practicing dga 09072010

8/10/2019 Practicing DGA 09072010

1/172

A BPL GLOBAL COMPANY

A BPL GLOBAL COMPANY1

8/10/2019 Practicing DGA 09072010

2/172


Serveron Corporation 2008. No use without permission.

2

Agenda

Serveron Corporate Background

Transformer Reliability

Dissolved Gas Analysis

DGA Interpretation

Accuracy & Gas levels

On-Line and Lab DGA Case Studies

8/10/2019 Practicing DGA 09072010

3/172



3

8/10/2019 Practicing DGA 09072010

4/172



4

Hillsboro, USA Headquarters Branch offices in Beijing and Brussels

Incorporated in 2001 EPRI and proprietary GC based

technology Siemens investment and re-

branding agreement-2006

Fully integrated operation Engineering Manufacturing Sales & Marketing Field Service Secure Data Center

Customer Base Over 1000 On-line DGA monitors Over 80 major utilities worldwide Most major transformer

manufacturers

8/10/2019 Practicing DGA 09072010

5/172



5

Michel Duval

Dr Michel Duval obtained a B.Sc. in chemical engineering in 1966and a Ph.D. in polymer chemistry in 1970. He has joined IREQ

(Institut de recherche dHydro Qubec) in 1970. Since then, he hasmade significant contributions in 3 main fields of R&D: dissolved gas-in-oil analysis (DGA), electrical insulating materials and lithium-polymer batteries.

In the field of DGA, M. Duval: Is well-known for his Triangle method of DGA interpretation, usedworldwide.

Has developed and promoted the use of gas-in-oil standards in the IECand ASTM standards

Has established the typical levels of gas formation observed in varioustypes of electrical equipment in service, now used as a reference by theindustry

Has been the Convenor of several IEC working groups and CIGRE taskforces and is the principal author of several IEC international standardson DGA (60567, 60599, 61181).

8/10/2019 Practicing DGA 09072010

6/172



6

Michel Duval

In the field of electrical insulating oils, M.Duval has researched andpublished in the areas of: Metal passivators

Oil reclamation timing

Unstable oil detection and prevention

Paraffinic content of oils

The characterization of XLPE in HV cables and of HV outdoors insulators

M. Duval holds 16 patents and is the author of more than 75 scientific

and technical papers and books , 5 international standards (IEC,

ASTM), and numerous technical reports and presentations in

conferences. He is a Fellow of IEEE and of the Chemical Institute ofCanada. He may be contacted at [email protected].

8/10/2019 Practicing DGA 09072010

7/172



7

8/10/2019 Practicing DGA 09072010

8/172


Serveron Corporation 2007. No use without permission.8

Transformer Asset Management

Transformers are a critical and costly element in the electrical grid

Unplanned failures at any point in the transformer lifecycle havemajor consequences

DGA condition assessment has been recognized for over 30 years forimproving reliability and lowering transformer asset maintenancecosts

TimeNew End of

Design Life

FailureRate

Bathtub Curve

8/10/2019 Practicing DGA 09072010

9/172



FailuresThey Happen!

8/10/2019 Practicing DGA 09072010

10/172



One Example

True Story: 520 MVA GSU Transformer

$3.5 million replacement cost

$0.5 million environmental cleanup

$1.5 million/day spot market buy

$17 million loss in

eight days!!

8/10/2019 Practicing DGA 09072010

11/172



Increasing Failure Rates

All catastrophic failures in any USA news media

Reported number of failures is increasing

Total % of explosive failures is increasing

0

50

100

150

200

250

79 80 82 84 86 88 90 92 94 96 98 00 02NumberofExplosions

SergiTransformer Explosions in the United States

8/10/2019 Practicing DGA 09072010

12/172



Increasing Failure Rates

William H. Bartley P.E., Hartford Steam Boiler Inspection & Insurance Co.,"Life Cycle Management of Utility Transformer Assets,"

0

1

2

3

4

5

1973 1978 1983 1988 1993 1998 2003 2008 2013

Transforme

r

FailureRate(%)

Projected rates will reach unacceptable levels if nothing is done to

improve the system;

Replacing the fleet is not an alternative

One solution is increased monitoring

Published reports in the news validate an increasing failure rate

8/10/2019 Practicing DGA 09072010

13/172



Factors Leading to Increased Failures

Stress experienced:

Mechanical

Thermal

Electrical

Chemical

Tight budgets:

Cutbacks in O&M

Deferred capital replacement

Stress with age: The average age of US power transformers is>42 yrs, increasing 0.7 yrs./yr.

Age itself is not a cause of failure

Increased energy demand: Transformer peak and average

loading has increased

8/10/2019 Practicing DGA 09072010

14/172



Modes of Stress

MECHANICAL STRESS

Failures due to internal stress, weather,and accidents:

Tank failure Radiator failure Winding buckling Core damage

Insulation damage LTC failures

Lead failures Bushing failures

ELECTRICAL STRESS

Failures initially categorized into areassuch as over-voltage or partial

discharge, typically accompanied bythermal or chemical failure:

Switching surges Lightning

THERMAL STRESS

Failures due to insulation destruction orconductor burn-through as the result of:

Short or long-term overloading Faults

Undersized leads Contact coking

Bad joints Loss of cooling Design issues

CHEMICAL STRESS

Failures due to:

Ingress of water

Ingress of oxygenLoss of insulating oil

Paper degradation due to heat or aging

8/10/2019 Practicing DGA 09072010

15/172



Types of Faults

Other

Lightning

Through FaultsInsulation

Deterioration

Inadequate

Maintenance

Moisture

Loose

Connections

Overloading

Workmanship

Source: Hartford Steam Boiler Inspection and Insurance Co.

8/10/2019 Practicing DGA 09072010

16/172

8/10/2019 Practicing DGA 09072010

17/172



8/10/2019 Practicing DGA 09072010

18/172



Monitoring of Gases in Transformers

As insulating material breaks

down due to stress, gases are

formed which dissolve in the

transformer oil

Levels and combinations of the

gases formed are used to detect

incipient faults

For over 50 years, DGA has been

the leading tool to assess

transformer condition

ISSOLVED

AS

NALYSIS
http://members.clipartinc.com/gallery/cgi-bin/imageFolio.cgi?direct=Animations/Mini/Alphabets/Liquid_Letters&img=18http://members.clipartinc.com/gallery/cgi-bin/imageFolio.cgi?direct=Animations/Mini/Alphabets/Liquid_Letters&img=18http://members.clipartinc.com/gallery/cgi-bin/imageFolio.cgi?direct=Animations/Mini/Alphabets/Liquid_Letters&img=18

8/10/2019 Practicing DGA 09072010

19/172



Monitoring of Gases in Transformers

Gas originates from many places:

Mineral insulating oil Conductor paper insulation

Pressboard barriers

Other materials

Gas generation is related to highmaterial temperatures (150C to

1,000C).

Gases are symptoms of:

Poor design or construction Too much electrical stress

Too much thermal stress

Too many short circuits

Overall poor condition

8/10/2019 Practicing DGA 09072010

20/172



Which Gases are Generated?

Insulating Oil

Eight key gases in transformer oil are associated with fault

conditions.

DGA detects the level of gases indicative of incipient faults

which may lead to transformer failure.

Acetylene

(C2H2)

Oxygen (O2)

CarbonDioxide (CO2)

CarbonMonoxide (CO)

Ethylene

(C2H4)

Hydrogen

(H2)

Methane

(CH4)

Ethane

(C2H6)

8/10/2019 Practicing DGA 09072010

21/172



The Importance of DGA to Reliability

DGA enables the detection of the presence and severity of faults:

Air leaks

Hot metal faults

Arcing and partial discharges

DGA may help to indirectly detect root cause of faults in:

Windings (short circuits, insulation failure)

Cleats and leads (high contact resistance, loose contacts)

Tanks (ground problems, circulating currents)

Tap changer (resistive contacts, leaks into main tank)

Core (magnetic flux problems)

Oxidation of materials (mostly CO, CO2)

8/10/2019 Practicing DGA 09072010

22/172



8/10/2019 Practicing DGA 09072010

23/172



Traditional DGA Methods

Samples of oil are manually taken

on a set schedule.Samples are sent to Accredited Labs

DGA labs use Gas Chromatography (GC)

8/10/2019 Practicing DGA 09072010

24/172

8/10/2019 Practicing DGA 09072010

25/172



Standards and Guidelines GoverningLaboratory DGA

ASTM D3612-2002 Standard Test Method for Analysis of Gases Dissolved inElectrical Insulating Oil by Gas Chromatography

IEC 60567-2005 Oil-filled electrical equipment - Sampling of gases and of oil

for analysis of free and dissolved gases

Some other ASTM standards are available;

D3613-1998 - Standard Practice for Sampling Insulating Liquids for Gas

Analysis and Determination of Water Content

D3305-95 (1999)Standard Practice for Sampling Small Gas Volume in

a Transformer

D2759-2000Standard Practice for Sampling Gas from a Transformer

under Positive Pressure

8/10/2019 Practicing DGA 09072010

26/172



Oil Sampling

Manual Sampling A small volume of oil (30 mL) is collected in a gas-tight syringe,

using a 3-way valve, then transported to the laboratory

ASTM method D3613 details procedures for oil sample handling

On-Line Sampling A small volume of oil is continuously circulated through the

monitor and then returned to the transformer

The oil is sampled and analyzed for gas content by the monitor

On-Line monitors offer a closed-loop repeatable oil sampling

process, with no possibility for contamination

8/10/2019 Practicing DGA 09072010

27/172



Oil Sampling Guidelines

References:

ASTM D923: Standard Practice for Sampling Electrical Insulating Oils for

Gas Analysis

IEC 60475: Method of Sampling Liquid Dielectrics

IEC 60567: Guide for the Sampling of Gases and Oil from Oil-filled

Electrical Equipment for the Analysis of Free Dissolved Gases

Sampling of Dielectric Liquids by Lance Lewand, Doble Client

Conference, 2003

8/10/2019 Practicing DGA 09072010

28/172



Gas Extraction

Dissolved gases are present in transformer oil atconcentrations from

8/10/2019 Practicing DGA 09072010

29/172



Gas Extraction

METHOD A (partial de-gassing method) Introduce the oil sample into a pre-evacuated vessel

The extracted gases are compressed to atmospheric pressure

and the total volume measured

The gases are then analyzed by gas chromatograph

METHOD B

The oil is sparged with a carrier gas on a stripper column

The gases are then flushed from the stripper column into a gas

chromatograph for analysis

8/10/2019 Practicing DGA 09072010

30/172



Gas Extraction

METHOD C (Headspace Method) Introduce an oil sample in a glass vial with a headspace gas

phase of argon above it.

Some of the gases migrate from the oil to the headspace and

equilibrate according to Henrys law.

Henrys Law - At a constant temperature, the amount of a given gas

dissolved in a given type & volume of liquid is directly proportional to the

partial pressure of that gas in equilibrium with that liquid.

At equilibrium, the headspace is transferred to a sample loop thenintroduced into a gas chromatograph

8/10/2019 Practicing DGA 09072010

31/172



Gas Chromatography

Mobile phase = gas

Stationary phase

= solid adsorbent or liquid

Mobile gas phase flows

through the column

under pressure0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0 50 100 150 200 250

Time (s)

Intensity

1 2 3

Mobile

phase

flow

Stationary phase

3 2 1

Gas Detector

8/10/2019 Practicing DGA 09072010

32/172



8/10/2019 Practicing DGA 09072010

33/172



Gas Solubility Coefficients

Some gases dissolve into the oil more than others

This is known as Solubility In the following graph, Ethane, C2H6dissolves the most and

Hydrogen, H2the least.

Solubility coefficients change over temperature Gases dissolve into oil in different amounts as temperature

changes

Solubility coefficients are required because all Laboratory gasmeasurement systems measure gas-in-gas and these concentrationsare different than the gas-in-oil

Note: Temperature of the Oil during a manual sample is required only for Lab

Moisture-in-Oil calculations

Temperature of the Oil during the gas extraction process is required todetermine the Solubility Coefficient

8/10/2019 Practicing DGA 09072010

34/172




The triangles and squares in the graph represent ASTM solubilitycoefficients at selected temperatures used in the Laboratory DGA process

8/10/2019 Practicing DGA 09072010

35/172




Solubility coefficients are applied to the laboratory measured gas-in-gasvalues to derive the true gas-in-oil concentration in transformer oil.

Example

Method C headspace sampling formula (from ASTM D3612)

CL= CG( K + VG/VL) CL= Gas-in-Oil value, ppm C

G= Gas-in-Gas value, ppm

K = Solubility Coefficient VG = Volume of Oil space, 7 mL* VL = Volume of Gas space, 15 mL*

(*)Typical Laboratory values

Example for Hydrogen, H2

CG= 40 ppm K = 0.0558 VG/VL = 7/15 = 0.467mL CL= 40 (0.0558 + 0.467) CL= 40 (0.5225)

CL= 20.9 ppm is the Gas-in-Oil value

8/10/2019 Practicing DGA 09072010

36/172



Manual DGA Process

Gas Chromatograph

Transformer GasTransformer Oil

Extraction Device

Vacuum (A)

"Stripper Column (B)

Headspace (C)

Oil Sampled From Syringe orBottle

Transformer

Oil Sampled into

Syringe or Bottle

Transportation

Manual Injection

Carrier Gas

Supply(8.2 psig)

A

1

09

8

7

65

4

3

2

1

1.0

mL

1.0mL

Vent

Porapack Column(A)

Molecular SievesColumn (B)

ThermalConductivity

Detector

Channel B

Channel A

Each time a sample

is handled, there is achance that error

may be introducedinto the

measurement

Gas Chromatograph

8/10/2019 Practicing DGA 09072010

37/172



Concerns with Laboratory DGA

Sampling:

If not done according to standard procedures by experienced personnel,the samples may lead to erroneous results.

Transport:

If procedures are not followed properly, transportation may introduce

contaminations.

Exposure to external factors may also affect the oil sample:

If low-quality syringes are used, additional O2 & CO2 may be introduced

or key gases (e.g. H2) may be lost.

8/10/2019 Practicing DGA 09072010

38/172



Good sample versus bad sample

It is sometimes very clear to the laboratory performing theanalysis if it was taken improperly.

The presence of free water or foreign objects such as

insects, pipe sealant, tape or putty are strong

indicators that the drain valve was not adequatelyflushed out prior to sampling.

Materials used for collecting samples;

Galvanic fittings (zinc coated) used in the drain valveassembly such as the drain plug, can create a galvanic

reaction with water, and cause very high levels of

hydrogen to be produced.

G d l b d l

8/10/2019 Practicing DGA 09072010

39/172



Good sample versus bad sample

Sample ports should have Brass Fittings or Stainless

Steel

Tubing - Tygon tubing or other compatible tubing is necessary for

sample collecting

Tubing should only be used once and then discarded as the wallsof the tubing have a memory (can hold gases, water and other

chemicals compounds in the wall of the tubing)

Incompatible tubing such as natural rubber or PVC will

contaminate a sample with unwanted materials.

8/10/2019 Practicing DGA 09072010

40/172



Laboratory handling is a concernthe sample may be contaminated

in the process of being analyzed

Gas Chromatography analysis in

the laboratory is another source

of errors if not done by

experienced personnel and not

following QA/QC procedures.

The accuracy required for laboratory analysis is +/- 15 % on gas

concentrations, but several laboratories are much less accurate than

that.

Concerns with Laboratory DGA

Diff A L b DGA R lt

8/10/2019 Practicing DGA 09072010

41/172



Differences Among Lab DGA Results

Independent studies have shown lab-to-lab DGA differences can belarge, with some results necessarily not accurate

This study consisted of Sixteen (16) samples from the sameTransformer - All samples were collected by the same individual usingthe same process.

All samples were sent to 3 different

laboratorys simultaneously.

Possible causes of differences are; Even with the same sampler

and sample process, contaminationcould have occurred

Transportation may have affectedthe samples

Laboratory accuracies are +15%on average; some are much worse

Methane

0

500

1000

1500

2000

2500

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Lab A Lab B Lab C

Hydrogen

0

500

1000

1500

2000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Lab A Lab B Lab C

8/10/2019 Practicing DGA 09072010

42/172



Questions

What is the current DGA process at your utility? How are samples taken?

What is the frequency of samples?

What method (A, B or C) is used in your lab?

Do you have your own lab or rely on third party labs?

How often is the Lab Equipment Calibrated?

Do you send a known sample to the Lab?

8/10/2019 Practicing DGA 09072010

43/172



8/10/2019 Practicing DGA 09072010

44/172



Standards and Guidelines Governing

8/10/2019 Practicing DGA 09072010

45/172



Standards and Guidelines Governingthe Interpretation of DGA

IEEE Std. C57.104.1991 IEEE Guide for the Interpretationof Gases Generated in Oil Immersed Transformers

IEEE PC57.104 Draft 11d IEEE Guidenot approved

IEC 60599-1999 Mineral Oil Impregnated Electrical

Equipment in Service: Guide to the Interpretation of

Dissolved and Free Gas Analysis.

IEC 60599-1999, Amendment 1, 04/2007

8/10/2019 Practicing DGA 09072010

46/172



8/10/2019 Practicing DGA 09072010

47/172



Gas Sources

Gases in oil always result from the decomposition ofelectrical insulation materials (oil or paper), as a result of

faults or chemical reactions in the equipment

For example: Oil is a molecule of hydrocarbons, i.e., containing hydrogen and

carbon atoms, linked by chemical bonds (C-H, C-C)

8/10/2019 Practicing DGA 09072010

48/172



Gas Formation

Some of these bonds may break and form H*,

CH3*, CH2* and CH* radicals.

8/10/2019 Practicing DGA 09072010

49/172



Gas Formation

All these radicals then recombine to form the fault gasesobserved in oil:

Gas Formation

8/10/2019 Practicing DGA 09072010

50/172



Gas Formation

In addition to these gases, the decomposition of paper

produces CO2, CO and H2O, because of the presence of

oxygen atoms in the molecule of cellulose:

8/10/2019 Practicing DGA 09072010

51/172



Hydrogen H2

Methane CH4

Ethane C2H6

Ethylene C2H4Acetylene C2H2

Carbon monoxide CO

Carbon dioxide CO2

Oxygen O2

Nitrogen N2

The Main Gases Analyzed by DGA

8/10/2019 Practicing DGA 09072010

52/172



Fault Gas Formation

Most of the time, all gases are present in DGA results,

however, some are formed in larger or smaller quantities

depending on the energy content of the fault

Example; Low energy faults such as Corona PartialDischarges in gas bubbles, or low temperature hot spots,

will form mainly Hydrogen, H2and Methane, CH4

8/10/2019 Practicing DGA 09072010

53/172



Fault Gas Formation

Faults of higher temperatures are necessary to form large quantities

of Ethylene, C2H4

Finally, it takes faults with a very high energy content, such as in

electrical arcs, to form large amounts of Acetylene, C2H2

By looking at the relative proportion of gases in the DGA results it is

possible to identify the type of fault occurring in a transformer in

service

8/10/2019 Practicing DGA 09072010

54/172



Gas Formation Patterns

Are related only to the materials used and faults involved

Are the same in all equipment where these materials are

used Sealed or air-breathing power transformers

Reactors Instrument transformers

LTCs

Etc.

8/10/2019 Practicing DGA 09072010

55/172



8/10/2019 Practicing DGA 09072010

56/172



8/10/2019 Practicing DGA 09072010

57/172



8/10/2019 Practicing DGA 09072010

58/172

A BPL GLOBAL COMPANY58 Serveron Corporation 2007. No use without permission.

8/10/2019 Practicing DGA 09072010

59/172


8/10/2019 Practicing DGA 09072010

60/172


8/10/2019 Practicing DGA 09072010

61/172


8/10/2019 Practicing DGA 09072010

62/172


Fault Gas Formations

8/10/2019 Practicing DGA 09072010

63/172



63

(Ref. S. D. Myers 1996)

8/10/2019 Practicing DGA 09072010

64/172



64

8/10/2019 Practicing DGA 09072010

65/172



65

8/10/2019 Practicing DGA 09072010

66/172



66

Fault Types

Partial discharges of the corona-type (PD)

Typical examples: Discharges in gas bubbles or voids trapped in paper A result of poor drying or poor oil-impregnation

Discharges of low energy (D1)

Typical examples: Partial discharges of the sparking-type Inducing carbonized punctures in paper Low-energy arcing, inducing surface tracking of paper and carbon

particles in oil

Ref. IEC 60599-1999

8/10/2019 Practicing DGA 09072010

67/172



67

Fault Types

Discharges of high energy (D2)

Typical Examples High Energy Arcing Flashovers Short Circuit with power follow through

These result in; Extensive damage to paper Large formation of carbon particles in oil Metal Fusion Tripping of the equipment or gas alarms

Ref. IEC 60599-1999

F l T

8/10/2019 Practicing DGA 09072010

68/172



68

Fault Types

Thermal faults of temperatures 200C) Black or carbonized (>300C)

Ref. IEC 60599-1999

F lt T

8/10/2019 Practicing DGA 09072010

69/172



69

Fault Types

Thermal faults of temperatures between 300 and700C (T2)

Typical Examples: Defective contacts

Defective welds Circulating currents

Evidenced by: Carbonization of paper

Formation of carbon particles

Ref. IEC 60599-1999

F lt T

8/10/2019 Practicing DGA 09072010

70/172



70

Fault Types

Thermal faults of temperatures >700C (T3)

Typical Examples: Large circulating currents in tank and core Short circuits in laminations

Evidenced by: extensive formation of carbon particles in oil metal coloration (800C) or metal fusion (>1000C)

Ref. IEC 60599-1999

8/10/2019 Practicing DGA 09072010

71/172



71

Diagnostic Tools for DGA

8/10/2019 Practicing DGA 09072010

72/172



72


Tool

Reference Standard

IEEE C57.

104-1991

IEEE PC57.

104 D11d

IEC 60599-

1999

Dornenburg Ratios

TDCG Procedure

Key Gas Procedure

TCG Procedure

Rogers Ratios

IEC Gas Ratios

Duval Triangle

CO2/CO Ratio

O2/N2 Ratio

C2H2/H2 Ratio


8/10/2019 Practicing DGA 09072010

73/172



73

Diagnostic Tools for DGAProcedures vs. Methods

Procedures are guidelines on what steps are necessary to

continue to examine the potential for a fault or problem

condition.

Methods are results that provide condition assessments

of potential faults directly

K G P d

8/10/2019 Practicing DGA 09072010

74/172



74

Key Gas Procedure

Complements the TDCG procedure in order to identify

faults

Focuses on the main (or key) gas formed

Limitations:

High tendency to return inconclusive results

Key Gas Procedure

8/10/2019 Practicing DGA 09072010

75/172


y

75


Key Gas Examples

8/10/2019 Practicing DGA 09072010

76/172



76

0

20

40

60

80

100

CO H2 CH4 C2H6 C2H4 C2H2

Thermal Oil and Cellulose Fault

0

20

40

60

80

100


Thermal Oil Fault

0

20

40

60

80

100


High Energy Arcing

0

20

40

60

80

100


Low Energy Partial Discharge

Rogers Ratio Method

8/10/2019 Practicing DGA 09072010

77/172



77

Rogers Ratio Method

IEC Gas Ratio Method

8/10/2019 Practicing DGA 09072010

78/172



78

IEC Gas Ratio Method

Summary of the Ratio Methods

8/10/2019 Practicing DGA 09072010

79/172



79

Summary of the Ratio Methods

The Basic, Rogers and Dornenburg methods use thesame 3 basic gas ratios: (CH4/H2, C2H2/C2H4and

C2H6/C2H4).

Depending on the values of these gas ratios, codes or

zones are defined for each type of fault

One drawback of these ratio methods is that no diagnosis

can be given in a significant number of cases, Dead

Zones - fall outside defined zones.

3-D IEC Gas & Rogers Ratio Methods

8/10/2019 Practicing DGA 09072010

80/172



80

g

8/10/2019 Practicing DGA 09072010

81/172



81

Triangle Method

8/10/2019 Practicing DGA 09072010

82/172



82

Triangle Method

The Triangle, developed empirically in the early 1970s,

and is used by the IEC.

Based upon the 3 gases (Methane, CH4, Ethylene, C2H4

and Acetylene, C2H2) corresponding to the increasing

energy levels of gas formation. One advantage of this method is that it always provides a

diagnosis, with a low percentage of wrong diagnoses.

There are no indeterminate diagnostics using the Triangle

method.

Triangle Method

8/10/2019 Practicing DGA 09072010

83/172



83

Triangle Method

The triangle method plots the relative % of the 3 gases on

each side of the triangle, from 0% to 100%.

The 6 main zones of faults are indicated in the triangle,

plus a DT zone (mixture of thermal and electrical faults)

Approximately 200+ inspected cases in service were used

to develop the Triangle

Triangle Method

8/10/2019 Practicing DGA 09072010

84/172



84

Triangle Method

8/10/2019 Practicing DGA 09072010

85/172



85

Triangle Method

How have fault zones been defined in the Triangle?

Answer:

They are based on a large number of cases of faulty

transformers in service which have been inspected visually.

The root cause of the failure was determined and matched to

the DGA data.

The Triangle was tested with all these cases and correctly

identifies the zone that matches the root cause of failure at a

very high percentage.

Triangle Method

8/10/2019 Practicing DGA 09072010

86/172



86

Cases of faults PD and D1

Tracking; Sparking; Small Arcing

Triangle Method

8/10/2019 Practicing DGA 09072010

87/172



87

Cases of faults D2

Triangle Method

8/10/2019 Practicing DGA 09072010

88/172



88

Circulating Currents; Laminations; Bad Contacts

Cases of thermal faults in oil only

Triangle Method

8/10/2019 Practicing DGA 09072010

89/172



89

Brownish Paper; Carbonized Paper; Not Mentioned

Cases of thermal faults in paper

Triangle Method FAQs

8/10/2019 Practicing DGA 09072010

90/172



90

Triangle Method FAQ s

How are corona PDs, which form a lot of Hydrogen, H2, identified in

the Triangle without using this gas?

Answer: In such faults, Methane, CH4 is indeed formed in

smaller amounts than Hydrogen, H2 (typically 10 to 20 times

less), but which can still be measured easily by DGA

0

20

40

60

80

100


Low Energy Partial Discharge

Triangle Method FAQs

8/10/2019 Practicing DGA 09072010

91/172



91

Triangle Method FAQ s

In the Triangle method, why not use Hydrogen, H2 rather

than Methane, CH4 to represent low energy faults? Answer: Because CH4 provides better overall diagnoses for all types of

faults

A possible explanation

Is that H2 diffuses muchmore rapidly than

hydrocarbon gases from

transformer oil.

This will affect gas ratiosusing H2 but not those

using hydrocarbon gases.

8/10/2019 Practicing DGA 09072010

92/172

Triangle Method Example

8/10/2019 Practicing DGA 09072010

93/172



93

Triangle Method Example

Using the Triangle Method

8/10/2019 Practicing DGA 09072010

94/172



94


The calculation of triangular coordinates can easily be done manually,

or with the help of a small algorithm or software

Errors are often made when developing such an algorithm, so check it

first with the free algorithm available. ([email protected])

For those familiar with computer graphics, it is also possible to

develop a software displaying the point and the fault zones

graphically in the triangle

Software from vendors is available for that purpose

An example of the software follows, Courtesy of Serveron

mailto:[email protected]:[email protected]

8/10/2019 Practicing DGA 09072010

95/172



95


Note: This is the same data as shown using Rogers Ratio example


8/10/2019 Practicing DGA 09072010

96/172



96


The Triangle, being a graphical method, allows you to

easily follow the evolution of faults with time

The previous example shows evolution from a thermal

fault to a potentially much more severe fault such as D1

- Fault Severity -

8/10/2019 Practicing DGA 09072010

97/172



97

Fault Severity

The most severe faults: Faults D2 in paper and in oil (high-energy arcing) Faults T2-T3 in paper (>300C) faults D1 in paper (tracking, arcing) faults T3 in oil (>700C)

The less severe faults: Faults PD/ D1 in oil (sparking) Faults T1 in paper (

8/10/2019 Practicing DGA 09072010

98/172



98

Fault Method Comparisons

% CorrectDiagnoses

% UnresolvedDiagnoses

% WrongDiagnoses

IEEE Key Gas

Method 42 0 58

IEEE RogersRatios 62 33 5

Doernenburg

Ratios 71 26 3

IEC Basic GasRatios 77 15 8

IEC Duval

Triangle 96 0 4

% CorrectDiagnoses

% UnresolvedDiagnoses

% WrongDiagnoses

IEEE Key Gas

Method 42 0 58

RogersRatio 62 33 5

Doernenburg

Ratios 71 26 3

IEC GasRatio 77 15 8

IEC Duval

Triangle 96 0 4

CO2/CO Ratio

8/10/2019 Practicing DGA 09072010

99/172



99

CO2/CO Ratio CO2/CO ratio used to detect paper related faults

If 200C

If >10, it indicates a fault of temperature T

8/10/2019 Practicing DGA 09072010

100/172



100

O2/N2: a decrease of this ratio indicatesexcessive heating

Acetylene/ Hydrogen, C2H

2/H

2; A ratio >3 in the main tank indicates contamination by the LTC

compartment

Gassing not related to faults in service

8/10/2019 Practicing DGA 09072010

101/172



101

g

Catalytic reactions on metal surfaces: Formation of H2only

Stray gassing of oil:

The unexpected gassing of oil at relatively lowtemperatures (80 to 200C):

Gassing of the T1 or T2 type

8/10/2019 Practicing DGA 09072010

102/172



102

Any questions about what was just presented?

8/10/2019 Practicing DGA 09072010

103/172



103

8/10/2019 Practicing DGA 09072010

104/172



104

Accuracy and Gas Levels

8/10/2019 Practicing DGA 09072010

105/172



105

y

IEC & CIGRE recommendation

DGA diagnosis should be attempted only if gas concentrations or

gassing rates are high enough to be considered significant

Low gas levels may be due to contamination or aging of

insulation, not necessarily to an actual fault.

Limitations to Accurate Diagnostics

First limit - LabAccuracy & Gas Concentrations

Second limit - Typical Values & Diagnosis

Accuracy and Gas Levels

8/10/2019 Practicing DGA 09072010

106/172



106

First limit: LAB ACCURACY & Gas Concentrations

Accuracy is defined here as the deviation from a prepared known

value

The accuracy of the average CIGRE /IEC lab (including in theUS) is ~15% at medium (routine) gas concentrations (> 10ppm for hydrocarbons)

Lab Accuracy & Gas Concentrations

8/10/2019 Practicing DGA 09072010

107/172



107

(medium gas concentrations of >10 ppm)

Lab Accuracy & Gas Concentrations(low gas concentrations of

8/10/2019 Practicing DGA 09072010

108/172



108

The accuracy of the average laboratory decreases to ~ 30% at 6ppm, and 100% near the lab detection limit (2 ppm)

Lab Accuracy & Diagnostics

8/10/2019 Practicing DGA 09072010

109/172



109

Effect of lab accuracies of 15% (in RED) and 30% (in BLUE) on DGAdiagnosis uncertainty

When an area of uncertainty crosses several fault zones, a reliable diagnosiscannot be given

This is particularly true for lab accuracies > 30%


8/10/2019 Practicing DGA 09072010

110/172



110

Diagnosis uncertainties corresponding to lab inaccuracies of15, 30, 50

and 75 % (at 10, 6, 4 and 3 ppm for the average lab):

This applies not only to the triangle but to all diagnosis methods


8/10/2019 Practicing DGA 09072010

111/172



111

The actual accuracy of your laboratory can be obtained byusing gas-in-oil standards, which are available

commercially for instance from Morgan Schaffer

(Truenorth)

Below 15% of lab accuracy, a calculation of diagnosis

uncertainty should be done, and commercial software are

available for that purpose, for instance from Delta-X

Research (TOA4)

Typical Values and Diagnosis

8/10/2019 Practicing DGA 09072010

112/172



112

Second limit: TYPICAL VALUES

Typical Values are not gas Limit Values

Typical Values are not the same for all Transformer Systems (may

change slightly from system to another).

Typical Values correspond to a given percentile of the populationof DGA results

Typical Values should be calculated for a specific Transformer

System (Utility specific)

Typical Values observed nationwide or worldwide may be used,

but with some caution.

Typical Values and Diagnosis

8/10/2019 Practicing DGA 09072010

113/172



113

Second limit: TYPICAL VALUES

Diagnosis only attempted when gas concentrations levels and

rates of gas increase is above Typical Values.

When gas levels are below Typical Values, gas concentrations

may be too close to detection limits.

Below Typical Values, the IEC recommends to use normal

sampling frequency (monthly, semi-annual, etc.,..) and not to

attempt a diagnosis.

Above Typical Values, the IEC recommends to use increasedsampling frequency (e.g., monthly or weekly) and a DGA

diagnosis may be attempted.

Typical Values and Diagnosis- 90% Typical Values -

8/10/2019 Practicing DGA 09072010

114/172



114

90% Typical Values

90% Typical Values are represented;

Typical Gas Concentration (TGC) levels

Typical Rate of Gas Increase (TRGI) levels.

TGC and TRGI values are the calculated values from a specific

population of transformers in service.

Reference - CIGRE Brochure # 296, Recent developments in DGA

Interpretation, Joint Task Force D1.01/A2.11, June 2006 (available

from [email protected])

Typical Values and Diagnosis- 90% TGC Values -

8/10/2019 Practicing DGA 09072010

115/172



115

C2H2 H2 CH4 C2H4 C2H6 CO CO2

All transformers 50-

150

30-

130

60-

280

20-

90

400-

600

3800-

14000

No OLTC 2-20

Communicating

OLTC

60-280

90% TGC Values -

Typical Gas Concentration (TGC) levels observed by CIGRE inpower transformers, in ppm:

NoOLTCrefers to a transformer not equipped withan OLTC or has a LTC not communicatingwith or

leaking to the Main Tank

Reference; Table 3, pg 5; CIGRE Brochure #296

The data shown here represents 25 Electrical Networks worldwide,including more than 20,000 Transformers

CommunicatingOLTCmeans that some oil and/or gas communication

is possible between the OLTC compartment and the Main Tank of a

Transformer.

Typical Values and Diagnosis- 90% TGC Values -

8/10/2019 Practicing DGA 09072010

116/172



116


All transformers 35-

132

10-

120

32-

146

5-

90

260-

1060

1700-

10,000

No OLTC 0-4

Communicating

OLTC21 - 37

90% TGC Values

Typical Rate of Gas Increase(TRGI) levels observed by CIGRE inpower transformers, in ppm/yr

NoOLTCrefers to a transformer not equipped withan OLTC or has a LTC not communicatingwith or

leaking to the Main Tank

Reference; Table 3, pg 5; CIGRE Brochure #296

The data shown here represents 25 Electrical Networks worldwide, includingmore than 20,000 Transformers

CommunicatingOLTCmeans that some oil and/or gas communication

is possible between the OLTC compartment and the Main Tank of a

Transformer.

Typical Values and Diagnosis- 90% TGC & TRGI Values -

8/10/2019 Practicing DGA 09072010

117/172



117

90% TGC & TRGI Values

The large ranges of TGC & TRGI values are an indication of the

differences in operation practices of individual Utilities.

Values in the preceding tables are from Core-Type XfmrShell type

values are likely to be higher.

Each individual Utility should preferably calculate its own specificTGC & TRGI values.

To calculate the TGC & TRGI values, separate out the different types

or sizes of transformers to be evaluated into specific groups or

classifications.

Each transformer group and classification will have its own TGC &

TRGI ranking or 90% values


8/10/2019 Practicing DGA 09072010

118/172



118

90% typical concentration value

1 10 100 1000 ppm

% 100

90

50

Cumulative curves of DGA results (concentrations orgassing rates)


8/10/2019 Practicing DGA 09072010

119/172



119

yp

Calculate TGC Concentration Values - Example

Hydrogen, H2 Methane, CH4 Ethylene, C2H4 Ethane, C2H6 CarbonMonoxide, CO

CarbonDioxide, CO2

Acetylene

0 0 0 0 0 177.3 0

0 0 0 0 0 258.6 0

0 0 0 0 21.5 487.4 0

1.1 0 0 0 26.3 494.9 0

1.7 0 0 0 29.1 662.3 0

2.3 0 0 0 52.9 1206 0

7.4 0 0 0 56 1263.6 0

7.6 1.5 0 0 121.5 1289.2 0

16.3 6.2 0 1.8 135.1 1748.4 0

18.8 9.4 2.9 2.5 143 2812.4 0

22.8 10.2 3.3 15.9 189.4 3369.8 0

23.6 19.9 5.4 39.5 222.1 4604.5 0

24 30.5 7 49 239.4 4639 0

30.6 49.5 9.9 76.2 282.3 6820.6 0

51.0 52.6 11.7 76.4 296.8 7244.1 0

53.8 63 18.4 76.7 306.9 8381.3 0

90% TGC 87.4 76.4 22.8 118.3 342.1 8426.1 2.1136.5 186.8 55.8 541.2 582.3 9457.3 3.2

138.6 1867.8 5930.9 1078.2 1760.1 19661.7 5.7

90% TGC

(CIGRE)

50 - 150 30 - 130 60 - 280 20 - 90 400 - 600 3800 - 14000 2 - 20


8/10/2019 Practicing DGA 09072010

120/172



120

Calculate TRGI Gassing Rates Values - Example

Hydrogen, H2 Carbon Monoxide, CO Carbon Dioxide, CO2 Acetylene, C2H21st 2nd 1st - 2nd 1st 2nd 1st - 2nd 1st 2nd 1st - 2nd 1st 2nd 1st - 2nd

0 0 0 24.8 26 1.2 0 0 0 0 0 0

22.3 22.6 0.3 47.5 51.5 4 0 0 0 0 0 0

6.7 5.8 0.9 262.2 248.9 13.3 0 0 0 0 0 0

81.7 82.8 1.1 4358.5 4319 39.5 133.9 135.1 1.2 0 0 0

30.1 28.5 1.6 123.7 168 44.3 30.3 32 1.7 0 0 0

28.1 30.1 2 2476.5 2563.5 87 22.8 20.9 1.9 0 0 0

0 2.1 2.1 134.1 223.8 89.7 145.5 142.9 2.6 0 0 0

0 2.4 2.4 138.1 229.3 91.2 291.8 294.9 3.1 0 0 0

3.7 0 3.7 31.1 161.3 130.2 303.4 328 24.6 0 0 0

0 4.8 4.8 2.9 177.6 174.7 1578.9 1615 36.1 0 0 0

11.5 16.3 4.8 316 494.8 178.8 517 573.8 56.8 0 0 0

24 32.4 8.4 9751.9 10095 343.1 304 367 63 2.2 2.3 0.1

15.4 2.8 12.6 607 1011 404 1160.2 1254.7 94.5 2.4 3.1 0.7

4.5 23.2 18.7 872.7 1281.4 408.7 152.8 298.8 146 7.3 6.6 0.7

28.6 50.2 21.6 17666.6 17120.6 546 6379.4 6627.5 248.1 1 2.2 1.2

144.3 172 27.7 4382 5069 687 326.2 668.2 342 1.5 0 1.5

48.3 11.7 37 6338.5 7110.4 772 489.2 1524.9 10367 2.2 0 2.2

140.2 91.8 48.4 6202.9 7296.9 1094 2324.6 5435.9 3111.3 6.4 1.4 5

186.2 134.1 52.1 6642.1 8622.1 1980 34 3426 3392 1.8 10.4 8.6


8/10/2019 Practicing DGA 09072010

121/172



121

yp

Influences on Typical Values:

Transformer Age Values are significantly higher in young equipment (suggesting some

unstable chemical bonds in new oil and paper)

Values are a bit higher in very old equipment

Transformer Type Values are higher in shell-type and shunt reactors (operating at

higher temperatures), lower in instrument transformers

Typical values are very similar in air-breathing and in sealed or

nitrogen blanketed equipment, contrary to a common belief in the

US

Transformer Oil Volume Values are not affected by oil volume (suggesting that larger faults

are formed in larger transformers)

90% Typical Concentration valuesobserved in the US

8/10/2019 Practicing DGA 09072010

122/172


122

H2 CH4 C2H4 C2H6 C2H2 CO CO2 TDCG

Condition 1 IEEE-2008 100 120 50 65 2 350 2500 720

Weidmann 96 88 57 79 3 613 5990 1058

APS 80 45 70 30 2 950 1250

GE 80 50 73 28 2 950 1183

IEC/CIGRE 100 80 170 55 3 500 8900 908


8/10/2019 Practicing DGA 09072010

123/172


123


123

yp

When DGA results in service reach typical values:

A diagnosis may be attempted to identify the fault (if

lab accuracy is good enough)

The equipment should not be considered at risk

However, it should be monitored more frequently by

DGA

8/10/2019 Practicing DGA 09072010

124/172


124


124

What are your current typical (normal) values for the

different gases?

8/10/2019 Practicing DGA 09072010

125/172


125


125

Gas Levels above Typical ValuesRisk of Failure in Service

8/10/2019 Practicing DGA 09072010

126/172


126


126

Probability of Failure in Service (PFS); Probability of having a failure related event in service, such as any

tripping, fault gas alarm, fire, etc. event

PFS has been defined as;

PFS = Number of DGA analyses followed by a failure

related event/Total number of Analyses@ each gas

concentration value

Gas Levels above Typical ValuesRisk of Failure in Service

8/10/2019 Practicing DGA 09072010

127/172


127


127

90 98 99 Norm, in %

ppm

PFS, in %

100 300 400

Probability of having a failure-related event (PFS, %) vs. the

concentration of C2H2 in ppm at HQ

Pre -FailureValue

200

Gas Levels above Typical ValuesPre-Failure Values

8/10/2019 Practicing DGA 09072010

128/172


128


128

The PFS remains almost constant below and above the

90% typical value, until it reaches an inflexion point on thecurve (Pre-Failure Value)

DGA monitoring should be done more and more

frequently as gas concentrations increase from typical toPre-Failure Value

Time to reach the Pre-Failure Value is unknown, could be

Hours, Days, Weeks or Months.

The IEC/CIGRE Approach(Gas Concentrations)

8/10/2019 Practicing DGA 09072010

129/172


129

Probability of having a failure-related event in service (PFS)

in %, vs. the concentration of all gases in ppm. T = 90%

typical value; P = pre-failure value

The IEC/CIGRE Approach(rates of Gas Increase)

8/10/2019 Practicing DGA 09072010

130/172


130

Probability of having a failure-related event in service

(PFS) in %, vs. the rate of increase of all gases in ppm/ yr.

Gas Levels above Typical ValuesPre-Failure Values

8/10/2019 Practicing DGA 09072010

131/172


131


131

Pre-Failure Value concentrations were found by CIGRE to

be surprisingly close on different networks:

This suggests that failure occurs when a critical amount of

insulation is destroyed

H2

CH4

C2H

4C

2H

6C

2H

2CO

240-

1320

270-

460

700-

990

750-

1800

310-

600

984-

3000

(in ppm)

Gas Levels above Typical ValuesPre-Failure Gassing Rates

8/10/2019 Practicing DGA 09072010

132/172


132


132

Pre-failure rates of gas increase in power transformers at

CIGRE, in ppm/day


0.5 3 5 5 11 NS NS

Note: This is an unpublished chart

Gas Levels above Typical ValuesCaution and Warning Values

8/10/2019 Practicing DGA 09072010

133/172


133


133

In-between typical and pre-failure values, specific caution and

warning values can be defined, depending on the tolerance to risk ofthe maintenance personnel, and on the maintenance budget available

For example, higher values may be used when the maintenance

budget is low, and lower values in the case of strategic equipment

CIGRE caution and warning values are significantly different from

those proposed by IEEE and are available upon request(from

[email protected])

Sample Intervals & Concentration Limits

8/10/2019 Practicing DGA 09072010

134/172


134

Sampling intervals and gas concentration limits in ppm

calculated for an average US power transformer

8/10/2019 Practicing DGA 09072010

135/172

Numerical values of sampling intervals and

rate of gas increase limits in ppm/ year

8/10/2019 Practicing DGA 09072010

136/172


136

calculated for an average CIGRE/ IEC power transformer

Gassing rate H2 CH4 C2H4 C2H6 C2H2 CO CO2 TDCG Sampling intervals

Nuclear Transmission

Typical 83 65 89 47 2 660 5850 946 6 months Yearly

Level 2 151 141 180 132 6 1408 12472 2049 Monthly

179 175 218 176 7 1737 15382 2539 Monthly

Level 3 247 266 319 308 13 2607 23067 3841 Weekly

280 313 369 382 17 3054 27012 4513 Weekly

Level 4 475 620 686 951 42 5940 52492 8896 Daily

509 679 745 1074 47 6491 57351 9738 Daily

Pre-failure 1095 1825 1825 4015 182 17000 150000 26000 Hourly Hourly

Sampling Intervals based on Combined Gas

Rate and Gas Concentration Levels of

8/10/2019 Practicing DGA 09072010

137/172


137

Individual Gases(IEC/CIGRE approach)

Sampling Intervals based on CombinedGas Rate and Concentration Levels

Rate

Level #

Conc.

Level #

Daily Weekly Monthly Quarterly Yearly

4 4 X

4 3 X

4 2 X

4 1 X

3 4 X

3 3 X

3 2 X

3 1 X

2 4 X

2 3 X

2 2 X

2 1 X

1 4 X

1 3 X

1 2 X

1 1 X

TDCG Procedure

8/10/2019 Practicing DGA 09072010

138/172


138


138

Focuses on Total Dissolved Combustible Gas (TDCG) content

TDCG, is the sum of all combustible gases, (H2+CH4+C2H2+

C2H4+C2H6+CO) dissolved in the oil

The table provides guidelines concerning gas levels in

transformers and suggests actions in terms of samplingintervals and operating procedures

IEEE Guide for the Interpretation of GasesIEEE Std C57.104-1991

8/10/2019 Practicing DGA 09072010

139/172


139

IEEE Guide for the Interpretation of GasesIEEE Std C57.104-1991

8/10/2019 Practicing DGA 09072010

140/172


140

8/10/2019 Practicing DGA 09072010

141/172


141


141

What are your current caution/ warning/ alarm values for

the different gases?

Triangle 2 for LTCsof the Conventional Oil-type

(UZB, UBB, URS, UTT, URT, 550, 394, TLH, TC, LR, LRT, etc)

8/10/2019 Practicing DGA 09072010

142/172


142

: Normal operation; : Severe coking; : Light coking;:

Heating; : Strong arcing D2; : Arcing D1

Triangle 2 for LTCsof the ConventionalVacuum Type (LRT, UVT)

8/10/2019 Practicing DGA 09072010

143/172


143

: no inspection made

Triangle 3 for Non-Mineral oils

8/10/2019 Practicing DGA 09072010

144/172


144

FR3 Mineral oil

Triangle 4 for Low-Temperature Faults inTransformers

8/10/2019 Practicing DGA 09072010

145/172


: Corona partial discharges; : Stray gassing at

120C; ; Stray gassing at 200C; ; : Hot spots with

carbonization of paper; ;: Overheating (T < 250C).

Triangle 5 for Low-Temperature Faults inTransformers

8/10/2019 Practicing DGA 09072010

146/172


: Corona partial discharges; : Stray gassing at

120C; ; Stray gassing at 200C; ; : Hot spots with

carbonization of paper; ;: Overheating (T < 250C).

Triangle 6 for Low-Temperature Faults in FR3

Oils

8/10/2019 Practicing DGA 09072010

147/172


stray gassing at T < 150 C;;

stray gassing at T > 200C;;

gas formation in Alliant transformers

8/10/2019 Practicing DGA 09072010

148/172

8/10/2019 Practicing DGA 09072010

149/172



149

How to proceed with DGA results coming from thelaboratory ?

8/10/2019 Practicing DGA 09072010

150/172



150

1. Examine DGA laboratory values for errors or

inconsistencies.

2. Attempt a fault diagnosis when possible

3. Evaluate the severity of the fault

4. Decide on appropriate actions on the equipment

#1 - Examine DGA laboratory values

(

8/10/2019 Practicing DGA 09072010

151/172



151

Eliminate zero values (replace them by the detection

limits of the laboratory, e.g., 2 or 1 ppm)

Compare with previous values in ppm on the same

transformer (DGA history)

Check for inconsistencies which might indicate

contamination during sampling or a laboratory analytical

error. Consider suspect values with caution or discard

them

#2 - Attempt a fault diagnosis

8/10/2019 Practicing DGA 09072010

152/172



152

Compare with in-house typical or normal values of

concentration and gassing rate, or compare with publishedvalues (e.g., from IEEE and CIGRE)

If measured DGA values are above typical/ normal values, a

fault diagnosis may be attempted

If measured values are below routine concentration values (10ppm for hydrocarbons), a fault diagnosis should be attempted

only after calculating the uncertainty on the diagnosis, basedon the accuracy of the laboratory at these low concentrationlevels

#3a - Evaluate the severity of the fault

8/10/2019 Practicing DGA 09072010

153/172



153

The severity of the fault will depend on:

The rate of gas formation

The concentration of gases

How far measured values are from typical values, and how closethey are to pre-failure values

The severity of the fault will also depend on:

The nature of the fault (electrical or thermal)

The location of the fault (paper or oil only)

#3b - Evaluate the severity of the fault

8/10/2019 Practicing DGA 09072010

154/172



154

Gas formation not related to faults (e.g., catalytic reactions of oil withmetals or stray gassing of oil) is of no particular concern

The CO2/CO ratio (incremented) can be used to determine paperinvolvement in the fault

The C2H2/ H2ratio can be used to detect contamination of the maintank by an LTC leak

To fully assess the severity of a fault, it may be useful at this stage toseek the advice of experienced DGA experts

8/10/2019 Practicing DGA 09072010

155/172

8/10/2019 Practicing DGA 09072010

156/172



156

Any questions about what was just presented?

8/10/2019 Practicing DGA 09072010

157/172



157

Benefits of On-Line DGA

Detects both gradual and sudden trends in all gases

8/10/2019 Practicing DGA 09072010

158/172



158

Detects both gradual and sudden trends in all gases

Correlates gassing events with external events such as transformer

load, oil temperature, LTC changes, etc.

Provides historic trail for delayed analysis of gassing events

May provide diagnosis on-line

Advantages of On-Line DGA

Oil i l d t ti ll i l d

8/10/2019 Practicing DGA 09072010

159/172



159

Oil is sampled automatically via closed-

loop GC:

No risk of human intervention

Repeatable sampling technique

No atmospheric exposure

Data is collected up to hourly:

Resulting in faster, more accuratedetermination of trends

All 8 fault gases + moisture are monitored

and correlated with oil temperature and

load

On-Line DGA

8/10/2019 Practicing DGA 09072010

160/172



160

Serveron Transformer Monitor

On-line DGA

8/10/2019 Practicing DGA 09072010

161/172



161

On-line gas monitors

Are best suited for measuring rates of gas increase (trends)

Will detect faults between regular oil samplings

May now also provide on-line diagnosis

8/10/2019 Practicing DGA 09072010

162/172



162

Serveron Story #11

8/10/2019 Practicing DGA 09072010

163/172



163

Three-phase 700 MVA 400 kV GSU transformer

Hints: - Increasing Methane (CH4), Ethane (C2H6), Ethylene (C2H4)

- Sudden increase in Acetylene (C2H2),- Transformer refurbished and re-commissioned 10/27/03

Serveron Story #11 Analysis:

8/10/2019 Practicing DGA 09072010

164/172



164

1. The Duval Triangle shows problem evolving from a

T3 Thermal Fault to D1 Discharge of low energy

2. Intermittent grounding was provided by the fastening bolt causing a transient potential

rise and subsequent discharges occurring between the corona ring and the main tank

ground point.

3. An on-site repair was performed and the transformer

was returned to service.

The Duval Triangle is a DGA tool included in the IEC 60599 Gas Guide.HV lead corona ring

shield connection(note burn mark on corona ring

material & eroded bolt)

Serveron Story #12

8/10/2019 Practicing DGA 09072010

165/172



165

3-phase, 1100 MVA, 345 kV GSU transformer

Serveron Story #12 Analysis:

8/10/2019 Practicing DGA 09072010

166/172



166

The Duval Triangle is a DGA tool included in the IEC

60599 Gas Guide.

Rogers Ratios are included in IEEE 57.104 Gas

Guide (similar to Basic Gas Ratios in IEC-60599)

1. Both the Duval Triangle and Rogers Ratio analysis shows the fault condition is in T2

indicating a thermal problem getting worse in the range of 300

C to 700

C

2. Combustible gas levels were rising very quickly, exceeding preset rate of change limits.

Transformer load reduction began approximately 32 hours after levels began to change

and was fully de-energized within approximately 52 hours

3. Root cause of the problem was insulation design issues around HV and LV leads

Serveron Story #12

8/10/2019 Practicing DGA 09072010

167/172



Serveron Story #12

8/10/2019 Practicing DGA 09072010

168/172



How to proceed with DGA results coming from the on-line gas monitor

No need to check for errors or inconsistencies in DGA

8/10/2019 Practicing DGA 09072010

169/172



169

No need to check for errors or inconsistencies in DGA

values Diagnosis may be available on-line using the main

diagnosis methods

What should still be evaluated is the reliability of the

diagnosis, depending on gas level, and the severity of thefault as in the case of manual DGA

Also, decide on appropriate actions on the equipment asin the case of manual DGA, once the diagnosis is

confirmed

Your Laboratory DGA

E i ti f DGA d t f l t d t f

8/10/2019 Practicing DGA 09072010

170/172



170

Examination of your DGA data from selected transformers

to get a diagnosis

The TM series monitors

8/10/2019 Practicing DGA 09072010

171/172


Transformer Mounted

8/10/2019 Practicing DGA 09072010

172/172

practicing dga 09072010

Documents