CN105842588B - A kind of method and system for correcting ultrasonic wave Partial Discharge Detection - Google Patents
A kind of method and system for correcting ultrasonic wave Partial Discharge Detection Download PDFInfo
- Publication number
- CN105842588B CN105842588B CN201610156063.5A CN201610156063A CN105842588B CN 105842588 B CN105842588 B CN 105842588B CN 201610156063 A CN201610156063 A CN 201610156063A CN 105842588 B CN105842588 B CN 105842588B
- Authority
- CN
- China
- Prior art keywords
- signal
- local discharge
- disaggregated model
- linear spectral
- pure noise
- 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.)
- Active
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
- G01R31/1209—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 using acoustic measurements
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
- G01R31/1263—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation
- G01R31/1272—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation of cable, line or wire insulation, e.g. using partial discharge measurements
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The present invention provides a kind of method for correcting ultrasonic wave Partial Discharge Detection, including extracting multiple local discharge signal linear spectral post-processings from historical data, the principal component feature of coefficient matrix and each local discharge signal is obtained, constructs the disaggregated model of local discharge signal and per a kind of linear spectral mean value；The linear spectral mean value for obtaining pure noise signal obtains the disaggregated model of pure noise ultrasonic signal according to coefficient matrix；Local discharge signal is added per a kind of linear spectral mean value with the linear spectral mean value of pure noise signal, the disaggregated model of local discharge signal is corrected；Suspect signal is obtained, the principal component feature of suspect signal is obtained according to coefficient matrix and is introduced into pure noise and the revised disaggregated model of local discharge signal, the signal that Euclidean distance minimum corresponds in disaggregated model is filtered out and exports.Implement the present invention, avoids artifact from interfering, introduced feature dimension is low, and training data is few, to reach the time saving and energy saving and objective purpose of testing result.
Description
Technical field
The present invention relates to local discharge signal detection technique field more particularly to a kind of amendment ultrasonic wave Partial Discharge Detections
Method and system.
Background technology
In order to ensure the safe and stable operation of equipment, insulating properties detection and fault diagnosis need to be carried out, and shelf depreciation is
The main reason for causing apparatus insulated accident to occur, thus it is real-time to equipment progress shelf depreciation using ultrasonic detection method
Detection, can grasp equipment real-time operating conditions comprehensively, and can to latter section of duration in state of insulation predict, simultaneously
Suitable maintenance and maintenance strategy are taken according to its insulation status, this is to ensureing that the reliable and stable operation of power supply system has very
Positive meaning.
The use currently, detection device based on ultrasound examination shelf depreciation technology has put goods on the market, common equipment have
Two kinds：One is by piezoelectric principle, after converting ultrasonic signal to current signal, sound is translated into through internal processor
Frequency signal, to which the presence or absence of local discharge signal can be judged by high-fidelity headphone to monitor the abnormal of audio signal；It is another
Kind is to realize detection by artificial given threshold and warning function, passes through dB values and shows the big of Processing of Partial Discharge Ultrasonic Signals
It is small.
Some following problems can be all encountered during above two equipment is actually detected at the scene：It was detected in the first equipment
Cheng Zhong, due to the difference of everyone auditory physiology characteristic, testing staff has not the audio signal in the earphone that listens to
The subjective initiative of same judgement, testing result and testing staff are closely bound up, be easy to cause erroneous judgement；It is detected in second of equipment
In the process, in the majority by experience due to breakdown judge, system detectio reliability is not high.
Therefore, in order to more accurately detect ultrasonic wave local discharge signal, some mode identification technology quilts based on statistics
It is applied in real system, but there is also following insufficient in the application for these technologies：(1) intrinsic dimensionality is more, needs a large amount of
Training data, and local discharge signal is usually not easy to collect；(2) when training environment and practical service environment mismatch,
Performance can be remarkably decreased.
Invention content
Technical problem to be solved of the embodiment of the present invention is, provides a kind of side for correcting ultrasonic wave Partial Discharge Detection
Method and system can avoid artifact from interfering, and introduced feature dimension is low, and training data is few, to reach it is time saving and energy saving and
The objective purpose of testing result.
In order to solve the above-mentioned technical problem, an embodiment of the present invention provides a kind of sides for correcting ultrasonic wave Partial Discharge Detection
Method, the method includes：
A, from the historical data of equipment high s/n ratio ultrasonic wave local discharge signal, multiple local discharge signals are extracted,
And multiple local discharge signals of the extraction are converted into after linear spectral and are handled by Principal Component Analysis, obtain P dimensions
Principal component coefficient matrix, and principal component coefficient matrix is tieed up according to the obtained P, obtain the P dimension masters of each local discharge signal
The P of obtained each local discharge signal dimension principal component features are further carried out Unsupervised clustering, obtained by composition characteristics
The disaggregated model of local discharge signal and its corresponding linear spectral mean value per one kind；Wherein, P is natural number；
B, pure noise ultrasonic signal is obtained, and the pure noise ultrasonic signal got is converted into linear spectral
Afterwards, the linear spectral mean value of pure noise ultrasonic signal is calculated, and principal component coefficient matrix is tieed up according to the obtained P, it will
The linear spectral mean value of the pure noise ultrasonic signal is converted to the P dimension principal component features of the pure noise ultrasonic signal
Afterwards, the disaggregated model of pure noise ultrasonic signal is obtained；
C, will surpass respectively with the pure noise per a kind of linear spectral mean value in the local discharge signal disaggregated model
The linear spectral mean value of acoustic signals disaggregated model is added, and the linear spectral mean value after being added per one kind, and root are obtained
Principal component coefficient matrix is tieed up according to the obtained P, the obtained linear spectral mean value after being added per one kind is converted to
After corresponding P dimensions principal component feature, obtained disaggregated model is as the revised disaggregated model of local discharge signal；
D, suspect signal is obtained, and the suspect signal got is converted into linear spectral, and is obtained according to described
P tie up principal component coefficient matrix, the linear spectral of the suspect signal is converted to the P Wei Zhuchengfente of the suspect signal
Sign；
E, the P of suspect signal dimension principal component features are introduced to the disaggregated model of the pure noise ultrasonic signal respectively
In the revised disaggregated model of the local discharge signal, calculates the suspect signal and arrive the pure noise ultrasonic wave respectively
The Euclidean distance of the disaggregated model of signal and the revised disaggregated model of the local discharge signal, and according to two obtained
A Euclidean distance filters out the corresponding disaggregated model of minimum value in described two Euclidean distances, further by institute's sifting sort mould
Corresponding signal is exported as the testing result of the suspect signal in type.
Wherein, the step a is specifically included：
From the historical data of equipment high s/n ratio ultrasonic wave local discharge signal, multiple local discharge signals are extracted, and
After multiple local discharge signals of the extraction are carried out framing as unit of certain time length, to each office after the framing
Portion's discharge signal carries out Fourier transformation, obtains the linear spectral of each local discharge signal；
Dimension-reduction treatment is carried out to the linear spectral of any local discharge signal by Principal Component Analysis, obtain P dimensions it is main at
Point coefficient matrix, and principal component coefficient matrix is tieed up according to the obtained P, to each local discharge signal of the extraction into
Row linear transformation obtains the P dimension principal component features of each local discharge signal；
The P of obtained each local discharge signal dimension principal component features are carried out without prison simultaneously using k-mean algorithms
Cluster is superintended and directed, multiple cluster centres are calculated and preserves the disaggregated model as local discharge signal；
The principle averaged after being added according to same class linear spectral, in the disaggregated model of the local discharge signal
It is calculated, is obtained in the disaggregated model of the local discharge signal per a kind of linear spectral per the linear spectral contained by one kind
Mean value.
Wherein, the step b is specifically included：
Before obtaining the suspect signal, pure noise ultrasonic signal of the actual environment for a period of time in range is obtained,
And after the pure noise ultrasonic signal got carried out framing as unit of certain time length, make an uproar to pure after the framing
Sound ultrasonic signal carries out Fourier transformation, obtains the linear spectral of the pure noise ultrasonic signal；
It averages after the linear spectral of the obtained pure noise ultrasonic signal is added, it is super as the pure noise
The linear spectral mean value of acoustic signals；
Principal component coefficient matrix is tieed up according to the obtained P, by the linear spectral mean value of the pure noise ultrasonic signal
After the P dimension principal component features for being converted to the pure noise ultrasonic signal, the disaggregated model for only including a cluster centre is obtained
And preserve disaggregated model as the pure noise ultrasonic signal.
Wherein, the step c is specifically included：
It determines per a kind of linear spectral mean value in the local discharge signal disaggregated model, and the pure noise is ultrasonic
The linear spectral mean value of wave Modulation recognition model is added with the linear spectral mean value per one kind of the determination respectively, is added
The linear spectral mean value per one kind afterwards；
Principal component coefficient matrix is tieed up according to the obtained P, by the obtained linear spectral after being added per one kind
After mean value is converted to corresponding P dimensions principal component feature, obtains the only disaggregated model comprising a cluster centre and preserve as institute
State the revised disaggregated model of local discharge signal.
Wherein, the step e is specifically included：
By the P of suspect signal dimension principal component feature introduce respectively the pure noise ultrasonic signal disaggregated model and
In the revised disaggregated model of local discharge signal, calculates the suspect signal and arrive the pure noise ultrasonic wave letter respectively
Number disaggregated model and the revised disaggregated model of the local discharge signal Euclidean distance；
Judge whether the suspect signal is less than institute to the Euclidean distance of the disaggregated model of the pure noise ultrasonic signal
State suspect signal to the revised disaggregated model of the local discharge signal Euclidean distance；
If it is, the disaggregated model of the screening is the disaggregated model of the pure noise ultrasonic signal, and will be described
Suspect signal is that pure noise ultrasonic signal is exported as testing result；
If it is not, then the disaggregated model of the screening is the revised disaggregated model of the local discharge signal, and by institute
It is that local discharge signal is exported as testing result to state suspect signal.
The embodiment of the present invention additionally provides a kind of system for correcting ultrasonic wave Partial Discharge Detection, the system comprises：
Local discharge signal analysis model acquiring unit, for going through from equipment high s/n ratio ultrasonic wave local discharge signal
In history data, multiple local discharge signals are extracted, and multiple local discharge signals of the extraction are converted into linear spectral
It is handled afterwards by Principal Component Analysis, obtains P dimension principal component coefficient matrixes, and principal component coefficient square is tieed up according to the obtained P
Battle array obtains the P dimension principal component features of each local discharge signal, further by the P of obtained each local discharge signal
It ties up principal component feature and carries out Unsupervised clustering, obtain the disaggregated model of local discharge signal and its corresponding linear spectral per one kind
Mean value；Wherein, P is natural number；
Pure noise ultrasound signal analyzing model acquiring unit, is obtained for obtaining pure noise ultrasonic signal, and by described
After the pure noise ultrasonic signal got is converted to linear spectral, the linear spectral mean value of pure noise ultrasonic signal is calculated,
And principal component coefficient matrix is tieed up according to the obtained P, the linear spectral mean value of the pure noise ultrasonic signal is converted to
After the P dimension principal component features of the pure noise ultrasonic signal, the disaggregated model of pure noise ultrasonic signal is obtained；
Local discharge signal analysis model amending unit is used for every one kind in the local discharge signal disaggregated model
Linear spectral mean value is added with the linear spectral mean value of the pure noise ultrasonic signal disaggregated model respectively, is added
The linear spectral mean value per one kind afterwards, and principal component coefficient matrix is tieed up according to the obtained P, after the obtained addition
Corresponding P dimensions principal component feature is converted to per a kind of linear spectral mean value after, obtained disaggregated model is as shelf depreciation
Disaggregated model after signal correction；
Suspect signal analytic unit is converted to linearly for obtaining suspect signal, and by the suspect signal got
Frequency spectrum, and principal component coefficient matrix is tieed up according to the obtained P, the linear spectral of the suspect signal is converted to described to be checked
The P of signal ties up principal component feature；
Suspect signal recognition unit, for the P dimension principal component features of the suspect signal to be introduced the pure noise respectively
In the disaggregated model of ultrasonic signal and the revised disaggregated model of the local discharge signal, the suspect signal point is calculated
Be clipped to the pure noise ultrasonic signal disaggregated model and the revised disaggregated model of the local discharge signal Euclidean away from
From, and according to two obtained the Euclidean distance, the corresponding disaggregated model of minimum value in described two Euclidean distances is filtered out,
Further exported signal corresponding in institute's sifting sort model as the testing result of the suspect signal.
Wherein, the local discharge signal analysis model acquiring unit includes：
Local discharge signal linear spectral conversion module, for going through from equipment high s/n ratio ultrasonic wave local discharge signal
In history data, multiple local discharge signals are extracted, and by multiple local discharge signals of the extraction with certain time length for singly
After position carries out framing, Fourier transformation is carried out to each local discharge signal after the framing, obtains each shelf depreciation
The linear spectral of signal；
Principal component analysis module, for being dropped to the linear spectral of any local discharge signal by Principal Component Analysis
Dimension processing obtains P dimension principal component coefficient matrixes, and ties up principal component coefficient matrix according to the obtained P, to the every of the extraction
One local discharge signal carries out linear transformation, obtains the P dimension principal component features of each local discharge signal；
Local discharge signal Clustering Model computing module, for using k-mean algorithms while to obtained each office
The P dimension principal component features of portion's discharge signal carry out Unsupervised clustering, calculate multiple cluster centres and preserve as shelf depreciation
The disaggregated model of signal；
Local discharge signal Clustering Model classification mean value computation module, for asking flat after being added according to same class linear spectral
The principle of mean value obtains institute to being calculated per the linear spectral contained by one kind in the disaggregated model of the local discharge signal
It states in the disaggregated model of local discharge signal per a kind of linear spectral mean value.
Wherein, the pure noise ultrasound signal analyzing model acquiring unit includes：
Pure noise ultrasonic signal linear spectral conversion module, it is practical for before obtaining the suspect signal, obtaining
The environment pure noise ultrasonic signal in range for a period of time, and by the pure noise ultrasonic signal got with a timing
Length is after unit carries out framing, to carry out Fourier transformation to the pure noise ultrasonic signal after the framing, obtain described pure
The linear spectral of noise ultrasonic signal；
Linear spectral mean value computation module, after being added the linear spectral of the obtained pure noise ultrasonic signal
It averages, the linear spectral mean value as the pure noise ultrasonic signal；
Pure noise ultrasonic signal Clustering Model computing module, the P for being obtained according to tie up principal component coefficient matrix,
The P that the linear spectral mean value of the pure noise ultrasonic signal is converted to the pure noise ultrasonic signal ties up principal component feature
Afterwards, it obtains the only disaggregated model comprising a cluster centre and preserves the disaggregated model as the pure noise ultrasonic signal.
Wherein, the local discharge signal analysis model amending unit includes：
Linear spectral accumulator module, it is equal per a kind of linear spectral in the local discharge signal disaggregated model for determining
Value, and by the linear spectral mean value of the pure noise ultrasonic signal disaggregated model respectively with the determination per the linear of one kind
Spectrum mean is added, and is obtained after being added per a kind of linear spectral mean value；
Local discharge signal Clustering Model correcting module, the P for being obtained according to ties up principal component coefficient matrix, by institute
After the linear spectral mean value after being added per one kind stated is converted to corresponding P dimensions principal component feature, obtain including only one
The disaggregated model of a cluster centre is simultaneously preserved as the revised disaggregated model of the local discharge signal.
Wherein, the suspect signal recognition unit includes：
Oldham distance calculating module, for the P dimension principal component features of the suspect signal to be introduced the pure noise respectively
In the disaggregated model of ultrasonic signal and the revised disaggregated model of the local discharge signal, the suspect signal point is calculated
Be clipped to the pure noise ultrasonic signal disaggregated model and the revised disaggregated model of the local discharge signal Euclidean away from
From；
Judgment module, for judge the suspect signal to the disaggregated model of the pure noise ultrasonic signal Euclidean away from
From whether be less than the suspect signal to the revised disaggregated model of the local discharge signal Euclidean distance；
First result output module, the disaggregated model for the screening are the classification mould of the pure noise ultrasonic signal
Type, and be that pure noise ultrasonic signal is exported as testing result using the suspect signal；
Second result output module, the disaggregated model for the screening are the revised classification of the local discharge signal
Model, and be that local discharge signal is exported as testing result using the suspect signal.
Implement the embodiment of the present invention, has the advantages that：
In embodiments of the present invention, the dimension drop of training data can be effectively reduced due to the use of principal component analytical method
Low data complexity, while use environment noise mean value is modified disaggregated model, can effectively improve detection method and exists
Performance under all kinds of noise circumstances, so as to avoid artifact from interfering, and introduced feature dimension is low, and training data is few, from
And reach the time saving and energy saving and objective purpose of testing result.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention, for those of ordinary skill in the art, without having to pay creative labor, according to
These attached drawings obtain other attached drawings and still fall within scope of the invention.
Fig. 1 is a kind of flow chart of method for correcting ultrasonic wave Partial Discharge Detection provided in an embodiment of the present invention；
Fig. 2 is the method flow diagram of step S1 in Fig. 1；
Fig. 3 is the method flow diagram of step S2 in Fig. 1；
Fig. 4 is the method flow diagram of step S3 in Fig. 1；
Fig. 5 is the method flow diagram of step S5 in Fig. 1；
Fig. 6 is a kind of structural schematic diagram of system for correcting ultrasonic wave Partial Discharge Detection provided in an embodiment of the present invention.
Specific implementation mode
To make the object, technical solutions and advantages of the present invention clearer, the present invention is made into one below in conjunction with attached drawing
Step ground detailed description.
As shown in Figure 1, in the embodiment of the present invention, a kind of method of the amendment ultrasonic wave Partial Discharge Detection provided, institute
The method of stating includes：
Step S1, from the historical data of equipment high s/n ratio ultrasonic wave local discharge signal, multiple shelf depreciations are extracted
Signal, and multiple local discharge signals of the extraction are converted into after linear spectral and are handled by Principal Component Analysis, it obtains
Principal component coefficient matrix is tieed up to P, and principal component coefficient matrix is tieed up according to the obtained P, obtains the P of each local discharge signal
Principal component feature is tieed up, the P of obtained each local discharge signal dimension principal component features are further subjected to Unsupervised clustering,
Obtain the disaggregated model of local discharge signal and its corresponding linear spectral mean value per one kind；Wherein, P is natural number；
Detailed process is, step S11, from the historical data of equipment high s/n ratio ultrasonic wave local discharge signal, extraction
Multiple local discharge signals, and after multiple local discharge signals of extraction are carried out framing as unit of certain time length, to point
Each local discharge signal after frame carries out Fourier transformation, obtains the linear spectral of each local discharge signal；
Step S12, dimension-reduction treatment is carried out to the linear spectral of any local discharge signal by Principal Component Analysis, obtained
P ties up principal component coefficient matrix, and ties up principal component coefficient matrix according to obtained P, to each local discharge signal of extraction into
Row linear transformation obtains the P dimension principal component features of each local discharge signal；
Step S13, the P dimension principal component features of obtained each local discharge signal are carried out simultaneously using k-mean algorithms
Unsupervised clustering calculates multiple cluster centres and preserves the disaggregated model as local discharge signal；
Step S14, the principle averaged after being added according to same class linear spectral, to the classification mould of local discharge signal
It is calculated, is obtained in the disaggregated model of local discharge signal per a kind of linear spectral per the linear spectral contained by one kind in type
Mean value.
In one embodiment, the first step, extraction signal-to-noise ratio are more than the ultrasonic wave local discharge signal of 20dB, to extraction
After frame length of multiple local discharge signals as unit of 1s divides, the DFT that Fast Fourier Transform (FFT) calculates signal is carried out, is obtained
To the linear spectral of multiple local discharge signals；
Second step carries out principal component analysis to linearly frequency spectrum, and signal is dropped and is tieed up at P, obtains P dimension principal component coefficient squares
Battle array.In principal component analysis, for a sample data, p variable is observed, the data information battle array of n sample is specific such as formula
(1) shown in：
Variable in formula (1) is the linear spectral of audio fragment.
At this point, p observational variable synthesis is become p new generalized variables by principal component analysis, it is variable to turn to formula (2) institute
Show：
In formula (2), F1-FnNew generalized variable, i.e., the principal component of original p variable of signal；F1Variance maximum be known as
One principal component is used, F2Variance time be known as Second principal component, greatly, and so on.
The coefficient of aforesaid equation (2) is calculated with principal component analysis (PCA) method, i.e.,：
A, initial data is standardized, as shown in formula (3)；
In formula (3),
B, sample correlation coefficient matrix R is calculated, as shown in formula (4)；
For convenience, it is assumed that still indicated with X after initial data standardization, then the related coefficient of normalized treated data
For:
C, the characteristic value (λ of correlation matrix R is sought with Jacobian technique1,λ2…λp) and corresponding feature vector ai=
(ai1, ai2... aip), i=1,2 ... p；To obtain the principal component coefficient matrix of high s/n ratio ultrasonic wave local discharge signal, i.e.,
As shown in formula (5)：
Third step, with principal component coefficient matrices A to all signals carries out linear transformation obtain P dimension principal component feature, with institute
Have the P of signal be main composition characteristics be input Unsupervised clustering, preserve its cluster centre and obtain M class ultrasonic wave local discharge signals
Disaggregated model.All audio signals are converted with principal component coefficient matrix, obtain P dimension principal component features, such as certain section
The linear spectral of ultrasonic wave discharge signal is x={ x1x2x3...x4xp, dimensionality reduction multiplies in principal component coefficient matrix Xp=Ax, obtains
To the P dimension principal component features X of the segment signalp.Using the P dimensional features of all signals as inlet flow Unsupervised clustering, using k-
Ultrasonic wave local discharge signal cluster is M classes by mean algorithms.Wherein, M=10；
4th step, the linear spectral mean value for calculating separately every one kind.I.e. to each in the disaggregated model of local discharge signal
The linear spectral that a class is included sums and is averaged, and obtains the linear spectral mean value of such signal.
Step S2, pure noise ultrasonic signal is obtained, and the pure noise ultrasonic signal got is converted into line
Property frequency spectrum after, calculate the linear spectral mean value of pure noise ultrasonic signal, and principal component coefficient square is tieed up according to the obtained P
Battle array, the P that the linear spectral mean value of the pure noise ultrasonic signal is converted to the pure noise ultrasonic signal tie up principal component
After feature, the disaggregated model of pure noise ultrasonic signal is obtained；
Detailed process is, step S21, before obtaining suspect signal, it is pure in range for a period of time to obtain actual environment
Noise ultrasonic signal, and after the pure noise ultrasonic signal got is carried out framing as unit of certain time length, to framing
Pure noise ultrasonic signal afterwards carries out Fourier transformation, obtains the linear spectral of pure noise ultrasonic signal；
Step S22, it averages after being added the linear spectral of obtained pure noise ultrasonic signal, it is super as pure noise
The linear spectral mean value of acoustic signals；
Step S23, according to obtained P dimension principal component coefficient matrixes, by the linear spectral mean value of pure noise ultrasonic signal
After the P dimension principal component features for being converted to pure noise ultrasonic signal, the only disaggregated model comprising a cluster centre and guarantor are obtained
Deposit the disaggregated model as pure noise ultrasonic signal.
In one embodiment, before actually detected, the ultrasonic signal of one section of pure noise in scene is enrolled, by this section of ultrasound
Wave signal carries out framing by frame length of 1s, seeks linear spectral to all frames and is averaged, obtains the linear spectral of pure noise signal
Mean value.The linear spectral mean value of pure noise signal is converted into P with obtained P dimension principal component coefficient matrixes and ties up principal component feature,
Due to pure noise signal linear spectral mean value only there are one, transformed p dimensions principal component feature is equivalent in a cluster
The heart, to obtain the disaggregated model of the only pure noise ultrasonic signal comprising a cluster centre.If linear spectral mean value is N,
It is N that then P, which ties up principal component,p=AN.
Step S3, it pure will make an uproar with described respectively per a kind of linear spectral mean value in the local discharge signal disaggregated model
The linear spectral mean value of sound ultrasonic signal disaggregated model is added, and the linear spectral mean value after being added per one kind is obtained,
And principal component coefficient matrix is tieed up according to the obtained P, the obtained linear spectral mean value after being added per one kind is turned
After being changed to corresponding P dimensions principal component feature, obtained disaggregated model is as the revised disaggregated model of local discharge signal；
Detailed process is step S31, to determine per a kind of linear spectral mean value in local discharge signal disaggregated model, and
The linear spectral mean value of pure noise ultrasonic signal disaggregated model is added with determining per a kind of linear spectral mean value respectively,
It obtains after being added per a kind of linear spectral mean value；
Step S32, after being added per a kind of linear spectral by what is obtained according to obtained P dimension principal component coefficient matrixes
After mean value is converted to corresponding P dimensions principal component feature, obtains the only disaggregated model comprising a cluster centre and preserve as institute
State the revised disaggregated model of local discharge signal.
Step S4, suspect signal is obtained, and the suspect signal got is converted into linear spectral, and according to described
Obtained P dimension principal component coefficient matrixes, the P that the linear spectral of the suspect signal is converted to the suspect signal tie up principal component
Feature；
Step S5, the P of suspect signal dimension principal component features are introduced to point of the pure noise ultrasonic signal respectively
In class model and the revised disaggregated model of the local discharge signal, calculates the suspect signal and arrive the pure noise respectively
The Euclidean distance of the disaggregated model of ultrasonic signal and the revised disaggregated model of the local discharge signal, and obtained according to described
Two Euclidean distances arrived, filter out the corresponding disaggregated model of minimum value in described two Euclidean distances, will further be screened
Corresponding signal is exported as the testing result of the suspect signal in disaggregated model.
Detailed process is that the P of suspect signal dimension principal component features step S51, are introduced pure noise ultrasonic signal respectively
Disaggregated model and the revised disaggregated model of local discharge signal in, calculate suspect signal and arrive the pure noise ultrasound respectively
The Euclidean distance of the revised disaggregated model of disaggregated model and local discharge signal of wave signal；
Step S52, judge that suspect signal is waited for whether the Euclidean distance of the disaggregated model of pure noise ultrasonic signal is less than
Euclidean distance of the inspection signal to the revised disaggregated model of local discharge signal；If it is, executing next step S53；If
It is no, then it redirects and executes step S54；
Step S53, the disaggregated model of screening is the disaggregated model of pure noise ultrasonic signal, and is pure make an uproar by suspect signal
Sound ultrasonic signal is exported as testing result；
Step S54, the disaggregated model of screening is the revised disaggregated model of local discharge signal, and is office by suspect signal
Portion's discharge signal is exported as testing result.
It should be noted that k-mean clustering algorithms and Euclidean distance algorithm belong to the algorithms most in use of art technology,
This is not repeated.
As shown in fig. 6, in the embodiment of the present invention, a kind of system of the amendment ultrasonic wave Partial Discharge Detection provided, institute
The system of stating includes：
Local discharge signal analysis model acquiring unit 610 is used for from equipment high s/n ratio ultrasonic wave local discharge signal
Historical data in, extract multiple local discharge signals, and multiple local discharge signals of the extraction are converted into linearly
It is handled by Principal Component Analysis after frequency spectrum, obtains P dimension principal component coefficient matrixes, and according to the obtained P dimension principal components system
Matrix number obtains the P dimension principal component features of each local discharge signal, further by obtained each local discharge signal
P dimension principal component features carry out Unsupervised clustering, obtain the disaggregated model of local discharge signal and its corresponding per a kind of linear
Spectrum mean；Wherein, P is natural number；
Pure noise ultrasound signal analyzing model acquiring unit 620, for obtaining pure noise ultrasonic signal, and will be described
After the pure noise ultrasonic signal got is converted to linear spectral, the linear spectral for calculating pure noise ultrasonic signal is equal
Value, and principal component coefficient matrix is tieed up according to the obtained P, the linear spectral mean value of the pure noise ultrasonic signal is converted
After the P dimension principal component features of the pure noise ultrasonic signal, the disaggregated model of pure noise ultrasonic signal is obtained；
Local discharge signal analysis model amending unit 630, being used for will be each in the local discharge signal disaggregated model
The linear spectral mean value of class is added with the linear spectral mean value of the pure noise ultrasonic signal disaggregated model respectively, is obtained
It is after being added to tie up principal component coefficient matrix per a kind of linear spectral mean value, and according to the obtained P, by the obtained phase
After the linear spectral mean value per one kind after adding is converted to corresponding P dimensions principal component feature, obtained disaggregated model is as part
The revised disaggregated model of discharge signal；
Suspect signal analytic unit 640 is converted to line for obtaining suspect signal, and by the suspect signal got
Property frequency spectrum, and principal component coefficient matrix is tieed up according to the obtained P, the linear spectral of the suspect signal is converted into described wait for
The P for examining signal ties up principal component feature；
Suspect signal recognition unit 650, for the P of suspect signal dimension principal component feature to be introduced described pure make an uproar respectively
In the disaggregated model of sound ultrasonic signal and the revised disaggregated model of the local discharge signal, the suspect signal is calculated
The Euclidean of the disaggregated model and the revised disaggregated model of the local discharge signal of the pure noise ultrasonic signal is arrived respectively
Distance, and according to two obtained the Euclidean distance, filter out the corresponding classification mould of minimum value in described two Euclidean distances
Type is further exported signal corresponding in institute's sifting sort model as the testing result of the suspect signal.
Wherein, the local discharge signal analysis model acquiring unit 610 includes：
Local discharge signal linear spectral conversion module 6101 is used for from equipment high s/n ratio ultrasonic wave local discharge signal
Historical data in, extract multiple local discharge signals, and by multiple local discharge signals of the extraction with certain time length
After carrying out framing for unit, Fourier transformation is carried out to each local discharge signal after the framing, obtains each part
The linear spectral of discharge signal；
Principal component analysis module 6102, for by Principal Component Analysis to the linear spectral of any local discharge signal into
Row dimension-reduction treatment obtains P dimension principal component coefficient matrixes, and ties up principal component coefficient matrix according to the obtained P, to the extraction
Each local discharge signal carry out linear transformation, obtain each local discharge signal P dimension principal component feature；
Local discharge signal Clustering Model computing module 6103, for using k-mean algorithms simultaneously to it is described obtain it is every
The P dimension principal component features of one local discharge signal carry out Unsupervised clustering, calculate multiple cluster centres and preserve as part
The disaggregated model of discharge signal；
Local discharge signal Clustering Model classification mean value computation module 6104, after being added according to same class linear spectral
The principle averaged is obtained to being calculated per the linear spectral contained by one kind in the disaggregated model of the local discharge signal
To every a kind of linear spectral mean value in the disaggregated model of the local discharge signal.
Wherein, the pure noise ultrasound signal analyzing model acquiring unit 620 includes：
Pure noise ultrasonic signal linear spectral conversion module 6201, for before obtaining the suspect signal, obtaining
The actual environment pure noise ultrasonic signal in range for a period of time, and by the pure noise ultrasonic signal got with one
Timing length is after unit carries out framing, to carry out Fourier transformation to the pure noise ultrasonic signal after the framing, obtain institute
State the linear spectral of pure noise ultrasonic signal；
Linear spectral mean value computation module 6202, for by the linear spectral phase of the obtained pure noise ultrasonic signal
It averages after adding, the linear spectral mean value as the pure noise ultrasonic signal；
Pure noise ultrasonic signal Clustering Model computing module 6203, the P for being obtained according to tie up principal component coefficient
Matrix, by the linear spectral mean value of the pure noise ultrasonic signal be converted to the pure noise ultrasonic signal P dimension it is main at
After dtex sign, obtains the only disaggregated model comprising a cluster centre and preserve the classification as the pure noise ultrasonic signal
Model.
Wherein, the local discharge signal analysis model amending unit 630 includes：
Linear spectral accumulator module 6301, for determining in the local discharge signal disaggregated model per a kind of linear frequency
Compose mean value, and by the linear spectral mean value of the pure noise ultrasonic signal disaggregated model respectively with the determination per a kind of
Linear spectral mean value is added, and is obtained after being added per a kind of linear spectral mean value；
Local discharge signal Clustering Model correcting module 6302, the P for being obtained according to tie up principal component coefficient matrix,
After the obtained linear spectral mean value after being added per one kind is converted to corresponding P dimensions principal component feature, only wrapped
Disaggregated model containing a cluster centre is simultaneously preserved as the revised disaggregated model of the local discharge signal.
Wherein, the suspect signal recognition unit 650 includes：
Oldham distance calculating module 6501, it is described pure for introducing the P dimension principal component features of the suspect signal respectively
In the disaggregated model of noise ultrasonic signal and the revised disaggregated model of the local discharge signal, the letter to be checked is calculated
The Europe of the disaggregated model and the revised disaggregated model of the local discharge signal of the pure noise ultrasonic signal number is arrived respectively
Family name's distance；
Judgment module 6502, for judging the suspect signal to the Europe of the disaggregated model of the pure noise ultrasonic signal
Whether family name's distance is less than the suspect signal to the Euclidean distance of the revised disaggregated model of the local discharge signal；
First result output module 6503, the disaggregated model for the screening are point of the pure noise ultrasonic signal
Class model, and be that pure noise ultrasonic signal is exported as testing result using the suspect signal；
Second result output module 6504, the disaggregated model for the screening are that the local discharge signal is revised
Disaggregated model, and be that local discharge signal is exported as testing result using the suspect signal.
Implement the embodiment of the present invention, has the advantages that：
In embodiments of the present invention, the dimension drop of training data can be effectively reduced due to the use of principal component analytical method
Low data complexity, while use environment noise mean value is modified disaggregated model, can effectively improve detection method and exists
Performance under all kinds of noise circumstances, so as to avoid artifact from interfering, and introduced feature dimension is low, and training data is few, from
And reach the time saving and energy saving and objective purpose of testing result.
It is worth noting that, in above system embodiment, included each system unit only according to function logic into
What row divided, but it is not limited to above-mentioned division, as long as corresponding function can be realized；In addition, each functional unit
Specific name is also only to facilitate mutually distinguish, the protection domain being not intended to restrict the invention.
One of ordinary skill in the art will appreciate that implement the method for the above embodiments be can be with
Relevant hardware is instructed to complete by program, the program can be stored in a computer read/write memory medium,
The storage medium, such as ROM/RAM, disk, CD.
The above disclosure is only the preferred embodiments of the present invention, cannot limit the right model of the present invention with this certainly
It encloses, therefore equivalent changes made in accordance with the claims of the present invention, is still within the scope of the present invention.
Claims (10)
1. a kind of method for correcting ultrasonic wave Partial Discharge Detection, which is characterized in that the method includes：
A, from the historical data of equipment high s/n ratio ultrasonic wave local discharge signal, multiple local discharge signals are extracted, and will
Multiple local discharge signals of the extraction are handled after being converted into linear spectral by Principal Component Analysis, obtain P dimension it is main at
Divide coefficient matrix, and principal component coefficient matrix is tieed up according to the obtained P, obtains the P dimension principal components of each local discharge signal
The P of obtained each local discharge signal dimension principal component features are further carried out Unsupervised clustering, obtain part by feature
The disaggregated model of discharge signal and its corresponding linear spectral mean value per one kind；Wherein, P is natural number；
B, pure noise ultrasonic signal is obtained, and after the pure noise ultrasonic signal got is converted to linear spectral,
The linear spectral mean value of pure noise ultrasonic signal is calculated, and principal component coefficient matrix is tieed up according to the obtained P, it will be described
After the linear spectral mean value of pure noise ultrasonic signal is converted to the P dimension principal component features of the pure noise ultrasonic signal, obtain
To the disaggregated model of pure noise ultrasonic signal；
C, by the local discharge signal disaggregated model per a kind of linear spectral mean value respectively with the pure noise ultrasonic wave
The linear spectral mean value of Modulation recognition model is added, and obtains the linear spectral mean value after being added per one kind, and according to institute
The obtained linear spectral mean value after being added per one kind is converted to correspondence by the P dimension principal component coefficient matrixes stated
P dimension principal component feature after, obtained disaggregated model is as the revised disaggregated model of local discharge signal；
D, suspect signal is obtained, and the suspect signal got is converted into linear spectral, and is tieed up according to the obtained P
Principal component coefficient matrix, the P that the linear spectral of the suspect signal is converted to the suspect signal tie up principal component feature；
E, the P of suspect signal dimension principal component features are introduced to disaggregated model and the institute of the pure noise ultrasonic signal respectively
It states in the revised disaggregated model of local discharge signal, calculates the suspect signal and arrive the pure noise ultrasonic signal respectively
Disaggregated model and the revised disaggregated model of the local discharge signal Euclidean distance, and according to two obtained the Europe
Family name's distance filters out the corresponding disaggregated model of minimum value in described two Euclidean distances, further will be in institute's sifting sort model
Corresponding signal is exported as the testing result of the suspect signal.
2. the method as described in claim 1, which is characterized in that the step a is specifically included：
From the historical data of equipment high s/n ratio ultrasonic wave local discharge signal, multiple local discharge signals are extracted, and by institute
State extraction multiple local discharge signals carry out framing as unit of certain time length after, each part after the framing is put
Electric signal carries out Fourier transformation, obtains the linear spectral of each local discharge signal；
Dimension-reduction treatment is carried out to the linear spectral of any local discharge signal by Principal Component Analysis, obtains P dimension principal components system
Matrix number, and principal component coefficient matrix is tieed up according to the obtained P, to each local discharge signal of the extraction into line
Property transformation, obtain each local discharge signal P dimension principal component feature；
The P of obtained each local discharge signal dimension principal component features are carried out simultaneously using k-mean algorithms unsupervised poly-
Class calculates multiple cluster centres and preserves the disaggregated model as local discharge signal；
The principle averaged after being added according to same class linear spectral, to each in the disaggregated model of the local discharge signal
Linear spectral contained by class is calculated, and is obtained equal per a kind of linear spectral in the disaggregated model of the local discharge signal
Value.
3. the method as described in claim 1, which is characterized in that the step b is specifically included：
Before obtaining the suspect signal, pure noise ultrasonic signal of the actual environment for a period of time in range is obtained, and will
It is super to the pure noise after the framing after the pure noise ultrasonic signal got carries out framing as unit of certain time length
Acoustic signals carry out Fourier transformation, obtain the linear spectral of the pure noise ultrasonic signal；
It averages after the linear spectral of the obtained pure noise ultrasonic signal is added, as the pure noise ultrasonic wave
The linear spectral mean value of signal；
Principal component coefficient matrix is tieed up according to the obtained P, the linear spectral mean value of the pure noise ultrasonic signal is converted
After the P dimension principal component features of the pure noise ultrasonic signal, the only disaggregated model comprising a cluster centre and a guarantor are obtained
Deposit the disaggregated model as the pure noise ultrasonic signal.
4. the method as described in claim 1, which is characterized in that the step c is specifically included：
It determines per a kind of linear spectral mean value in the local discharge signal disaggregated model, and the pure noise ultrasonic wave is believed
The linear spectral mean value of number disaggregated model is added with the determination per a kind of linear spectral mean value respectively, is obtained after being added
Per a kind of linear spectral mean value；
Principal component coefficient matrix is tieed up according to the obtained P, by the obtained linear spectral mean value after being added per one kind
After being converted to corresponding P dimensions principal component feature, obtains the only disaggregated model comprising a cluster centre and preserve as the office
The revised disaggregated model of portion's discharge signal.
5. the method as described in claim 1, which is characterized in that the step e is specifically included：
The P of suspect signal dimension principal component feature is introduced into the disaggregated model of the pure noise ultrasonic signal and described respectively
In the revised disaggregated model of local discharge signal, calculates the suspect signal and arrive the pure noise ultrasonic signal respectively
The Euclidean distance of disaggregated model and the revised disaggregated model of the local discharge signal；
Judge whether the suspect signal is less than described wait for the Euclidean distance of the disaggregated model of the pure noise ultrasonic signal
Euclidean distance of the inspection signal to the revised disaggregated model of the local discharge signal；
If it is, the disaggregated model of the screening is the disaggregated model of the pure noise ultrasonic signal, and will be described to be checked
Signal is that pure noise ultrasonic signal is exported as testing result；
If it is not, then the disaggregated model of the screening is the revised disaggregated model of the local discharge signal, and waited for described
It is that local discharge signal is exported as testing result to examine signal.
6. a kind of system for correcting ultrasonic wave Partial Discharge Detection, which is characterized in that the system comprises：
Local discharge signal analysis model acquiring unit, for the history number from equipment high s/n ratio ultrasonic wave local discharge signal
In, multiple local discharge signals are extracted, and lead to after multiple local discharge signals of the extraction are converted into linear spectral
Principal Component Analysis processing is crossed, P dimension principal component coefficient matrixes is obtained, and principal component coefficient matrix is tieed up according to the obtained P, obtains
To each local discharge signal P tie up principal component feature, further by the P of obtained each local discharge signal dimension it is main at
Dtex sign carries out Unsupervised clustering, obtains the disaggregated model of local discharge signal and its corresponding linear spectral mean value per one kind；
Wherein, P is natural number；
Pure noise ultrasound signal analyzing model acquiring unit, gets for obtaining pure noise ultrasonic signal, and by described
Pure noise ultrasonic signal be converted to linear spectral after, calculate the linear spectral mean value of pure noise ultrasonic signal, and root
Principal component coefficient matrix is tieed up according to the obtained P, the linear spectral mean value of the pure noise ultrasonic signal is converted to described
After the P dimension principal component features of pure noise ultrasonic signal, the disaggregated model of pure noise ultrasonic signal is obtained；
Local discharge signal analysis model amending unit, for by the local discharge signal disaggregated model per a kind of linear
Spectrum mean is added with the linear spectral mean value of the pure noise ultrasonic signal disaggregated model respectively, is obtained after being added
Tie up principal component coefficient matrix per a kind of linear spectral mean value, and according to the obtained P, by it is described obtain it is after being added every
After a kind of linear spectral mean value is converted to corresponding P dimensions principal component feature, obtained disaggregated model is as local discharge signal
Revised disaggregated model；
Suspect signal analytic unit is converted to linear spectral for obtaining suspect signal, and by the suspect signal got,
And principal component coefficient matrix is tieed up according to the obtained P, the linear spectral of the suspect signal is converted into the suspect signal
P tie up principal component feature；
Suspect signal recognition unit, for the P dimension principal component features of the suspect signal to be introduced the pure noise ultrasound respectively
In the disaggregated model of wave signal and the revised disaggregated model of the local discharge signal, calculates the suspect signal and arrive respectively
The Euclidean distance of the disaggregated model and the revised disaggregated model of the local discharge signal of the pure noise ultrasonic signal, and
According to two obtained the Euclidean distance, the corresponding disaggregated model of minimum value in described two Euclidean distances is filtered out, into one
Step is exported signal corresponding in institute's sifting sort model as the testing result of the suspect signal.
7. system as claimed in claim 6, which is characterized in that the local discharge signal analysis model acquiring unit includes：
Local discharge signal linear spectral conversion module, for the history number from equipment high s/n ratio ultrasonic wave local discharge signal
In, extract multiple local discharge signals, and by multiple local discharge signals of the extraction as unit of certain time length into
After row framing, Fourier transformation is carried out to each local discharge signal after the framing, obtains each local discharge signal
Linear spectral；
Principal component analysis module, for being carried out at dimensionality reduction to the linear spectral of any local discharge signal by Principal Component Analysis
Reason obtains P dimension principal component coefficient matrixes, and ties up principal component coefficient matrix according to the obtained P, to each office of the extraction
Portion's discharge signal carries out linear transformation, obtains the P dimension principal component features of each local discharge signal；
Local discharge signal Clustering Model computing module, for being put simultaneously to obtained each part using k-mean algorithms
The P dimension principal component features of electric signal carry out Unsupervised clustering, calculate multiple cluster centres and preserve as local discharge signal
Disaggregated model；
Local discharge signal Clustering Model classification mean value computation module, for averaging after being added according to same class linear spectral
Principle, in the disaggregated model of the local discharge signal per one kind contained by linear spectral calculate, obtain the office
Per a kind of linear spectral mean value in the disaggregated model of portion's discharge signal.
8. system as claimed in claim 6, which is characterized in that the pure noise ultrasound signal analyzing model acquiring unit packet
It includes：
Pure noise ultrasonic signal linear spectral conversion module, for before obtaining the suspect signal, obtaining actual environment
Pure noise ultrasonic signal within the scope of a period of time, and be with certain time length by the pure noise ultrasonic signal got
After unit carries out framing, Fourier transformation is carried out to the pure noise ultrasonic signal after the framing, obtains the pure noise
The linear spectral of ultrasonic signal；
Linear spectral mean value computation module, for asking flat after being added the linear spectral of the obtained pure noise ultrasonic signal
Mean value, the linear spectral mean value as the pure noise ultrasonic signal；
Pure noise ultrasonic signal Clustering Model computing module, the P for being obtained according to ties up principal component coefficient matrix, by institute
State pure noise ultrasonic signal linear spectral mean value be converted to the pure noise ultrasonic signal P dimension principal component feature after,
It obtains the only disaggregated model comprising a cluster centre and preserves the disaggregated model as the pure noise ultrasonic signal.
9. system as claimed in claim 6, which is characterized in that the local discharge signal analysis model amending unit includes：
Linear spectral accumulator module, for determining per a kind of linear spectral mean value in the local discharge signal disaggregated model,
And by the linear spectral mean value of the pure noise ultrasonic signal disaggregated model respectively with the determination per a kind of linear frequency
It composes mean value to be added, obtain after being added per a kind of linear spectral mean value；
Local discharge signal Clustering Model correcting module, the P for being obtained according to tie up principal component coefficient matrix, will be described
To it is after being added be converted to corresponding P dimensions principal component feature per a kind of linear spectral mean value after, obtain only poly- comprising one
The disaggregated model at class center is simultaneously preserved as the revised disaggregated model of the local discharge signal.
10. system as claimed in claim 6, which is characterized in that the suspect signal recognition unit includes：
Oldham distance calculating module, for the P dimension principal component features of the suspect signal to be introduced the pure noise ultrasound respectively
In the disaggregated model of wave signal and the revised disaggregated model of the local discharge signal, calculates the suspect signal and arrive respectively
The Euclidean distance of the disaggregated model and the revised disaggregated model of the local discharge signal of the pure noise ultrasonic signal；
Judgment module, for judging that the suspect signal is to the Euclidean distance of the disaggregated model of the pure noise ultrasonic signal
The no Euclidean distance less than the suspect signal to the revised disaggregated model of the local discharge signal；
First result output module, the disaggregated model for the screening are the disaggregated model of the pure noise ultrasonic signal,
And the suspect signal is exported for pure noise ultrasonic signal as testing result；
Second result output module, the disaggregated model for the screening are the revised classification mould of the local discharge signal
Type, and be that local discharge signal is exported as testing result using the suspect signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610156063.5A CN105842588B (en) | 2016-03-18 | 2016-03-18 | A kind of method and system for correcting ultrasonic wave Partial Discharge Detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610156063.5A CN105842588B (en) | 2016-03-18 | 2016-03-18 | A kind of method and system for correcting ultrasonic wave Partial Discharge Detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105842588A CN105842588A (en) | 2016-08-10 |
| CN105842588B true CN105842588B (en) | 2018-09-28 |
Family
ID=56587940
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610156063.5A Active CN105842588B (en) | 2016-03-18 | 2016-03-18 | A kind of method and system for correcting ultrasonic wave Partial Discharge Detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105842588B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108333484B (en) * | 2018-01-23 | 2020-06-30 | 国网河北省电力有限公司电力科学研究院 | Method for detecting partial discharge of electrical equipment |
| CN108896878B (en) * | 2018-05-10 | 2020-06-19 | 国家电网公司 | Partial discharge detection method based on ultrasonic waves |
| CN111352006A (en) * | 2020-03-27 | 2020-06-30 | 国网甘肃省电力公司电力科学研究院 | External insulation equipment real-time discharge intensity quantification and evaluation system based on ultraviolet imaging |
| CN115166453B (en) * | 2022-09-08 | 2023-01-24 | 国网智能电网研究院有限公司 | Partial discharge continuous monitoring method and device based on edge real-time radio frequency pulse classification |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004061358A (en) * | 2002-07-30 | 2004-02-26 | Kawatetsu Advantech Co Ltd | Detection method and system for partial discharge in insulator of power apparatus |
| CN102426835A (en) * | 2011-08-30 | 2012-04-25 | 华南理工大学 | Method for identifying local discharge signals of switchboard based on support vector machine model |
| CN102628917A (en) * | 2012-04-25 | 2012-08-08 | 广州供电局有限公司 | Partial discharge recognition method and system |
| CN103558519A (en) * | 2013-11-02 | 2014-02-05 | 国家电网公司 | GIS partial discharge ultrasonic signal identification method |
| CN103675610A (en) * | 2013-09-29 | 2014-03-26 | 国家电网公司 | Method for extracting characteristic factors in online local discharge detection |
| WO2015070513A1 (en) * | 2013-11-14 | 2015-05-21 | 国家电网公司 | Pattern recognition method for partial discharge of three-phase in one enclosure type ultrahigh voltage gis |
-
2016
- 2016-03-18 CN CN201610156063.5A patent/CN105842588B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004061358A (en) * | 2002-07-30 | 2004-02-26 | Kawatetsu Advantech Co Ltd | Detection method and system for partial discharge in insulator of power apparatus |
| CN102426835A (en) * | 2011-08-30 | 2012-04-25 | 华南理工大学 | Method for identifying local discharge signals of switchboard based on support vector machine model |
| CN102628917A (en) * | 2012-04-25 | 2012-08-08 | 广州供电局有限公司 | Partial discharge recognition method and system |
| CN103675610A (en) * | 2013-09-29 | 2014-03-26 | 国家电网公司 | Method for extracting characteristic factors in online local discharge detection |
| CN103558519A (en) * | 2013-11-02 | 2014-02-05 | 国家电网公司 | GIS partial discharge ultrasonic signal identification method |
| WO2015070513A1 (en) * | 2013-11-14 | 2015-05-21 | 国家电网公司 | Pattern recognition method for partial discharge of three-phase in one enclosure type ultrahigh voltage gis |
Non-Patent Citations (1)
| Title |
|---|
| 基于支持向量机的GIS超高频局部放电模式识别;司良奇 等;《高压电器》;20141130;第50卷(第11期);第1-6页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105842588A (en) | 2016-08-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105842588B (en) | A kind of method and system for correcting ultrasonic wave Partial Discharge Detection | |
| US20230123117A1 (en) | Method and Apparatus for Inspecting Wind Turbine Blade, And Device And Storage Medium Thereof | |
| CN108734192B (en) | Voting mechanism-based support vector machine mechanical fault diagnosis method | |
| CN108229382A (en) | Vibration signal characteristics extracting method, device, storage medium and computer equipment | |
| CN112486097A (en) | Cutting platform wear state monitoring method and system based on model algorithm and readable storage medium | |
| CN103048593A (en) | Identification method of insulation defect type of gas-insulated switchgear | |
| CN105356886B (en) | Power line noise compression method and device based on compressed sensing | |
| CN111918196A (en) | Method, device and equipment for diagnosing recording abnormity of audio collector and storage medium | |
| CN110020637A (en) | A kind of analog circuit intermittent fault diagnostic method based on more granularities cascade forest | |
| CN109932053A (en) | A kind of state monitoring apparatus and method for high-voltage shunt reactor | |
| CN102664772A (en) | Multi-scale detecting method of network flow anomaly in dynamic environment | |
| EP2391895B1 (en) | Precision measurement of waveforms | |
| CN108594156B (en) | Improved current transformer saturation characteristic identification method | |
| CN110609088A (en) | Post insulator fault identification method and device | |
| US8583423B2 (en) | Method and arrangement for processing of speech quality estimate | |
| CN114137410A (en) | Hydraulic mechanism circuit breaker fault identification method based on voiceprint detection technology | |
| CN116386669A (en) | Machine running acoustic state monitoring method and system based on block automatic encoder | |
| CN116482526A (en) | Analysis system for non-fault phase impedance relay | |
| Serrà et al. | Nonlinear audio recurrence analysis with application to genre classification | |
| CN108846106A (en) | A kind of method and apparatus for judging to whether there is identical audio in multiple audios | |
| Zivanovic | Detection of non-stationary sinusoids by using joint frequency reassignment and null-to-null bandwidth | |
| CN103438983A (en) | Data processing method of signal random average spectrum | |
| CN112908347A (en) | Noise detection method and terminal | |
| CN112485616A (en) | Cable insulation aging detection method and device, storage medium and processor | |
| CN112114215A (en) | Transformer aging evaluation method and system based on error back propagation algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |