CN110045248A - A kind of oil-immersed sleeve pipe ageing state appraisal procedure based on image recognition - Google Patents
A kind of oil-immersed sleeve pipe ageing state appraisal procedure based on image recognition Download PDFInfo
- Publication number
- CN110045248A CN110045248A CN201910347131.XA CN201910347131A CN110045248A CN 110045248 A CN110045248 A CN 110045248A CN 201910347131 A CN201910347131 A CN 201910347131A CN 110045248 A CN110045248 A CN 110045248A
- Authority
- CN
- China
- Prior art keywords
- oil
- sleeve pipe
- thermography
- immersed sleeve
- image
- 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
- 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
- G01R31/1218—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing using optical methods; using charged particle, e.g. electron, beams or X-rays
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
- G01R31/1263—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation
- G01R31/1281—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation of liquids or gases
-
- 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
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
- G01R31/62—Testing of transformers
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Power Engineering (AREA)
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
Abstract
The oil-immersed sleeve pipe ageing state appraisal procedure based on image recognition that the invention discloses a kind of, it include: to build experiment porch, then the oil-immersed sleeve pipe sample of different ageing states is prepared, obtain the thermography basic data of sample, after carrying out pretreatment and characteristic parameter extraction to thermography image, it establishes neural network model and carries out learning training, then model parameter is corrected using live oil-immersed sleeve pipe test result, finally the ageing state of oil-immersed sleeve pipe is assessed.Oil-immersed sleeve pipe ageing state appraisal procedure based on image recognition is provided through the invention, can effectively assess the ageing state of oil-immersed sleeve pipe.
Description
Technical field
The invention belongs to casing Condition assessment of insulation fields, and in particular to a kind of oil-immersed sleeve pipe based on image recognition is old
Change state evaluating method.
Background technique
Oil-immersed sleeve pipe is the main accessories of high-power transformer, and be that inside transformer is connect with external electrical network is unique logical
Road.Casing is in the process of running, long-term to bear electricity, heat and mechanical stress, is the weak link of electric system, insulation performance is straight
Connect the stable operation for influencing electric system.Insulating thermal aging is the main factor of oil-immersed sleeve pipe insulation degradation, long-term burnt
The deterioration that effect of having burning ears acts on lower paper oil insulation material causes insulation performance to reduce, and effective assessment of casing state of insulation has weight
Big meaning.
With the propulsion of China's smart grid, more and more substations are using the crusing robot for carrying thermal infrared imager
For monitoring the temperature of electrical equipment for fault detection.Infrared Thermogram utilizes infrared imagery technique detection oil-immersed sleeve pipe
Infra-red radiation generates the temperature distribution image of casing, can find the heat generating spot of oil-immersed sleeve pipe in time, and then remove morning in time
The failure of phase improves the reliability of power supply.Infrared Thermal-imaging Diagnostic Technique is by the operating condition of casing and environment temperature etc. at present
The influence of factor lacks failure TuPu method in reliable history picture library, existing to be mainly using the evaluation method of infrared thermal imagery
According to related regulation, method is relatively simple, cannot give warning in advance for failure, and when often finding the problem, problem has compared at this time
It is more serious.Therefore it is badly in need of a kind of oil-immersed sleeve pipe ageing state appraisal procedure based on image recognition.
Summary of the invention
In order to effectively assess the ageing state of oil-immersed sleeve pipe, the present invention provides a kind of oil based on image recognition
Immersion casing ageing state appraisal procedure, comprising the following steps:
Step 1: building experiment porch
Oil-immersed sleeve pipe sample thermography experiment porch is built, mainly by current source (1), conducting rod top (2), conducting rod
Lower part (3), oil-immersed sleeve pipe sample (4), insulation fuel tank (5), insulating oil (6), PTC constant temperature heating plate (7), voltage source (8), the
One temperature sensor (9a), second temperature sensor (9b), temperature collecting device (10), infrared thermal imager (11), thermography
Acquisition device (12), PC computer (13), electric blender (14) form, and insulation fuel tank (5) is provided with insulating oil (6), insulating oil (6)
Liquid level is located at insulation fuel tank (5) top 7/8, and oil-immersed sleeve pipe sample (4) is put vertically from insulation fuel tank (5) top center
Enter in the fuel tank that insulate (5), conducting rod lower part (3) distance insulation fuel tank (5) bottom 30cm, electric blender (14) distance insulation fuel tank
(5) bottom 10cm, PTC constant temperature heating plate (7) are placed in insulation fuel tank (5) bottom center, infrared thermal imager (11), first
Temperature sensor (9a) is placed at oil-immersed sleeve pipe sample (4) 5m, and infrared thermal imager (11) can scan oil immersed type set
The complete thermography of pipe sample (4) entirety, second temperature sensor (9b) are placed at insulation fuel tank (5) Left-side center；It is conductive
Bar top (2) and conducting rod lower part (3) are connected to current source (1)；Electric blender (14), PTC constant temperature heating plate (7) are connected to voltage
Source (8)；First temperature sensor (9a), second temperature sensor (9b) are connected to temperature collecting device (10), temperature acquisition dress
It sets (10) and is connected to PC computer (13)；Infrared thermal imager (11) is connected to thermography acquisition device (12), thermography acquisition dress
It sets (12) and is connected to PC computer (13)；
Step 2: the sample preparation of different ageing states is obtained with basic data
Preparation insulation paper polymerization degree is respectively 200,500,650,750,900,1050 oil-immersed sleeve pipe sample (4), benefit
Environment temperature T is recorded with second temperature sensor (9b)0, test temperature T is set, suitable PTC constant temperature heating plate (7) is selected,
Cut-in voltage source (8) makes electric blender (14), PTC constant temperature heating plate (7) work, and second temperature sensor (9b) tests insulating oil
(6) temperature；When temperature keeps stablizing, firing current source (1) is small to oil-immersed sleeve pipe sample (4) load sample rated current 2
When, after closing current source (1) 15 minute, thermography is carried out to oil-immersed sleeve pipe sample (4) using infrared thermal imager (11) and is swept
It retouches and takes pictures 10 times, interval time is 15 minutes, obtains different ageing state oil-immersed sleeve pipe samples (4) at different temperatures not
Thermography in the same time, thermography pixel are 512 × 512 image；
Step 3: image preprocessing and characteristic parameter extraction
Thermography image is moved, sleeved conducting rod is located at the center of image vertical direction, and casing flange is located at water
Prosposition is set at 3/4 position from top to bottom, and excess pixel is cut, and obtains 256 × 256 images；
Gray value processing, gray value processing method are carried out to thermography image are as follows: calculate the rgb color point of thermography image
Amount, is weighted RGB three-component by formula (1) to obtain the gray level image f of the m times thermography of taking picturesm(x,y)
fm(x, y)=0.299Rm(x,y)+0.578Gm(x,y)+0.114Bm(x,y) (1)
In formula, Rm(x,y)、Gm(x,y)、Bm(x, y) is respectively the red, green, blue color component of thermography image, x, y difference
For thermography picture position, x=1,2 ..., 256, y=1,2 ..., 256, m=1,2 ..., 10；
Gray value is divided into 0-255 totally 256 grades；
Discrete Fourier transform F is carried out to image grayscale figurem(u,v)
In formula, j is imaginary unit；
Calculate modulus E, E0And S
In formula, σ is the standard deviation of the 1st thermography grayscale image of taking pictures；
Calculate characteristic parameter K0, calculation are as follows:
Step 4: neural network learning training
Learning training is carried out to neural network, firstly, establishing the sample set for training depth convolutional neural networks, sample
Integrate as thermography image Fourier transformation spectrogram and characteristic parameter K0, wherein the cutoff frequency of Fourier transformation spectrogram be
256Hz；Secondly, constructing neural network, and the neural network of construction is initialized；The neural network, structure include: defeated
Enter layer, hidden layer and output layer；For the single neuron with multi input, output is using Gaussian function as activation letter
Number；The threshold value for initializing each neuron is a random number between 0 to 1；Finally, neural network is trained, first
By preceding weighting parameter is updated using gradient descent method, training error is made to reach minimum to transmitting, then back transfer error；
Save trained neural network model；
Step 5: live oil-immersed sleeve pipe test is corrected with model parameter
The real-time current before the stoppage in transit of completely new oil-immersed sleeve pipe in 2 hours is recorded, is denoted as I (t), casing rated current is
I1, and calculating current virtual value I0, current effective value I0Calculation are as follows:
Using oil-immersed sleeve pipe as the center of circle, radius is that two same type infrared thermal imagers are placed in the position of 5m respectively, two
The angle formed between infrared thermal imager and casing is π/6, utilizes infrared thermal imager to oil immersed type set after casing stoppage in transit 15min
Pipe is scanned takes pictures with infrared ray, and primary, test 10 times is tested at interval of 15min, and each infrared thermal imager tests institute
The thermography image that oil-immersed sleeve pipe is obtained in different moments is one group of test result, obtains two groups of oil-immersed sleeve pipes in different moments
Thermography image measurement result；Image preprocessing is carried out to the thermography that live oil-immersed sleeve pipe is tested according to third step；
Calculate characteristic parameter K1
Trained neural network is loaded, one group of test result is identified with trained neural network, according to knowledge
Other result optimizes trained neural network, is tested and is continuouslyd optimize to the model after optimization with test result,
Until the recognition result error of two groups of test results is less than 0.5%；
Step 6: oil-immersed sleeve pipe ageing state is assessed
According to the heat for the oil-immersed sleeve pipe that the thermography test method test of the live oil-immersed sleeve pipe of the 5th step need to be assessed
As figure, it is loaded and optimized after neural network model, thermography is identified and exports result.
The present invention has the advantages that
The oil-immersed sleeve pipe ageing state appraisal procedure based on image recognition that the present invention provides a kind of, by building experiment
Then platform prepares the oil-immersed sleeve pipe sample of different ageing states, the thermography basic data of sample is obtained, by thermal imagery
Figure image carries out pretreatment with after characteristic parameter extraction, establishes neural network model and carries out learning training, then utilizes scene
Oil-immersed sleeve pipe test result corrects model parameter, finally assesses the ageing state of oil-immersed sleeve pipe.By this hair
Bright oil-immersed sleeve pipe ageing state appraisal procedure of the offer based on image recognition, can effectively assess the aging shape of oil-immersed sleeve pipe
State.
Detailed description of the invention
Oil-immersed sleeve pipe ageing state appraisal procedure flow chart of the Fig. 1 based on image recognition.
Oil-immersed sleeve pipe ageing state appraisal procedure experiment porch of the Fig. 2 based on image recognition.
Specific embodiment
Invention is further explained with reference to the accompanying drawing.
A kind of oil-immersed sleeve pipe ageing state appraisal procedure based on image recognition, includes the following steps:
Step 1: building experiment porch
Oil-immersed sleeve pipe sample thermography experiment porch is built, mainly by current source (1), conducting rod top (2), conducting rod
Lower part (3), oil-immersed sleeve pipe sample (4), insulation fuel tank (5), insulating oil (6), PTC constant temperature heating plate (7), voltage source (8), the
One temperature sensor (9a), second temperature sensor (9b), temperature collecting device (10), infrared thermal imager (11), thermography
Acquisition device (12), PC computer (13), electric blender (14) form, and insulation fuel tank (5) is provided with insulating oil (6), insulating oil (6)
Liquid level is located at insulation fuel tank (5) top 7/8, and oil-immersed sleeve pipe sample (4) is put vertically from insulation fuel tank (5) top center
Enter in the fuel tank that insulate (5), conducting rod lower part (3) distance insulation fuel tank (5) bottom 30cm, electric blender (14) distance insulation fuel tank
(5) bottom 10cm, PTC constant temperature heating plate (7) are placed in insulation fuel tank (5) bottom center, infrared thermal imager (11), first
Temperature sensor (9a) is placed at oil-immersed sleeve pipe sample (4) 5m, and infrared thermal imager (11) can scan oil immersed type set
The complete thermography of pipe sample (4) entirety, second temperature sensor (9b) are placed at insulation fuel tank (5) Left-side center；It is conductive
Bar top (2) and conducting rod lower part (3) are connected to current source (1)；Electric blender (14), PTC constant temperature heating plate (7) are connected to voltage
Source (8)；First temperature sensor (9a), second temperature sensor (9b) are connected to temperature collecting device (10), temperature acquisition dress
It sets (10) and is connected to PC computer (13)；Infrared thermal imager (11) is connected to thermography acquisition device (12), thermography acquisition dress
It sets (12) and is connected to PC computer (13)；
Step 2: the sample preparation of different ageing states is obtained with basic data
Preparation insulation paper polymerization degree is respectively 200,500,650,750,900,1050 oil-immersed sleeve pipe sample (4), benefit
Environment temperature T is recorded with second temperature sensor (9b)0, test temperature T is set, suitable PTC constant temperature heating plate (7) is selected,
Cut-in voltage source (8) makes electric blender (14), PTC constant temperature heating plate (7) work, and second temperature sensor (9b) tests insulating oil
(6) temperature；When temperature keeps stablizing, firing current source (1) is small to oil-immersed sleeve pipe sample (4) load sample rated current 2
When, after closing current source (1) 15 minute, thermography is carried out to oil-immersed sleeve pipe sample (4) using infrared thermal imager (11) and is swept
It retouches and takes pictures 10 times, interval time is 15 minutes, obtains different ageing state oil-immersed sleeve pipe samples (4) at different temperatures not
Thermography in the same time, thermography pixel are 512 × 512 image；
Step 3: image preprocessing and characteristic parameter extraction
Thermography image is moved, sleeved conducting rod is located at the center of image vertical direction, and casing flange is located at water
Prosposition is set at 3/4 position from top to bottom, and excess pixel is cut, and obtains 256 × 256 images；
Gray value processing, gray value processing method are carried out to thermography image are as follows: calculate the rgb color point of thermography image
Amount, is weighted RGB three-component by formula (1) to obtain the gray level image f of the m times thermography of taking picturesm(x,y)
fm(x, y)=0.299Rm(x,y)+0.578Gm(x,y)+0.114Bm(x,y) (9)
In formula, Rm(x,y)、Gm(x,y)、Bm(x, y) is respectively the red, green, blue color component of thermography image, x, y difference
For thermography picture position, x=1,2 ..., 256, y=1,2 ..., 256, m=1,2 ..., 10；
Gray value is divided into 0-255 totally 256 grades；
Discrete Fourier transform F is carried out to image grayscale figurem(u,v)
In formula, j is imaginary unit；
Calculate modulus E, E0And S
In formula, σ is the standard deviation of the 1st thermography grayscale image of taking pictures；
Calculate characteristic parameter K0, calculation are as follows:
Step 4: neural network learning training
Learning training is carried out to neural network, firstly, establishing the sample set for training depth convolutional neural networks, sample
Integrate as thermography image Fourier transformation spectrogram and characteristic parameter K0, wherein the cutoff frequency of Fourier transformation spectrogram be
256Hz；Secondly, constructing neural network, and the neural network of construction is initialized；The neural network, structure include: defeated
Enter layer, hidden layer and output layer；For the single neuron with multi input, output is using Gaussian function as activation letter
Number；The threshold value for initializing each neuron is a random number between 0 to 1；Finally, neural network is trained, first
By preceding weighting parameter is updated using gradient descent method, training error is made to reach minimum to transmitting, then back transfer error；
Save trained neural network model；
Step 5: live oil-immersed sleeve pipe test is corrected with model parameter
The real-time current before the stoppage in transit of completely new oil-immersed sleeve pipe in 2 hours is recorded, is denoted as I (t), casing rated current is
I1, and calculating current virtual value I0, current effective value I0Calculation are as follows:
Using oil-immersed sleeve pipe as the center of circle, radius is that two same type infrared thermal imagers are placed in the position of 5m respectively, two
The angle formed between infrared thermal imager and casing is π/6, utilizes infrared thermal imager to oil immersed type set after casing stoppage in transit 15min
Pipe is scanned takes pictures with infrared ray, and primary, test 10 times is tested at interval of 15min, and each infrared thermal imager tests institute
The thermography image that oil-immersed sleeve pipe is obtained in different moments is one group of test result, obtains two groups of oil-immersed sleeve pipes in different moments
Thermography image measurement result；Image preprocessing is carried out to the thermography that live oil-immersed sleeve pipe is tested according to third step；
Calculate characteristic parameter K1
Trained neural network is loaded, one group of test result is identified with trained neural network, according to knowledge
Other result optimizes trained neural network, is tested and is continuouslyd optimize to the model after optimization with test result,
Until the recognition result error of two groups of test results is less than 0.5%；
Step 6: oil-immersed sleeve pipe ageing state is assessed
According to the heat for the oil-immersed sleeve pipe that the thermography test method test of the live oil-immersed sleeve pipe of the 5th step need to be assessed
As figure, it is loaded and optimized after neural network model, thermography is identified and exports result.
Claims (1)
1. a kind of oil-immersed sleeve pipe ageing state appraisal procedure based on image recognition, which comprises the following steps:
Step 1: building experiment porch
Oil-immersed sleeve pipe sample thermography experiment porch is built, mainly by current source (1), conducting rod top (2), conducting rod lower part
(3), oil-immersed sleeve pipe sample (4), insulation fuel tank (5), insulating oil (6), PTC constant temperature heating plate (7), voltage source (8), the first temperature
Spend sensor (9a), second temperature sensor (9b), temperature collecting device (10), infrared thermal imager (11), thermography acquisition
Device (12), PC computer (13), electric blender (14) form, and insulation fuel tank (5) is provided with insulating oil (6), insulating oil (6) liquid level
Height is located at insulation fuel tank (5) top 7/8, and oil-immersed sleeve pipe sample (4) is put into absolutely vertically from insulation fuel tank (5) top center
In edge fuel tank (5), conducting rod lower part (3) distance insulation fuel tank (5) bottom 30cm, electric blender (14) distance insulation fuel tank (5)
Bottom 10cm, PTC constant temperature heating plate (7) are placed in insulation fuel tank (5) bottom center, infrared thermal imager (11), the first temperature
Degree sensor (9a) is placed at oil-immersed sleeve pipe sample (4) 5m, and infrared thermal imager (11) can scan oil-immersed sleeve pipe
The complete thermography of sample (4) entirety, second temperature sensor (9b) are placed at insulation fuel tank (5) Left-side center；Conducting rod
Top (2) and conducting rod lower part (3) are connected to current source (1)；Electric blender (14), PTC constant temperature heating plate (7) are connected to voltage source
(8)；First temperature sensor (9a), second temperature sensor (9b) are connected to temperature collecting device (10), temperature collecting device
(10) PC computer (13) are connected to；Infrared thermal imager (11) is connected to thermography acquisition device (12), thermography acquisition device
(12) PC computer (13) are connected to；
Step 2: the sample preparation of different ageing states is obtained with basic data
Preparation insulation paper polymerization degree is respectively 200,500,650,750,900,1050 oil-immersed sleeve pipe sample (4), utilizes the
Two temperature sensors (9b) record environment temperature T0, test temperature T is set, suitable PTC constant temperature heating plate (7) is selected, is opened
Voltage source (8) makes electric blender (14), PTC constant temperature heating plate (7) work, and second temperature sensor (9b) tests insulating oil (6)
Temperature；When temperature keeps stablizing, firing current source (1) loads oil-immersed sleeve pipe sample (4) sample rated current 2 hours,
After closing current source (1) 15 minute, thermographic scan is carried out simultaneously to oil-immersed sleeve pipe sample (4) using infrared thermal imager (11)
Take pictures 10 times, interval time is 15 minutes, obtain different ageing state oil-immersed sleeve pipe samples (4) it is different at different temperatures when
The thermography at quarter, thermography pixel are 512 × 512 image；
Step 3: image preprocessing and characteristic parameter extraction
Thermography image is moved, sleeved conducting rod is located at the center of image vertical direction, and casing flange is located at horizontal position
It sets at 3/4 position from top to bottom, cuts excess pixel, obtain 256 × 256 images；
Gray value processing, gray value processing method are carried out to thermography image are as follows: the rgb color component of thermography image is calculated,
RGB three-component is weighted by formula (1) to obtain the gray level image f of the m times thermography of taking picturesm(x,y)
fm(x, y)=0.299Rm(x,y)+0.578Gm(x,y)+0.114Bm(x,y) (1)
In formula, Rm(x,y)、Gm(x,y)、Bm(x, y) is respectively the red, green, blue color component of thermography image, and x, y are respectively heat
Picture figure picture position, x=1,2 ..., 256, y=1,2 ..., 256, m=1,2 ..., 10；
Gray value is divided into 0-255 totally 256 grades；
Discrete Fourier transform F is carried out to image grayscale figurem(u,v)
In formula, j is imaginary unit；
Calculate modulus E, E0And S
In formula, σ is the standard deviation of the 1st thermography grayscale image of taking pictures；
Calculate characteristic parameter K0, calculation are as follows:
Step 4: neural network learning training
Learning training is carried out to neural network, firstly, establishing the sample set for training depth convolutional neural networks, sample set is
Thermography image Fourier transformation spectrogram and characteristic parameter K0, wherein the cutoff frequency of Fourier transformation spectrogram be
256Hz；Secondly, constructing neural network, and the neural network of construction is initialized；The neural network, structure include: defeated
Enter layer, hidden layer and output layer；For the single neuron with multi input, output is using Gaussian function as activation letter
Number；The threshold value for initializing each neuron is a random number between 0 to 1；Finally, neural network is trained, first
By preceding weighting parameter is updated using gradient descent method, training error is made to reach minimum to transmitting, then back transfer error；
Save trained neural network model；
Step 5: live oil-immersed sleeve pipe test is corrected with model parameter
The real-time current before the stoppage in transit of completely new oil-immersed sleeve pipe in 2 hours is recorded, I (t) is denoted as, casing rated current is I1, and count
Calculate current effective value I0, current effective value I0Calculation are as follows:
Using oil-immersed sleeve pipe as the center of circle, radius is that two same type infrared thermal imagers are placed in the position of 5m respectively, two infrared
The angle formed between thermal imaging system and casing is π/6, after casing stoppage in transit 15min using infrared thermal imager to oil-immersed sleeve pipe into
Row scanning and infrared ray are taken pictures, and are tested once at interval of 15min, are tested 10 times, each infrared thermal imager test gained oil
Thermography image of the immersion casing in different moments is one group of test result, obtains two groups of oil-immersed sleeve pipes in the heat of different moments
As figure image measurement result；Image preprocessing is carried out to the thermography that live oil-immersed sleeve pipe is tested according to third step；
Calculate characteristic parameter K1
Trained neural network is loaded, one group of test result is identified with trained neural network, is tied according to identification
Fruit optimizes trained neural network, is tested and is continuouslyd optimize to the model after optimization with test result, until
The recognition result error of two groups of test results is less than 0.5%；
Step 6: oil-immersed sleeve pipe ageing state is assessed
The thermography for the oil-immersed sleeve pipe that need to be assessed is tested according to the thermography test method of the live oil-immersed sleeve pipe of the 5th step,
Neural network model after loaded and optimized, identifies thermography and exports result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910347131.XA CN110045248B (en) | 2019-04-27 | 2019-04-27 | Oil-immersed bushing aging state evaluation method based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910347131.XA CN110045248B (en) | 2019-04-27 | 2019-04-27 | Oil-immersed bushing aging state evaluation method based on image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110045248A true CN110045248A (en) | 2019-07-23 |
| CN110045248B CN110045248B (en) | 2020-04-24 |
Family
ID=67279808
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910347131.XA Active CN110045248B (en) | 2019-04-27 | 2019-04-27 | Oil-immersed bushing aging state evaluation method based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110045248B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114062232A (en) * | 2021-09-30 | 2022-02-18 | 国高材高分子材料产业创新中心有限公司 | Oven, and automatic measuring system and method for thermal-oxidative aging life of polymer material |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408580A (en) * | 2008-11-21 | 2009-04-15 | 重庆大学 | Method for evaluating oil paper insulation ageing state based on local discharge characteristic parameter |
| JP2017011263A (en) * | 2015-06-25 | 2017-01-12 | ミツビシ・エレクトリック・アールアンドディー・センター・ヨーロッパ・ビーヴィＭｉｔｓｕｂｉｓｈｉ Ｅｌｅｃｔｒｉｃ Ｒ＆Ｄ Ｃｅｎｔｒｅ Ｅｕｒｏｐｅ Ｂ．Ｖ． | Method and system for monitoring electrolytic capacitor state online |
| CN108594089A (en) * | 2018-04-28 | 2018-09-28 | 广东电网有限责任公司电力科学研究院 | The degree of polymerization detection method and device of insulating paper in a kind of transformer |
| CN108875262A (en) * | 2018-07-05 | 2018-11-23 | 国网北京市电力公司 | The detection method and device of transformer oil ageing state, storage medium, processor |
-
2019
- 2019-04-27 CN CN201910347131.XA patent/CN110045248B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408580A (en) * | 2008-11-21 | 2009-04-15 | 重庆大学 | Method for evaluating oil paper insulation ageing state based on local discharge characteristic parameter |
| JP2017011263A (en) * | 2015-06-25 | 2017-01-12 | ミツビシ・エレクトリック・アールアンドディー・センター・ヨーロッパ・ビーヴィＭｉｔｓｕｂｉｓｈｉ Ｅｌｅｃｔｒｉｃ Ｒ＆Ｄ Ｃｅｎｔｒｅ Ｅｕｒｏｐｅ Ｂ．Ｖ． | Method and system for monitoring electrolytic capacitor state online |
| CN108594089A (en) * | 2018-04-28 | 2018-09-28 | 广东电网有限责任公司电力科学研究院 | The degree of polymerization detection method and device of insulating paper in a kind of transformer |
| CN108875262A (en) * | 2018-07-05 | 2018-11-23 | 国网北京市电力公司 | The detection method and device of transformer oil ageing state, storage medium, processor |
Non-Patent Citations (1)
| Title |
|---|
| 李和阳等: "基于图像匹配的变压器套管工况识别方法研究", 《计算机工程与应用》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114062232A (en) * | 2021-09-30 | 2022-02-18 | 国高材高分子材料产业创新中心有限公司 | Oven, and automatic measuring system and method for thermal-oxidative aging life of polymer material |
| CN114062232B (en) * | 2021-09-30 | 2024-02-02 | 国高材高分子材料产业创新中心有限公司 | Automatic measuring system and method for thermal oxidative aging life of baking oven and polymer material |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110045248B (en) | 2020-04-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Deep learning based module defect analysis for large-scale photovoltaic farms | |
| CN103487729B (en) | Based on the power equipments defect detection method that ultraviolet video and infrared video merge | |
| CN103901072B (en) | Method for diagnosing equipment overheating defects by utilizing infrared spectrum analysis | |
| CN110567964B (en) | Method and device for detecting defects of power transformation equipment and storage medium | |
| CN109870617A (en) | A kind of intelligent power plant electrical equipment fault diagnosis method based on width study and infrared image space-time characteristic | |
| CN109945977B (en) | Thermal fault diagnosis method and system for dry-type transformer | |
| CN112949198B (en) | Method and device for detecting faults of photovoltaic string and computer storage medium | |
| CN111342391A (en) | Power transmission line insulator and line fault inspection method and inspection system | |
| CN116342952B (en) | Transformer bushing abnormality identification method and system | |
| Liao et al. | Using Matlab real-time image analysis for solar panel fault detection with UAV | |
| CN109345586A (en) | Electrical equipment discharge characteristic extracting method based on ultraviolet imagery technology | |
| CN110335260A (en) | A kind of power cable damage testing method based on light convolutional neural networks | |
| CN110045248A (en) | A kind of oil-immersed sleeve pipe ageing state appraisal procedure based on image recognition | |
| CN113820017A (en) | Generator carbon brush temperature monitoring system based on infrared image and temperature prediction method | |
| CN116054417A (en) | Monitoring system and method for transformer substation | |
| CN109086891B (en) | Hot spot fault diagnosis method based on fuzzy logic reasoning | |
| CN113139730B (en) | Power equipment state evaluation method and system based on digital twin model | |
| Han et al. | Transformer defects detection method based on visible and infrared fusion images | |
| CN110610136B (en) | Substation equipment identification module and method based on deep learning | |
| Liu et al. | Insulator contamination diagnosis method based on deep learning convolutional neural network | |
| CN113284103B (en) | Substation equipment defect online detection method based on space transformation fast R-CNN model | |
| CN108492292A (en) | Conducting wire broken lot detection method based on infrared image processing | |
| CN115393747A (en) | Photovoltaic fault detection method based on deep learning | |
| CN114325274A (en) | Converter transformer valve side sleeve partial discharge detection system and method | |
| Xu et al. | Autonomous diagnosis of overheating defects in cable accessories based on Faster RCNN and Mean-Shift algorithm |
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 |