CN107133378A - Become the oil-filled transformer fault early warning method of point estimation based on Higher Dimensional Linear Models - Google Patents
Become the oil-filled transformer fault early warning method of point estimation based on Higher Dimensional Linear Models Download PDFInfo
- Publication number
- CN107133378A CN107133378A CN201710193920.3A CN201710193920A CN107133378A CN 107133378 A CN107133378 A CN 107133378A CN 201710193920 A CN201710193920 A CN 201710193920A CN 107133378 A CN107133378 A CN 107133378A
- Authority
- CN
- China
- Prior art keywords
- mrow
- mtd
- msub
- mtr
- mtable
- 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
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000008569 process Effects 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000000844 transformation Methods 0.000 claims description 2
- 239000007789 gas Substances 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 5
- 238000004868 gas analysis Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 238000001931 thermography Methods 0.000 description 3
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 230000007257 malfunction Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- RSWGJHLUYNHPMX-UHFFFAOYSA-N Abietic-Saeure Natural products C12CCC(C(C)C)=CC2=CCC2C1(C)CCCC2(C)C(O)=O RSWGJHLUYNHPMX-UHFFFAOYSA-N 0.000 description 1
- 229910000805 Pig iron Inorganic materials 0.000 description 1
- KHPCPRHQVVSZAH-HUOMCSJISA-N Rosin Natural products O(C/C=C/c1ccccc1)[C@H]1[C@H](O)[C@@H](O)[C@@H](O)[C@@H](CO)O1 KHPCPRHQVVSZAH-HUOMCSJISA-N 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009529 body temperature measurement Methods 0.000 description 1
- 239000004020 conductor Substances 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007654 immersion Methods 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000010358 mechanical oscillation Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000013021 overheating Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000004861 thermometry Methods 0.000 description 1
- KHPCPRHQVVSZAH-UHFFFAOYSA-N trans-cinnamyl beta-D-glucopyranoside Natural products OC1C(O)C(O)C(CO)OC1OCC=CC1=CC=CC=C1 KHPCPRHQVVSZAH-UHFFFAOYSA-N 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
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
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
Abstract
The invention discloses a kind of fault early warning method for the oil-filled transformer for becoming point estimation based on Higher Dimensional Linear Models, group's drag-line estimation that the Higher Dimensional Linear Models set up by several oil-filled transformer failure indexs of correlation to selection are adapted to, the height position in Higher Dimensional Linear Models is obtained, the position occurred using height realizes the early warning to transformer fault as the position of oil-filled transformer fault pre-alarming.Relative to other failure prediction methods, the present invention only needs to the characteristic gas concentration of dissolving in oil-filled transformer insulating oil, without other extra equipment.Meanwhile, applicability of the present invention extensively, goes for the various fault types produced inside oil-filled transformer.
Description
Technical field
The present invention relates to oil-filled transformer fault pre-alarming field, and in particular to one kind is estimated based on Higher Dimensional Linear Models height
The oil-filled transformer fault early warning method of meter.
Background technology
The most failure of oil-filled transformer will not occur within the extremely short time, be a process slowly accumulated,
When accumulation reaches a certain amount, the generation of failure will be caused.Therefore we need a kind of failure of oil-filled transformer pre-
Alarm method so that we can be within the period that oil-filled transformer failure is accumulated, to the running status of oil-filled transformer
Make early warning.So that we carry out necessary maintenance and repair to oil-filled transformer, the generation of failure is prevented, and can save
Substantial amounts of manpower, financial resources and material resources loss.
Whether the failure predication of oil-filled transformer is the feature according to failure, be able to can be broken down progress for transformer
Prediction.At present, there is following conventional method both at home and abroad：
1) dissolved gas analysis method in oil
The ratio method of dissolved gas analysis is one of transformer fault prediction most common method.Due in transformer
The different failure in portion can produce different internal environments, so as to produce different gases.Therefore, we are by analyzing oil dissolved gas
The index such as content, relative percentage, can reach the purpose that transformer fault diagnosis and prediction failure occur.
2) vibration analysis method
The failures such as short circuit, the insulation ag(e)ing of inside transformer are likely to make the winding of transformer be deformed or make to draw
The structure of line produces skew or disturbed.Therefore, using these fault messages, early warning can also be carried out to the failure of transformer.Shake
Dynamic analytic approach is exactly a kind of method for monitoring this transformer fault, and this method is by the vibration signal for transformer
It is monitored and analyzes, reaches the purpose of monitoring running state of transformer.
3) IR thermometry
Infrared Thermography Technology refers to receiving the infrared radiation signal that measured target is given out by infrared detector,
Then by the processing of amplification, the vision signal of standard is converted into, and then shows by monitor infrared thermography.Such as
Fruit inside transformer exist the bad contact of lead, have rosin joint at the conductor tab of coil, overload a series of feelings such as operation
During condition, it may all cause the loop hot-spot of inner conductive, meanwhile, if there is multipoint earthing of iron core situation, it is also possible to produce
Pig iron core is overheated.Infrared temperature-test technology is exactly to utilize the infrared waves produced in these situations, carries out failure predication.At present,
The temperature measurement accuracy of Infrared Thermography Technology is higher, can be analyzed in actual applications using infrared image, monitor transformer component
In each several part temperature, temperature data is then carried out longitudinal direction and lateral comparison, running state of transformer judged and must
The early warning wanted.
However, there are some defects in above-mentioned existing method.Dissolved gas analysis side in such as current existing oil
In method, the static threshold that several characteristic gas indexs of fixation are in accordance with greatly is pre-warning signal, and the accuracy rate of prediction is relatively low.Shake
Dynamic analytic approach is larger by external environment influence, and larger interference, Er Qiebian may be produced under relatively noisy working environment
The failure of some types inside depressor can not well be showed by the form of mechanical oscillation.Infrared measurement of temperature rule must
More accurate infrared monitoring equipment need additionally be bought, it is necessary to put into more financial resources, next there are some inside transformers
Failures such as failure, such as shelf depreciation etc., can not be showed well by infrared temperature-test technology.Lacking based on the above method
Fall into, a kind of fault early warning method of accurate quick of oil-filled transformer field exigence.
The content of the invention
In view of above-mentioned, the present invention proposes a kind of oil-filled transformer fault early warning method for becoming point estimation based on higher-dimension,
This method is based on the various dissolving characteristic gas index in oil-filled transformer insulating oil, to various features gas index institute
(height refers to a certain position or moment to the height of the Higher Dimensional Linear Models of composition, and front and rear observation or data follow two herein
The point of individual different model) estimated, so as to obtain the fault pre-alarming moment.This method can be according to a variety of transformers
Failure, chooses a variety of different characteristic gas indexs, carries out the prediction of transformer fault, accurate with stronger applicability and prediction
Exactness.
A kind of oil-filled transformer fault early warning method for being become point estimation based on Higher Dimensional Linear Models of the present invention, is chosen first
Several oil-filled transformer failure indexs of correlation, then using these Index Establishment Higher Dimensional Linear Models, pass through the group of adaptation
Drag-line method of estimation, estimates the height position in Higher Dimensional Linear Models, and oil immersion is used as using the time location that height occurs
The position of formula transformer fault early warning.The oil-filled transformer fault early warning method objective and fair, it is simple and easy to apply.
A kind of fault early warning method for the oil-filled transformer for becoming point estimation based on Higher Dimensional Linear Models, including following step
Suddenly：
(1) p kind oil-filled transformer failure indexs of correlation are chosen, y is designated as respectively1,y2,...,yi,...,yp, wherein, yi
=(yi,1,yi,2,..,yi,t,..,yi,n)T, yi,tRepresent value of i-th kind of oil-filled transformer failure index of correlation in t；
(2) according to the p kind oil-filled transformer failure indexs of correlation selected, the Higher Dimensional Linear Models of many heights are set up such as
Under：
Wherein, vector xitIt is d dimension explanatory variables, vectorial βjIt is the d dimension unknown parameters of non-zero, eitIt is error term, snIt is not
Know variable point numeral；aj, j=1,2 ..., snRepresent the position of unknown height；
yi,tRepresent the value in i-th kind of failure index of correlation of t, the common p of index of correlationtKind, ptRepresent the failure in moment t
The number of index of correlation, because failure index of correlation there may be missing values, there are p kinds at the not necessarily all moment；
Do sign reversingY=
(y(1),...,y(n))T, e=(e(1),...,e(n))T,θ(1)=β1,J=1 ..., sn(Meaning
Taste parameter betaj+1≠βj, i.e. ajFor the height position of model),
The Higher Dimensional Linear Models of many power transformations are converted into matrix form y=X θ+e；
(3) by adaptability group's drag-line estimation (Adaptive Group Lasso Estimator) to the higher-dimension of many heights
Parameter θ in linear model y=X θ+e is estimated, it is possible to obtain the position of height in the Higher Dimensional Linear Models of many heights
Put；θ group's drag-line estimator is：
Wherein,
| | | | the theorem in Euclid space norm of expression, d is explanatory variable xitDimension；Explanatory variable is transformer itself
Endogenous variable, for explaining dependent variable yi,tλnIt is regulation parameter with v；Parameter v=1, parameter lambdanPass through bayesian information criterion
(Bayesian information criterion) chooses, i.e.,
Wherein,
Described group's drag-line estimates that the method for parameter θ is：
A, note X=(X(1),...,X(n)), wherein
B, definitionT=1 ..., n；(ν is Greek alphabet)
C, initializes the value of parameter, and wherein s=0 is iteration count parameter, and r is iteration parameter,
D, for arbitrary t=1 ..., n, does following iteration：
(d-1) calculate
(d-2) update
(d-3) update
Until r convergences, otherwise update s=s+1, repeat step (d-1)~step (d-3) process；
E, after convergence, adaptability group's drag-line estimation parameter of parameter θBecome
Point set is combined into
After height position is obtained, in sample data, to be not later than last height position of failure appearance as change
Depressor fault pre-alarming position, the time between transformer fault early warning position and abort situation is the average pre-warning time of failure.
Compared with prior art, the present invention has advantages below：
Oil-filled transformer fault early warning method of the present invention is related by several oil-filled transformer failures to selection
Group's drag-line estimation that the Higher Dimensional Linear Models that index is set up are adapted to, has obtained the height position in Higher Dimensional Linear Models,
The position occurred using height realizes the early warning to transformer fault as the position of oil-filled transformer fault pre-alarming.Relatively
In other failure prediction methods, the present invention only needs to the characteristic gas concentration of dissolving in oil-filled transformer insulating oil, without
Want other extra equipment.Meanwhile, applicability of the present invention extensively, goes for the various events produced inside oil-filled transformer
Hinder type.
Brief description of the drawings
Fig. 1 is transformer CH in embodiment 14And C2H6Concentration tendency and height and abort situation figure；
Fig. 2 is transformer CO/CO in embodiment 12、CH4/(CH4+C2H6+C2H4) and C2H4/(H2+C2H6+C2H4) tendency with
Height and abort situation figure.
Embodiment
In order to more specifically describe the present invention, below in conjunction with the accompanying drawings and embodiment is to technical scheme
It is described in detail.
Embodiment 1
The present embodiment have chosen 5 kinds of characteristic gas indexs by taking the failure transformer that Zhejiang power grid numbering is 895058 as an example,
The CH4 concentration of dissolving respectively in oil-filled transformer insulating oil, C2H6 concentration, CO/CO2 ratio, CH4 relative to C2H6,
Relative concentration values C2H4/ (H2+s of the C2H4 relative concentration values CH4/ (CH4+C2H6+C2H4) and C2H4 relative to H2, C2H6
C2H6+C2H4 value).
First, it regard the later moment in time of above-mentioned 5 kinds of indexs as failure phase relative to the absolute value of the growth rate of previous moment
Close index.Assuming that λ1,λ2,...,λ5For above-mentioned five kinds of characteristic gas indexs, wherein λi=(λi,1,λi,2,..,λi,n)T, λi,tFor
Value of the i kind characteristic gas indexs t-th of moment.Failure index of correlation is designated as y1,y2,...,y5, wherein yi=(yi,1,
yi,2,..,yi,n)T, yi,tRepresent value of i-th kind of oil-filled transformer failure index of correlation t-th of moment.Order：
Then, according to 5 kinds of failure indexs of correlation obtained above, the Higher Dimensional Linear Models for setting up many heights are as follows：
Wherein yi,tRepresent the value in i-th kind of failure index of correlation of t (totally 5 kinds)；Vector xitIt is d dimension explanatory variables, to
Measure βjIt is the d dimension unknown parameters of non-zero, makes d=1, i.e. explanatory variable only comprising intercept here, so that xit=1.eitIt is error
, snIt is unknown variable point numeral；aj, j=1,2 ..., snRepresent the position of unknown height.
Do sign reversingY=
(y(1),...,y(n))T, e=(e(1),...,e(n))T,θ(1)=β1,J=1 ..., sn(Meaning
Taste parameter betaj+1≠βj, i.e. ajFor the height position of model),
Then the Higher Dimensional Linear Models of many heights can be converted into matrix form y=X θ+e.
Finally, the parameter θ in the Higher Dimensional Linear Models y=X θ+e of many heights is estimated by adaptability group's drag-line estimation
Meter, it is possible to obtain the position of height in the Higher Dimensional Linear Models of many heights.θ group's drag-line estimator is：
Wherein,
| | | | the theorem in Euclid space norm of expression, λnIt is regulation parameter with v.Parameter v=1, parameter lambdanBelieved by Bayes
Cease criterion to choose, i.e.,
Wherein
Parameter θ group drag-line estimation method be：
A, note X=(X(1),...,X(n)), wherein
B, definitionT=1 ..., n；(ν is Greek alphabet)
C, initializes the value of parameter, and wherein s=0 is iteration count parameter, and r is iteration parameter,
D, for arbitrary t=1 ..., n, does following iteration：
(d-1) calculate
(d-2) update
(d-3) update
Until r convergences, otherwise update s=s+1, repeat step (d-1)~step (d-3) process.
E, after convergence, adaptability group's drag-line estimation parameter of parameter θBecome
Point set is combined into
And then the height position estimated can be regard as the fault pre-alarming position of oil-filled transformer.
Fig. 1,2 illustrate five kinds of failure indexs of correlation (every by taking the failure transformer that Zhejiang power grid numbering is 895058 as an example
One curve is a kind of failure index of correlation, because CH4、C2H6Concentration and CO/CO2,CH4/(CH4+C2H6+C2H4) and C2H4/(H2
+C2H6+C2H4) ordinate ratio gap is excessive, therefore divide into two figures to represent) relation of variation tendency and height position.
With the place of perpendicular solid marks it is the height position that model is estimated on axis of abscissas, the position with perpendicular dashed lines labeled is failure hair
Raw position.If using be not later than failure appearance last height as fault pre-alarming position, four-headed arrow mark distance
Exactly 895058 transformers are converted into the early warning duration of malfunction from fault pre-alarming, i.e., No. 895058 failure transformers
The early warning duration is 119 days.
Embodiment 2
With reference to 14 high temperature overheating fault transformers of Zhejiang power grid and 24 normal transformer belt electro-detection data, use
Same method in embodiment 1, the height situation to all 38 transformers is estimated.To be not later than the last of failure appearance
One height is transformer fault early warning position, and estimated result is as shown in table 1.
Transformer fault early warning situation of the table 1 based on many height Higher Dimensional Linear Models
| Failure transformer | Normal transformer | |
| Average variable point numeral | 5.2 | 2.3 |
| It is not later than the early warning rate of failure generation | 78.6% | / |
| Failure is to failure average time | 193 days | / |
As can be seen from Table 1, the average variable point numeral of failure transformer is apparently higher than the average variable point numeral of normal transformer.In failure
In transformer, failure transformer can be accounted in the transformer for being not later than failure appearance and carrying out fault pre-alarming effect by becoming point estimation
Ratio is 78.6%.If to be not later than last height of failure appearance as fault pre-alarming position, from fault pre-alarming to former
The average transformation time of barrier state, that is, the average pre-warning time of failure is 193 days.Therefore, by failure index of correlation
Many height Higher Dimensional Linear Models in height position estimation, using model change point as malfunction early-warning point, it is possible to achieve to become
The dynamic early-warning of depressor operation conditions.So as to can have certain lead, to that may break down, oil-filled transformer enters
The necessary maintenance and repair of row, and then the generation of failure is prevented, save substantial amounts of manpower, financial resources, material resources consumption.
Technical scheme and beneficial effect are described in detail above-described embodiment, Ying Li
Solution is to the foregoing is only presently most preferred embodiment of the invention, is not intended to limit the invention, all principle models in the present invention
Interior done any modification, supplement and equivalent substitution etc. are enclosed, be should be included in the scope of the protection.
Claims (2)
1. a kind of fault early warning method for the oil-filled transformer for being become point estimation based on Higher Dimensional Linear Models, is comprised the following steps：
(1) p kind oil-filled transformer failure indexs of correlation are chosen, y is designated as respectively1,y2,...,yi,...,yp, wherein, yi=
(yi,1,yi,2,..,yi,t,..,yi,n)T, yi,tRepresent value of i-th kind of oil-filled transformer failure index of correlation in t；
(2) according to the p kind oil-filled transformer failure indexs of correlation selected, the Higher Dimensional Linear Models for setting up many heights are as follows：
<mrow>
<msub>
<mi>y</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>t</mi>
</mrow>
</msub>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msup>
<msub>
<mi>x</mi>
<mrow>
<mi>i</mi>
<mi>t</mi>
</mrow>
</msub>
<mi>T</mi>
</msup>
<msub>
<mi>&beta;</mi>
<mn>1</mn>
</msub>
<mo>+</mo>
<msub>
<mi>e</mi>
<mrow>
<mi>i</mi>
<mi>t</mi>
</mrow>
</msub>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mn>1</mn>
<mo>&le;</mo>
<mi>t</mi>
<mo><</mo>
<msub>
<mi>a</mi>
<mn>1</mn>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msup>
<msub>
<mi>x</mi>
<mrow>
<mi>i</mi>
<mi>t</mi>
</mrow>
</msub>
<mi>T</mi>
</msup>
<msub>
<mi>&beta;</mi>
<mn>2</mn>
</msub>
<mo>+</mo>
<msub>
<mi>e</mi>
<mrow>
<mi>i</mi>
<mi>t</mi>
</mrow>
</msub>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<msub>
<mi>a</mi>
<mn>1</mn>
</msub>
<mo>&le;</mo>
<mi>t</mi>
<mo><</mo>
<msub>
<mi>a</mi>
<mn>2</mn>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mtable>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
</mtable>
</mtd>
<mtd>
<mtable>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
</mtable>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msup>
<msub>
<mi>x</mi>
<mrow>
<mi>i</mi>
<mi>t</mi>
</mrow>
</msub>
<mi>T</mi>
</msup>
<msub>
<mi>&beta;</mi>
<mrow>
<mi>j</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>+</mo>
<msub>
<mi>e</mi>
<mrow>
<mi>i</mi>
<mi>t</mi>
</mrow>
</msub>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<msub>
<mi>a</mi>
<mi>j</mi>
</msub>
<mo>&le;</mo>
<mi>t</mi>
<mo><</mo>
<msub>
<mi>a</mi>
<mrow>
<mi>j</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mtable>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
</mtable>
</mtd>
<mtd>
<mtable>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
</mtable>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msup>
<msub>
<mi>x</mi>
<mrow>
<mi>i</mi>
<mi>t</mi>
</mrow>
</msub>
<mi>T</mi>
</msup>
<msub>
<mi>&beta;</mi>
<mrow>
<msub>
<mi>s</mi>
<mi>n</mi>
</msub>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>+</mo>
<msub>
<mi>e</mi>
<mrow>
<mi>i</mi>
<mi>t</mi>
</mrow>
</msub>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<msub>
<mi>a</mi>
<msub>
<mi>s</mi>
<mi>n</mi>
</msub>
</msub>
<mo>&le;</mo>
<mi>t</mi>
<mo><</mo>
<mi>n</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
Wherein, vector xitIt is d dimension explanatory variables, vectorial βjIt is the d dimension unknown parameters of non-zero, eitIt is error term, snIt is unknown change
Points；aj, j=1,2 ..., snRepresent the position of unknown height；
Do sign reversingY=
(y(1),...,y(n))T, e=(e(1),...,e(n))T,Meaning
Taste parameter betaj+1≠βj, i.e. ajFor the height position of model),
<mrow>
<mi>X</mi>
<mo>=</mo>
<mfenced open = "(" close = ")">
<mtable>
<mtr>
<mtd>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</msub>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mo>...</mo>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
</mtr>
<mtr>
<mtd>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</msub>
</mtd>
<mtd>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</msub>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mo>...</mo>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
</mtr>
<mtr>
<mtd>
<mtable>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
</mtable>
</mtd>
<mtd>
<mtable>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
</mtable>
</mtd>
<mtd>
<mtable>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
</mtable>
</mtd>
<mtd>
<mtable>
<mtr>
<mtd>
<mrow></mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow></mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow></mrow>
</mtd>
</mtr>
</mtable>
</mtd>
<mtd>
<mtable>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
</mtable>
</mtd>
</mtr>
<mtr>
<mtd>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
</msub>
</mtd>
<mtd>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
</msub>
</mtd>
<mtd>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
</msub>
</mtd>
<mtd>
<mo>...</mo>
</mtd>
<mtd>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
</msub>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
Then the Higher Dimensional Linear Models of many power transformations are converted into matrix form y=X θ+e；
(3) parameter θ in the Higher Dimensional Linear Models y=X θ+e of many heights is estimated by adaptability group's drag-line estimation, obtained
The height position into the Higher Dimensional Linear Models of many heights；θ group's drag-line estimator is：
<mrow>
<mover>
<mi>&theta;</mi>
<mo>^</mo>
</mover>
<mo>=</mo>
<mi>arg</mi>
<munder>
<mrow>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
</mrow>
<mi>&theta;</mi>
</munder>
<mo>{</mo>
<mo>|</mo>
<mo>|</mo>
<mi>y</mi>
<mo>-</mo>
<mi>X</mi>
<mi>&theta;</mi>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mi>n</mi>
</msub>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>t</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<mfrac>
<msqrt>
<mi>d</mi>
</msqrt>
<mrow>
<mo>|</mo>
<mo>|</mo>
<msub>
<mover>
<mi>&theta;</mi>
<mo>~</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mi>v</mi>
</msup>
</mrow>
</mfrac>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>&theta;</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</msub>
<mo>|</mo>
<mo>|</mo>
<mo>}</mo>
</mrow>
Wherein,
<mrow>
<msub>
<mover>
<mi>&theta;</mi>
<mo>~</mo>
</mover>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<msqrt>
<mi>d</mi>
</msqrt>
</mfrac>
<msup>
<mrow>
<mo>(</mo>
<msubsup>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msubsup>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<msubsup>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msubsup>
<msub>
<mi>y</mi>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</msub>
</mrow>
<mrow>
<msub>
<mover>
<mi>&theta;</mi>
<mo>~</mo>
</mover>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<msqrt>
<mi>d</mi>
</msqrt>
</mfrac>
<msup>
<mrow>
<mo>(</mo>
<msubsup>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msubsup>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<msubsup>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msubsup>
<msub>
<mi>y</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</msub>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<msqrt>
<mi>d</mi>
</msqrt>
</mfrac>
<msup>
<mrow>
<mo>(</mo>
<msubsup>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msubsup>
<msub>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<msubsup>
<mi>X</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msubsup>
<msub>
<mi>y</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</msub>
</mrow>
| | | | the theorem in Euclid space norm of expression, d is explanatory variable xitDimension；λnIt is regulation parameter with v；Parameter v=1,
Parameter lambdanChosen by bayesian information criterion：
<mrow>
<mi>&lambda;</mi>
<mo>=</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>L</mi>
<mi>L</mi>
</msub>
<mo>,</mo>
<mn>..</mn>
<mo>,</mo>
<msub>
<mi>&lambda;</mi>
<mi>n</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mi>arg</mi>
<munder>
<mrow>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
</mrow>
<mi>&lambda;</mi>
</munder>
<mrow>
<mo>(</mo>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mo>(</mo>
<mrow>
<mfrac>
<mn>1</mn>
<mi>N</mi>
</mfrac>
<mo>|</mo>
<mo>|</mo>
<mi>y</mi>
<mo>-</mo>
<mi>X</mi>
<mover>
<mi>&theta;</mi>
<mo>^</mo>
</mover>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
<mo>)</mo>
<mo>+</mo>
<mfrac>
<mrow>
<mi>d</mi>
<mi>f</mi>
</mrow>
<mi>N</mi>
</mfrac>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mo>(</mo>
<mi>N</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
</mrow>
Wherein,
(4) after height position is obtained, in sample data, to be not later than last height position of failure appearance as transformation
Device fault pre-alarming position, the time between transformer fault early warning position and abort situation is the average pre-warning time of failure.
2. the fault early warning method of the oil-filled transformer as claimed in claim 1 for becoming point estimation based on Higher Dimensional Linear Models,
Characterized in that, the process to parameter θ progress group's drag-line estimation is：
A, note X=(X(1),...,X(n)), wherein
B, definition
C, initializes the value of parameter, and wherein s=0 is iteration count parameter, and r is iteration parameter,
D, for arbitrary t=1 ..., n, does following iteration：
(d-1) calculate
(d-2) update
(d-3) update
Until r convergences, otherwise update s=s+1, repeat step (d-1)~step (d-3) process；
E, after convergence, adaptability group's drag-line estimation parameter of parameter θBecome point set
For
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710193920.3A CN107133378B (en) | 2017-03-28 | 2017-03-28 | Oil-immersed transformer fault early warning method based on the estimation of Higher Dimensional Linear Models height |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710193920.3A CN107133378B (en) | 2017-03-28 | 2017-03-28 | Oil-immersed transformer fault early warning method based on the estimation of Higher Dimensional Linear Models height |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107133378A true CN107133378A (en) | 2017-09-05 |
| CN107133378B CN107133378B (en) | 2019-07-23 |
Family
ID=59721474
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710193920.3A Expired - Fee Related CN107133378B (en) | 2017-03-28 | 2017-03-28 | Oil-immersed transformer fault early warning method based on the estimation of Higher Dimensional Linear Models height |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107133378B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI777724B (en) * | 2021-08-12 | 2022-09-11 | 中華電信股份有限公司 | Transformer maintenance management system, method and computer readable medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101587155A (en) * | 2009-06-08 | 2009-11-25 | 浙江大学 | Oil soaked transformer fault diagnosis method |
| CN105629109A (en) * | 2015-12-30 | 2016-06-01 | 西安工程大学 | ARTI-neural network-based oil-immersed transformer fault diagnosis method |
| CN106447202A (en) * | 2016-09-29 | 2017-02-22 | 国网山东省电力公司 | Power equipment state evaluation method based on data mining and principal component analysis |
-
2017
- 2017-03-28 CN CN201710193920.3A patent/CN107133378B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101587155A (en) * | 2009-06-08 | 2009-11-25 | 浙江大学 | Oil soaked transformer fault diagnosis method |
| CN105629109A (en) * | 2015-12-30 | 2016-06-01 | 西安工程大学 | ARTI-neural network-based oil-immersed transformer fault diagnosis method |
| CN106447202A (en) * | 2016-09-29 | 2017-02-22 | 国网山东省电力公司 | Power equipment state evaluation method based on data mining and principal component analysis |
Non-Patent Citations (4)
| Title |
|---|
| BACH F .: "Consistency of the group Lasso and multiple kernel learning", 《 JOURNAL OF MACHINE LEARNING RESEARCH》 * |
| 孙文星，杨贤，李朝晖，孟天男: "一种高可靠的变压器绝缘状态监测方法", 《变压器》 * |
| 李峰: "变压器油中溶解气体在线监测预警策略研究", 《广东电力》 * |
| 石少伟，王可等: "基于模糊综合评价和贝叶斯判别的电力变压器状态判别和预警", 《电力自动化设备》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI777724B (en) * | 2021-08-12 | 2022-09-11 | 中華電信股份有限公司 | Transformer maintenance management system, method and computer readable medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107133378B (en) | 2019-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109740687B (en) | Fermentation process fault monitoring method based on DLAE | |
| Li et al. | A power system disturbance classification method robust to PMU data quality issues | |
| CN104764985A (en) | Method for estimating low-frequency dielectric loss of paper oil insulation system based on parameter identification | |
| CN107203690A (en) | A kind of GM（1,1）Model Transformer failure prediction method | |
| CN107133476B (en) | Test excitation and test point collaborative optimization method based on response aliasing measurement | |
| CN106526468A (en) | Breaker state detection method based on waveform characteristics identification | |
| CN104953583A (en) | Method used for online monitoring of low-frequency oscillation of electric power system and based on combination of change-point detection and Prony method | |
| CN107133378A (en) | Become the oil-filled transformer fault early warning method of point estimation based on Higher Dimensional Linear Models | |
| CN105223481A (en) | Based on the partial-discharge ultrahigh-frequency signal initial time defining method of difference energy function | |
| Liu et al. | Dual-channel convolutional network-based fault cause identification for active distribution system using realistic waveform measurements | |
| CN109635430B (en) | Power grid transmission line transient signal monitoring method and system | |
| Bachri et al. | Cognitive Artificial Intelligence method for interpreting transformer condition based on maintenance data | |
| CN212749115U (en) | Fault single-end accurate positioning system for high-voltage GIL equipment | |
| Paul et al. | Copula based bivariate modelling of DGA and breakdown voltage in high voltage transformers and reactors | |
| CN113762507B (en) | Semi-supervised deep learning arc voltage anomaly detection method based on phase space reconstruction | |
| CN107122829B (en) | Method for diagnosing transformer fault by training neural network through virtual sample | |
| Zhang et al. | A Novel Diagnosis Method Based on the Fea-ture Selection Algorithm and Convolution Neural Network for SF6 Gas-Insulated Equipment | |
| US11320492B2 (en) | Systems and methods of power system state estimation | |
| Mitiche et al. | An evaluation of total variation signal denoising methods for partial discharge signals | |
| CN114462521A (en) | Efficient intelligent classification and detection method for power equipment | |
| Shingare et al. | Efficient Fault Detection and Location in Extra High Voltage Networks: An Artificial Neural Network (ANN)-Based Approach | |
| CN113418632A (en) | Conceptual drift detection method for oil temperature prediction of oil-immersed transformer | |
| Zhang et al. | Transformer anomaly detection method based on mds and lof algorithm | |
| CN112630638B (en) | Method and system for identifying flashover of porcelain bushing of circuit breaker | |
| JP2006229148A (en) | Method of diagnosing deterioration of oil-immersed transformer |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190723 |