CN109828181A - A kind of transformer winding minor failure detection method based on MODWT - Google Patents
A kind of transformer winding minor failure detection method based on MODWT Download PDFInfo
- Publication number
- CN109828181A CN109828181A CN201910020911.3A CN201910020911A CN109828181A CN 109828181 A CN109828181 A CN 109828181A CN 201910020911 A CN201910020911 A CN 201910020911A CN 109828181 A CN109828181 A CN 109828181A
- Authority
- CN
- China
- Prior art keywords
- fault
- turn
- transformer
- modwt
- minor failure
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The transformer winding minor failure detection method based on MODWT that the present invention relates to a kind of is included under simulated conditions and emulates to the slight shorted-turn fault of transformer generation and turn-to-turn Arcing fault；Electrical quantity is extracted, is distinguished using the difference of equivalent instant excitation inductance of the transformer under different operating conditions；In order to effectively extract fault characteristic value, the fault eigenvalue under various operating conditions is extracted using Maximum overlap wavelet transform.Fault detection and classification are carried out using the characteristic quantity of extraction, so as to realize the detection to various fault conditions.Prove that the algorithm proposed accurately can detect and distinguish excitation surge current, slight shorted-turn fault and turn-to-turn Arcing fault, so as to improve the reliability of electric system and the stability of operation finally by emulation.Compared with prior art, the present invention has many advantages, such as accuracy height, and arithmetic speed is fast.
Description
Technical field
The present invention relates to transformer winding fault detection technique fields, more particularly, to a kind of transformer based on MODWT
Winding minor failure detection method.
Background technique
Power transformer is one of expensive device in electric system, and any failure all may cause its long-time and interrupt fortune
Row, causes serious impact to power grid.And its cost of repairs is high, time-consuming.Therefore the protection of transformer and presence are examined
Survey is problem most important in electric system.
During transformer station high-voltage side bus, by various mechanical forces, the cumulative effect of mechanical force be will lead to for electrical winding and iron core
Electrical winding insulation runs down, and eventually leads to winding inter-turn failure.Find that these failures are usually with single turn by investigation
Failure starts, but if undiscovered in the initial stage, then failure the number of turns can be more, causes destructive malfunction.For above-mentioned feelings
Condition should be diagnosed early and be out of order, adopt an effective measure.Gas is depended at present to the identification of minor failure in fuel tank to protect
Shield, come determine whether occur in fuel tank slightly discharge, insulation oil temperature it is whether excessively high situations such as.But due to reflection insulation oil decomposition
The high oil pressure non-electric quantity generated, action sensitivity is low, poor reliability is often reported by mistake, has buried hidden danger for major accident.
Based on the proposition that equivalent instant excitation inductance defines, it can effectively identify transformer excitation flow to simulating, verifying, and in this base
Protection criteria for transformer fault identification is constructed on plinth, has achieved certain achievement in tranformer protection field.Artificial intelligence
Energy method is widely used to transformer fault diagnosis, and the characteristic quantity extracted from voltage and current signals is utilized to combine intelligence point
Analysis method detects failure.Failure is carried out using intelligent methods such as support vector machines, artificial neural network, Bayes classifiers
Classification and Identification.In addition, wavelet transform and Maximum overlap wavelet transform, S-transformation and Chirplet transformation etc. are extensively
For extracting feature vector to train classifier.But at present few documents propose based on equivalent instant excitation inductance this without exception
The slight turn-to-turn fault of identification transformer and turn-to-turn arc fault are read, to realize the online inspection to transformer winding minor failure
It surveys.There are larger difference, Maximum overlap discrete wavelet transformers for equivalent instant excitation inductance when transformer is operated normally and broken down
Change can effectively extract characteristic quantity, and due to not having down-sampling process, have rapidity.
Summary of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide a kind of changes based on MODWT
Depressor winding minor failure detection method, the difference based on equivalent instant excitation inductance under various operating conditions, in Matlab/
The slight shorted-turn fault model of transformer and turn-to-turn Arcing fault model are established in Simulink, it is slight to various degree
Failure carries out simulated extraction difference current and false voltage, seeks equivalent instant excitation inductance.Utilize Maximum overlap discrete wavelet
Transform analysis extract fault feature vector, training and test value as decision tree, realize failure modes, be suitable for transformer around
Group minor failure state-detection.
The purpose of the present invention can be achieved through the following technical solutions:
A kind of transformer winding minor failure detection method based on MODWT, comprising the following steps:
S1: transformer winding minor failure model is established；
S2: it extracts the electrical quantity of fault model and carries out equivalent instant excitation inductance calculating；
S3: extracting fault characteristic value to equivalent instant excitation inductance, carries out the classification of fault characteristic value for practical event
Barrier detection.
Further, the transformer winding minor failure model in the step S1, describes formula are as follows:
In formula, up(k) and up(k+1) be respectively kT He (k+1) T moment transformer primary side port voltage, r be it is equivalent
Resistance, T are sampling period, the transformer one that Δ i (k), Δ i (k+1), Δ i (k-1), Δ i (k+2) are inscribed when being respectively corresponding
The difference current of secondary side and secondary side, L are equivalent instant excitation inductance.
Further, the calculation formula of the equivalent instant excitation inductance in the step S2 are as follows:
It further, further include the braking electricity being arranged for preventing transformer differential protection from malfunctioning in the step S2
Stream, the calculation formula of the stalling current are as follows:
In formula, ithrFor stalling current, iCT1And iCT2Respectively current transformer is once held and the transient current of secondary terminals,
kresFor restraint coefficient.
Further, the step S3 include it is following step by step:
S301: the Maximum overlap wavelet transform mathematical modulo that equivalent instant excitation inductance extracts fault characteristic value is established
Type；
S302: classification processing is carried out to fault characteristic value.
Further, the Maximum overlap wavelet transform mathematical model in the step S301, describes formula are as follows:
In formula,For the energy of signal,For the approximation coefficient energy on scale J,For the wavelet coefficient energy on scale J, k and j are natural number.
Compared with prior art, the invention has the following advantages that
(1) accuracy is high, the present invention is based on difference of the equivalent instant excitation inductance under various operating conditions,
The slight shorted-turn fault model of transformer and turn-to-turn Arcing fault model are established in Matlab/Simulink, to various
Degree minor failure carries out simulated extraction difference current and false voltage, seeks equivalent instant excitation inductance.Utilize Maximum overlap
Fault feature vector is extracted in wavelet transform analysis, and training and test value as decision tree are realized failure modes, be suitable for
Transformer winding minor failure state-detection, accuracy are high.
(2) judge that speed is fast, the present invention utilizes the equivalent instant excitation inductance under the various operating conditions of calculating transformer, proposes base
In the Maximum overlap discrete wavelet transformer of db4 wavelet function bring effectively extract fault signature, using the Maximum overlap based on db4 from
Scattered wavelet transformation can effectively extract fault eigenvalue, to realize the accurate inspection to transformer winding fault using traditional decision-tree
It surveys.
Detailed description of the invention
Fig. 1 is transformer T-type equivalent circuit diagram provided by the invention；
Fig. 2 is fault detection provided by the invention and classification process figure；
Fig. 3 is system simulation model figure provided by the invention；
Fig. 4 is transformer no-load switchon surge waveform diagram provided by the invention；
Fig. 5 is normal transformer equivalent instant excitation inductance waveform diagram provided by the invention；
Fig. 6 is that the MODWT provided by the invention based on db4 wavelet function decomposes (1.92% shorted-turn fault) figure,
In, Fig. 6 (a) is 1.92% shorted-turn fault equivalent instant excitation inductance waveform diagram, and Fig. 6 (b) is 1.92% turn-to-turn short circuit event
L1 waveform diagram after barrier decomposes, Fig. 6 (c) are l2 waveform diagram after 1.92% shorted-turn fault decomposes, and Fig. 6 (d) is 1.92% turn-to-turn
L3 waveform diagram after short trouble decomposes, Fig. 6 (e) are l4 waveform diagram after 1.92% shorted-turn fault decomposes；
Fig. 7 is that the MODWT provided by the invention based on db4 wavelet function decomposes (3.85% shorted-turn fault) figure,
In, Fig. 7 (a) is 3.85% shorted-turn fault equivalent instant excitation inductance waveform diagram, and Fig. 7 (b) is 3.85% turn-to-turn short circuit event
L1 waveform diagram after barrier decomposes, Fig. 7 (c) are l2 waveform diagram after 3.85% shorted-turn fault decomposes, and Fig. 7 (d) is 3.85% turn-to-turn
L3 waveform diagram after short trouble decomposes, Fig. 7 (e) are l4 waveform diagram after 3.85% shorted-turn fault decomposes；
Fig. 8 is that the MODWT provided by the invention based on db4 wavelet function decomposes (l=5mm turn-to-turn Arcing fault)
Figure, wherein Fig. 8 (a) is l=5mm turn-to-turn Arcing fault equivalent instant excitation inductance waveform diagram, and Fig. 8 (b) is l=5mm circle
Between Arcing fault decompose after l1 waveform diagram, Fig. 8 (c) be l=5mm turn-to-turn Arcing fault decompose after l2 waveform diagram, Fig. 8
(d) l3 waveform diagram after decomposing for l=5mm turn-to-turn Arcing fault, Fig. 8 (e) are the decomposition of l=5mm turn-to-turn Arcing fault
L4 waveform diagram afterwards；
Fig. 9 is decision tree classification structure chart provided by the invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiment is a part of the embodiments of the present invention, rather than whole embodiments.Based on this hair
Embodiment in bright, those of ordinary skill in the art's every other reality obtained without making creative work
Example is applied, all should belong to the scope of protection of the invention.
Embodiment
One: equivalent instant excitation inductance
Transformer T-type equivalent circuit is as shown in Figure 1.
Transformer T-type equivalent circuit loop equation are as follows:
In formula, up、ip、Lp、RpRespectively primary side winding end voltage, electric current, leakage inductance and resistance；Rm、imRespectively excitation electricity
Resistance and exciting current.
Transformer station high-voltage side bus when (normal operation, malfunction, zero load and load running), is based on equivalent under various operating conditions
The transformer equation of instantaneous excitation inductance is as follows:
In formula: upFor transformer primary side port voltage, r is equivalent resistance, and L is equivalent instant excitation inductance, △ i=ip+
isFor the difference current of transformer primary side and secondary side.
According to transformer T-type equivalent circuit, the continuous differential equation is turned to the discrete differential equation according to trapezoidal method, this
When consider the influence of equivalent resistance, improve the accuracy of equivalent instant excitation inductance calculating, the then system of kT and (k+1) T moment
One equation has following form:
In formula, up(k) and up(k+1) be respectively kT He (k+1) T moment transformer primary side port voltage, r be it is equivalent
Resistance, T are sampling period, the transformer one that Δ i (k), Δ i (k+1), Δ i (k-1), Δ i (k+2) are inscribed when being respectively corresponding
The difference current of secondary side and secondary side, L are equivalent instant excitation inductance.
Equivalent resistance is eliminated, kT moment equivalent instantaneous inductor is obtained:
Two: Maximum overlap wavelet transform
Wavelet transform (DWT) is commonly used in the time of analysis non-stationary signal and the conversion of spectral characteristic.It is discrete small
Wave variation is easier to realize compared with continuous wavelet transform.In the operation of wavelet transform, equivalent instant excitation inductance is logical
It crosses low-pass filter and obtains approximate signal, detail signal is obtained by high-pass filter.In a subsequent step, approximate letter
Number processing mode it is identical as main signal, that is, obtain second decompose after approximation and detail signal.Wavelet transform is determined
Adopted formula is as follows:
In formula: xm,nFor discrete morther wavelet.
Maximum overlap wavelet transform is similar to wavelet transform, but compared with DWT, MODWT does not have down-sampling
Process can thus detect the transition in electrical quantity to quickly detect failure quickly.Meanwhile MODWT can apply
The sample of any size.Detail coefficients and approximation coefficient after MODWT decomposition have translation invariance, and its all Decomposition order
Time having the same rate respectively.One important step of pattern-recognition is to extract important feature from various interference informations.For
The performance of classifier is improved, the present invention using MODWT analyzes sampled signal.
MODWT is obtained by flexible formula, can be obtained:
In formula: φ (x) and ψ (x) is respectively father's small echo and morther wavelet.Low-pass filter coefficients gkWith high-pass filter coefficient
hk, it is as follows:
The coefficient of high-pass filter coefficient and low-pass filter is related with selected morther wavelet.
The calculation formula of detail coefficients and approximation coefficient under first scale is as follows:
According to Parseval's theorem it is found that the energy of signal can be decomposed at each scale j={ 1,2 ..., J }
Wavelet coefficient energy and the approximation coefficient energy on scale J, as follows:
In formula,For the energy of signal,For the approximation coefficient energy on scale J,For the wavelet coefficient energy on scale J, k and j are natural number.
As it appears from the above, MODWT can be used to obtain the Energy Decomposition based on scale.According in the past researched and analysed it is found that
It can be realized the accurate detection to transition in electric power networks using Daubechies 4 (db4), there is maturation in the power system
Application.The present invention chooses the db4 wavelet function in small echo family, and MODWT is applied to initial data.
Extracting effective feature facilitates subsequent study and classification, while can also reduce the dimension of data.It carries out maximum
Be overlapped wavelet transform when, select wavelet transformation appropriate that can help to extract required feature vector, to classification method into
The accurate training of row, more accurately detects the operating status of studied transformer, and improve system reliability.
On this basis, the present invention builds transformer winding minor failure and turn-to-turn electric arc in Matlab/Simulink
Discharge fault model extracts transformer station high-voltage side bus difference current and original edge voltage under various operating conditions, calculates Equivalent Instantaneous excitation
Inductance.Then, using the MODWT based on db4 wavelet basis function, detail signal is obtained in best frequency domain region.Finally by thin
Section coefficient extracts characteristic value for classifying.
Characteristic value is extracted by MODWT, for the detection and sort research to transformer minor failure.The present invention is mentioned
The flow chart of algorithm out is as shown in Figure 2.Transformer differential protection malfunctions in order to prevent, introduces stalling current, following institute
Show:
In formula, ithrFor stalling current, iCT1And iCT2Respectively current transformer is once held and the transient current of secondary terminals,
kresFor restraint coefficient.
Three: classification
The purpose of classification method is the class that identification belongs to object from some descriptive characteristics.Decision tree be for classify and
The nonparametric supervised learning method of recurrence.Its fundamental difference of decision tree classification based on algorithms of different is Split Attribute selection
Judgment criteria.
ID3 algorithm selects Split Attribute by information gain, and information gain is the difference of variation front and back entropy, and entropy is information
Desired value.If a given sample S and probability distribution P=(p1,p2,...,pn), i.e., the information that this distribution carries is known as
Entropy.Its definition is as follows:
The entropy of sample set after being divided by attribute A are as follows:
In formula: m is m sub- sample sets being divided by the attribute value of attribute A, | Sj| indicate that j-th of subsample is concentrated
Sample size, | S | total sample number amount in data set before indicating to divide.
The information gain of sample set at this time are as follows:
InfoGain (S, A)=Entropie (S)-EntropieA(S)
C4.5 algorithm is to select Split Attribute by information gain-ratio, overcomes and is selected to be biased to choosing when attribute with information gain
Select the deficiency of the more attribute of value.C4.5 algorithm uses following methods:
Selection: for dividing training data (for classification based training data)；
Termination condition: determine when stop dividing (termination condition for determining segmentation)；
Pruning algorithms: the algorithm attempts to prevent overfitting；
The information gain-ratio of sample set is defined as follows after being divided by attribute A:
In formula,Wherein, training dataset S is divided by the attribute value of attribute A
M Sub Data Set, | Sj| indicate sample size in j-th of Sub Data Set, | S | total sample number in data set before indicating to divide
Amount.
C4.5 algorithm using the pessimistic beta pruning method in rear pruning method, this method be according to the error rate before and after beta pruning come
Determine whether to carry out the trimming of subtree.It is to generate decision tree using training set training set is recycled to carry out beta pruning, i.e., such side
Method not want independent beta pruning collection.
In the present invention, the decision tree classifier based on C4.5 algorithm is for illustrating to assume.If problem includes N number of spy
Sign, then the maximum height of decision tree is N.One advantage of C4.5 algorithm is that tree will be trained to be converted to a series of if-then rules.
In addition, this method does not need Feature Extraction Technology.And its attribute type for being capable of handling discrete type and continuous type, it is capable of handling tool
There is the training data of missing attribute values.
Four: simulation analysis
Simulation model is established in Matlab/Simulink, carries out the on-line checking of transformer winding state, system emulation
Model is as shown in Figure 3.The rated capacity of single-phase two-winding transformer is 150MVA, voltage rating 500kV/220kV.When emulation
Sample frequency be set as 3.6kHz, simulation time is set as 0.2s, then number of sampling points be 720.Circuit Fault on Secondary Transformer around
When shorted-turn fault occurs for group, short-circuit resistance value is fixed value resistance.When turn-to-turn arc discharge occurs for secondary side winding, resistance
Value is replaced using Mayr Arc Modelling (nonlinear time-varying Arc Modelling), establishes the slight turn-to-turn short circuit of winding and turn-to-turn electric arc respectively
Discharge fault model carries out simulation analysis.
Primary side current waveform diagram when Fig. 4 is normal transformer during no-load closing, closing time 1/30s.Normal transformer
It is biased to time shaft side containing a large amount of aperiodic component in excitation surge current when idle-loaded switching-on, meeting constantly decays at any time and waveform
There are interval angles.Equivalent instant excitation inductance, such as Fig. 5 when transformer operates normally can be acquired according to the electric current of emulation and voltage
It is shown.
In t=0.1s, shorted-turn fault occurs for inside transformer, by the short-circuited winding circle for changing transformer winding
Number emulates the shorted-turn fault of 1.92%~5.77% degree.Failure corresponds to sampled point when occurring be 360, chooses a week
Equivalent instant excitation inductance in wave is analyzed.
When 1.92% slight shorted-turn fault occurs for step down side winding, the maximum based on db4 small echo is utilized
Overlapping wavelet transform is decomposed.Fig. 6 (a)~(e) and Fig. 7 (a)~(e) is respectively that 1.92% and 3.85% turn-to-turn occurs
The waveform after equivalent instant excitation inductance and decomposition when short trouble.Characteristic value is extracted, using them as feature vector application
In classifier, to carry out Accurate classification to various operating statuses.
It is free, thermal inertia and Re Ping based on heat using Mayr Simulation of Arc Models winding inter-turn Arcing fault
A kind of dynamic arc model that the three kinds of principles that weigh are established, the characteristic Simulation suitable for low current electric arc.If assuming, arc length is
L, Mayr Arc Modelling equation are as follows:
Wherein: dg/dt is the conductance change rate of unit length arc length；G is the electric arc conductance of unit length.U=el is electricity
Arc voltage；P0=PLL is the heat radiation power of electric arc.Length by changing electrical discharge arc in model distinguishes simulated arc length and is
The turn-to-turn Arcing fault of 5mm~10mm.It is illustrated in figure 8 and occurs corresponding to the turn-to-turn Arcing fault that arc length is 5mm
Equivalent instant excitation inductance and its decomposition after figure.
The turn-to-turn electric arc that 5mm~10mm occurs to transformer respectively emulates, and extracts electrical quantity and various works are calculated
Equivalent instant excitation inductance under condition carries out above-mentioned decomposition and extracts fault characteristic value.Extracted fault characteristic value is used for event
Barrier classification.
When being classified using classifier to fault type, input includes the various events extracted from above-mentioned simulation analysis
Hinder feature vector, export as various fault types: slight shorted-turn fault (C1), turn-to-turn Arcing fault (C2) and excitation
Shove (C3).When corresponding to different switching angles under 25%, 50%, 90% load and no-load condition excitation surge current occurs for emulation
Equivalent instant excitation inductance as classification samples.Simulation analysis is carried out using simulation model, every kind of fault type takes 30 groups of numbers
According to sharing 90 groups of data in sample set, i.e. sample point can be distributed evenly in various fault types.In training dataset,
Using 70% data set as training sample, i.e. 21 groups of data in every kind of operating status are as training sample, totally 63 groups of training
Sample.Remaining 30% is used as test sample, totally 27 groups of test samples.The algorithm proposed is used to detect the operation shape of transformer
State.The characteristic value of extraction is imported into 8.1 software of Rapidminer Studio, which has abundant data mining analysis
And algorithm function.When in use, the training method of selection division verifying, i.e., 70% data set is training set, and 30% is sample
Collection.Finally, having obtained the classification accuracy of each classification on this basis.Fig. 9 is the decision tree structure of classification.
According to Fig. 9 it will be seen that decision tree utilizes the characteristic value l1- extracted from Maximum overlap wavelet transform
L4 classifies, and comes differentiating transformer exciting surge, slight shorted-turn fault and turn-to-turn partial discharges fault.Table 1 lists
Testing result when using LM-BP training neural network and using C4.5 training decision tree.
From table 1 it will be seen that when using the decision tree based on C4.5 algorithm, for the verification and measurement ratio of three kinds of failures
It is all up to 90% or more, the detection for excitation surge current situation, accuracy rate has reached 100%.For using artificial neural network
When being detected, for detecting 22 groups of samples in 30 groups of samples of C1 (slight shorted-turn fault), accuracy is
70.33%；Verification and measurement ratio correct for C2 classification is 70%.Overall, the verification and measurement ratio of ANN is 76.78%, and the inspection of DT
Survey rate is 96.67%.The characteristic value l1-l4 extracted using the MODWT based on db4 wavelet function, as ANN and decision tree
The training set and test set of classification are able to achieve to the slight shorted-turn fault of transformer winding, turn-to-turn Arcing fault and encourage
The detection and identification that magnetic shoves, the Decision-Tree Method accuracy with higher based on C4.5 algorithm, can accurately examine
Transformer winding fault is measured, to realize failure modes.
1 failure modes result of table
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can readily occur in various equivalent modifications or replace
It changes, these modifications or substitutions should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with right
It is required that protection scope subject to.
Claims (6)
1. a kind of transformer winding minor failure detection method based on MODWT, which comprises the following steps:
S1: transformer winding minor failure model is established；
S2: it extracts the electrical quantity of fault model and carries out equivalent instant excitation inductance calculating；
S3: extracting fault characteristic value to equivalent instant excitation inductance, carries out the classification of fault characteristic value to examine for physical fault
It surveys.
2. a kind of transformer winding minor failure detection method based on MODWT according to claim 1, feature exist
In transformer winding minor failure model in the step S1 describes formula are as follows:
In formula, up(k) and up(k+1) be respectively kT He (k+1) T moment transformer primary side port voltage, r is equivalent resistance,
T is the sampling period, the transformer primary side that Δ i (k), Δ i (k+1), Δ i (k-1), Δ i (k+2) are inscribed when being respectively corresponding with
The difference current of secondary side, L are equivalent instant excitation inductance.
3. a kind of transformer winding minor failure detection method based on MODWT according to claim 1, feature exist
In the calculation formula of the equivalent instant excitation inductance in the step S2 are as follows:
4. a kind of transformer winding minor failure detection method based on MODWT according to claim 1, feature exist
In, further include in the step S2 be arranged for prevent transformer differential protection malfunction stalling current, the stalling current
Calculation formula are as follows:
In formula, ithrFor stalling current, iCT1And iCT2Respectively current transformer is once held and the transient current of secondary terminals, kresFor
Restraint coefficient.
5. a kind of transformer winding minor failure detection method based on MODWT according to claim 1, feature exist
In, the step S3 include it is following step by step:
S301: the Maximum overlap wavelet transform mathematical model that equivalent instant excitation inductance extracts fault characteristic value is established；
S302: classification processing is carried out to fault characteristic value.
6. a kind of transformer winding minor failure detection method based on MODWT according to claim 5, feature exist
In Maximum overlap wavelet transform mathematical model in the step S301 describes formula are as follows:
In formula,For the energy of signal,For the approximation coefficient energy on scale J,
For the wavelet coefficient energy on scale J, k and j are natural number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910020911.3A CN109828181A (en) | 2019-01-09 | 2019-01-09 | A kind of transformer winding minor failure detection method based on MODWT |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910020911.3A CN109828181A (en) | 2019-01-09 | 2019-01-09 | A kind of transformer winding minor failure detection method based on MODWT |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109828181A true CN109828181A (en) | 2019-05-31 |
Family
ID=66861550
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910020911.3A Pending CN109828181A (en) | 2019-01-09 | 2019-01-09 | A kind of transformer winding minor failure detection method based on MODWT |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109828181A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110350481A (en) * | 2019-07-18 | 2019-10-18 | 杭州电力设备制造有限公司 | A kind of differential protecting method of transformer, system, equipment and storage medium |
| CN111339872A (en) * | 2020-02-18 | 2020-06-26 | 国网信通亿力科技有限责任公司 | Power grid fault classification method based on classification model |
| CN112304207A (en) * | 2020-10-20 | 2021-02-02 | 上海电力大学 | Transformer winding deformation online detection method by using leakage inductance parameter change estimation |
| CN112327208A (en) * | 2020-11-02 | 2021-02-05 | 国网江苏省电力有限公司电力科学研究院 | Fault diagnosis method and device for turn-to-turn short circuit of phase modulator rotor winding |
| CN112819059A (en) * | 2021-01-26 | 2021-05-18 | 中国矿业大学 | Rolling bearing fault diagnosis method based on popular retention transfer learning |
| CN112986868A (en) * | 2021-04-16 | 2021-06-18 | 成都工百利自动化设备有限公司 | Transformer state monitoring method |
| CN115047240A (en) * | 2022-05-17 | 2022-09-13 | 国网湖北省电力有限公司黄冈供电公司 | Transformer magnetizing inrush current discrimination method using wavelet detail component change characteristics |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070191725A1 (en) * | 2006-02-10 | 2007-08-16 | Nelson Alex T | Wavelet transform and pattern recognition method for heart sound analysis |
| EP2265977A1 (en) * | 2008-04-07 | 2010-12-29 | Chevron U.S.A. Inc. | Lithofacies classification system and method |
| CN102879671A (en) * | 2012-09-17 | 2013-01-16 | 华北电力大学 | Method for judging inrush current locking of equivalent instantaneous inductance for protection of extra-high voltage regulating transformer |
| CN103413142A (en) * | 2013-07-22 | 2013-11-27 | 中国科学院遥感与数字地球研究所 | Remote sensing image land utilization scene classification method based on two-dimension wavelet decomposition and visual sense bag-of-word model |
| CN103633622A (en) * | 2013-12-11 | 2014-03-12 | 国家电网公司 | Method and system for ultra-high voltage regulating transformer excitation inrush current identification |
| CN107451557A (en) * | 2017-07-29 | 2017-12-08 | 吉林化工学院 | Transmission line short-circuit fault diagnostic method based on experience wavelet transformation and local energy |
| CN107677904A (en) * | 2017-09-21 | 2018-02-09 | 广东电网有限责任公司电力科学研究院 | A kind of voltage dip origin cause of formation discrimination method and system |
| CN107727344A (en) * | 2017-09-12 | 2018-02-23 | 国网天津市电力公司电力科学研究院 | A kind of transformer based on wavelet energy method collides record data analysis method |
| US20180166089A1 (en) * | 2013-03-15 | 2018-06-14 | Vios Medical Singapore Pte. Ltd. | Method and system for signal decomposition, analysis and reconstruction |
-
2019
- 2019-01-09 CN CN201910020911.3A patent/CN109828181A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070191725A1 (en) * | 2006-02-10 | 2007-08-16 | Nelson Alex T | Wavelet transform and pattern recognition method for heart sound analysis |
| EP2265977A1 (en) * | 2008-04-07 | 2010-12-29 | Chevron U.S.A. Inc. | Lithofacies classification system and method |
| CN102879671A (en) * | 2012-09-17 | 2013-01-16 | 华北电力大学 | Method for judging inrush current locking of equivalent instantaneous inductance for protection of extra-high voltage regulating transformer |
| US20180166089A1 (en) * | 2013-03-15 | 2018-06-14 | Vios Medical Singapore Pte. Ltd. | Method and system for signal decomposition, analysis and reconstruction |
| CN103413142A (en) * | 2013-07-22 | 2013-11-27 | 中国科学院遥感与数字地球研究所 | Remote sensing image land utilization scene classification method based on two-dimension wavelet decomposition and visual sense bag-of-word model |
| CN103633622A (en) * | 2013-12-11 | 2014-03-12 | 国家电网公司 | Method and system for ultra-high voltage regulating transformer excitation inrush current identification |
| CN107451557A (en) * | 2017-07-29 | 2017-12-08 | 吉林化工学院 | Transmission line short-circuit fault diagnostic method based on experience wavelet transformation and local energy |
| CN107727344A (en) * | 2017-09-12 | 2018-02-23 | 国网天津市电力公司电力科学研究院 | A kind of transformer based on wavelet energy method collides record data analysis method |
| CN107677904A (en) * | 2017-09-21 | 2018-02-09 | 广东电网有限责任公司电力科学研究院 | A kind of voltage dip origin cause of formation discrimination method and system |
Non-Patent Citations (3)
| Title |
|---|
| R. P. MEDEIROS 等: "Differential protection of power transformers using the wavelet transform", 《2014 IEEE PES GENERAL MEETING | CONFERENCE & EXPOSITION》 * |
| 孙玉胜 等: "电力变压器差动保护研究现状与发展趋势", 《电工电气》 * |
| 邓祥力 等: "基于等效励磁电感波形相似度的变压器绕组轻微故障检测方法研究", 《高压电器》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110350481A (en) * | 2019-07-18 | 2019-10-18 | 杭州电力设备制造有限公司 | A kind of differential protecting method of transformer, system, equipment and storage medium |
| CN111339872A (en) * | 2020-02-18 | 2020-06-26 | 国网信通亿力科技有限责任公司 | Power grid fault classification method based on classification model |
| CN112304207A (en) * | 2020-10-20 | 2021-02-02 | 上海电力大学 | Transformer winding deformation online detection method by using leakage inductance parameter change estimation |
| CN112327208A (en) * | 2020-11-02 | 2021-02-05 | 国网江苏省电力有限公司电力科学研究院 | Fault diagnosis method and device for turn-to-turn short circuit of phase modulator rotor winding |
| CN112819059A (en) * | 2021-01-26 | 2021-05-18 | 中国矿业大学 | Rolling bearing fault diagnosis method based on popular retention transfer learning |
| CN112819059B (en) * | 2021-01-26 | 2022-03-29 | 中国矿业大学 | Rolling bearing fault diagnosis method based on popular retention transfer learning |
| CN112986868A (en) * | 2021-04-16 | 2021-06-18 | 成都工百利自动化设备有限公司 | Transformer state monitoring method |
| CN112986868B (en) * | 2021-04-16 | 2021-08-31 | 成都工百利自动化设备有限公司 | Transformer state monitoring method |
| CN115047240A (en) * | 2022-05-17 | 2022-09-13 | 国网湖北省电力有限公司黄冈供电公司 | Transformer magnetizing inrush current discrimination method using wavelet detail component change characteristics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109828181A (en) | A kind of transformer winding minor failure detection method based on MODWT | |
| Afrasiabi et al. | Integration of accelerated deep neural network into power transformer differential protection | |
| Wang et al. | ArcNet: Series AC arc fault detection based on raw current and convolutional neural network | |
| CN110082640B (en) | Distribution network single-phase earth fault identification method based on long-time memory network | |
| Atthapol et al. | Internal Fault Classification in Transformer Windings using Combination of Discrete Wavelet-Transforms and Back-propagation Neural Networks | |
| CN106707094A (en) | Classification recognition method of arc fault of low voltage power supply and distribution line | |
| CN112381668B (en) | Information extraction method for power grid faults | |
| CN112462193B (en) | Automatic reclosing judgment method for power distribution network based on real-time fault filtering data | |
| CN106203382A (en) | A kind of excitation surge current based on kernel function extreme learning machine and fault current recognition methods | |
| Gomes et al. | High-impedance faults in power distribution systems: A narrative of the field’s developments | |
| CN113850330A (en) | Power distribution network fault cause detection method based on short-time Fourier transform and convolutional neural network | |
| Moradzadeh et al. | Location of disk space variations in transformer winding using convolutional neural networks | |
| CN107765065B (en) | Fundamental wave attenuation factor-based power distribution network magnetizing inrush current identification method | |
| Li et al. | A fault pattern and convolutional neural network based single-phase earth fault identification method for distribution network | |
| CN112986754A (en) | Small current grounding system fault identification method and device based on data driving | |
| Jiang et al. | Machine learning approach to detect arc faults based on regular coupling features | |
| CN111999591B (en) | Method for identifying abnormal state of primary equipment of power distribution network | |
| CN113848471B (en) | Intelligent fault positioning method and system for relay protection system | |
| CN113610119A (en) | Method for identifying power transmission line developmental fault based on convolutional neural network | |
| CN113376474A (en) | Neural network fault arc identification system and method based on generalized S transformation | |
| Adly et al. | A novel wavelet packet transform based fault identification procedures in HV transmission line based on current signals | |
| Wu et al. | Transmission line fault cause identification method based on transient waveform image and MCNN-LSTM | |
| Afrasiabi et al. | A comparative analysis of artificial intelligence for power transformer differential protection | |
| Murali et al. | Forecasting and Classification Of Power Quality Disturbance In Smart Grid Using Hybrid Networks | |
| Alsafasfeh | Pattern recognition for fault detection, classification, and localization in electrical power systems |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190531 |