Condition Monitoring of Power Condition Monitoring of Power Transformers using Dissolved Gas Transformers using Dissolved Gas

( G )( G )Analysis (DGA)Analysis (DGA)

Ahmad Shahsiah Ph D P EAhmad Shahsiah, Ph.D., P.E.Exponent Failure Analysis Associates

March 16, 2009

00M23SF.000 0000 0309 AS00

OverviewOverview

n DGA - Why Do We Care?n DGA Why Do We Care?n DGA Background n Potential Source of Errorn Potential Source of Error n Previous Works

Thn Theoryn Simulations and Resultsn Experimentsn Summary and Conclusions

2

DGA DGA -- Why Do We Care?Why Do We Care?

n Incipient faults could exist in power transformerstransformers

n Early detection of the faults could significantly reduce the cost of repair and loss of servicep

n Industry standards do not always give the right answer

n Understanding the dynamic nature of evolution and migration of characteristic gases improves our interpretation of DGA analysis

3

DGA BackgroundDGA Background-- History History fn Breife History

0Late nineteenth century: Mineral oils have been used for electrical insulation and cooling purposes01928: Attempts to diagnose the type of failure from

l d f il i dgases evolved from oil-immersed power transformers started by Buchholtz1970: Dorneneburg differentiated faults of thermal01970: Dorneneburg differentiated faults of thermal or electrical origin (Brown-Boveri Review publication)publication)01973: Halstead made a thermodynamic

assessment the formation of gaseous hydrocarbonsassessment the formation of gaseous hydrocarbons in Mineral Oil

4

DGA BackgroundDGA Background-- History History

n Briefe History- continued 1975: Rogers suggested the ratio method01975: Rogers suggested the ratio method along with a coding system to recognize the type of the faultthe type of the fault 01977: Modified Rogers method also

included in IEC document 10A 5301980’s and early 1990’s: attempts to

incorporate expert systems in fault p p ydiagnosis based on DGA01999-03: IEC 60599 latest edition.

5

DGA BackgroundDGA Background-- Chemical ReactionsChemical Reactionsn Decomposition of Oil0Hydrogen, Methane, Ethane, Ethylene, Acetylene0Mechanism of formation:

Scission of C-H bonds: low energy faults such as partial di h f ld l (C )discharge of cold plasma (Corona)Scission of C-C, C=C and C≡C bonds and recombination: more and more energy or higherrecombination: more and more energy or higher temperature needed.

0Hydrogen can be generated as a result of variety of reactions

Large quantities of hydrogen have been reported in t f th t h d b i dsome transformers that had never been energized

6

DGA BackgroundDGA Background-- Chemical ReactionsChemical Reactions

n Decomposition of Cellulose Insulation Mostly Carbon Monoxide and Carbon Dioxide0Mostly Carbon Monoxide and Carbon Dioxide

0Minor amounts of hydrocarbons and Furanic compoundscompounds0Mechanism of formation:

C-O bonds in Cellulose polymer chains are thermallyC-O bonds in Cellulose polymer chains are thermally less stable than hydrocarbon bonds in oilScission of cellulose polymer chains higher than 105°C Complete decomposition and carbonization above 300°C300 C

7

DGA BackgroundDGA Background-- IEC 60599IEC 60599n Electrical discharge - visually detectable faults

defined by IEC 605990Partial discharge (PD): Possible X-wax deposition,

pin holes or carbonized perforation in paper0Discharge of low energy (D1): Large carbonized

perforation through paper, carbonization of paper f b ti l i ilsurface, or carbon particles in oil

0Discharge of high energy (D2): Extensive destruction and carbonization of paper metaldestruction and carbonization of paper, metal fusion at the discharge location, extensive carbonization in oil in some cases tripping ofcarbonization in oil, in some cases tripping of equipment

8

DGA BackgroundDGA Background-- IEC 60599IEC 60599n Thermal faults - visually detectable faults

defined by IEC 60599y0Thermal faults, in oil or/and in paper, below 300°C

if the paper has turned brownish (T1) and above 300°C if the paper has carbonized (T2)0Thermal faults of temperature above 700°C (T3) if

there is strong evidence of carbonization of the oil, metal coloration(800°C) or metal fusion (>1000°C)

n Suggested ratios: 0C2H2/C2H4, CH4/H2, C2H4/C2H6

9

DGA BackgroundDGA Background-- IEC 60599IEC 60599

Range of 90% typical concentration values observed in power transformers (all types) (ppm in volume)

TX type H2 CO CO2 CH4 C2H6 C2H4 C2H2ypNo LTC 60-150 540-900 5 100-13 000 40-110 50-90 60-280 3-50

Communicating LTC

75-150 400-850 5 300-12 000 35-130 50-70 110-250 80-270

Incremental ratios of CO2/CO less than 3 are generally considered as an indication of probable paper involvement in a fault

10

Potential Source of ErrorPotential Source of Error

n Gases Fluctuate in Healthy Transformers.n Gases Migrate Between the Oil and Paper.n This Could Cause Problems when Interpreting

DGA R ltDGA Results.n No Published Information about the Phenomenon

T D tTo Date.n A Physical Model can Help with Interpretation of

DGA Analysis Data Especially in the Case of CODGA Analysis Data Especially in the Case of CO2and CO.

11

Review of Previous WorksReview of Previous Works

n IEC 60599 and ANSI/IEEE C57.104.n Detroit Edison Company Studied the Gasn Detroit Edison Company Studied the Gas

Migration in HPFF Cables.Kh d Mi t St di d th Mi tin Khan and Miyamoto Studied the Migration of CO2 and CO.

12

Literature DataLiterature Data

n Diffusivity Data by Detroit Edison Co.n ASTM D2779 - Gas Solubility in the Oil.yn CO2 in Paper Pulp.n Oommen’s Methodn Oommen s Method.

13

D i M d lD i M d lDynamic ModelDynamic Model)ln(2 TDccvcD

tc

∇∇+∇⋅−∇=∂∂ αDiffusion Equation:

Z

Vz

Oil Semi

Approximations:• Soret Effect is Ignored.

t∂

X

Oil Semi infinite paper

• Diffusion in the x Direction.

• Velocity in the z Direction.

I i l Oil D t

zcv

xcD

tc

z ∂∂

−∂∂

=∂∂

2

2

In a single Oil Duct:

Transport in Oil is Faster than Diffusion:

2cDc ∂=

∂2x

Dt ∂=

∂

SimulationSimulation ResultsResultsSimulation Simulation Results Results (animation)(animation)

COCO22 Equilibrium CurvesEquilibrium Curves

n ASTM D2779 - Gas Solubility in the Oil.yn CO2 in Paper Pulp by D.J. Salley.n Oommen’s Methodn Oommen s Method.

ASTM D2779 ASTM D2779 –– Solubility of Solubility of Selected GasesSelected Gases inin the Oilthe OilSelected Gases Selected Gases in in the Oilthe Oil

1000000

1200000

1400000

1600000

vol.

200000

400000

600000

800000

ppm

by

v

CO2

180000

0273 288 311 339 355

T (K)

80000100000120000140000160000

by v

ol. Hydrogen

NitrogenCO

020000400006000080000

ppm

COOxygen

0273 288 311 339 355

T (K)

COCO22 Equilibrium CurvesEquilibrium Curves

40000

50000

m b

y vo

l.)

20000

30000

on in

pap

er (p

pm 0 C15 C27 C38 C56 C

0

10000

Con

cent

ratio

56 C80 C

00 50000 100000 150000 200000 250000

Concentration in oil (ppm by vol.)

COCO22 Equilibrium Curves at Lower Equilibrium Curves at Lower ConcentrationsConcentrationsConcentrationsConcentrations

Model ParametersModel Parameters

Headspace sampling

Reservoirs

1000 ml 1000 ml

ports

Inert gas

2000 ml 2000 ml

Relaxation chambers

Oil sampling ports

Precision pump

PressboardMain container

1000 ml Heater tape

20

Model ValidationModel Validation

1000 l

Headspace sampling port

Reservoir

Inert

ml

Relaxation chamber

Pressboard

gas2000 ml

chamber

Oil flow

Oil sampling port

Precision pump

1000 ml

Heater tape

Main container

21

Main container

Experimental DataExperimental Data

1000

10000

m b

y vo

l.)

CO CO2

100

Con

cent

ratio

n (p

pm

Introducing gas and temp. step up (25-70C)

Temp. step down (70C to 25C)

Temp. step up (25C to 70C)

Temp. step down (70C to 25C)

Temp. step up (25C to 70C)

100.0 200.0 400.0 600.0 800.0 1000.0 1200.0 1400.0 1600.0

Time (hours)

10000 C2H4 C2H6 C2H2

1000

ation (p

pm by vo

l.)

Introducing gas and temp. step up (25-70C)

Temp. step down (70C to 25C)

Temp. step up (25C to 70C)

Temp. step down (70C - 25C) Temp. step up

(25C to 70C)

1000.0 200.0 400.0 600.0 800.0 1000.0 1200.0 1400.0 1600.0

Con

centra

22

0.0 200.0 400.0 600.0 800.0 1000.0 1200.0 1400.0 1600.0

Time (hours)

Experimental Comparison Experimental Comparison –– COCO22

23

Summary and ConclusionsSummary and Conclusionsn Concentrations of CO and CO in Transformer Oil cann Concentrations of CO2 and CO in Transformer Oil can

Change Considerably due to a Temperature Change Because of the Migration Process.

Thi L d t Si ifi t E h I t ti DGA0This can Lead to Significant Errors when Interpreting DGA Analysis Data

n Migration Phenomenon of Characteristic Gases can be E l i d b Diff i PExplained by Diffusion Process.

n Application of the Model Works Best in the Case of Carbon Dioxide and Carbon Monoxide.

n Diffusion Time Constant and Steady-State Ratio of Gases in an Oil-Paper System have been determined.

n Steady state Ratio of CO in Oil / Paper System wasn Steady-state Ratio of CO2 in Oil / Paper System was obtained using Oommen’s Method Similar to the Case of Moisture.

24

nThis Project Received the Endorsement of DEIS Liquids Technical Committee in 2004.

nThis Project Received the DEIS Fellowship jAward in 2005.

nSix Technical Papers, a Thesis and a BookSix Technical Papers, a Thesis and a Book have been Published based on the Methodology and Results of this ResearchMethodology and Results of this Research

Questions?Questions?

25