CN107356843A - The partial discharge of transformer method for diagnosing faults of small echo is synchronously extruded based on gradient threshold - Google Patents

The partial discharge of transformer method for diagnosing faults of small echo is synchronously extruded based on gradient threshold Download PDF

Info

Publication number
CN107356843A
CN107356843A CN201710249329.5A CN201710249329A CN107356843A CN 107356843 A CN107356843 A CN 107356843A CN 201710249329 A CN201710249329 A CN 201710249329A CN 107356843 A CN107356843 A CN 107356843A
Authority
CN
China
Prior art keywords
partial discharge
transformer
swt
signal
formula
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201710249329.5A
Other languages
Chinese (zh)
Other versions
CN107356843B (en
Inventor
王文波
王斌
余东
汪祥莉
喻敏
赵彦超
晋云雨
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University of Science and Engineering WUSE
Original Assignee
Wuhan University of Science and Engineering WUSE
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan University of Science and Engineering WUSE filed Critical Wuhan University of Science and Engineering WUSE
Priority to CN201710249329.5A priority Critical patent/CN107356843B/en
Publication of CN107356843A publication Critical patent/CN107356843A/en
Application granted granted Critical
Publication of CN107356843B publication Critical patent/CN107356843B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
    • G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials

Abstract

The partial discharge of transformer method for diagnosing faults provided by the invention that small echo is synchronously extruded based on gradient threshold, is had the feature that, comprised the following steps：Step 1, to transformer partial discharge signal f (t) to carry out SWT decomposition, obtain SWT decomposition coefficientsI=1,2, L, I；Step 2, utilize formulaEstimate noise criteria difference σn；Step 3, to the layer coefficients of SWT i-thBy fixed step size to ca(i)Progressively value, the minimum value of mean square error is calculated by mean square error formula, determines ca(i)Optimal value；Step 4, calculate gradient threshold γa(i)=ca(i)σn, the de-noising of i-th layer of SWT coefficient is realized by i-th layer of SWT coefficient formula；Step 5, each layer coefficients after denoising are reconstructed using noise reduction formula, obtain the de-noising transformer partial discharge signal y (t) after de-noising；And step 6, feature extraction is carried out to the de-noising transformer partial discharge signal y (t) after denoising, fault diagnosis is carried out according to extracted feature.

Description

The partial discharge of transformer method for diagnosing faults of small echo is synchronously extruded based on gradient threshold
Technical field
The present invention relates to power system signal process field, the more particularly to transformation based on gradient threshold synchronization extruding small echo Device partial discharges fault diagnostic method.
Background technology
Shelf depreciation refer in insulation system under electric field action only have regional area discharge, rather than large area or Through the electric discharge of whole conductor.The insulation system of large-scale power transformer is more complicated, and the material used is varied, whole exhausted The distribution of edge system field is very uneven.Make to contain air gap in insulation system due to improving not to the utmost in design or manufacturing process, or The humidified insulation in During Process of Long-term Operation, moisture is decomposed under electric field action to be produced gas and forms bubble.Because air Dielectric constant it is smaller than the dielectric constant of insulating materials, even if insulating materials under less high electric field action, air gap, bubble portion The field strength that position is born also can be very high, and shelf depreciation will occur when field strength reaches more than certain value.Other insulating inner is present Defect is mixed into various impurity, or has in insulation system that some electrical connections are bad etc., and Dou Huishi internal fields concentrate, It is possible to that solid insulation surface-discharge and floating potential discharge occurs in the place that electric field is concentrated.In the oilpaper of power transformer In insulation system, on the one hand shelf depreciation makes oil decomposition go out gas, on the other hand may generate sludge deposition in solid insulation again On material, more violent electric discharge is formed in this place, and makes to turn into heat spot at this, promotes insulation damages.
The sustainable development of shelf depreciation progressively expands the deterioration for making insulation damage, finally shorten insulation ordinary life, Dielectric strength reduces in short-term, in some instances it may even be possible to makes whole insulation breakdown.Because transformer is equipment mostly important in power system, In order to realize on-line monitoring and fault diagnosis to the state of insulation of transformer, both at home and abroad to the part of reflection transformer insulation state Electric discharge electric characteristic amount has carried out numerous studies.But local discharge signal is very faint, and because transformer is in complexity Electromagnetic interference environment in so that original very faint local discharge signal was submerged among very strong various interference, so as to very Hardly possible obtains real useful information so that the extraction and pattern-recognition of local discharge characteristic amount are very difficult, can not Accurate Diagnosis The real insulation status of equipment.The feature extraction of transformer partial discharge signal is to carry out the key of insulation tube arrester, and to part It is the basis and premise that local discharge characteristic accurately extracts that discharge signal, which carries out denoising,.
Currently, it is wavelet threshold denoising method and empirical modal for transformer partial discharge signal, main denoising method Decompose (empirical mode decomposition) Denoising Algorithm.In wavelet threshold denoising method, the selection of wavelet basis is to denoising The distortion of signal has close relationship, but to how optimum choice wavelet basis is still without clear and definite research method；And utilize During threshold method denoising, the coefficient that amplitude is more than to threshold value all retains, and the coefficient that amplitude is less than threshold value is all deleted, unavoidably The less signal coefficient of partial amplitude is deleted, and the larger noise coefficient of partial amplitude is retained, and have impact on transformation The denoising effect of device local discharge signal and EMD Denoising Algorithms extract the signal after denoising by iteration, cause speed very slow, when Between spend it is larger；And noise it is stronger when, EMD methods can not be disturbed larger, and the error of signal is larger after denoising.Due to these Shortcoming, Wavelet-denoising Method and EMD Denoising Algorithms is adapted to different types of local discharge signal, be unfavorable for actual production ring The application of transformer partial discharge signal denoising in border.
Synchronous extruding wavelet transformation (Synchrosqueezed Wavelet Transform, SWT) is become in continuous wavelet A kind of new Time-Frequency Analysis Method .SWT to grow up on the basis of changing by continuous wavelet transform time-frequency figure in frequency side To extruding, can obtain the time-frequency curve of higher precision, and cross term is not present between the time-frequency curve extracted, be effectively improved Aliasing .SWT during non-linear sophisticated signal time frequency analysis to noise has very strong robustness, when signal is strong During noise pollution, SWT can still obtain clearly time-frequency curve and the decomposition result being basically unchanged.Therefore, SWT is especially suitable for non- The denoising of linear signal, test result indicates that, compared with wavelet transformation and EMD decompose, SWT can obtain the more preferable denoising knot of effect Fruit.
But existing research relates only to the Denoising Study of SWT individual layer threshold values, for the optimization structure of SWT gradient thresholds Make, especially the denoising of SWT gradient thresholds is applied in partial discharge of transformer fault diagnosis, almost related to without the patent of correlation And.
The content of the invention
The present invention is that adaptivity is strong, and it is an object of the present invention to provide a kind of autgmentability is strong in order to solving the above problems and carry out , the partial discharge of transformer method for diagnosing faults based on gradient threshold synchronization extruding small echo.
The partial discharge of transformer method for diagnosing faults provided by the invention that small echo is synchronously extruded based on gradient threshold, is had Such feature, comprises the following steps：
Step 1, to transformer partial discharge signal f (t) to carry out SWT decomposition, obtain SWT decomposition coefficients
Step 2, utilize formulaEstimate noise criteria difference σn；
Step 3, to the layer coefficients of SWT i-thBy fixed step size to ca(i)Progressively value, pass through mean square error formula The minimum value of mean square error is calculated, determines ca(i)Optimal value；
Step 4, calculate gradient threshold γa(i)=ca(i)σn, i-th layer of SWT coefficient is realized by i-th layer of SWT coefficient formula De-noising；
Step 5, each layer coefficients after denoising are reconstructed using noise reduction formula, obtain the de-noising transformer after de-noising Local discharge signal y (t)；And
Step 6, feature extraction is carried out to the de-noising transformer partial discharge signal y (t) after denoising, according to being carried Feature is taken to carry out fault diagnosis.
The partial discharge of transformer method for diagnosing faults provided by the invention that small echo is synchronously extruded based on gradient threshold, is also had There is such feature：Wherein, the step 1, it is assumed that the transformer partial discharge signal f (t) is
fk(t)=Ak(t)cos[2πφk(t)], Ak(t) signal amplitude, φ are representedk(t) signal phase is represented, n (t) is represented White Gaussian noise,
First, continuous wavelet transform is carried out to the transformer partial discharge signal
ψ (t) represents wavelet basis function used, and parameter a represents scale factor during wavelet transformation, and parameter b represents small echo Shift factor during conversion,
Then, the instantaneous frequency of signal is calculated on the basis of continuous wavelet transform
Finally, to wavelet coefficient Wf(a, b) carries out the synchronous extruding change that threshold value is γ, precision is δ, after synchronous extruding Coefficient is
Aγ,f(b)={ a ∈ R+；|Wf(a, b) | ＞ γ } represent yardstick set.
The partial discharge of transformer method for diagnosing faults provided by the invention that small echo is synchronously extruded based on gradient threshold, is also had There is such feature：Wherein, the mean square error formula is
The partial discharge of transformer method for diagnosing faults provided by the invention that small echo is synchronously extruded based on gradient threshold, is also had There is such feature：Wherein, i-th layer of SWT coefficient formula is
The partial discharge of transformer method for diagnosing faults provided by the invention that small echo is synchronously extruded based on gradient threshold, is also had There is such feature：Wherein, the noise reduction formula is
Invention effect and effect
According to it is involved in the present invention based on gradient threshold synchronously extruding small echo partial discharge of transformer method for diagnosing faults, With very strong autgmentability and adaptivity, preferable effect can be still obtained under strong noise environment；When realizing without artificial Intervene, all parameters can adaptive polo placement；Around the denoising of transformer partial discharge signal, overall procedure has complete compile Code, can be used the denoising of the transformer partial discharge signal under different situations, is effectively improved transformer partial discharge signal Feature extraction and fault diagnosis effect, there is wide promotional value.
Brief description of the drawings
Fig. 1 is that the partial discharge of transformer failure that based on gradient threshold is synchronously extruded small echo of the present invention in embodiment is examined The flow chart of disconnected method；
Fig. 2 is analogue transformer local discharge signal partial discharge through this patent method denoising after of the present invention in embodiment Emulate signal graph；
Fig. 3 is that analogue transformer local discharge signal of the present invention in embodiment is noisy after this patent method denoising Partial discharge emulates signal graph；
Fig. 4 is analogue transformer local discharge signal tradition through this patent method denoising after of the present invention in embodiment SWT method denoising figures；
Fig. 5 is analogue transformer local discharge signal simulation through this patent method denoising after of the present invention in embodiment Denoising figure；
Fig. 6 is actual measurement transformer partial discharge signal actual measurement office through context of methods denoising after of the present invention in embodiment Discharge signal figure；
Fig. 7 is actual measurement transformer partial discharge signal tradition through context of methods denoising after of the present invention in embodiment SWT method denoising figures；And
Fig. 8 is that actual measurement transformer partial discharge signal actual measurement through context of methods denoising after of the present invention in embodiment is gone Make an uproar figure.
Embodiment
Referring to the drawings reality and apply example to it is involved in the present invention based on gradient threshold synchronously extruding small echo transformer Partial discharges fault diagnostic method is explained in detail.
Embodiment
Fig. 1 is that the partial discharge of transformer failure that based on gradient threshold is synchronously extruded small echo of the present invention in embodiment is examined The flow chart of disconnected method.
As shown in figure 1, based on gradient threshold synchronously extruding small echo partial discharge of transformer method for diagnosing faults have with Lower step：
Step 1：SWT decomposition is carried out to transformer partial discharge signal f (t), obtains SWT decomposition coefficients Into step 2.
Assuming that transformer partial discharge signal f (t) is
Wherein fk(t)=Ak(t)cos[2πφk(t)], Ak(t) signal amplitude, φ are representedk(t) signal phase, n (t) are represented Represent white Gaussian noise.
First, continuous wavelet transform is carried out to transformer partial discharge signal
Wherein ψ (t) represents wavelet basis function used, and parameter a represents scale factor during wavelet transformation, and parameter b is represented Shift factor during wavelet transformation.
Then, the instantaneous frequency of signal is calculated on the basis of continuous wavelet transform
Finally, to wavelet coefficient Wf(a, b) carries out the synchronous extruding change that threshold value is γ, precision is δ, after synchronous extruding Coefficient is
Wherein, Aγ,f(b)={ a ∈ R+；|Wf(a, b) | ＞ γ } represent yardstick set.
Step 2：Utilize formulaEstimate noise criteria difference σn, into step 3.
For transformer partial discharge signal f (t), its component signal fk(t) can be by SWT Perfect Reconstructions, therefore can pass through Following formula realizes noise reductions of the SWT to signal
In SWT in cancellation process, that is, consider signal amplitude information (| Wf(a, t) | ＜ γ), it is contemplated that signal With basic frequency curve φkBetween ' (t) relativeness (| ω (a, t)-φk' (t) | ＜ ε).
If the yardstick in SWT decomposition used in continuous wavelet transform is a={ a (i), i=1,2, L, I }, then noisy chaos is believed The i-th layer coefficients number after SWT is converted are
From above formula, amplitude informations of the SWT according to signal coefficient and the relative position relation between basic frequency curve The coefficient that should retain after denoising is determined, i.e.,
Wherein γa(i)Amplitude thresholds during SWT denoisings are represented, ε represents the existing SWT denoisings of frequency threshold during denoising In method, amplitude thresholds γa(i)Value be global threshold, i.e., to each layer of SWT coefficient, γa(i)All take identical value：
γa(i)=γ=c σn (7)
Wherein,Representing that noise criteria is poor, N represents signal length,It is normal for one 2 points of deficiencies be present as single threshold value in SWT denoisings, using the threshold value in formula (7) in number：1) partial discharge of transformer is believed Number after SWT is decomposed, each layer coefficientsThe intensity of middle institute's Noise simultaneously differs, and suppresses to make an uproar using identical threshold value Sound, the effect of de-noising will necessarily be influenceed；2) when local discharge signal length N is larger, threshold gamma can be excessive, and when N is smaller, Threshold gamma again can be too small, lack necessary adaptivity.For the deficiency of single amplitude thresholds in existing SWT denoisings, herein from The least mean-square error of chaotic signal sets out after denoising, and according to signal coefficient and the distribution character of noise coefficient, construction has one Determine the layering amplitude thresholds of adaptivity
γa(i)=ca(i)σn, i=1,2, L, I.
It is so as to the denoising model of the gradient threshold transformer partial discharge signal based on SWT：
Step 3：To the layer coefficients of SWT i-thBy fixed step size to ca(i)Progressively value, pass through mean square error formula The minimum value of mean square error is calculated, determines ca(i)Optimal value, into step 4.
The SWT of transformer partial discharge signal is transformed toWherein WithRepresent respectively signals and associated noises y, actual signal f and noise n SWT coefficients set the coefficient after threshold denoising as
Purpose to chaotic signal de-noising is to make the signal after de-noising as close as actual signal, even if also As close asTherefore the purpose of invention is just to determine suitable ca(i)So that using gradient threshold γa(i)=ca (i)σnCoefficient after de-noisingWithMean square error between true chaotic signal coefficient reaches minimum, i.e., Reach minimum, and
Because noise is separate with signal, thereforeInstitute's above formula can abbreviation be
From de-noising formula (6),
Therefore,Substitution formula (8)
In view of the true chaotic signals of f (t),With de-noising threshold parameter ca(i), ε it is unrelated, therefore mean square error letter NumberIt can be written as
In SWT chaos de-noisings, frequency threshold ε takes constant, thereforeThat is mean square error function
Step 4：Calculate gradient threshold γa(i)=ca(i)σn, pass through i-th layer of SWT coefficient formula The de-noising of i-th layer of SWT coefficient is realized, into step 5.
Step 5：Each layer coefficients after denoising are reconstructed using noise reduction formula, obtain the de-noising transformer after de-noising Local discharge signal y (t), into step 6.
Seek optimal amplitude thresholds γa(i)=ca(i)σn, equivalent to seeking e (ca(i)) it is minimum when ca(i)Value, that is, ask
In formula (11)Can directly it be obtained by SWT coefficients, in order to askOnly need to calculateI.e. Can.From formula (9), work as signal coefficientWhen being zeroed out, noise coefficientAlso it is zeroed out, it is assumed thatIn have k Individual point is zeroed out, then accordinglyIn also there is k point to be zeroed out, remaining point withIf the identical of value it is rightWithEnter rearrangement from small to large according to absolute value, then understandPreceding k point be 0, and N-k below WithIt is identical, therefore
Due toTherefore
When signals and associated noises coefficientWhen, noise coefficientSet up almost everywhere, therefore
Due to noise coefficientApproximate Normal Distribution, and mixed coefficintApproximation obeys Laplace distributions, i.e., Laplace(0,σy) so
Wherein γa(i)=ca(i)σn,Understood by formula (11), (12), (13), (14),
The i-th layer coefficients after being decomposed for SWT, different c are calculated using formula (15)a(i)E (c during valuea(i)) value, so as to really Determine e (ca(i)) threshold parameter c when reaching minimuma(i)Value.In this patent, ca(i)Span be set as ca(i)∈[1.5, 6.5], according to 0.01 step-length value successively.
Step 6：Feature extraction is carried out to the de-noising transformer partial discharge signal y (t) after denoising, according to being carried Feature is taken to carry out fault diagnosis.
Using SWT gradient thresholds method to each layer component denoising of local discharge signal after, largely eliminate and make an uproar Interference of the sound to discharge fault feature, entropy feature extraction is carried out to the signal after each layer denoising in this patent, calculates its 6 kinds respectively Synchronous extruding wavelet transformation (SWT) entropy, it is respectively：SWT Energy Spectrum Entropies (SEE), SWT time entropys (STE), SWT singular entropies (SSE), SWT Time-frequency entropy (STFE), SWT mean entropies (SAE), SWT Distance entropies (SDE).Because partial discharge of transformer fault diagnosis characteristic is complicated, So the selection of identification model is particularly important after feature extraction.We select the Hopfield with feedback function to move in this patent State neutral net, partial discharge of transformer failure is diagnosed on the basis of SWT entropy features.Hopfield dynamical feedbacks god The real process of system is more nearly through network, possesses dynamic and memory link, for the multiple input/multiple output transformation of complexity When device partial discharges fault is recognized, more preferable recognition effect can be obtained.
The effect of embodiment and effect
The partial discharge of transformer fault diagnosis side of small echo is synchronously extruded based on gradient threshold according to involved by the present embodiment Method, there is very strong autgmentability and adaptivity, preferable effect can be still obtained under strong noise environment；When realizing without people Work intervention, all parameters can adaptive polo placements；Around the denoising of transformer partial discharge signal, overall procedure has complete Coding, can be used the denoising of the transformer partial discharge signal under different situations, is effectively improved transformer partial discharge signal Feature extraction and fault diagnosis effect, there is wide promotional value.
Above-mentioned embodiment is the preferred case of the present invention, is not intended to limit protection scope of the present invention.

Claims (5)

  1. A kind of 1. partial discharge of transformer method for diagnosing faults that small echo is synchronously extruded based on gradient threshold, it is characterised in that bag Include following steps：
    Step 1, to transformer partial discharge signal f (t) to carry out SWT decomposition, obtain SWT decomposition coefficientsI=1,2, L,I；
    Step 2, utilize formulaEstimate noise criteria difference σn；
    Step 3, to the layer coefficients of SWT i-thBy fixed step size to ca(i)Progressively value, calculated by mean square error formula The minimum value of mean square error, determines ca(i)Optimal value；
    Step 4, calculate gradient threshold γa(i)=ca(i)σn, disappearing for i-th layer of SWT coefficient is realized by i-th layer of SWT coefficient formula Make an uproar；
    Step 5, each layer coefficients after denoising are reconstructed using noise reduction formula, obtain the part of the de-noising transformer after de-noising Discharge signal y (t)；And
    Step 6, feature extraction is carried out to the de-noising transformer partial discharge signal y (t) after denoising, according to extracting spy Sign carries out fault diagnosis.
  2. 2. the partial discharge of transformer fault diagnosis side according to claim 1 that small echo is synchronously extruded based on gradient threshold Method, it is characterised in that：
    Wherein, the step 1, it is assumed that the transformer partial discharge signal f (t) is
    fk(t)=Ak(t)cos[2πφk(t)], Ak(t) signal amplitude, φ are representedk(t) signal phase is represented, n (t) represents Gauss White noise,
    First, continuous wavelet transform is carried out to the transformer partial discharge signal
    ψ (t) represents wavelet basis function used, and parameter a represents scale factor during wavelet transformation, and parameter b represents wavelet transformation When shift factor,
    Then, the instantaneous frequency of signal is calculated on the basis of continuous wavelet transform
    Finally, to wavelet coefficient Wf(a, b) carries out the synchronous extruding change that threshold value is γ, precision is δ, the coefficient after synchronous extruding For
    Aγ,f(b)={ a ∈ R+；|Wf(a, b) | ＞ γ } represent yardstick set.
  3. 3. the partial discharge of transformer fault diagnosis side according to claim 1 that small echo is synchronously extruded based on gradient threshold Method, it is characterised in that：
    Wherein, the mean square error formula is
  4. 4. the partial discharge of transformer fault diagnosis side according to claim 1 that small echo is synchronously extruded based on gradient threshold Method, it is characterised in that：
    Wherein, i-th layer of SWT coefficient formula is
  5. 5. the partial discharge of transformer fault diagnosis side according to claim 1 that small echo is synchronously extruded based on gradient threshold Method, it is characterised in that：
    Wherein, the noise reduction formula is
CN201710249329.5A 2017-04-17 2017-04-17 Transformer partial discharge fault diagnosis method based on hierarchical threshold synchronous extrusion wavelet Expired - Fee Related CN107356843B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710249329.5A CN107356843B (en) 2017-04-17 2017-04-17 Transformer partial discharge fault diagnosis method based on hierarchical threshold synchronous extrusion wavelet

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710249329.5A CN107356843B (en) 2017-04-17 2017-04-17 Transformer partial discharge fault diagnosis method based on hierarchical threshold synchronous extrusion wavelet

Publications (2)

Publication Number Publication Date
CN107356843A true CN107356843A (en) 2017-11-17
CN107356843B CN107356843B (en) 2020-05-08

Family

ID=60271486

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710249329.5A Expired - Fee Related CN107356843B (en) 2017-04-17 2017-04-17 Transformer partial discharge fault diagnosis method based on hierarchical threshold synchronous extrusion wavelet

Country Status (1)

Country Link
CN (1) CN107356843B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108918994A (en) * 2018-06-11 2018-11-30 北京印刷学院 A kind of duration power quality disturbances method of auto-adapted fitting Lifting Wavelet packet Reduction Analysis
CN108983051A (en) * 2018-07-25 2018-12-11 武汉科技大学 Based on the synchronous shelf depreciation kind identification method for squeezing wavelet transformation
CN110244202A (en) * 2019-06-21 2019-09-17 三峡大学 Based on synchronous compression wavelet transformed domain partial discharge of transformer denoising method
CN111044865A (en) * 2019-12-30 2020-04-21 杭州国洲电力科技有限公司 Partial discharge wavelet denoising method for direct-current gas insulation electrical equipment
CN112098102A (en) * 2020-09-04 2020-12-18 常州工学院 Internal combustion engine abnormal sound identification and diagnosis method based on EWT-SCWT

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6253175B1 (en) * 1998-11-30 2001-06-26 International Business Machines Corporation Wavelet-based energy binning cepstal features for automatic speech recognition
CN102129672B (en) * 2011-03-15 2013-06-12 西安电子科技大学 SWT domain improved particle filter-based SAR image despeckling method
CN105572501A (en) * 2015-12-17 2016-05-11 西安理工大学 Power quality disturbance identification method based on SST conversion and LS-SVM
CN105606892A (en) * 2015-12-17 2016-05-25 西安理工大学 Power grid harmonic and inter-harmonic analysis method based on SST transformation
CN105787655A (en) * 2016-02-24 2016-07-20 西安工业大学 Superhigh-layer structure modal parameter identification method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6253175B1 (en) * 1998-11-30 2001-06-26 International Business Machines Corporation Wavelet-based energy binning cepstal features for automatic speech recognition
CN102129672B (en) * 2011-03-15 2013-06-12 西安电子科技大学 SWT domain improved particle filter-based SAR image despeckling method
CN105572501A (en) * 2015-12-17 2016-05-11 西安理工大学 Power quality disturbance identification method based on SST conversion and LS-SVM
CN105606892A (en) * 2015-12-17 2016-05-25 西安理工大学 Power grid harmonic and inter-harmonic analysis method based on SST transformation
CN105787655A (en) * 2016-02-24 2016-07-20 西安工业大学 Superhigh-layer structure modal parameter identification method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
汪祥莉 等: "混沌背景下非平稳谐波信号的自适应同步挤压小波变换提取", 《物理学报》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108918994A (en) * 2018-06-11 2018-11-30 北京印刷学院 A kind of duration power quality disturbances method of auto-adapted fitting Lifting Wavelet packet Reduction Analysis
CN108918994B (en) * 2018-06-11 2021-12-10 北京印刷学院 Electric energy quality disturbance detection method for improving wavelet packet noise reduction analysis through self-adaptive fitting
CN108983051A (en) * 2018-07-25 2018-12-11 武汉科技大学 Based on the synchronous shelf depreciation kind identification method for squeezing wavelet transformation
CN108983051B (en) * 2018-07-25 2021-02-09 武汉科技大学 Partial discharge type identification method based on synchronous extrusion wavelet transform
CN110244202A (en) * 2019-06-21 2019-09-17 三峡大学 Based on synchronous compression wavelet transformed domain partial discharge of transformer denoising method
CN110244202B (en) * 2019-06-21 2021-07-20 三峡大学 Transformer partial discharge denoising method based on synchronous compression wavelet transform domain
CN111044865A (en) * 2019-12-30 2020-04-21 杭州国洲电力科技有限公司 Partial discharge wavelet denoising method for direct-current gas insulation electrical equipment
CN112098102A (en) * 2020-09-04 2020-12-18 常州工学院 Internal combustion engine abnormal sound identification and diagnosis method based on EWT-SCWT
CN112098102B (en) * 2020-09-04 2022-12-20 常州工学院 Internal combustion engine abnormal sound identification and diagnosis method based on EWT-SCWT

Also Published As

Publication number Publication date
CN107356843B (en) 2020-05-08

Similar Documents

Publication Publication Date Title
CN107356843A (en) The partial discharge of transformer method for diagnosing faults of small echo is synchronously extruded based on gradient threshold
CN107832777B (en) Electric energy quality disturbance identification method adopting time domain compression multi-resolution rapid S-transform feature extraction
CN103728535B (en) A kind of extra-high-voltage direct-current transmission line fault location based on wavelet transformation transient state energy spectrum
CN103630808B (en) A kind of partial discharge signal denoising method based on lifting wavelet transform
CN109142851B (en) Novel power distribution network internal overvoltage identification method
Wang et al. Noise suppression of corona current measurement from HVdc transmission lines
CN104730423A (en) Island effect detecting method of grid-connected photovoltaic power system
CN111239549A (en) Power distribution fault rapid positioning method based on discrete wavelet transform
CN109164362A (en) A kind of recognition methods and system of direct current cables shelf depreciation defect failure
Hussein et al. Denoising different types of acoustic partial discharge signals using power spectral subtraction
CN111914705A (en) Signal generation method and device for improving health state evaluation accuracy of reactor
Tabatabaei et al. A traveling-wave fault location technique for three-terminal lines based on wavelet analysis and recurrent neural network using GPS timing
CN114265010A (en) Transformer fault sound source positioning method and system based on microphone array
CN111723684A (en) Method for identifying transient overvoltage type in offshore wind farm
Liu et al. Dual-channel convolutional network-based fault cause identification for active distribution system using realistic waveform measurements
Ganjun et al. Partial discharge pattern recognition of high voltage cables based on the stacked denoising autoencoder method
CN105893976A (en) Parameterization identification method of traveling wave signal
Geng et al. Mechanical fault diagnosis of power transformer by GFCC time-frequency map of acoustic signal and convolutional neural network
Zhu et al. Complex disturbances identification: A novel PQDs decomposition and modeling method
CN113514743B (en) Method for constructing GIS partial discharge pattern recognition system based on multidimensional features
CN113642417A (en) Improved wavelet algorithm-based denoising method for partial discharge signals of insulated overhead conductor
CN117289013A (en) Data processing method and system for pulse current test
Leśniak et al. Application of Kalman filter to noise reduction in multichannel data
CN114157023B (en) Distribution transformer early warning information acquisition method
Ye et al. Fault classification of high voltage transmission line based on convolutional neural network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20200508

CF01 Termination of patent right due to non-payment of annual fee