separation for obtaining every class discharge mode according to training sample and support vector description algorithm surpasses
Spheroid, now each separate in suprasphere only comprising the whole training samples under its corresponding class discharge mode, afterwards again according to
The radius for separating suprasphere is optimized according to adaptive algorithm, obtains optimal suprasphere, the optimal suprasphere can be as far as possible
Comprising the whole actual samples under its corresponding class discharge mode.It can be seen that, feature of the present invention without the concern for Partial discharge signal
The regularity of distribution of parameter, it is only necessary to judge which optimal suprasphere it is in, you can judge the electric discharge mould of actual Partial discharge signal
Formula, therefore the present invention is for the occasion of different sample datas, accuracy of identification is attained by requiring, can be to big-sample data at
Reason, versatility is stronger.Meanwhile, grader leakage point of the invention, wrong point of situation are few, and the classification degree of accuracy and accuracy of identification are high.
Also, the adaptive algorithm fast convergence rate for using, can have computing capability higher for different research objects.
Present invention also offers a kind of solid insulation ring main unit PD Pattern Recognition device, shown in Figure 2, Fig. 2
It is a kind of structural representation of solid insulation ring main unit PD Pattern Recognition device that the present invention is provided.The device includes：
Sample setup module 1, the training sample of the characteristic parameter for setting multiple Partial discharge signals；
The originally determined module 2 of grader, for determining the electric discharge per class according to training sample and support vector description algorithm
The separation suprasphere of pattern, wherein, each separates the whole training only comprising itself corresponding class discharge mode in suprasphere
Sample；
Optimal correction module 3, for the current radius of each separation suprasphere to be carried out into self adaptation as initial value respectively
Treatment is calculated, and obtains each optimal radius for separating suprasphere, according to optimal radius, is adjusted each separation suprasphere and is obtained every class
The corresponding optimal suprasphere of discharge mode；
Actual characteristic identification module 4, for gathering actual Partial discharge signal, obtains the characteristic parameter in actual Partial discharge signal；
Classification judge module 5, for determining the optimal suprasphere residing for the characteristic parameter in actual Partial discharge signal, this is optimal
The corresponding discharge mode of suprasphere is the discharge mode of actual Partial discharge signal.
Preferably, classification judge module 5 also includes：
Interference judging unit, for determine the characteristic parameter in actual Partial discharge signal whether in all optimal supraspheres it
Outward, if so, then judging that actual Partial discharge signal is interference signal.
The invention provides a kind of solid insulation ring main unit PD Pattern Recognition device, multiple Partial discharge signals are obtained
After the training sample of characteristic parameter, the separation for obtaining every class discharge mode according to training sample and support vector description algorithm surpasses
Spheroid, now each separate in suprasphere only comprising the whole training samples under its corresponding class discharge mode, afterwards again according to
The radius for separating suprasphere is optimized according to adaptive algorithm, obtains optimal suprasphere, the optimal suprasphere can be as far as possible
Comprising the whole actual samples under its corresponding class discharge mode.It can be seen that, feature of the present invention without the concern for Partial discharge signal
The regularity of distribution of parameter, it is only necessary to judge which optimal suprasphere it is in, you can judge the electric discharge mould of actual Partial discharge signal
Formula, therefore the present invention can be processed big-sample data, identification accuracy is higher, and versatility is stronger.
Each embodiment is described by the way of progressive in this specification, and what each embodiment was stressed is and other
The difference of embodiment, between each embodiment identical similar portion mutually referring to.For device disclosed in embodiment
For, because it is corresponded to the method disclosed in Example, so description is fairly simple, related part is said referring to method part
It is bright.
Also, it should be noted that in this manual, such as first and second or the like relational terms be used merely to by
One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation
Between there is any this actual relation or order.And, term " including ", "comprising" or its any other variant meaning
Covering including for nonexcludability, so that process, method, article or equipment including a series of key elements not only include that
A little key elements, but also other key elements including being not expressly set out, or also include for this process, method, article or
The intrinsic key element of equipment.In the absence of more restrictions, the key element limited by sentence "including a ...", does not arrange
Except also there is other identical element in the process including the key element, method, article or equipment.
The foregoing description of the disclosed embodiments, enables professional and technical personnel in the field to realize or uses the present invention.
Various modifications to these embodiments will be apparent for those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, the present invention
The embodiments shown herein is not intended to be limited to, and is to fit to and principles disclosed herein and features of novelty phase one
The scope most wide for causing.
Claims (6)
1. a kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method, it is characterised in that including：
The training sample of the characteristic parameter of multiple Partial discharge signals is set；
Determine the separation suprasphere per class discharge mode according to the training sample and support vector description algorithm, wherein, often
Whole training samples of itself corresponding class discharge mode are only included in the individual separation suprasphere；
Each described current radius for separating suprasphere is carried out into self-adaptive processing calculating as initial value respectively, each institute is obtained
The optimal radius for separating suprasphere are stated, according to the optimal radius, each described separation suprasphere is adjusted and is obtained every class electric discharge mould
The corresponding optimal suprasphere of formula；
The actual Partial discharge signal of collection, obtains the characteristic parameter in the actual Partial discharge signal；
Determine the optimal suprasphere residing for the characteristic parameter in the actual Partial discharge signal, the corresponding electric discharge mould of the optimal suprasphere
Formula is the discharge mode of the actual Partial discharge signal.
2. method according to claim 1, it is characterised in that the characteristic parameter is Fn、m、g、p=Δ Q, Δ T, Δ U,
Au,Ku,mcc,Sk, Γ }, wherein, Δ Q is discharge charge amount, Δ T is neighbouring discharge time interval, Δ U is adjacent partial discharge pulse width
Value different information, AuIt is maximum discharge pulse amplitude ratio, SkIt is degree of skewness, KuIt is steepness, mccIt is amendment coefficient correlation, Γ
It is relative magnitude dispersion.
3. method according to claim 1, it is characterised in that also include：
If the characteristic parameter in the actual Partial discharge signal is in outside all optimal supraspheres, the actual Partial discharge signal is
Interference signal.
4. method according to claim 1, it is characterised in that the process that the self-adaptive processing is calculated is specially：
Step s201：Using the current radius of the separation suprasphere as initial value input adaptive processor, first is obtained defeated
Go out result y (n), the relational expression of the output radius that self-adaptive processing is calculated is：
R (n)=d (n)-y (n)
Minimal error is ε (n), and its relational expression is：
Wherein, λ is exponential weighting factor；D (n) is Expected Response；Marked according to adaptive algorithm minimum mean square error criterion
Quasi- relational expression：
Step s202：Used as initial value, return to step s201 is computed repeatedly the output radius that will be obtained, until the error for obtaining
Meet error precision, the output radius for now obtaining are the optimal radius.
5. a kind of solid insulation ring main unit PD Pattern Recognition device, it is characterised in that including：
Sample setup module, the training sample of the characteristic parameter for setting multiple Partial discharge signals；
The originally determined module of grader, for determining the mould that discharged per class according to the training sample and support vector description algorithm
The separation suprasphere of formula, wherein, each described whole instruction separated in suprasphere only comprising itself corresponding class discharge mode
Practice sample；
Optimal correction module, for respectively carrying out at self adaptation each described current radius for separating suprasphere as initial value
Reason is calculated, and obtains each optimal radius for separating suprasphere, according to the optimal radius, adjusts each separation hypersphere
Body obtains the corresponding optimal suprasphere of every class discharge mode；
Actual characteristic identification module, for gathering actual Partial discharge signal, obtains the characteristic parameter in the actual Partial discharge signal；
Classification judge module, for determining the optimal suprasphere residing for the characteristic parameter in the actual Partial discharge signal, this is optimal
The corresponding discharge mode of suprasphere is the discharge mode of the actual Partial discharge signal.
6. device according to claim 5, it is characterised in that the classification judge module also includes：
Interference judging unit, for determine the characteristic parameter in the actual Partial discharge signal whether in all optimal supraspheres it
Outward, if so, then judging that the actual Partial discharge signal is interference signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710174238.XA CN106771938B (en) | 2017-03-22 | 2017-03-22 | Method and device for identifying partial discharge mode of solid insulation ring main unit |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710174238.XA CN106771938B (en) | 2017-03-22 | 2017-03-22 | Method and device for identifying partial discharge mode of solid insulation ring main unit |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106771938A true CN106771938A (en) | 2017-05-31 |
| CN106771938B CN106771938B (en) | 2019-12-10 |
Family
ID=58967278
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710174238.XA Active CN106771938B (en) | 2017-03-22 | 2017-03-22 | Method and device for identifying partial discharge mode of solid insulation ring main unit |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106771938B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110672989A (en) * | 2019-09-17 | 2020-01-10 | 国网江西省电力有限公司电力科学研究院 | Self-adaptive partial discharge spectrogram phase correction method and correction system |
| CN111626374A (en) * | 2020-06-02 | 2020-09-04 | 上海电力大学 | Switch cabinet fault classification method based on semi-supervised learning |
| CN113064032A (en) * | 2021-03-26 | 2021-07-02 | 云南电网有限责任公司电力科学研究院 | Partial discharge mode identification method based on map features and information fusion |
| CN113156284A (en) * | 2021-04-28 | 2021-07-23 | 西安西电开关电气有限公司 | Method and device for processing interference signals of GIS equipment switching action |
| CN113933660A (en) * | 2021-08-31 | 2022-01-14 | 华能澜沧江水电股份有限公司 | Generator insulation fault degradation monitoring method based on partial discharge characteristics |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102540028A (en) * | 2012-01-13 | 2012-07-04 | 广东电网公司电力科学研究院 | Automatic pattern recognition method for partial discharge defects of generator stator insulation |
| CN102707203A (en) * | 2012-02-16 | 2012-10-03 | 安徽理工大学 | Discriminating and measuring method for partial discharge modes of transformer |
| CN103077402A (en) * | 2012-12-28 | 2013-05-01 | 湖北省电力公司电力科学研究院 | Transformer partial-discharging mode recognition method based on singular value decomposition algorithm |
| CN103076547A (en) * | 2013-01-24 | 2013-05-01 | 安徽省电力公司亳州供电公司 | Method for identifying GIS (Gas Insulated Switchgear) local discharge fault type mode based on support vector machines |
| CN104316846A (en) * | 2014-08-28 | 2015-01-28 | 国家电网公司 | Identification method, apparatus, and system of partial discharger mode of electrical equipment |
-
2017
- 2017-03-22 CN CN201710174238.XA patent/CN106771938B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102540028A (en) * | 2012-01-13 | 2012-07-04 | 广东电网公司电力科学研究院 | Automatic pattern recognition method for partial discharge defects of generator stator insulation |
| CN102707203A (en) * | 2012-02-16 | 2012-10-03 | 安徽理工大学 | Discriminating and measuring method for partial discharge modes of transformer |
| CN103077402A (en) * | 2012-12-28 | 2013-05-01 | 湖北省电力公司电力科学研究院 | Transformer partial-discharging mode recognition method based on singular value decomposition algorithm |
| CN103076547A (en) * | 2013-01-24 | 2013-05-01 | 安徽省电力公司亳州供电公司 | Method for identifying GIS (Gas Insulated Switchgear) local discharge fault type mode based on support vector machines |
| CN104316846A (en) * | 2014-08-28 | 2015-01-28 | 国家电网公司 | Identification method, apparatus, and system of partial discharger mode of electrical equipment |
Non-Patent Citations (1)
| Title |
|---|
| 张建文等: "基于自回归模型和超球面支持向量机的局部放电模式识别", 《电工电能新技术》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110672989A (en) * | 2019-09-17 | 2020-01-10 | 国网江西省电力有限公司电力科学研究院 | Self-adaptive partial discharge spectrogram phase correction method and correction system |
| CN110672989B (en) * | 2019-09-17 | 2021-12-31 | 国网江西省电力有限公司电力科学研究院 | Self-adaptive partial discharge spectrogram phase correction method and correction system |
| CN111626374A (en) * | 2020-06-02 | 2020-09-04 | 上海电力大学 | Switch cabinet fault classification method based on semi-supervised learning |
| CN113064032A (en) * | 2021-03-26 | 2021-07-02 | 云南电网有限责任公司电力科学研究院 | Partial discharge mode identification method based on map features and information fusion |
| CN113064032B (en) * | 2021-03-26 | 2022-08-02 | 云南电网有限责任公司电力科学研究院 | Partial discharge mode identification method based on map features and information fusion |
| CN113156284A (en) * | 2021-04-28 | 2021-07-23 | 西安西电开关电气有限公司 | Method and device for processing interference signals of GIS equipment switching action |
| CN113933660A (en) * | 2021-08-31 | 2022-01-14 | 华能澜沧江水电股份有限公司 | Generator insulation fault degradation monitoring method based on partial discharge characteristics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106771938B (en) | 2019-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106771938A (en) | A kind of solid insulation ring main unit Partial Discharge Pattern Recognition Method and device | |
| Brown et al. | Characterizing and quantifying noise in PMU data | |
| EP3832553A1 (en) | Method for identifying energy of micro-energy device on basis of bp neural network | |
| CN102185735B (en) | Network security situation prediction method | |
| CN105911476B (en) | A kind of battery energy storage system SOC prediction techniques based on data mining | |
| CN104155585A (en) | GIS partial discharge type identification method based on GK fuzzy clustering | |
| WO2006135175A1 (en) | Apparatus and method for testing state of charge in battery | |
| CN110850164A (en) | Multi-harmonic responsibility division method considering background harmonic voltage fluctuation and impedance change | |
| CN107220907B (en) | Harmonic pollution user grading method adopting rank-sum ratio comprehensive evaluation | |
| CN107064759A (en) | A kind of extra-high voltage equipment insulation defect type judgement method and system | |
| CN113138322A (en) | Method, device, equipment and storage medium for detecting high-resistance earth fault of power distribution network | |
| Karthikeyan et al. | Conception of complex probabilistic neural network system for classification of partial discharge patterns using multifarious inputs | |
| CN113408138B (en) | Lithium battery SOH estimation method and system based on secondary fusion | |
| Anderson et al. | Kullback-Leibler Divergence (KLD) based anomaly detection and monotonic sequence analysis | |
| CN106251861A (en) | A kind of abnormal sound in public places detection method based on scene modeling | |
| CN112636328A (en) | Medium-voltage distribution network equipment load state identification method | |
| Stepień | A comparison of classical and Bayesian interval estimation for long-term indicators of road traffic noise | |
| CN109031057A (en) | A kind of high-tension electricity apparatus local discharge mode identification method and device | |
| CN114912356A (en) | Construction method of segmented probability harmonic model of impact load | |
| CN106788816A (en) | A kind of channel status detection method and device | |
| Sundaram et al. | Characterization and localization of power quality disturbances based on S-transform and fuzzy expert system | |
| KR20050037242A (en) | System and method for detecting the failures in gas insulated switchgear using fuzzy inference | |
| CN115859690B (en) | Multi-level QMU (quality-of-the-tube) evaluation method and system for electromagnetic threat of equipment | |
| Chang et al. | Detection of latent fault in medium voltage distribution cables based on guiding learning model | |
| Butcher et al. | Pruning reservoirs with random static projections |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |