CN108828440A - High-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy - Google Patents
High-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy Download PDFInfo
- Publication number
- CN108828440A CN108828440A CN201810599664.2A CN201810599664A CN108828440A CN 108828440 A CN108828440 A CN 108828440A CN 201810599664 A CN201810599664 A CN 201810599664A CN 108828440 A CN108828440 A CN 108828440A
- Authority
- CN
- China
- Prior art keywords
- wavelet
- defect
- voltage circuitbreaker
- energy
- packet
- 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
- G01R31/327—Testing of circuit interrupters, switches or circuit-breakers
- G01R31/3271—Testing of circuit interrupters, switches or circuit-breakers of high voltage or medium voltage devices
- G01R31/3275—Fault detection or status indication
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The present invention relates to a kind of high-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy.This method includes：The current signal monitored is carried out coarse extraction, filters out the wave regions with notable feature, then obtain more pure divide-shut brake current waveform through denoising by the first step；Second step carries out three layers of wavelet packet to the original signal containing characteristic information and decomposes to extract the original feature vector that each frequency band energy is constituted；Third step carries out dimensionality reduction to original feature vector using principal component analysis, and prominent differences between samples characteristic further promotes data separability；Data with label are input in support vector machines by the 4th step, construct high-voltage circuitbreaker defect estimation model.The present invention can effectively check equipment deficiency situation, be able to satisfy high-voltage circuitbreaker defect diagonsis multiple requirements, high to the defect nicety of grading and detection accuracy of high-voltage circuitbreaker.
Description
Technical field
The present invention relates to the monitoring control technology of equipment for power transmission and distribution, specifically a kind of high pressure open circuit based on wavelet-packet energy
Device defect diagnostic method.
Background technique
High-voltage circuitbreaker with its huge quantity, extensive operation strategies and in the power system on-off load, excision therefore
The control and protection function of barrier, becomes important component indispensable in China's power grid.Therefore, to evade all kinds of accidents
Generation, the safety and stability of further safeguards system reduces the O&M cost of smart grid, promotes the development of power grid new technology,
It then widens the new approaches of high-voltage circuitbreaker feature extraction, failure (defect) diagnostic techniques is promoted to grow with each passing hour, mention for repair based on condition of component
It is necessary for more structurally sound diagnostic message.
Traditional failure (defect) diagnostic techniques includes method based on analytic modell analytical model, time-domain and frequency-domain analysis method, polynary
Statistical method and Knowledge based engineering method etc..In the selection of main monitoring signals, mainly there are vibration signal, divide-shut brake coil electricity
Stream, contact displacement or angle of eccentricity and power equipment image etc..Wherein, the conduct of divide-shut brake coil current covers high-voltage circuitbreaker and exists
The important symbol object of a large amount of key features of operation, not only monitoring is convenient, and compares other several signals, it includes it is disconnected
Road device failure (defect) information is more comprehensive, relatively broad by the identifiable failure of the signal and defect type, such as：It controls back
Road failure, iron core bite, coil aging etc., while the defect level of breaker more can also be distinguished meticulously, so using
Divide-shut brake coil current is as the main monitoring quantity of the present invention.It is diagnosed currently based on the feature extraction of the signal and failure (defect)
It is mostly to be realized by intelligent algorithm, specific extraction approach has respectively with diagnostic method, and time domain is asked extreme point and classification tree, is based on
Spline interpolation and multiple dimensioned linear fit and normal characteristics comparison, gray relative analysis method etc..
The above research achievement can be used with Example Verification and there are certain theoretical value and more practical values, but with science and technology
Progress, breaker type is various, feature extracting method mostly be the breaker for being directed to specific model, diagnostic method does not have
Universality is unable to satisfy the status monitoring of all types of breakers.Meanwhile in recent years operating mechanism as mechanical breakdown in breaker
The highest component of rate is the key object of status monitoring, and the purpose of status monitoring, is not only the essence to mechanical defect and failure
Really judgement, should more pay attention to trend and degree that defect state is changed react and identification, therefore be directed to breaker operator machine
The extent of disease severity and failure (defect) diagnostic method of structure defect are all the major issue for being difficult to solve with optimization for a long time.
Summary of the invention
To solve the above problems, the present invention, which provides one kind, can effectively check equipment deficiency situation, be able to satisfy height
Voltage breaker defect diagonsis multiple requirements are high based on wavelet packet to the defect nicety of grading and detection accuracy of high-voltage circuitbreaker
The high-voltage circuitbreaker defect diagnostic method of energy.
High-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy of the invention, includes the following steps：
Step 1, the current signal monitored carry out coarse extraction, filter out the wave regions with notable feature, then through disappearing
Processing of making an uproar obtains more pure divide-shut brake current waveform；
Step 2 carries out three layers of wavelet packet to the original signal containing characteristic information and decomposes to extract each frequency band energy structure
At original feature vector；
Step 3 carries out dimensionality reduction to original feature vector using principal component analysis, and prominent differences between samples characteristic further mentions
Rise data separability；
Data with label are input in support vector machines by step 4, construct high-voltage circuitbreaker defect estimation model.
Further, in the step 1, in high-voltage circuit-breaker switching on-off coil current signal, the wave with notable feature
The extraction in shape region is extracted and the variation to current value captures, and extracted data are after wavelet noise is handled
Unwanted noise is filtered out.
Further, in the step 2, three layers of wavelet packet is carried out to the original signal containing characteristic information and is decomposed to mention
The original feature vector that each frequency band energy is constituted is taken out to include the following steps：
(21) original signal containing characteristic information is set as m (t), then its energy is represented by：
M (t) is carried out by 3 layers of decomposition using db3 wavelet mother function, 8 small subband frequencies are obtained in third layer terminal node, frequency band
Energy distribution is different, then different frequency bands energy can be calculated as：Wherein, W (c, x) is wavelet systems
Number, Em(c, x) indicates the energy of x-th of subband of the c level of WAVELET PACKET DECOMPOSITION；
(22) after three layers of wavelet packet are decomposed, high-voltage circuitbreaker defective data is classified simultaneously according to the type of experiment
It numbers respectively, then each frequency band energy feature vector of one group of high-voltage circuit-breaker switching on-off coil current can be set asWherein, a table
Show in simulated experiment, a kind operating status (a=1,2 ..., q of high-voltage circuitbreaker；Q >=2 and q ∈ N+)；B indicates a state
B group data (b=1,2,3,4,5..., p；P >=5 and p ∈ N+), then the high-voltage circuitbreaker based on wavelet-packet energy is respectively run
The feature samples E of statemFor
Further, in the step 3, carrying out dimensionality reduction to original feature vector using principal component analysis includes following step
Suddenly：
(31) first by feature samples space EmDimension is eliminated in standardization
Wherein, i=1,2 ... s；J=1,2,3 ..., u；N is sample size,For the sample average of j-th of variable, σj
For the standard deviation of the variable；
(32) correlation matrix is set as X, meets xi,j=xj,iAnd xi,i=1, then
(33) it calculates the characteristic value of X and sorts from large to small
λ1≥λ2≥λ3≥L≥λj≥0 (5)
Then corresponding feature vector is represented by l1,l2,l3,…,lj, and it is arranged by the descending of characteristic value；
(34) variance contribution ratio α is calculatedj
To αjDescending sort simultaneously calculates accumulative variance contribution ratioIt sets and works as in the present inventionWhen greater than 90%, then only
Retain the 1st to h-th of principal component, and thinks that wherein included enough keys for reflecting high-voltage circuitbreaker defect are special
Reference breath, principal component is taken out and is set as transformation matrix Pt, then EmFinal feature vector Em' be represented by
Em'=Em×Pt (7)。
The present invention is based on the high-voltage circuitbreaker defect diagnostic methods of wavelet-packet energy to have the advantages that：
1, wavelet-packet energy is replaced into the characteristic value in original waveform, eliminates the cumbersome mistake for extracting current signal extreme value
Journey effectively eliminates because extracting characteristic value inaccuracy due to bring Error Diagnostics, improves diagnostic accuracy；2, PCA algorithm draws
Entering to realize simplifying for data volume and being effectively retained for defect information, reduces Euclidean distance in class, Euclidean distance increases between class,
That is the association of same kind data is even closer, and different data distribution more disperses, and improves the separability of data, same in processing
Its advantage is embodied in the different set of metadata of similar data of seed type defect, defect level, experimental data proves that the algorithm is high diagnosis essence
The effective guarantee means of degree；3, the present invention is based on divide-shut brake coil current signal, in conjunction with wavelet packet analysis, PCA and SVM algorithm,
Building integrates the high-voltage circuitbreaker defect diagonsis model of the thinkings such as feature extraction, data-optimized and classification diagnosis, is analyzing
In high-voltage circuitbreaker defect and defect level, accuracy and application value with higher；4, smart grid fortune is improved
Dimension is horizontal, has widened thinking of the high-voltage circuitbreaker defect diagonsis in terms of defect severity, has helped to formulate more reasonable inspection
Scheme is repaired to reduce operation expense on ordinary days, the requirement of high-voltage circuitbreaker defect diagonsis various aspects can be met, theoretical value is answered
It is huge with value and prospect,
Detailed description of the invention
Fig. 1 a is high-voltage circuitbreaker normal operating condition and two types defective waveform comparison diagram；
Fig. 1 b is high-voltage circuitbreaker normal operating condition and analog control loop defective waveform comparison diagram；
Fig. 1 c is high-voltage circuitbreaker normal operating condition and simulation iron core bite defective waveform comparison diagram；
Fig. 2 is three layers of decomposition diagram of wavelet packet；
Fig. 3 is the measured waveform under 6 kinds of operating statuses
Wherein：Fig. 3 a is normal operating condition；
Fig. 3 b is 50 Europe of control loop series connection；
Fig. 3 c is 100 Europe of control loop series connection；
Fig. 3 d is that iron core end hangs weight m1；
The e iron core end Fig. 3 hangs weight m2；
Fig. 3 f is that iron core end hangs weight m3
Fig. 4 is divide-shut brake coil current waveform comparison diagram before and after data prediction
Fig. 5 is to handle SVM diagnostic result without PCA；
Fig. 6 is by PCA treated SVM diagnostic result；
Present invention is further described in detail with specific embodiment with reference to the accompanying drawing：
Specific embodiment
The present invention provides a kind of high-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy, for high-voltage circuitbreaker
Defect classification accuracy is high, improves smart grid O&M level and has widened high-voltage circuitbreaker defect diagonsis in defect severity
The thinking of aspect can effectively check equipment deficiency, help to formulate more reasonable maintenance solution to reduce operation and maintenance on ordinary days
Cost etc., can meet the requirement of high-voltage circuitbreaker defect diagonsis various aspects, and theoretical value, application value and prospect are huge.
As shown, the present invention provides a kind of high-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy, including
Following steps：
Step 1, the current signal monitored carry out coarse extraction, filter out the wave regions with notable feature, then through disappearing
Processing of making an uproar obtains more pure divide-shut brake current waveform；
Step 2 carries out three layers of wavelet packet to the original signal containing characteristic information and decomposes to extract each frequency band energy structure
At original feature vector；
Step 3 carries out dimensionality reduction to original feature vector using principal component analysis, and prominent differences between samples characteristic further mentions
Rise data separability；
Data with label are input in support vector machines by step 4, construct high-voltage circuitbreaker defect estimation model.
In above-mentioned steps one, in high-voltage circuit-breaker switching on-off coil current signal, the wave regions with notable feature
Extraction is extracted and the variation to current value captures, and extracted data are made an uproar after wavelet noise is handled by extra
Sound filters out, the characteristic partial enlargement comparison diagram in high-voltage circuit breaker lock coil current waveform as shown in 1a- Fig. 1 c, with
Normal operating condition is compared, and coil current reduces when control loop defect, and iron core movement is sluggish, and opening time is with series resistance
The increase of value and increase；And influence of the iron core bite to opening coil current value is smaller, but opening time is hung with iron core end
The increase of weight and increase.
Further, the present invention in step 2, to containing characteristic information original signal carry out three layers of wavelet packet decompose with
The original feature vector that each frequency band energy is constituted is extracted to include the following steps：
(21) original signal containing characteristic information is set as m (t), then its energy is represented by：
M (t) is carried out by 3 layers of decomposition using db3 wavelet mother function, 8 small subband frequencies, wavelet packet is obtained in third layer terminal node
Three layers of decomposition diagram are as shown in Fig. 2, the Energy distribution of frequency band is different, then different frequency bands energy can be calculated as：Wherein, W (c, x) is wavelet coefficient, EmThe xth of (c, x) expression c-th of level of WAVELET PACKET DECOMPOSITION
The energy of a subband.
(22) after three layers of wavelet packet are decomposed, high-voltage circuitbreaker defective data is classified simultaneously according to the type of experiment
It numbers respectively, then each frequency band energy feature vector of one group of high-voltage circuit-breaker switching on-off coil current can be set asWherein, a table
Show in simulated experiment, a kind operating status (a=1,2 ..., q of high-voltage circuitbreaker；Q >=2 and q ∈ N+)；B indicates a state
B group data (b=1,2,3,4,5..., p；P >=5 and p ∈ N+), then the high-voltage circuitbreaker based on wavelet-packet energy is respectively run
The feature samples E of statemFor
In above-mentioned steps three, dimensionality reduction is carried out to original feature vector using principal component analysis and is included the following steps：
(31) first by feature samples space EmDimension is eliminated in standardization
Wherein, i=1,2 ... s；J=1,2,3 ..., u；N is sample size,For the sample average of j-th of variable, σj
For the standard deviation of the variable.
(32) correlation matrix is set as X, meets xi,j=xj,iAnd xi,i=1, then
(33) it calculates the characteristic value of X and sorts from large to small
λ1≥λ2≥λ3≥L≥λj≥0 (5)
Then corresponding feature vector is represented by l1,l2,l3,…,lj, and it is arranged by the descending of characteristic value.
(34) variance contribution ratio α is calculatedj
To αjDescending sort simultaneously calculates accumulative variance contribution ratioIt sets and works as in the present inventionWhen greater than 90%, then only
Retain the 1st to h-th of principal component, and thinks that wherein included enough keys for reflecting high-voltage circuitbreaker defect are special
Reference breath, principal component is taken out and is set as transformation matrix Pt, then EmFinal feature vector Em' be represented by
Em'=Em×Pt (7)
Embodiment
The present invention is directed to the 10kV breaker of certain electrical equipment company production, is based on its spring operating mechanism, builds experiment
Platform simultaneously carries out defects simulation experiment, the data source by acquisition divide-shut brake coil current as defect diagonsis.The present invention is total
Design two major classes typical defect, the respectively aging of divide-shut brake coil and iron core bite.Under its coil aging again be divided into slightly with
Two kinds of degree of agings of moderate；Iron core bite is divided into slight, three kinds of bite degree of moderate and severe.In experiment, divide-shut brake coil
Aging realized by sealing in adjustable resistor in divide-shut brake coil control circuit, and by resistance value is adjusted to respectively 50 with
100 Ω simulate the coil aging of slight and moderate.Iron core bite is then by divide-shut brake iron-core coil underhung weight
It is simulated with hindering its movement in divide-shut brake, and by gradually increasing weight quality to m1(0.1kg)、m2
(0.2kg)、 m3(0.3kg), with simulate iron core bite it is light, in and severe these three defect levels.According to table 1, experimental data
Including breaker normal operation, 2 kinds of different degrees of coil agings and 3 kinds of different degrees of iron core bites, amount to 6 kinds of operation shapes
State, measured waveform are as shown in Figure 3.When practical operation experiment, algorithm design are with data verification, switching on and off experimental data is
Verify feasible, diagnostic method has versatility, therefore is only analyzed herein with separating brake data instance.
Defect type, defect severity, analogy method and the tag along sort of circuit breaker operation mechanism in the experiment of table 1
Data prediction is carried out first.Treatment process is：The crude sampling for being first 125000*1 by obtained data dimension
Signal (sample frequency 3.2*10-6) according to threshold method, there will be the waveform extracting of feature to come out, then by the signal extracted through small
After wave soft-threshold de-noising, it is unified for same dimension, guarantees that all raw data forms are consistent, to carry out subsequent analysis.With just
For divide-shut brake coil current under normal state, treatment process is as shown in Figure 4.
According to the data obtained after pretreatment, each 50 groups, total 300 groups of experimental datas progress are chosen under different operating statuses
Three layers of wavelet packet decomposition, each band energy constitute 8*300 primitive character sample Em。
After original current signal is by coarse extraction and wavelet noise, each band energy constitutes 8* after three layers of wavelet packet are decomposed
300 primitive character sample EmAs shown in table 2.
2 primitive character sample E of tablem
By EmIt is all sent into dimensionality reduction in PCA and obtains Em', and contribution rate of accumulative total is calculated, the h=3 when it reaches 90%, therefore
The principal component in original sample space is as shown in table 3.
3 dimensionality reduction data E of tablem’
240 groups of dimensionality reduction data composing training collection for having been defined as training sample are extracted, remaining 60 groups are used as test set,
Actual sample type and SVM the identification sample type comparing result for optimizing without PCA and optimizing through PCA are as shown in Figure 5 and Figure 6.
Experimental result shows that, when total sample number is 300 groups according to amount, the wavelet-packet energy extracted from current data is handled by PCA
Afterwards, SVM can reach the accuracy rate of diagnosis of normal condition and coil aging and iron core bite this two major classes defect type
100%, defect extent of disease severity accuracy rate also can reach 100%, illustrate that the above two class defect characteristics are extracted efficiently and answer
With svm classifier policy selection is proper.And the data sample without PCA optimization, it is 100% in the accuracy rate of diagnosis of defect type,
But in defect extent of disease severity, second with the third operating status, i.e., the slight aging of coil and coil mittlere alterung this
Two kinds of defect levels get the wrong sow by the ear, and mittlere alterung is mistaken for slight aging, so that the recognition correct rate of defect level is reduced to
98.33%, it is slightly poorer than the diagnosis effect through the processed sample of PCA.Further, it calculates with the result of analysis Euclidean distance such as
Shown in table 4, Euclidean distance is the Euclidean distance of the operating status and normal operating condition between the class in table, and Euclidean distance is in class
The operating status and the Euclidean distance between itself.Through data comparison it is found that Euclidean distance reduces in class, Euclidean distance increases between class
Greatly, i.e. the association of same kind data is even closer, and different data distribution more disperses, and also the opposite PCA that demonstrates is to raising number
According to the influence of separability.
Table 4 calculates influence of the Euclidean distance verifying PCA to data separability
The above described is only a preferred embodiment of the present invention, being not the limit for making any other form to the present invention
System, and made any modification or equivalent variations according to the technical essence of the invention, still fall within present invention model claimed
It encloses.
Claims (4)
1. a kind of high-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy, it is characterised in that：Include the following steps,
Step 1, the current signal that monitors carry out coarse extraction, filter out the wave regions with notable feature, then through de-noising at
Reason obtains more pure divide-shut brake current waveform；
Step 2 carries out what three layers of wavelet packet decomposition were constituted to extract each frequency band energy to the original signal containing characteristic information
Original feature vector；
Step 3 carries out dimensionality reduction to original feature vector using principal component analysis, and prominent differences between samples characteristic further promotes number
According to separability；
Data with label are input in support vector machines by step 4, construct high-voltage circuitbreaker defect estimation model.
2. a kind of high-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy according to claim 1, feature exist
In：In the step 1, in high-voltage circuit-breaker switching on-off coil current signal, the extraction of the wave regions with notable feature is
It is extracted and the variation to current value captures, extracted data filter unwanted noise after wavelet noise is handled
It removes.
3. a kind of high-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy according to claim 1, feature exist
In：In the step 2, three layers of wavelet packet are carried out to the original signal containing characteristic information and is decomposed to extract each frequency band energy
The original feature vector of composition includes the following steps：
(21) original signal containing characteristic information is set as m (t), then its energy is expressed as：Using db3
M (t) is carried out 3 layers of decomposition by wavelet mother function, and 8 small subband frequencies, the energy point of frequency band is obtained in third layer terminal node
Cloth is different, then different frequency bands energy balane is：Wherein, W (c, x) is wavelet coefficient, Em(c,x)
Indicate the energy of x-th of subband of c-th of level of WAVELET PACKET DECOMPOSITION；
(22) after three layers of wavelet packet are decomposed, high-voltage circuitbreaker defective data is classified and is distinguished according to the type of experiment
Number, then each frequency band energy feature vector of one group of high-voltage circuit-breaker switching on-off coil current can be set asWherein, a indicates mould
In draft experiment, a kind operating status (a=1,2 ..., q of high-voltage circuitbreaker；Q >=2 and q ∈ N+)；The b of b expression a state
Group data (b=1,2,3,4,5..., p；P >=5 and p ∈ N+), then each operating status of the high-voltage circuitbreaker based on wavelet-packet energy
Feature samples EmFor
4. a kind of high-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy according to claim 1, feature exist
In：In the step 3, dimensionality reduction is carried out to original feature vector using principal component analysis and is included the following steps：
(31) first by feature samples space EmDimension is eliminated in standardization
Wherein, i=1,2 ... s；J=1,2,3 ..., u；N is sample size,For the sample average of j-th of variable, σjFor this
The standard deviation of variable；
(32) correlation matrix is set as X, meets xi,j=xj,iAnd xi,i=1, then
(33) it calculates the characteristic value of X and sorts from large to small
λ1≥λ2≥λ3≥L≥λj≥0 (5)
Then corresponding feature vector is represented by l1,l2,l3,…,lj, and it is arranged by the descending of characteristic value；
(34) variance contribution ratio α is calculatedj
To αjDescending sort simultaneously calculates accumulative variance contribution ratioIt sets and works as in the present inventionWhen greater than 90%, then only retain
1st to h-th of principal component, and think wherein it is included it is enough reflect high-voltage circuitbreaker defect key feature letter
Breath, principal component is taken out and is set as transformation matrix Pt, then EmFinal feature vector Em' be expressed as
Em'=Em×Pt (7)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810599664.2A CN108828440A (en) | 2018-06-12 | 2018-06-12 | High-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810599664.2A CN108828440A (en) | 2018-06-12 | 2018-06-12 | High-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108828440A true CN108828440A (en) | 2018-11-16 |
Family
ID=64144705
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810599664.2A Pending CN108828440A (en) | 2018-06-12 | 2018-06-12 | High-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108828440A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110207974A (en) * | 2019-07-10 | 2019-09-06 | 西南交通大学 | Circuit breaker failure recognition methods based on vibration signal time-frequency energy-distributing feature |
| CN113589166A (en) * | 2021-07-15 | 2021-11-02 | 上海电力大学 | Data-drive-based variable frequency motor end insulation state online monitoring method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104237777A (en) * | 2014-08-14 | 2014-12-24 | 国家电网公司 | Support vector machine high-voltage circuit breaker fault diagnosis method based on core principal component analysis |
| CN105891707A (en) * | 2016-05-05 | 2016-08-24 | 河北工业大学 | Opening-closing fault diagnosis method for air circuit breaker based on vibration signals |
| CN106482937A (en) * | 2016-09-30 | 2017-03-08 | 南方电网科学研究院有限责任公司 | A kind of monitoring method of mechanical state of high-voltage circuit breaker |
| CN107329079A (en) * | 2017-07-28 | 2017-11-07 | 河海大学 | A kind of primary cut-out on-line monitoring and synthetic fault diagnosis system |
| CN107505133A (en) * | 2017-08-10 | 2017-12-22 | 滁州学院 | The probability intelligent diagnosing method of rolling bearing fault based on adaptive M RVM |
-
2018
- 2018-06-12 CN CN201810599664.2A patent/CN108828440A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104237777A (en) * | 2014-08-14 | 2014-12-24 | 国家电网公司 | Support vector machine high-voltage circuit breaker fault diagnosis method based on core principal component analysis |
| CN105891707A (en) * | 2016-05-05 | 2016-08-24 | 河北工业大学 | Opening-closing fault diagnosis method for air circuit breaker based on vibration signals |
| CN106482937A (en) * | 2016-09-30 | 2017-03-08 | 南方电网科学研究院有限责任公司 | A kind of monitoring method of mechanical state of high-voltage circuit breaker |
| CN107329079A (en) * | 2017-07-28 | 2017-11-07 | 河海大学 | A kind of primary cut-out on-line monitoring and synthetic fault diagnosis system |
| CN107505133A (en) * | 2017-08-10 | 2017-12-22 | 滁州学院 | The probability intelligent diagnosing method of rolling bearing fault based on adaptive M RVM |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110207974A (en) * | 2019-07-10 | 2019-09-06 | 西南交通大学 | Circuit breaker failure recognition methods based on vibration signal time-frequency energy-distributing feature |
| CN113589166A (en) * | 2021-07-15 | 2021-11-02 | 上海电力大学 | Data-drive-based variable frequency motor end insulation state online monitoring method |
| CN113589166B (en) * | 2021-07-15 | 2023-11-24 | 上海电力大学 | Method for online monitoring of insulation state of end part of variable frequency motor based on data driving |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109444656B (en) | Online diagnosis method for deformation position of transformer winding | |
| CN110926778B (en) | Mechanical fault diagnosis method for gas insulated switchgear assembly based on abnormal vibration | |
| CN112014047B (en) | Mechanical fault diagnosis method for on-load tap-changer | |
| CN109633368A (en) | The method of duration power quality disturbances containing distributed power distribution network based on VMD and DFA | |
| CN103968937B (en) | A kind of distribution switch mechanical state diagnostic method based on EMD Sample Entropies and FCM | |
| CN105891707A (en) | Opening-closing fault diagnosis method for air circuit breaker based on vibration signals | |
| CN110488164B (en) | High-voltage cable insulation aging state comprehensive assessment early warning method and system | |
| CN105424395A (en) | Method and device for determining equipment fault | |
| CN106526468B (en) | Circuit-breaker status detection method based on wave character identification | |
| CN108828440A (en) | High-voltage circuitbreaker defect diagnostic method based on wavelet-packet energy | |
| CN112069930A (en) | Vibration signal processing method and device for improving GIS equipment fault diagnosis accuracy | |
| CN110320467A (en) | A kind of low-voltage direct circuit breaker failure diagnostic method | |
| CN112462193B (en) | Automatic reclosing judgment method for power distribution network based on real-time fault filtering data | |
| CN108898274A (en) | A kind of power scheduling log defect classification method | |
| CN111239672B (en) | Machine learning algorithm-based gradient fault prediction method for optical fiber current transformer | |
| CN114298175A (en) | Power equipment state monitoring and fault early warning method and system based on edge calculation | |
| CN105241643A (en) | High-voltage circuit breaker mechanical state monitoring method employing HS transformation and single-type support vector machine | |
| CN113745049B (en) | Vacuum degree monitoring method and system in vacuum arc extinguish chamber | |
| CN111999591A (en) | Method for identifying abnormal state of primary equipment of power distribution network | |
| CN111474471B (en) | Method for extracting current characteristic parameters of opening and closing coil of high-voltage alternating-current circuit breaker | |
| CN113191247A (en) | GIS equipment mechanical fault diagnosis method and system | |
| CN112132226A (en) | Method for diagnosing defects of +/-1100 kV composite insulator | |
| CN111948531A (en) | On-load tap-changer live-line detection method based on multi-parameter extraction | |
| CN110929673A (en) | Transformer winding vibration signal identification method based on ITD (inverse discrete cosine transformation) permutation entropy and CGWO-SVM (Carrier-support vector machine) | |
| CN110132567A (en) | A kind of OLTC method for diagnosing faults based on LCD and arrangement entropy |
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 |