CN102662113A - Comprehensive diagnosis method of oil-immersed transformer based on fault tree - Google Patents

Abstract

The invention discloses a comprehensive diagnosis method of an oil-immersed transformer based on a fault tree, which includes steps: (1) collecting correlated condition characteristic parameter values of gas dissolved in oil, moisture in oil, local discharge, iron core grounding current, top oil temperature and winding optical fiber temperature measurement of the transformer in real time, (2) normalizing by aid of diagnosed fault types of gas dissolved in oil, (3) counting fault diagnose accuracy of all diagnose methods, (4) obtaining state quantity of measured gas dissolved in oil for respective diagnose, (5) obtaining a fault confidence digit group, and (6) integrating to obtain a fault diagnose result of the transformer. The comprehensive diagnosis method combines advantages of various oil chromatographic diagnose methods, overcomes defects of a single diagnose method, adopts gas dissolved in oil as a diagnose core, is supplemented by fault tree diagnose, masters operation health conditions of the transformer macroscopically, can achieve fault positioning of devices, and has important practical meaning for finding potential hidden fault danger of devices and ensuring safe and stable operation of an electrical power system.

Description

Oil-filled transformer error comprehensive diagnosis method based on fault tree

Technical field

The present invention relates to belong to converting equipment fault diagnosis technology field, the oil-filled transformer error comprehensive diagnosis method that is based on fault tree that is specifically related to.

Background technology

Power transformer is the hinge of electrical network electric energy transmitting; It is the main equipment of operation of power networks; Its reliability of operation is the safety of electric system and stable essential condition; Along with the raising of generating set capacity and the rising of transmission voltage grade, also corresponding thereupon raising of the capacity of transformer and electric pressure and rising, the reliability to transformer has also proposed requirements at the higher level simultaneously.So on-line monitoring and fault diagnosis through transformer are found potential potential faults early, have important and practical meanings.

Formed ripe relatively transformer online monitoring technology both at home and abroad, like shelf depreciation, oil chromatography analysis, iron core grounding current and top-oil temperature etc.Yet method for diagnosing faults on this basis relatively lags behind; Not only all only carry out fault diagnosis like IEC three-ratio method, David's triangulation method, neural network algorithm and rough set etc. to single monitoring type; And the good operation health status of grasp macroscopical transformer; And various diagnostic methods have corresponding shortcoming, the situation of " do not have coding " can occur like three-ratio method, and neural network algorithm needs a large amount of historical failure data as training sample etc.

Summary of the invention

The objective of the invention is to overcome the shortcoming of the single diagnostic method of traditional oils chromatogram; Utilize multiple diagnostic method comprehensive diagnos to improve the accuracy of fault diagnosis; Introducing monitoring variables such as top-oil temperature, shelf depreciation, Wei Shui, iron core grounding current and winding optical fiber temperature-measurement simultaneously assesses transformer operation health status; In order to definite scope of failure, and excavate relevant diagnosis rule, for repair based on condition of component provides the plan foundation.

The technical scheme that the present invention adopts is: the oil-filled transformer error comprehensive diagnosis method based on fault tree, it is characterized in that, and comprise the following steps:

(1), gathers little water, shelf depreciation, iron core grounding current, top-oil temperature, winding optical fiber temperature-measurement correlation behavior characteristic parameter value in the oil dissolved gas, oil of transformer in real time through on-Line Monitor Device;

(2) can turn to following 8 kinds through the fault type specification of oil dissolved gas diagnosis: low-temperature heat fault (t＜300 ℃), in warm fault (300 ℃≤t＜700 ℃), elevated temperature heat fault (t>=700 ℃), shelf depreciation, low-yield discharge, that low-yield discharge is held concurrently is overheated, high-energy discharge, that high-energy discharge is held concurrently is overheated;

(3) choose the improvement three-ratio method, method, IEC60599 method, no compiling method and five kinds of methods of David's triangulation method diagnostic method as oil dissolved gas grinds in improvement electricity association; And utilize the historical sample data, all kinds of diagnostic method tracing trouble accuracy are added up;

(4) five quantity of states of hydrogen, methane, ethane, ethene, acetylene that obtain the oil dissolved gas that monitoring device records (cross but insulating oil is not outgased for fault handling and handle; get rate of change as deal with data) use five kinds of methods in (3) to diagnose respectively, confirm corresponding nature of trouble;

(5) obtain little water in the oil, shelf depreciation, iron core grounding current, top-oil temperature, the relevant monitoring variable of winding optical fiber temperature-measurement, utilize fault tree diagnosis mechanism to obtain fault and put the letter array;

(6) diagnostic result in comprehensive step (4) and the step (5) carries out analysis-by-synthesis, draws the fault diagnosis result of this transformer;

(7) operations staff is according to the real fail result feedback of various test check method afford systems, and system's each minute diagnostic method carries out accuracy to be upgraded.

The comprehensive oil dissolved gas diagnostic method of transformer in the above-mentioned steps (1), its method is the mode that the diagnostic method that combines various maturations combines with correct probability, confirms the weight coefficient of various kinds of equipment likelihood of failure.Be about to various diagnostic methods and carry out weight assignment, draw a comprehensive oil dissolved gas diagnostic result, concrete steps are following:

Step 1 is got hydrogen (H in the transformer oil dissolved gas ₂), methane (CH ₄), ethane (C ₂H ₆), ethene (C ₂H ₄), acetylene (C ₂H ₂) oil in concentration value (or changing value), calculate the gas production rate of each gas;

Step 2 judges whether gas concentration value or gas production rate surpass demand value (referring to DL/T722-2000); If surpass demand value, then execution in step 3; Otherwise stop.

Step 3; Grind method, no compiling method and five kinds of comparatively ripe methods of David's triangulation method according to improvement three-ratio method, IEC60599, improvement electricity association the data of step 1 are done fault diagnosis, and give accuracy weight (each diagnostic method accuracy of diagnosis sees table 5 for details) the result.

Step 4, do normalization according to the result of step 3 and accuracy weight and handle and provide all kinds of probabilities of malfunction.

According to above-mentioned comprehensive oil dissolved gas diagnostic result, utilize the fault tree principle to combine monitoring variable data such as top-oil temperature, Wei Shui, shelf depreciation, iron core grounding current, winding optical fiber temperature-measurement, transformer health status is done comprehensive diagnosis.

Comprehensive diagnos result, CO and CO in conjunction with above-mentioned monitoring variable data ₂Concentration the solid insulating material (insulating paper etc.) that whether relates to of fault is diagnosed and N ₂And O ₂Relation judge the oxidation situation, provide the state description of transformer comprehensive health at last.

Above-mentioned CO and CO ₂Concentration to carry out diagnostic method be according to CO to the solid insulating material that whether relates to of fault ₂The value of/CO is judged the fault of solid insulating material, if CO ₂/ CO＞7 solid insulating material is aging; If CO ₂/ CO＜3, then fault relates to solid insulating material; If O ₂/ N ₂＜0.3, then there is oxidative phenomena in insulating oil, causes oxygen extremely to be consumed.

The present invention combines the advantage of various oil chromatography diagnostic methods and overcomes the shortcoming of single diagnostic method, as the diagnosis core, and is aided with fault tree diagnosis with oil dissolved gas.Utilize oil dissolved gas, top-oil temperature, shelf depreciation, Wei Shui, iron core grounding current and the multi-parameter integrated diagnosis of winding optical fiber temperature-measurement simultaneously; The operation health condition of not only can macroscopic view controlling transformer through this method; Can also realize the localization of fault of equipment; To the potential potential faults of discovering device, guarantee that the operation of power system safety and stability has important and practical meanings.

Description of drawings

Fig. 1 is the transformer comprehensive diagnos process flow diagram that the present invention is based on fault tree;

Fig. 2 is many monitoring variables of transformer fault tree of the present invention;

Fig. 3 is David's triangle fault shape diagnostic graph of the present invention.

Embodiment

For technological means, creation characteristic that the present invention is realized, reach purpose and effect and be easy to understand and understand, below in conjunction with embodiment, further set forth the present invention.

Referring to Fig. 1, a kind of oil-filled transformer error comprehensive diagnosis method of the present invention based on fault tree, concrete diagnosis algorithm is following:

(1) obtains the value of the relevant monitoring variable of transformer through oil dissolved gas, shelf depreciation, top-oil temperature, Wei Shui, iron core grounding current and winding optical fiber temperature-measurement monitoring device, should get increment as diagnostic value for fault and thoroughly repaired transformer.The monitoring variable that various monitoring types are gathered is following:

Oil dissolved gas: hydrogen, methane, ethane, ethene, acetylene, carbon monoxide, carbon dioxide, oxygen, nitrogen, total hydrocarbon;

Shelf depreciation: discharge capacity, discharge position, pulse number;

Little water: moisture;

Iron core grounding current: iron core total current;

Top-oil temperature: top-oil temperature;

Winding optical fiber temperature-measurement: measuring point temperature.

(2) can standardize specific as follows through the transformer fault type that oil dissolved gas is diagnosed out:

0101 oil chromatography diagnosis cryogenic overheating;

Temperature is overheated in the diagnosis of 0102 oil chromatography;

0103 oil chromatography diagnosis hyperthermia and superheating;

0104 oil chromatography diagnosis shelf depreciation;

0105 oil chromatography is diagnosed low-yield discharge;

It is overheated that 0106 oil chromatography diagnoses low-yield discharge to hold concurrently;

0107 oil chromatography diagnosis high-energy discharge;

It is overheated that 0108 oil chromatography diagnosis high-energy discharge is held concurrently;

(3) judge according to table 1 whether hydrogen, acetylene and methane concentration value exceed standard.

Table 1 gas concentration demand value unit: ppm

Electric pressure	Total hydrocarbon	Acetylene	Hydrogen
				VoltageLeve?l	TotalHydrocarbon	C 2H 2	H 2
≥330kV	150	1	150
				≤220kV	150	5	150

If do not exceed standard, then judge corresponding gas gas production rate according to table 2:

Table 2 gas production rate demand value unit: mL/d

Equipment	Total hydrocarbon	Acetylene	Hydrogen	Carbon monoxide	Carbon dioxide
						Equipmenttype	TotalHydrocarbon	C 2H 2	H 2	CO	CO 2
Open	6	0.1	5	50	100
						Diaphragm type	12	0.2	10	100	200

As meet above condition and then grind method, IEC60599 method, no compiling method and five kinds of oil chromatography diagnostic methods of David's triangulation method the oil dissolved gas monitoring variable in (1) is diagnosed according to improvement three-ratio method, improvement electricity association.Concrete diagnostic flow chart is seen accompanying drawing 1.

(4) give corresponding accuracy rate according to the corresponding historical accuracy of diagnosis of five kinds of methods, merge the same fault type and also give the diagnosis diagnosis, directly give 0, draw final fault probability array for the fault type of not diagnosing out, shape as: $[{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^1,{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^2,{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^3,{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^4,{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^5,{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^6,{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^7,{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^8].$

(5) according to the monitoring variable of shelf depreciation, Wei Shui, iron core grounding current, top-oil temperature, winding optical fiber temperature-measurement,, draw the operation health status of transformer based on the diagnosis principle of fault tree.Concrete fault tree sees accompanying drawing 2.

(6) according to the fault tree diagnostic result of (5), must be out of order and put the letter array, concrete as [n1; N2, n3, n4; N5, n6] represent respectively the low-temperature heat fault, in warm fault, elevated temperature heat fault, shelf depreciation, low-yield discharge and high-energy discharge take place put the letter number of times.

Monitor type corresponding corresponding failure type in following measured value range for top-oil temperature, winding optical fiber temperature-measurement and iron core grounding current:

Table 3 monitoring type and hot stall eigenwert

The monitoring type	The low-temperature heat fault	In warm fault	The elevated temperature heat fault
				Top-oil temperature	≥85℃	≥95℃	≥105℃
The winding optical fiber temperature-measurement	≥100℃	≥110℃	≥120℃
				Iron core grounding current	≥100mA	≥300mA	≥1000mA

For partial discharge monitoring type corresponding corresponding failure type in following measured value range:

Table 4 shelf depreciation type feature value

The monitoring type	Shelf depreciation	Low-yield discharge	High-energy discharge
				Partial discharge monitoring	≥300pC	≥500pC	≥1000pC

For little water in the oil:

The transformer voltage grade is 220kV and following, and demand value is: 25mg/L; More than electric pressure 330kV reached, demand value was: 15mg/L.

(7) according to the fault of the oil dissolved gas diagnostic result of (4) and (6) a situation arises array, the probability of malfunction array is revised according to following method:

When 1≤X≤5, then $p_x^{\&prime;}=(1+\mathrm{NX}/3)\&times;p_0^X;$

When X=6, $p_6^{\&prime;}=(1+(\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">n</figure-callout>}1+n2+n3)/3)\&times;(1+n5/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^6;$

When X=7, $P_7^{\&prime;}=(1+n6/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^7;$

When X=8, $p_8^{\&prime;}=(1+(\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">n</figure-callout>}1+n2+n3)/3)\&times;(1+n6/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^8;$

Its concrete correction formula is following:

$1){\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">p</figure-callout>}}_1^{\&prime;}=(1+\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">n</figure-callout>}1/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^1;$

$2)p_2^{\&prime;}=(1+n2/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^2;$

$3)p_3^{\&prime;}=(1+n3/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^3;$

$4)p_4^{\&prime;}=(1+n4/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^4;$

$5)p_5^{\&prime;}=(1+n5/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^5;$

$6)p_6^{\&prime;}=(1+(\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">n</figure-callout>}1+n2+n3)/3)\&times;(1+n5/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^6;$

$7)p_7^{\&prime;}=(1+n6/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^7;$

$8)p_8^{\&prime;}=(1+(\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">n</figure-callout>}1+n2+n3)/3)\&times;(1+n6/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^8;$

Obtain the probability of malfunction array for through above step

(8) the probability of malfunction array of (7) is carried out probability normalization, concrete steps are following: the probability of malfunction array be [p ' ₁, p ' ₂, p ' ₃, p ' ₄, p ' ₅, p ' ₆, p ' ₇, p ' ₈]

$\mathrm{Pi}=\frac{P_i^{\&prime;}}{{\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_1^{\&prime;}+P_2^{\&prime;}+\&CenterDot;\&CenterDot;\&CenterDot;+P_8^{\&prime;}}\&times;100\%$

I=1 in the formula, 2 ..., 8;

Obtain normalized probability of malfunction array [p at last ₁, p ₂, p ₃, p ₄, p ₅, p ₆, p ₇, p ₈];

(9) according to CO ₂The value of/CO is judged the fault of solid insulating material, if CO ₂/ CO＞7 solid insulating material is aging; If CO ₂/ CO＜3, then fault relates to solid insulating material.

(10) if O ₂/ N ₂＜0.3, then there is oxidative phenomena in insulating oil, causes oxygen extremely to be consumed.

(11) combine the probability of malfunction array of (8) and the fault diagnosis result in (5), provide transformer operation health status and describe.And notify corresponding personnel to overhaul investigation, and fail result is returned.Based on fail result the correct counting rate meter of diagnostic method is upgraded, revise by following formula, and fault data is done the diagnosis rule storage for the accuracy of diagnosis correct method.

Diagnosis correct method accuracy correction formula:

The method accuracy correction formula of DE: wherein, p is the accuracy of diagnostic method.

Be the technical scheme of further explanation this law, lift following specific embodiment at present:

(1), at first oil dissolved gas, shelf depreciation, Wei Shui, iron core grounding current, top-oil temperature, the monitoring device of winding optical fiber temperature-measurement monitoring type of transformer significant points are delivered to the shutdown of state Access Network through bus with the quantity of state information of obtaining, and resolve and deposit in the corresponding database table.

(2), utilizing sample that method, no compiling method and David's triangulation method are ground by improvement three-ratio method, IEC60599 method, improvement electricity association adds up the accuracy of diagnosis of various method for diagnosing faults and draws initial examining

The disconnected correct counting rate meter of method.As shown in table 5:

The correct counting rate meter of table 5 diagnostic method

Fault 1

Fault 2

Fault 3

Fault 4

Fault 5

Fault 6

Fault 7

Fault 8

Diagnostic method 1

70％

86％

76％

72％

80％

-

80％

-

Diagnostic method 2

74％

79％

78％

75％

88％

84％

62％

70％

Diagnostic method 3

68％

83％

92％

64％

88％

-

85％

-

Diagnostic method 4

?81％

75％

92％

?-

43％

75％

83％

73％

Diagnostic method 5

74％

62％

84％

67％

62％

82％

87％

68％

Wherein

(1) on behalf of improvement three-ratio method, improvement electricity association, diagnostic method 1, diagnostic method 1, diagnostic method 2, diagnostic method 3, diagnostic method 4 and diagnostic method 5 grind method, IEC60599 method, no compiling method and David's triangulation method respectively.

(2) fault 1, fault 2, fault 3, fault 4, fault 5, fault 6, fault 7 and fault 8 represent respectively low-temperature heat fault (t＜300 ℃), in warm fault (300 ℃≤t＜700 ℃), elevated temperature heat fault (t>=700 ℃), shelf depreciation, low-yield discharge, that low-yield discharge is held concurrently is overheated, high-energy discharge and high-energy discharge are held concurrently is overheated.

(3) Pij in the literary composition of back; I ∈ 1,2,3,4, and 5}, { 1,2,3,4,5,6,7,8} representes that diagnostic method i diagnoses the accuracy of the j that is out of order to j ∈.

(3), tentatively judge through following table:

At first the demand value with total hydrocarbon, acetylene, hydrogen and table 1 compares; If do not exceed standard; Then judge according to table 2 whether corresponding gas production rate exceeds standard, five kinds of gas concentrations of methane, hydrogen, ethane, ethene, acetylene of oil dissolved gas carried out fault diagnosis for the following 5 kinds of diagnostic methods of oil dissolved gas utilization that surpass demand value:

1. improve three-ratio method ethane (C ₂H ₆), ethene (C ₂H ₄), acetylene (C ₂H ₂)

Table 6 coding rule table

Table 7 coding-fault type table

2. method is ground by improvement electricity association

Table 8 coding rule table

Table 9 coding-fault type table

3. IEC-60599 method

Table 10IEC-60699 fault analysis table

Code	Fault	C 2H 2/C 2H 4	CH 4/H2	C 2H 4/C 2H 6
					PD	Shelf depreciation	NS	＜0.1	＜0.2
D1	Low-yield discharge	＞1	0.1-0.5	＞1
					D2	High-energy discharge	0.6-2.5	0.1-1	＞2
T1	Hot stall (＜300 ℃)	NS	＞1	＜1
					T2	Hot stall (300 ℃-700 ℃)	＜0.1	＞1	1-4
T3	Hot stall (＞700 ℃)	＜0.2	＞1	＞4

Annotate: NS representes that whatsoever numerical value is all meaningless

4. there is not compiling method

The no compiling method fault type of table 11 table

Fault type	C 2H 2/C 2H 4	C 2H 4/C 2H 6	CH 4/H2
				Hot stall (＜300 ℃)	＜0.1	＜1	Irrelevant
Hot stall (300 ℃-700 ℃)	＜0.1	1-3	Irrelevant
				Hot stall (＞700 ℃)	＜0.1	＞3	Irrelevant
High-energy discharge	0.1-3	Irrelevant	＜1
				High-energy discharge is held concurrently overheated	0.1-3	Irrelevant	＞1
Low-yield discharge	＞3	Irrelevant	＜1
				It is overheated that low-yield discharge is held concurrently	＞3	Irrelevant	＞1

5. David's triangulation method (seeing accompanying drawing 3)

Among Fig. 3:

X=[C ₂H ₂] unit is: ppm;

Y=[C ₂H ₄] unit is: ppm;

Z=[CH ₄] unit is: ppm;

Zone among the figure: PD: shelf depreciation;

D1: low energy discharge;

D2: high-energy discharge;

T1: hot stall: t＜300 ℃;

T2: hot stall: 300 ℃≤t＜700 ℃;

T3: hot stall: t＞700 ℃;

D+T: overheated double discharge.

Each piece zone is divided by following table ratio among the figure:

Table 12 zone boundary numerical tabular

PD	98％CH 4
					D1	23％C 2H 4	13％C 2H 2
D2	23％C 2H 4	13％C 2H 2	38％C 2H 4	29％C 2H 2
					T1	4％C 2H 2	10％C 2H 4
T2	4％C 2H 2	10％C 2H 4	50％C 2H 4
					T3	15％C 2H 2	50％C 2H 4

[022] (four), can obtain the diagnostic result of following form through (three):

Table 13 diagnostic method and table as a result

Diagnostic method	Improve three ratios	Method is ground by improvement electricity association	IEC60599	No compiling method	David's triangulation method
						Diagnostic result	Fault type i1	Fault type i2	Fault type i3	Fault type i4	Fault type i5

Annotate: i1, i 2, i 3, i 4 and i 5 can equate.

Draw probability of malfunction array { (fault type i1, the P of this transformer _{1, i1}), (fault type i2, P _{2, i2}), (fault type i3, P _{3, i3}), (fault type i4, P _{4, i4}), (fault type i5, P _{5, i5}), merge same fault type (the direct addition of accuracy), obtain the probability of malfunction array:

(5), utilize fault tree respectively the monitoring variable of shelf depreciation, Wei Shui, iron core grounding current, top-oil temperature, winding optical fiber temperature-measurement to be carried out comprehensive diagnos (seeing accompanying drawing 2 for details), draw transformer fault and put the letter array, form is following:

[cryogenic overheating, middle temperature is overheated, hyperthermia and superheating, shelf depreciation, low-yield discharge, high-energy discharge]=[n 3 for n1, n2, n4, n5, n6].

(6), two arrays drawing according to (four) and (five) carry out the correction of probability of malfunction array, specifically mode is following: $1){\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">p</figure-callout>}}_1^{\&prime;}=(1+\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">n</figure-callout>}1/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^1;$

$2)p_2^{\&prime;}=(1+n2/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^2;$

$3)p_3^{\&prime;}=(1+n3/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^3;$

$4)p_4^{\&prime;}=(1+n4/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^4;$

$5)p_5^{\&prime;}=(1+n5/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^5;$

$6)p_6^{\&prime;}=(1+(\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">n</figure-callout>}1+n2+n3)/3)\&times;(1+n5/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^6;$

$7)p_7^{\&prime;}=(1+n6/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^7;$

$8)p_8^{\&prime;}=(1+(\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">n</figure-callout>}1+n2+n3)/3)\&times;(1+n6/3)\&times;{\mathrm{<figure-callout\; id="0"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_0^8;$

Through above step obtain the probability of malfunction array for [p ' ₁, p ' ₂, p ' ₃, p ' ₄, p ' ₅, p ' ₆, p ' ₇, p ' ₈]; Then the above probability of malfunction array that draws is standardized:

$\frac{P_i^{\&prime;}}{{\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="HDA0000154087210000012.png"\; state="\{\{state\}\}">P</figure-callout>}}_1^{\&prime;}+P_2^{\&prime;}+\&CenterDot;\&CenterDot;\&CenterDot;+P_8^{\&prime;}}\&times;100\%$ (i=1 wherein, 2 ..., 8)

Obtain normalized probability of malfunction array [p at last ₁, p ₂, p ₃, p ₄, p ₅, p ₆, p ₇, p ₈];

(6) according to CO ₂/ CO and O ₂/ N ₂The oxidation situation that must be out of order respectively and whether relate to insulating material and insulating oil.

(7) 8 probability in normalized probability of malfunction array are sorted, and get preceding 3, provide the comprehensive description of transformer fault:

Show this transformer P through comprehensive diagnos _iThe probability i that breaks down; P _jThe probability j that breaks down; P _kThe probability k that breaks down; (annotate: P _i, P _j, P _kFor arranging the probability of first three in the normalization probability of malfunction array, below select according to monitoring type monitoring result) for option

1. have electric discharge phenomena at the A place, discharge capacity reaches N1pC, belongs to the B electric discharge type;

2. top-oil temperature reaches N ₂℃, it is overheated to reach C;

3. iron core grounding current reaches N3mA, reaches the D overheated condition;

4. winding measuring point E temperature reaches N4 ℃, exists E overheated;

5. little water reaches N5mg/L in the oil, has surpassed demand value, causes hydrogen content too high, or even causes the reason of shelf depreciation;

6. fault relates to solid insulating material, has the creepage fault of screen;

7. the insulating oil oxidative phenomena is serious, and oxygen extremely consumes.

Wherein A, N1, N2, N3, N4, N5 are the amount of all kinds of monitoring type monitorings, the corresponding failure type that B, C, D, E draw according to the fault tree diagnosis.

The present invention combines the advantage of various oil chromatography diagnostic methods and overcomes the shortcoming of single diagnostic method, as the diagnosis core, and is aided with fault tree diagnosis with oil dissolved gas.Utilize oil dissolved gas, top-oil temperature, shelf depreciation, Wei Shui, iron core grounding current and the multi-parameter integrated diagnosis of winding optical fiber temperature-measurement simultaneously; The operation health condition of not only can macroscopic view controlling transformer through this method; Can also realize the localization of fault of equipment; To the potential potential faults of discovering device, guarantee that the operation of power system safety and stability has important and practical meanings.

The above is merely embodiments of the invention; Be not so limit claim of the present invention; Every equivalent structure or equivalent flow process conversion that utilizes instructions of the present invention and accompanying drawing content to be done; Or directly or indirectly be used in other relevant technical fields, all in like manner be included in the scope of patent protection of the present invention.

Claims (10)

1. based on the oil-filled transformer error comprehensive diagnosis method of fault tree, it is characterized in that, comprise the following steps:

(1), gathers little water, shelf depreciation, iron core grounding current, top-oil temperature, winding optical fiber temperature-measurement correlation behavior characteristic parameter value in the oil dissolved gas, oil of transformer in real time through on-Line Monitor Device;

(2) can turn to following 8 kinds through the fault type specification of oil dissolved gas diagnosis: t＜300 ℃ for low-temperature heat fault, 300 ℃≤t＜700 ℃ be in warm fault, t>=700 ℃ for elevated temperature heat fault, shelf depreciation, low-yield discharge, that low-yield discharge is held concurrently is overheated, high-energy discharge, that high-energy discharge is held concurrently is overheated; Wherein, t is a temperature;

(3) choose the improvement three-ratio method, method, IEC60599 method, no compiling method and five kinds of methods of David's triangulation method diagnostic method as oil dissolved gas grinds in improvement electricity association; And utilize the historical sample data, all kinds of diagnostic method tracing trouble accuracy are added up;

(4) five quantity of states of hydrogen, methane, ethane, ethene, acetylene that obtain the oil dissolved gas that monitoring device records use five kinds of methods in the step (3) to diagnose respectively, confirm corresponding nature of trouble;

(5) obtain little water in the oil, shelf depreciation, iron core grounding current, top-oil temperature, the relevant monitoring variable of winding optical fiber temperature-measurement, utilize fault tree diagnosis mechanism to obtain fault and put the letter array;

(6) diagnostic result in comprehensive step (4) and the step (5) carries out analysis-by-synthesis, draws the fault diagnosis result of this transformer;

(7) operations staff is according to the real fail result feedback of various test check method afford systems, and system's each minute diagnostic method carries out accuracy to be upgraded.

2. the oil-filled transformer error comprehensive diagnosis method based on fault tree according to claim 1 is characterized in that, in the said step (1), the monitoring variable that various monitoring types are gathered is following:

Oil dissolved gas: hydrogen, methane, ethane, ethene, acetylene, carbon monoxide, carbon dioxide, oxygen, nitrogen, total hydrocarbon;

Shelf depreciation: discharge capacity, discharge position, pulse number;

Little water: moisture;

Iron core grounding current: iron core total current;

Top-oil temperature: top-oil temperature;

Winding optical fiber temperature-measurement: measuring point temperature;

In the said step (4), when obtaining the oil dissolved gas that monitoring device records, cross but insulating oil is not outgased for fault handling and handle, get rate of change as deal with data.

3. the oil-filled transformer error comprehensive diagnosis method based on fault tree according to claim 1; It is characterized in that; In the said step (3); Corresponding historical accuracy of diagnosis according to five kinds of methods is given corresponding accuracy rate; Merge the same fault type and give the diagnosis diagnosis, directly give 0, draw final fault probability array

for the fault type of not diagnosing out

4. the oil-filled transformer error comprehensive diagnosis method based on fault tree according to claim 3 is characterized in that, in the said step (4); Must being out of order according to the fault tree diagnostic result, to put the letter array be n1, n2, n 3, n4, n5, n6, wherein, and n1; N2, n3, n4; N5, n6 represent respectively the low-temperature heat fault, in warm fault, elevated temperature heat fault, shelf depreciation, low-yield discharge and high-energy discharge take place put the letter number of times; According to the oil dissolved gas diagnostic result of step (4) and the final fault probability array of step (3), the probability of malfunction array is revised then according to following method:

When 1≤X≤5, then

When X=6,

When X=7,

When X=8,

5. the oil-filled transformer error comprehensive diagnosis method based on fault tree according to claim 4 is characterized in that, above-mentioned probability of malfunction array is carried out the probability normalization, and concrete steps are following:

The probability of malfunction array is

I=1 in the formula, 2 ..., 8;

Obtain normalized probability of malfunction array [p at last ₁, p ₂, p ₃, p ₄, p ₅, p ₆, p ₇, p ₈].

6. the oil-filled transformer error comprehensive diagnosis method based on fault tree according to claim 5 is characterized in that, according to above-mentioned probability of malfunction array and fault diagnosis result, provides transformer operation health status and describes; And notify corresponding personnel to overhaul investigation, and fail result is returned; According to fail result the correct counting rate meter of diagnostic method is upgraded, revise by following formula, and fault data is done the diagnosis rule storage for the accuracy of diagnosis correct method;

Diagnosis correct method accuracy correction formula:

The method accuracy correction formula of DE: wherein, p is the accuracy of diagnostic method.

7. the oil-filled transformer error comprehensive diagnosis method based on fault tree according to claim 1; It is characterized in that the comprehensive oil dissolved gas diagnostic method of transformer in the above-mentioned steps (1), the mode of its method for combining various diagnostic methods to combine with correct probability; Confirm the weight coefficient of various kinds of equipment likelihood of failure; Be about to various diagnostic methods and carry out weight assignment, draw a comprehensive oil dissolved gas diagnostic result, concrete steps are following:

Step 1 is got concentration value or changing value in the oil of hydrogen, methane, ethane, ethene, acetylene in the transformer oil dissolved gas, calculates the gas production rate of each gas;

Step 2 judges whether gas concentration value or gas production rate surpass demand value; If surpass demand value, then execution in step 3; Otherwise stop.

Step 3 is ground method, no compiling method and five kinds of methods of David's triangulation method the data of step 1 is done fault diagnosis, and give the accuracy weight to the result according to improvement three-ratio method, IEC60599, improvement electricity association;

Step 4 is done the normalization processing and is provided all kinds of probabilities of malfunction according to the result and the accuracy weight of step 3.

8. the oil-filled transformer error comprehensive diagnosis method based on fault tree according to claim 7; It is characterized in that; Give the accuracy weight according to above-mentioned to the result; Utilize the fault tree principle to combine the monitoring variable data of top-oil temperature, Wei Shui, shelf depreciation, iron core grounding current and winding optical fiber temperature-measurement, transformer health status is done comprehensive diagnosis.

9. the oil-filled transformer error comprehensive diagnosis method based on fault tree according to claim 8 is characterized in that, according to transformer health status being the result of comprehensive diagnostic and being combined CO and CO ₂Concentration the solid insulating material that whether relates to of fault is diagnosed and N ₂And O ₂Relation judge the oxidation situation, provide the state description of transformer comprehensive health at last.

10. the oil-filled transformer error comprehensive diagnosis method based on fault tree according to claim 9 is characterized in that above-mentioned CO and CO ₂Concentration to carry out diagnostic method be according to CO to the solid insulating material that whether relates to of fault ₂The value of/CO is judged the fault of solid insulating material, if CO ₂/ CO＞7 solid insulating material is aging; If CO ₂/ CO＜3, then fault relates to solid insulating material; If O ₂/ N ₂＜0.3, then there is oxidative phenomena in insulating oil, causes oxygen extremely to be consumed.

Images