CN109033612A - A kind of Diagnosis Method of Transformer Faults based on vibration noise and BP neural network - Google Patents
A kind of Diagnosis Method of Transformer Faults based on vibration noise and BP neural network Download PDFInfo
- Publication number
- CN109033612A CN109033612A CN201810805455.9A CN201810805455A CN109033612A CN 109033612 A CN109033612 A CN 109033612A CN 201810805455 A CN201810805455 A CN 201810805455A CN 109033612 A CN109033612 A CN 109033612A
- Authority
- CN
- China
- Prior art keywords
- transformer
- neural network
- vibration
- neuron
- vibration signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Geometry (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Testing Electric Properties And Detecting Electric Faults (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a kind of Diagnosis Method of Transformer Faults based on vibration noise and BP neural network, it is related to transformer fault processing technology field, the following steps are included: S1, acquiring by Noise Sources Identification module the vibration noise sound pressure signal of transformer each region, and the region where maximum noise source is obtained according to the vibration noise sound pressure signal；S2, by vibration measurement module to the region where the maximum noise source, acquire vibration signal；S3, transformer fault diagnosis is carried out to the vibration signal using BP neural network algorithm.The present invention passes through acquisition of the S1 and S2 to transformer noise sound pressure signal, carry out the combination of transformer fault diagnosis to the vibration signal by BP neural network algorithm again, the precision of fault diagnosis is greatly improved, to solve the disadvantage by the vibration signal of transformer state information to transformer fault diagnosis difficulty.
Description
Technical field
The invention belongs to transformer fault processing technology fields, more particularly to one kind to be based on vibration noise and BP neural network
Diagnosis Method of Transformer Faults.
Background technique
Power transformer is one of most important equipment in electric system, once transformer breaks down, it will to power grid
Cause tremendous influence.Therefore, extremely important to the maintenance of transformer, detection work.However, transformer device structure is complicated, maintenance
Heavy workload, difficulty are higher.And regular dismounting transformer, is also easily damaged component, and reduce transformer station high-voltage side bus can
By property.Band electric-examination fault diagnosis art seems especially urgent for transformer fault diagnosis.In the method for transformer fault diagnosis,
Vibratory drilling method is developed in recent years more rapidly and with the good reliable transformer fault diagnosis technology for using prospect.Vibratory drilling method passes through
The mailbox vibration signal of transformer is analyzed to carry out fault diagnosis.Transformer vibration signal includes the fault message of transformer, but
It is that the vibration signal of comprehensive transformer state information is obtained since volume of transformer is huge is a big difficulty.Therefore, power grid is anxious
A kind of more efficient way is needed to carry out fault diagnosis to transformer.
Summary of the invention
The purpose of the present invention is to provide one kind, to solve the existing vibration signal pair by transformer state information
The disadvantage of transformer fault diagnosis difficulty.
To achieve the above object, the present invention provides a kind of transformer faults based on vibration noise and BP neural network to examine
Disconnected method, comprising the following steps:
S1, pass through Noise Sources Identification module, acquire the vibration noise sound pressure signal of transformer each region, and according to described
Vibration noise sound pressure signal obtains the region where maximum noise source；
S2, by vibration measurement module to the region where the maximum noise source, acquire vibration signal；
S3, transformer fault diagnosis is carried out to the vibration signal using BP neural network algorithm.
Further, the S1 includes:
S10, the vibration noise sound pressure signal is subjected to Fourier's variation, obtains sound pressure level maximum frequency range；
S11, the corresponding vibration noise sound pressure signal of the sound pressure level maximum frequency range is subjected to beamforming algorithm calculating, obtained
To the location of the maximum noise source of vibration noise acoustic pressure in each sound source region in transformer-cabinet surface；
S12, region locating for maximum noise source is determined according to the location of described maximum noise source.
Further, the vibration measurement module adopts the vibration signal using acceleration transducer array
Collection.
Further, the S3 includes:
S30, the training sample data and each group training sample data that multiple groups vibration signal is obtained according to the vibration signal
Corresponding transformer fault type；
S31, it handles the progress of training sample data described in each group FFT to obtain the characteristic quantity of vibration signal, the characteristic quantity
As the neuron of BP neural network input layer, the quantity of the characteristic quantity is that input layer quantity is m；
S32, characteristic quantity described in each group is normalized, obtains the normalization data of each group characteristic quantity；
S33, the transformer fault type is encoded, obtains the quantity of the transformer fault type, the change
Depressor fault type is the neuron of BP neural network output layer, and the quantity of the transformer fault type is output layer neuron
Quantity n；
S34, to be that m and output layer neuron quantity n obtains BP neural network by the input layer quantity implicit
Layer neuronal quantity h；
It S35, is m, output layer neuron quantity n and BP neural network hidden layer mind according to the input layer quantity
Initial BP neural network is constructed through first quantity h；
Characteristic quantity and transformer fault type corresponding with characteristic quantity described in each group described in S36, foundation each group, to the BP
Neural network is trained, the BP neural network after being trained；
S37, the characteristic quantity for extracting transformer vibration signal to be diagnosed are carried out using the BP neural network obtained after training
Diagnosis, obtains its fault type.
Further, the characteristic quantity of the vibration signal includes: fundamental frequency specific gravity, basic amplitude, dominant frequency specific gravity, dominant frequency amplitude
And vibrational entropy.
Further, the transformer fault type includes: fault-free, iron core failure and winding failure.
Further, the S34 includes:
S3400, hidden layer neuron quantity is calculated using following formula:
In formula (6), h is hidden layer neuron quantity, and regulating constant of a between 1-10, m is input layer number
Amount, n are output layer neuron quantity；
S3401, from hminStart, increases neuron number one by one until hmax, by [hmin,hmax] be trained test respectively
Card；
S3402, neuron number corresponding to optimal verification result, the neuron number in training verification result are chosen
For BP neural network hidden layer neuron quantity h.
Further, in the S36 BP neural network training the following steps are included:
S3600, the characteristic quantity that step S31 is obtained to each group training sample data input the BP neural network, obtain defeated
The error of result and output result out；
S3601, judge that the error is no less than presetting threshold value, if the error is less than presetting threshold value, into
Enter step S3602；If the error is greater than presetting threshold value, S3603 is entered step；
S3602, the initial BP neural network carry out back transfer using gradient descent method, along relative error quadratic sum
Direction of steepest descent, the continuous weight and threshold value for adjusting network reversely counted using output layer, hidden layer and input layer
Calculate, export each layer as a result, return step S35；
S3603, the BP neural network after training is constructed using the output result.
Compared with prior art, the invention has the following beneficial effects:
1, a kind of Diagnosis Method of Transformer Faults based on vibration noise and BP neural network provided by the present invention, passes through
Noise Sources Identification module obtains the region where transformer maximum noise source, acquires signal by vibration measurement module, most
Transformer fault is carried out to the vibration signal using BP neural network algorithm by BP neural network fault diagnosis module afterwards to examine
It is disconnected.By using various features amount in BP neural network fault diagnosis module, BP neural network is carried out by various data
It calculates, to solve the disadvantage by the vibration signal of transformer state information to transformer fault diagnosis difficulty；By making an uproar
The combination of identification of sound source module, vibration measurement module and BP neural network fault diagnosis module, greatly improves failure
The precision of diagnosis.
2, vibration measurement module provided by the present invention is using acceleration transducer array to where maximum noise source
Region acquire vibration signal, acceleration transducer array is arranged in nearest from vibration noise source where maximum noise source
At transformer case, the most abundant vibration signal of information can be collected, provides data basis for fault diagnosis.
Detailed description of the invention
It, below will be to attached drawing needed in embodiment description in order to illustrate more clearly of technical solution of the present invention
It is briefly described, it should be apparent that, the accompanying drawings in the following description is only one embodiment of the present of invention, general for this field
For logical technical staff, without creative efforts, it is also possible to obtain other drawings based on these drawings.
Fig. 1 is a kind of flow chart of the Diagnosis Method of Transformer Faults based on vibration noise and BP neural network of the present invention.
Specific embodiment
With reference to the attached drawing in the embodiment of the present invention, the technical solution in the present invention is clearly and completely described,
Obviously, described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Based in the present invention
Embodiment, those of ordinary skill in the art's every other embodiment obtained without creative labor,
It shall fall within the protection scope of the present invention.
As shown in Figure 1, the Diagnosis Method of Transformer Faults provided by the present invention based on vibration noise and BP neural network
The following steps are included:
S1, microphone array is used by Noise Sources Identification module, acquires the vibration noise acoustic pressure of transformer each region
Signal Pn, and the region where maximum noise source is obtained according to vibration noise sound pressure signal.It can using Noise Sources Identification module
Position and the intensity in vibration noise source are presented by the formal intuition of image.S1 the following steps are included:
S10, by vibration noise sound pressure signal PnFourier's variation is carried out, sound pressure level maximum frequency range f is obtainedmax；
S11, by sound pressure level maximum frequency range fmaxCorresponding vibration noise sound pressure signal carries out beamforming algorithm calculating, obtains
To the location of the maximum noise source of vibration noise acoustic pressure in each sound source region in transformer-cabinet surface；
The calculation formula of beamforming algorithm are as follows:
In formula (1), B (t, θ) is sound pressure level maximum frequency range fmaxThe position of corresponding vibration noise sound pressure signal, PnFor sound
Arbitrarily downgrade maximum frequency range fmaxCorresponding vibration noise sound pressure signal；θ is sound source focus direction；knFor sound pressure level maximum frequency range fmaxIt is right
The weight vectors for the vibration noise sound pressure signal answered, take k heren=1；N=1,2......N, N are microphone number, and τ is to mend
Repay time delay, τ=lnsinθ/c0, lnIt is the distance between microphone and reference microphone of serial number n, c0Jie is being propagated for sound wave
Spread speed in matter, t are the time of measurement；
S12, region W locating for maximum noise source is determined according to the location of maximum noise source.
S2, acceleration transducer array is used by vibration measurement module, the region W where maximum noise source is adopted
Collect vibration signal P, the transformer nearest from vibration noise source acceleration transducer array being arranged in where maximum noise source
At shell, the most abundant vibration signal of information can be collected.
S3, transformer fault diagnosis is carried out to vibration signal using BP neural network algorithm；Itself the following steps are included:
S30, the training sample data that multiple groups vibration signal is obtained according to vibration signal and each group training sample data are corresponding
Transformer fault type.
S31, it handles each group training sample data progress FFT to obtain the characteristic quantity of vibration signal, by resulting characteristic quantity
As the neuron of BP neural network input layer, the quantity of characteristic quantity is that input layer quantity is m；Characteristic quantity includes fundamental frequency
It is special when specific gravity, basic amplitude, dominant frequency specific gravity, dominant frequency amplitude and vibrational entropy, i.e. BP neural network input layer quantity m=5
The calculation formula of sign amount is as follows:
Fundamental frequency amplitude: the fundamental frequency f of vibration signal is 2 times of current signal fundamental frequency, so the fundamental frequency f of vibration signal is
100Hz obtains fundamental frequency amplitude A after FFT is handled100；
Fundamental frequency specific gravity:
In formula (2), P100The fundamental frequency specific gravity that fundamental frequency f for vibration signal is 100Hz, f are the fundamental frequency of vibration signal, fmaxFor
The maximum value of the fundamental frequency of vibration signal, AfFor the corresponding amplitude in the place frequency f in rumble spectrum；
Dominant frequency amplitude: choosing the integral multiple of the fundamental frequency f of vibration signal, and fundamental frequency f is 100Hz, the amplitude of selection such as: 200Hz,
300Hz、400Hz......fmaxHz, choosing amplitude maximum is dominant frequency amplitude Amain, i.e. Amain=max { A200,A300,
A400......Amax}；
Dominant frequency specific gravity:
In formula (3), PmainThe dominant frequency specific gravity that fundamental frequency f for vibration signal is 100；
Vibrational entropy:
In formula (4), PfThe specific gravity that fundamental frequency for vibration signal is f, TVE are the vibrational entropy that the fundamental frequency of vibration signal is f.
Wherein, embodiment 1, by Noise Sources Identification module and vibration measurement module to the iron core and winding of transformer
In failure and it is normal when two states when detected, obtain iron core failure and it is normal when vibration signal and winding failure and
Vibration signal when normal, by rapid S31 to iron core failure and it is normal when vibration signal and winding failure and it is normal when vibration
Dynamic signal calculates, obtained winding failure characteristic quality of sample and iron core fault sample characteristic quantity, as shown in Table 1 and Table 2.
Table 1: winding failure characteristic quality of sample
| Fundamental frequency amplitude | Fundamental frequency specific gravity | Dominant frequency amplitude | Dominant frequency specific gravity | Vibrational entropy | |
| Normally | 0.159 | 3% | 0.698 | 64% | 1.6091 |
| Failure | 0.299 | 7% | 0.957 | 75% | 1.3177 |
Table 2: iron core fault sample characteristic quantity
| Fundamental frequency amplitude | Fundamental frequency specific gravity | Dominant frequency amplitude | Dominant frequency specific gravity | Vibrational entropy | |
| Normally | 0.60 | 3% | 1.4322 | 57% | 2.0457 |
| Failure | 0.7895 | 3.1% | 2.1659 | 75% | 1.3537 |
S32, each group characteristic quantity is normalized, obtains the normalization data of each group characteristic quantity；Normalized
Formula are as follows:
In formula (5), xiIt is the data after normalization,It is characteristic quality of sample,WithIt is characteristic quality of sample respectively
Minimum data and maximum data.
S33, transformer fault type is encoded, obtains the quantity of transformer fault type；Transformer fault type
For the neuron of BP neural network output layer, output layer neuron quantity n.Transformer fault type includes: fault-free yk1, iron core
Failure yk2And winding failure yk3, then BP neuroid output layer neuron quantity n=3.Transformer fault type is compiled
The mode of code are as follows: when transformer is in i-th kind of state, yki=1, ykj=0 (j ≠ i), network output are Yk=[yk1,yk2,
yk3].In order to make BP neural network have better generalization ability, training sample exports y when actual operationki=0.9, ykj
=0.1 (j ≠ i)；
Wherein, the characteristic quality of sample for the Tables 1 and 2 that embodiment 1 obtains is sequentially input into BP neural network, respectively obtain around
The BP neural network output malfunction coding of group sample and the BP neural network of iron core fault sample export malfunction coding, such as 4 He of table
Shown in table 5.
Table 4: the BP neural network of winding sample exports malfunction coding
Table 5: the BP neural network of iron core fault sample exports malfunction coding
It S34, is that m and output layer neuron quantity n obtains BP neural network hidden layer mind by input layer quantity
Through first quantity h comprising following steps:
S3400, hidden layer neuron quantity is calculated using following formula:
In formula (6), h is hidden layer neuron quantity, and regulating constant of a between 1-10, m is input layer number
Amount, n are output layer neuron quantity；
S3401, from hminStart, increases neuron number one by one until hmax, by [hmin,hmax] be trained test respectively
Card；
S3402, neuron number corresponding to optimal verification result, gained neuron number in training verification result are chosen
For BP neural network hidden layer neuron quantity h；Wherein, the calculating of optimal verification result is as follows, such as:
1. being trained firstly the need of to sample database (known transformer fault type)；
2. each of sample database sample have one known to, fixed fault type coding, i.e. java standard library；
The important parameter h of 3.BP neural network, range are [hmin,hmax], h is exactly seen in calculating by training process, i.e.,
Step S3401；
4. h1 is seen in step S3401 iteration, the corresponding volume of each sample is calculated through BP neural network in sample database
Code, i.e. output library 1, export and have a deviation E1 between library 1 and java standard library；Similarly, it when h2 sees in iteration, can be exported
Library 2, the deviation for exporting library 2 and java standard library is E2；And so on, each h is corresponding to obtain a deviation E, minimum deflection E is chosen,
The corresponding h of minimum deflection E is exactly optimal verification result.
Embodiment 2, vibration measurement module obtain the original vibration letter of 100 groups of transformers using acceleration transducer measurement
Number, wherein 30 groups are iron core fault vibration signal, 30 groups of winding failure vibration signals, 40 groups are fault-free vibration signal.Pass through
Diagnosis Method of Transformer Faults based on vibration noise and BP neural network, the framework for obtaining BP neural network is 5-8-3, i.e., defeated
Entering layer has 5 neurons, and output layer has 3 neurons, and hidden layer has 8 neurons.
S35, foundation input layer quantity are m, output layer neuron quantity n, BP neural network hidden layer neuron
Weight ω between quantity h, input layer i to hidden layer neuron jijAnd hidden layer neuron j to output layer neuron k
Between weight ωjkConstruct initial BP neural network；
Wherein, the activation primitive of BP neural network uses S type function f (x):
The output of hidden layer neuron j is sj:
In formula (8), ωijFor the weight between input layer i to hidden layer neuron j, θ1, θ2For amount of bias, xiFor
I-th of input quantity of output layer；
The output of output layer neuron k is yk:
In formula (9), ωjkFor the weight between hidden layer neuron j to output layer neuron k, θ1, θ2For amount of bias.
Characteristic quantity and transformer fault type corresponding with characteristic quantity described in each group described in S36, foundation each group, to the BP
Neural network is trained, the BP neural network after being trained, BP neural network training the following steps are included:
S3600, the characteristic quantity that step S31 is obtained to each group training sample data input initial BP neural network, obtain defeated
Out as a result, obtaining the error of output result by exporting result；The calculation formula of overall error E are as follows:
In formula (10), dijThe output valve encoded for i-th j-th of sample, yijThe expectation encoded for i-th j-th of sample
Value；
S3601, error in judgement are no to be less than presetting threshold value, if error is less than presetting threshold value, enters step
S3602；If error is greater than presetting threshold value, S3603 is entered step；
S3602, initial BP neural network carry out back transfer using gradient descent method, most along relative error quadratic sum
Fast descent direction, the continuous weight and threshold value for adjusting network carry out retrospectively calculate using output layer, hidden layer and input layer,
Export each layer as a result, return step S35；
S3603, the BP neural network after training is constructed using output result.
S37, the characteristic quantity for extracting transformer vibration signal to be diagnosed are carried out using the BP neural network obtained after training
Diagnosis, obtains its fault type.
Above disclosed is only a specific embodiment of the invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, can readily occur in variation or modification,
It is covered by the protection scope of the present invention.
Claims (8)
1. a kind of Diagnosis Method of Transformer Faults based on vibration noise and BP neural network, it is characterised in that: including following step
It is rapid:
S1, pass through Noise Sources Identification module, acquire the vibration noise sound pressure signal of transformer each region, and according to the vibration
Noise sound pressure signal obtains the region where maximum noise source；
S2, by vibration measurement module to the region where the maximum noise source, acquire vibration signal；
S3, transformer fault diagnosis is carried out to the vibration signal using BP neural network algorithm.
2. Diagnosis Method of Transformer Faults according to claim 1, it is characterised in that: the S1 includes:
S10, the vibration noise sound pressure signal is subjected to Fourier's variation, obtains sound pressure level maximum frequency range；
S11, the corresponding vibration noise sound pressure signal of the sound pressure level maximum frequency range is subjected to beamforming algorithm calculating, is become
The location of the maximum noise source of vibration noise acoustic pressure in each sound source region of depressor tank surface；
S12, region locating for maximum noise source is determined according to the location of described maximum noise source.
3. Diagnosis Method of Transformer Faults according to claim 1, it is characterised in that: the vibration measurement module is adopted
The vibration signal is acquired with acceleration transducer array.
4. Diagnosis Method of Transformer Faults according to claim 1, it is characterised in that: the S3 includes:
S30, the training sample data that multiple groups vibration signal is obtained according to the vibration signal and each group training sample data are corresponding
Transformer fault type；
S31, it handles the progress of training sample data described in each group FFT to obtain the characteristic quantity of vibration signal, the characteristic quantity conduct
The neuron of BP neural network input layer, the quantity of the characteristic quantity are that input layer quantity is m；
S32, characteristic quantity described in each group is normalized, obtains the normalization data of each group characteristic quantity；
S33, the transformer fault type is encoded, obtains the quantity of the transformer fault type, the transformer
Fault type is the neuron of BP neural network output layer, and the quantity of the transformer fault type is output layer neuron quantity
n；
It S34, is that m and output layer neuron quantity n obtains BP neural network hidden layer mind by the input layer quantity
Through first quantity h；
S35, according to the input layer quantity be m, output layer neuron quantity n and BP neural network hidden layer neuron
Quantity h constructs initial BP neural network；
Characteristic quantity and transformer fault type corresponding with characteristic quantity described in each group described in S36, foundation each group, to the BP nerve
Network is trained, the BP neural network after being trained；
S37, the characteristic quantity for extracting transformer vibration signal to be diagnosed are examined using the BP neural network obtained after training
It is disconnected, obtain its fault type.
5. Diagnosis Method of Transformer Faults according to claim 4, it is characterised in that: the characteristic quantity packet of the vibration signal
It includes: fundamental frequency specific gravity, basic amplitude, dominant frequency specific gravity, dominant frequency amplitude and vibrational entropy.
6. Diagnosis Method of Transformer Faults according to claim 4, it is characterised in that: the transformer fault type packet
It includes: fault-free, iron core failure and winding failure.
7. Diagnosis Method of Transformer Faults according to claim 4, it is characterised in that: the S34 includes:
S3400, hidden layer neuron quantity is calculated using following formula:
In formula (6), h is hidden layer neuron quantity, and regulating constant of a between 1-10, m is input layer quantity, and n is
Output layer neuron quantity；
S3401, from hminStart, increases neuron number one by one until hmax, by [hmin,hmax] it is trained verifying respectively；
S3402, neuron number corresponding to optimal verification result in training verification result is chosen, the neuron number is BP
Neural network hidden layer neuron quantity h.
8. Diagnosis Method of Transformer Faults according to claim 4, it is characterised in that: the BP neural network in the S36
Training the following steps are included:
S3600, the characteristic quantity that step S31 is obtained to each group training sample data input the BP neural network, obtain output knot
The error of fruit and output result；
S3601, judge that the error is no less than presetting threshold value, if the error is less than presetting threshold value, enter step
Rapid S3602；If the error is greater than presetting threshold value, S3603 is entered step；
S3602, the initial BP neural network carry out back transfer using gradient descent method, most along relative error quadratic sum
Fast descent direction, the continuous weight and threshold value for adjusting network carry out retrospectively calculate using output layer, hidden layer and input layer,
Export each layer as a result, return step S35；
S3603, the BP neural network after training is constructed using the output result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810805455.9A CN109033612B (en) | 2018-07-20 | 2018-07-20 | Transformer fault diagnosis method based on vibration noise and BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810805455.9A CN109033612B (en) | 2018-07-20 | 2018-07-20 | Transformer fault diagnosis method based on vibration noise and BP neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109033612A true CN109033612A (en) | 2018-12-18 |
| CN109033612B CN109033612B (en) | 2023-05-05 |
Family
ID=64644882
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810805455.9A Active CN109033612B (en) | 2018-07-20 | 2018-07-20 | Transformer fault diagnosis method based on vibration noise and BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109033612B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110174255A (en) * | 2019-06-03 | 2019-08-27 | 国网上海市电力公司 | A kind of transformer vibration signal separation method based on radial base neural net |
| CN110689069A (en) * | 2019-09-25 | 2020-01-14 | 贵州电网有限责任公司 | Transformer fault type diagnosis method based on semi-supervised BP network |
| CN111026095A (en) * | 2019-12-30 | 2020-04-17 | 太原科技大学 | Fault diagnosis method with noise label based on recurrent neural network |
| CN111175698A (en) * | 2020-01-18 | 2020-05-19 | 国网山东省电力公司菏泽供电公司 | Transformer noise source positioning method, system and device based on sound and vibration combination |
| CN111638028A (en) * | 2020-05-20 | 2020-09-08 | 国网河北省电力有限公司电力科学研究院 | High-voltage parallel reactor mechanical state evaluation method based on vibration characteristics |
| CN112098066A (en) * | 2020-09-21 | 2020-12-18 | 陕西师范大学 | High-voltage shunt reactor fault diagnosis method and system based on gate control circulation unit |
| CN112150744A (en) * | 2020-09-24 | 2020-12-29 | 国网山东省电力公司临沂供电公司 | Transformer vibration noise abnormity analysis early warning system |
| CN112307918A (en) * | 2020-10-21 | 2021-02-02 | 华北电力大学 | Diagnosis method for transformer direct-current magnetic biasing based on fuzzy neural network |
| CN112665707A (en) * | 2020-12-15 | 2021-04-16 | 国网天津市电力公司电力科学研究院 | Cumulative effect after short circuit impact of transformer and diagnosis method |
| CN112879278A (en) * | 2021-01-11 | 2021-06-01 | 苏州欣皓信息技术有限公司 | Pump station unit fault diagnosis method based on noise signal A weighting analysis |
| CN113283310A (en) * | 2021-05-07 | 2021-08-20 | 国网浙江省电力有限公司武义县供电公司 | System and method for detecting health state of power equipment based on voiceprint features |
| CN113392511A (en) * | 2021-05-28 | 2021-09-14 | 广西电网有限责任公司电力科学研究院 | On-load tap-changer mechanical state monitoring method based on frequency spectrum envelope symbol entropy |
| CN114235366A (en) * | 2021-12-14 | 2022-03-25 | 国网天津市电力公司电力科学研究院 | Transformer fault identification method based on transformer noise vibration characteristics |
| CN115358110A (en) * | 2022-07-25 | 2022-11-18 | 国网江苏省电力有限公司淮安供电分公司 | Transformer fault detection system based on acoustic sensor array |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101246043A (en) * | 2008-03-28 | 2008-08-20 | 清华大学 | On-line monitoring method for vibration and noise of AC power transformer influenced by DC magnetic biasing |
| CN103218662A (en) * | 2013-04-16 | 2013-07-24 | 郑州航空工业管理学院 | Transformer fault diagnosis method based on back propagation (BP) neural network |
| CN103245524A (en) * | 2013-05-24 | 2013-08-14 | 南京大学 | Acoustic fault diagnosis method based on neural network |
| CN105353255A (en) * | 2015-11-27 | 2016-02-24 | 南京邮电大学 | Transformer fault diagnosis method based on neural network |
| JP2016051850A (en) * | 2014-09-01 | 2016-04-11 | 株式会社ダイヘン | Stationary induction apparatus |
| CN106066468A (en) * | 2016-05-25 | 2016-11-02 | 哈尔滨工程大学 | A kind of based on acoustic pressure, the vector array port/starboard discrimination method of vibration velocity Mutual spectrum |
| CN107271022A (en) * | 2017-05-31 | 2017-10-20 | 国网山西省电力公司电力科学研究院 | A kind of transformer vibration noise Integrated Measurement System |
| CN107748314A (en) * | 2017-10-18 | 2018-03-02 | 国网重庆市电力公司北碚供电分公司 | Transformer Faults Analysis system based on sound wave shock detection |
-
2018
- 2018-07-20 CN CN201810805455.9A patent/CN109033612B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101246043A (en) * | 2008-03-28 | 2008-08-20 | 清华大学 | On-line monitoring method for vibration and noise of AC power transformer influenced by DC magnetic biasing |
| CN103218662A (en) * | 2013-04-16 | 2013-07-24 | 郑州航空工业管理学院 | Transformer fault diagnosis method based on back propagation (BP) neural network |
| CN103245524A (en) * | 2013-05-24 | 2013-08-14 | 南京大学 | Acoustic fault diagnosis method based on neural network |
| JP2016051850A (en) * | 2014-09-01 | 2016-04-11 | 株式会社ダイヘン | Stationary induction apparatus |
| CN105353255A (en) * | 2015-11-27 | 2016-02-24 | 南京邮电大学 | Transformer fault diagnosis method based on neural network |
| CN106066468A (en) * | 2016-05-25 | 2016-11-02 | 哈尔滨工程大学 | A kind of based on acoustic pressure, the vector array port/starboard discrimination method of vibration velocity Mutual spectrum |
| CN107271022A (en) * | 2017-05-31 | 2017-10-20 | 国网山西省电力公司电力科学研究院 | A kind of transformer vibration noise Integrated Measurement System |
| CN107748314A (en) * | 2017-10-18 | 2018-03-02 | 国网重庆市电力公司北碚供电分公司 | Transformer Faults Analysis system based on sound wave shock detection |
Non-Patent Citations (4)
| Title |
|---|
| 吴桂娇: ""波束形成原理在噪声源识别技术中的应用"", 《第25届全国振动与噪声高技术及应用会与论文选集》 * |
| 姚明镜，唐璇: ""基于振动声幅射的机械设备故障检测技术研究"", 《机械制造》 * |
| 许敬成，陈长征: ""BP神经网络在齿轮箱故障诊断中的应用"", 《噪声与振动控制》 * |
| 马裕超等: ""220kV油浸式变压器振动与噪声试验研究"", 《变压器》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110174255A (en) * | 2019-06-03 | 2019-08-27 | 国网上海市电力公司 | A kind of transformer vibration signal separation method based on radial base neural net |
| CN110174255B (en) * | 2019-06-03 | 2021-04-27 | 国网上海市电力公司 | Transformer vibration signal separation method based on radial basis function neural network |
| CN110689069A (en) * | 2019-09-25 | 2020-01-14 | 贵州电网有限责任公司 | Transformer fault type diagnosis method based on semi-supervised BP network |
| CN111026095A (en) * | 2019-12-30 | 2020-04-17 | 太原科技大学 | Fault diagnosis method with noise label based on recurrent neural network |
| CN111175698B (en) * | 2020-01-18 | 2022-12-20 | 国网山东省电力公司菏泽供电公司 | Transformer noise source positioning method, system and device based on sound and vibration combination |
| CN111175698A (en) * | 2020-01-18 | 2020-05-19 | 国网山东省电力公司菏泽供电公司 | Transformer noise source positioning method, system and device based on sound and vibration combination |
| CN111638028A (en) * | 2020-05-20 | 2020-09-08 | 国网河北省电力有限公司电力科学研究院 | High-voltage parallel reactor mechanical state evaluation method based on vibration characteristics |
| CN112098066A (en) * | 2020-09-21 | 2020-12-18 | 陕西师范大学 | High-voltage shunt reactor fault diagnosis method and system based on gate control circulation unit |
| CN112150744A (en) * | 2020-09-24 | 2020-12-29 | 国网山东省电力公司临沂供电公司 | Transformer vibration noise abnormity analysis early warning system |
| CN112307918A (en) * | 2020-10-21 | 2021-02-02 | 华北电力大学 | Diagnosis method for transformer direct-current magnetic biasing based on fuzzy neural network |
| CN112665707A (en) * | 2020-12-15 | 2021-04-16 | 国网天津市电力公司电力科学研究院 | Cumulative effect after short circuit impact of transformer and diagnosis method |
| CN112665707B (en) * | 2020-12-15 | 2023-03-03 | 国网天津市电力公司电力科学研究院 | Cumulative effect after short circuit impact of transformer and diagnosis method |
| CN112879278A (en) * | 2021-01-11 | 2021-06-01 | 苏州欣皓信息技术有限公司 | Pump station unit fault diagnosis method based on noise signal A weighting analysis |
| CN113283310A (en) * | 2021-05-07 | 2021-08-20 | 国网浙江省电力有限公司武义县供电公司 | System and method for detecting health state of power equipment based on voiceprint features |
| CN113392511A (en) * | 2021-05-28 | 2021-09-14 | 广西电网有限责任公司电力科学研究院 | On-load tap-changer mechanical state monitoring method based on frequency spectrum envelope symbol entropy |
| CN114235366A (en) * | 2021-12-14 | 2022-03-25 | 国网天津市电力公司电力科学研究院 | Transformer fault identification method based on transformer noise vibration characteristics |
| CN115358110A (en) * | 2022-07-25 | 2022-11-18 | 国网江苏省电力有限公司淮安供电分公司 | Transformer fault detection system based on acoustic sensor array |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109033612B (en) | 2023-05-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109033612A (en) | A kind of Diagnosis Method of Transformer Faults based on vibration noise and BP neural network | |
| CN106405339B (en) | Based on the associated transmission line malfunction reason discrimination method of low-and high-frequency wavelet character | |
| CN106646096B (en) | Transformer fault classification and recognition methods based on vibration analysis method | |
| CN105424364A (en) | Diagnostic method and device of train bearing failure | |
| CN105424395B (en) | The definite method and apparatus of equipment fault | |
| CN100485342C (en) | Integrated supporting vector machine mixed intelligent diagnosing method for mechanical fault | |
| CN102122133B (en) | Self-adaption wavelet neural network abnormity detection and fault diagnosis classification system and method | |
| CN104052612B (en) | A kind of Fault Identification of telecommunication service and the method and system of positioning | |
| CN202404166U (en) | On-line monitoring system for vibration performance of transformer | |
| CN110490071A (en) | A kind of substation's Abstraction of Sound Signal Characteristics based on MFCC | |
| CN105547465A (en) | Transformer vibration signal winding state feature extraction method | |
| CN108646149A (en) | Fault electric arc recognition methods based on current characteristic extraction | |
| CN109948597A (en) | A kind of Mechanical Failure of HV Circuit Breaker diagnostic method | |
| CN103064008B (en) | A kind of Nolinear analog circuit soft fault diagnostic method based on Hilbert-Huang transform | |
| CN113077005B (en) | Anomaly detection system and method based on LSTM self-encoder and normal signal data | |
| CN101907437A (en) | Wavelet difference algorithm-based cable fault localization method | |
| CN109323754A (en) | A kind of train wheel polygon fault diagnosis detection method | |
| CN106447039A (en) | Non-supervision feature extraction method based on self-coding neural network | |
| CN110308002A (en) | A kind of municipal rail train suspension method for diagnosing faults based on ground detection | |
| CN110044623A (en) | The rolling bearing fault intelligent identification Method of empirical mode decomposition residual signal feature | |
| CN102797671A (en) | Fault detection method and device of reciprocating compressor | |
| CN108693437B (en) | Method and system for judging deformation of transformer winding | |
| CN107036808B (en) | Gearbox of wind turbine combined failure diagnostic method based on support vector machines probability Estimation | |
| CN113283310A (en) | System and method for detecting health state of power equipment based on voiceprint features | |
| CN106203531A (en) | A kind of transmission line fault sorting technique based on the sparse autocoder of convolution |
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 |