Diagnostic Techniques of Power

Transformers

ISEI 2010 –Short Course San Diego – June 6th 2010

Ali Naderian, P.Eng., PhD.Kinectrics Inc.

Toronto, Canadaali.naderian@kinectrics.com

+1-416-797-3724

1

Introduction• Myself:

ITC ( Siemens Power Transformer) 1998-2000 Designed a 3-step cascade Testing Transformer

(2000) Rebuilding High voltage Lab of the University of

Waterloo (2005) Kinectrics (Ontario Hydro Research) since 2007

working at the HV Lab, as well as Testing on Transformers/Cables

2

Objectives

• Identify the diagnostic and condition assessment methods

• Explain the selected conventional testing methods

• Identify innovative techniques of transformer diagnostics

• Interpret the test results of case studies3

Contents• Routine (conventional)

Winding/Insulation Diagnostics Tests:Polarization Index (insulation resistance)Turn ratioDC resistanceShort circuit ImpedanceExcitation CurrentPower Factor, capacitance

• Oil Tests:DGAOil quality tests (Dielectric, Power factor, …)

4

Contents• Advanced Condition Assessment

Techniques:

Partial Discharge (PD) :Electrical/Acoustic

Dielectric Frequency Response (DFR)

Frequency Response Analysis (FRA)

5

Polarization Index

• Propose: Tests : Winding Insulation + Oil Overall integrity of the winding insulation Verify that the state of dryness of insulation

• Definition:Polarization Index : PI=R10-min /R1-min

Trend Measure of dielectric deterioration

Absorption ratio : AR=R60s /R15s

• Test Method: H-LG, L-HG, HL-G Utest=5 kV-DCTime= 15 s, 60 s, 10 min

6

7

Insulation Resistance Model

iL iC

ReqRA is not a fix resistance:

8

Test Arrangement

kV 20 °C 30 °C 40 °C 50 °C 60 °C

6.6 400 200 100 50 25

6.6-19 800 400 200 100 50

22-45 1000 500 250 125 65

≥ 66 1200 600 300 100 75

Typical Insulation Resistance MΩ

When to use Guard?

9

Interpretation• R Needs temp correction

k=1.5 for oil-filled transformer

k=30 for untanked or dry-type transformers

• 1.3≤AR≤3.0: Dry Transformer

• PI (Large Power transformers)

1- Make sure transformer is grounded before and after test.

2- The energy stored must be discharged safely by short-circuiting at least x4 of the test period.

Winding

Winding

KVA

kUR 60

PI ConditionLess than 1 Dangerous

1.0 - 1.1 Poor (wet or poor dielectric)

1.1 - 1.25 Questionable1.25 - 2.0 AcceptableAbove 2.0 Very Good

Turn Ratio (TTR)

• r= Np/Ns=Ep/Es

• Deviation indicates problem in either of windings High-resistance connections in the lead circuitry or

high contact resistance in tap changers, open circuits Low resistance: shorted turn-to-turn

• Minimum accuracy : 0.1%

10

IEEE 62, IEEE C57.12.00 IEC 60076-1

+/-0.5% nameplate ratio The lesser of +/-0.5% of declared voltage ratio or 0.1*Uk%

Test Arrangement

11

• Auxiliary Transformer 230kV/13.8kV

12

Test Case

Nothing wrong with transformer!Tap changer index is off by one position. Position 1 is 2, position 2 is 3, …, position 5 is 1.

DC Resistance

13

• Measure of winding resistance • For temp-rise=55°C corrected to 75°C• For temp-rise=65°C corrected to 85°C• Before test:

(IEEE)different between top and bottom temp ≤5°C

(IEC) 3hours rest time

• Test current ≤10% of rated current• DC resistance should be ≤ 2% factory

test

DC ResistanceCan detect:• Shorted-turns• Loose connection on bushing• Loose connections or high-contact

resistance on tap changers• Broken winding strands• The above issues leads to hot-

spots, generates gases DGA

14

Short circuit impedance• Applications:

Investigate winding deformation confirm the name plate values To check shipping as a receiving

(pre- commissioning test) • If possible run three-phase if not

feasible it can be done single-phase , then average the results

15

Short circuit impedance

16

Apply voltage and measure current in HV side while shorting out the dual leg of the LV side.

Applied voltage is 100 V-500 V Watch the current in LV side LTC is in neutral position

3-phase Single phase

Short circuit impedance• If deviation exists it could be due to:

Type of excitation (1-phase versus 3-phase) Different instrumentation Winding deformation

17

• Deviations of over ± 3% from the benchmark Per Phase Tests could be related to winding deformations.

• Even 3-phase test result may be different from nameplate because of the Vtest.

Excitation Current • Can possibly detect: Core problems such as:

Shorted core laminations poor joints

Winding problems such as: short circuit (turn to turn) open circuit poor connections

LTC problems such as: high hesitance connection open circuitCoking and wear of LTC and DETC contacts

18

Excitation Current Test

• Factory tested at rated voltage (no-load)

• If possible run three-phase if not ,can be done single-phase.

• Perform Excitation Test before any DC test: DC test leaves residual magnetism in the core.

• Voltage is applied to HV :5kV or 10 kV• LTC set to: 1- Neutral , 2- 1 step up, 3- 1 step down

4- Full raise , 5- Full lower

19

Excitation Current Setup

20

Excitation Current Pattern

21

H1 H2 H3

ExcitationH1-H0

Similar to H3-H0

H2-H0

Higher R results higher

excitation

current

Excitation Current Results

22

Possible patterns

Characteristics of

H-L-H • 3 leg core-type (Most common)• 5 leg core or shell type with a delta secondary

L-H-L • 3 leg core-type ,Y secondary, inaccessible neutral

All equal

• 5 leg with Y secondary• 3 single phase connected as 3- phase• Shell-type with Y-connected secondary

All different

• Can be due to magnetized core• Defects, Faults

Excitation current

Criteria to detect defect

Iext<50 mA The difference between 2 higher current >10%

Iext >50 mA The difference between 2 higher current >5%

Excitation Current with LTC

23

When the Autotransformer is in the bridging position the excitation current goes up.

I

IC IR

Power Factor TestingDissipation Factor or is

the ratio of the resistive current to capacitive current

24

Dissipation factor : tan=IR/IC

I=IC+IR

Power Factor : P.F.= IR/I

To check the condition of the capacitive insulation: Between windingsBetween winding and coreBetween winding and tank

Capacitances

2-Winding Transformer

1 phase of 3 phases shown

25

• CH: HV bushing + HV winding +Oil• CL: LV bushing + LV winding + Oil• CHL: Both windings+ barriers + Oil

Test Arrangement3 test modes: Ungrounded Specimen Test (UST)

Grounded Specimen Test (GST)

Grounded Specimen with Guard (GST-g)

26

Test Arrangement

One of the most common source of measurement error: Neutral is not properly connected.

27

Test Mode Energize Ground Guard Winding

1 UST HV - - CHL

2 UST LV - - CHL

3 GST HV LV - CH+CHL

4 GST-g HV - LV CH

5 GST LV HV - CL+CHL

6 GST-g LV - HV CL

Usually 6 tests are performed to confirm the values:

P.F. Interpretation Power Factor Insulation ConditionAbove 1.0% Dangerous

wet transformer0.7 – 1.0 Investigate0.5 – 0.7 Deteriorated

Less than 0.5 Good

28

3 winding transformer

Shielding between LV/HV

IEEE Std 62: Service aged transformers : P.F.<2%

Oil Test: DGA

• Rogers Ratio• Doernenburg Ratio• Duval Triangle • IEC1 - (1st Edition 1978)• IEC2 – (2nd Edition 1999)• ANSI/IEEE (C57.104-1991)•CIGRE Method• Laborelec• Japanese Method• Russian Method

29

Key Gas Method

30

Key Gas

Secondary Gas Fault Pattern Possible Root cause

H2

CH4 and minor C2H6

and C2H4

Low energy Partial

Discharge

Aging of insulation, possible carbon particles in oil, poor grounding of metal objects, loosed lead, floating metal or contamination

C2H4

(ethylene)

CH4 and minor H2 and C2H6(ethane)

Oil overheatingPaper insulation destroyed. Metal discoloration. Oil heavily carbonized.

C2H2

H2 and minor CH4 and C2H4

High Energy Arcing

Poor contacts in leads, weakened insulation from aging, carbonized oil.

CO, CO2

If the fault involves and oil-impregnated

structure CH4 and C2H4

Conductor Overheating

Overloading or cooling problem, bad connection in leads, stray magnetic flux, discoloration of paper.

Duval Triangle

31

%C2H2=100x/(x+y+z);

%C2H4 = 100y/(x+y+z);

%CH4 = 100z/(x+y+z),

T1: Low-range thermal fault (below 300 C)T2 :Medium-range thermal fault (300-700 C) T3 :High-range thermal fault (above 700 C) D1: Low-energy electrical dischargeD2 :High-energy electrical dischargeDT: Indeterminate - thermal fault or electrical discharge.

x = (C2H2); y = (C2H4); z = (CH4), in ppm

IEEE Std C57.104

32

• Four-condition DGA guide to classify risks to transformers with no previous problems.

• Uses combinations of individual gases and total dissolved combustible gas concentration (TDCG).

The three-ratio version of Rogers

ratio Method uses the

following ratios: R1= C2H2/ C2H4 ,

R2= CH4/H2 , R3= C2H4/ C2H6

R2<0.1 0.1<R1<1 R3<1

1<R3<3

R1>1

R1<0.1 R3<1

1<R2<3

Gas Inputs R1,R2, R3

1<R3<3

R3>3

R3>30.1<R1<1

Case 0No fault or it cannot be detected

Case 3Low Temp Thermal Overloading

Case4Thermal <700 oC

Case5Thermal >700 oC

Case1PD low energy

Case2High energy arcing

Y

Y YY

Y Y

Y

Y

Y

Y

Y

Y

YN

N N

N

N

Rogers Fault Tree

IEEE Std C57.104

33

• Not all techniques were applicable in all cases

•CIGRE: CO2/CO <3 Excessive paper degradation

• If most of Gas levels < 1 and one or two is high, the error of regular methods might be high. Call your judgment, do not forget Key Gas Table.

Status H2 CH4 C2H2 C2H4 C2H6 CO CO2 TDCGCondition 1

100 120 35 50 65 350 2500 720

Condition 2

101-700 121-400 36-50 51-100 66-100 351-570

2500-4000

721-1920

Condition 3

701-1800

401-1000 51-80 101-200 101-150

571-1400

4001-10000

1921-4630

Condition 4

>1800 >1000 >80 >200 >150 >1400 >10000 >4630

Case 1:750 MVA,500kV

34

Date H2 CH4 C2H2 C2H4 C2H6 CO2 CO TCGppm/day

April 2010

10 5 2 6 2 1398 277 47,800

May 2009

10 5 2 2 2 155 71 0.13

Key Gas IEC IEEE Roger’s ratio

IEEE Duval Method

Doernenb-urg Ratio

Suggestion

Conductor Overheating

Partial discharge or power discharge

Not conclusive

TCG levels indicate excessive decomposition

D1:Disch--arge of low energy

No diagnosis

DGA every day, On-line PD test (electric follow up with acoustic) , temporary removal from service for off-line tests

Case 2:300 MVA, 420 kV

35

• The content of hydrogen detected by the on-line monitoring system increased continuously.• Concentration of hydrogen is only dependent on the oil temperature.

• Oil used was uninhibited oil considered as “moderately stray gassing”• Avoid misleading of DGA pattern (Stray Gassing) caused by catalytic effects of zinc surfaces and oil .

Oil Quality TestsRecommended oil quality tests for insulating oil

[IEEE Std C57.106-2006]

36

Test Dielectric Strength

Dissipation Factor

Interfacial tension (IFT)

Neutralization number (acidity)

PCB

Standard ASTM D1816 -97 (1 mm gap)ASTM 877[min]

ASTM D924-99 @ 25 0C[max]

ASTM D-971-91[min]

ASTM D974-92[max]

ASTM 4059-91[max]

Limit (new oil)

1816:23kV877: 26 kV

0.1% 35 mN/m 0.03 mg KOH/g

2

Limit (service aged oil)

1mm gap:23kV

0.5% 24 mN/m 0.2 mg KOH/g

50

Post Failure Test Interpretation

37

Suggested Problem Category

Test Data1st priority 2nd priority 3rd priority

Turn-to-turn fault Out of tolerance ratio

Low winding resistance

Excitation increase

Damage to major insulation

High power factor Low insulation resistance

Abnormal DGA

Lead and terminal issues

Abnormal DGA, trend inclines

DC resistance Excitation increases if load

increasesThrough fault mechanical

damage

FRA shows different pattern for faulty phase

Deviation of exciting current

Change in impulse

Core heating Abnormal DGA Low core ground resistance

Excitation increases

Moisture High insulation power factor

Low dielectric oil test

Low insulation resistance

Time for Break!

38

Selected Advanced Techniques

• Partial Discharge

• Dielectric Frequency Response (or Frequency Domain Spectroscopy)

• Frequency Response Analysis

39

Partial Discharge

40

Draft IEEE PC57.113™

IEEE Std C57.127

Partial Discharge-IEC 60270

41

• PD should be done along with induced voltage test• Background noise < 100 pC

max (pC)=100 at 110%Urated max (pC)=300 at 130%Urated max (pC)=500 at 150%Urated

Wide-Band PD measurement Narrow-band PD measurement

Partial Discharge Measurement

42

• Signal from coupling capacitor or bushing tap

coupling device in series with the coupling capacitor frequently used circuit in test laboratories

Via Bushing Tap often applied in on-site offline /online PD investigations

Bushing Tap Sensors

43

To get down the voltage to U, it is necessary to use capacitance (Cz)

Possible power that can be taken from the test tap:

To avoid bushing damage:

Partial Discharge Patterns

44

Corona at HV electrode Corona at Ground

Floated metal object Creeping discharges

1 Cycle screenshot 1 Cycle screenshot

1 Cycle screenshot PD Resolved Pattern

UHF Measurement Via Oil Valve

45

• PD-signals :UHF frequency range (300 MHz – 1 GHz)

• Sensor application at oil valves, which are available e.g. for oil filling or draining.

Case1 Case2

Courtesy of LDS GmbH (Lemke)

Drawback: High attenuation (high frequency); cannot determine which phase?; Hard to use for ball valves.

How to distinguish PD?

46

• External noises appear often independent from the applied AC test voltage level.

• Pulse-shaped noises may appear unsynchronized with the applied AC test voltage, whereas PD pulses occur always phase-correlated.

• Rise-time of PD is different from noise signals are different from pulse.

• Pattern recognition techniques help a lot!

T-F Classification Map

47

Acoustic PD :Off-line/On-line

48

1. All-Acoustic (minimum 3 sensors)2. Acoustic with Electrical PD trigger

Velocity of sound in oil :1413 m/s at 20 °C

Acoustic PD

49

• Vibration noises (core,fan,pump) <50kHZ• Band Pass filter: f1=50 kHz, f2=350 kHz• Propagation path: direct , indirect

Sensor locations when phase of PD source is not known

Acoustic PD

50

• Before placing the AE transducer, wipe the area (dirt, oil, bugs,…).

• Acoustic couplant needed for enhancing the mechanical and acoustical coupling between the transducer and the tank surface.

• A sound transmitting epoxy to be used if the mounting location is non-magnetic.

• Magnetic tank shielding causes extra signal attenuation.

• LTC operation contains a high electromechanical energy that usually propagates through the entire transformer. To be distinguished in post-analysis.

• Initial place to start: one sensor in the bottom connection of each bushing.

Application of Acoustic PD

51

• When electrical PD is detected, for confirmation and source location

• When DGA indicates the possible presence of PD

• For PD detection during factory impulse testing

• When static electrification is suspected

Acoustic PD

52

• Electrical PD has a threshold 300 (500)pC. There is no similar threshold for acoustic systems.

• A strong signal buried deep within a winding may be very weak by the time it reaches the acoustic sensor.

• Sometimes longer monitoring period is necessary: weeks

• Only high PD levels can be detected.• The correlation is weak between measured and real

PD level due to attenuation.

Wall propagationDirect signal

Test Case:

53

Single phase autotransformer: 500/230/13.8 kV,146/194/243 MVA

• After two years in operation started gassing.

• Acoustic monitoring for 5 days :significant activity

Load is minimumVoltage is maximum

Test Case:3-D Plot

54

Dielectric Frequency Response

55

• “Dielectric Frequency Response” also called “Frequency Domain Spectroscopy”

• DFR is Power factor measured in a wide range of frequency (mHZ-kHZ) unlike conventional 60Hz measurement.

Power Factor @60Hz DFR

How to Measure DFR

• Measurement modes: UST, GST, GTS-g

56

Application

• Originally to Estimate the Moisture Content

• Now extend to find out other problems : Insulation Contamination Degradation of overall insulation system High resistance contacts Bushings issues ….

57

DFR to Estimate Moisture

58Using DFR curve, the Wt% can be estimated.

How DFR Estimates Moisture?• X-Y Model of Insulation

59)(

1

)()(

1

)(

1

boardoilspacercomb CCCC

How DFR Estimates Moisture? Oil

– Non-polar liquid; r = 2.2 Pressboard, oil impregnated cellulose

– More polar; r = 4.5

– Sensitive to moisture

60

Why moisture is important?

61

Conventional Moisture Estimation

• Low accuracy at low temperatures because the water has migrated to the paper

62

MOISTURE EQUILIBRIUM BETWEEN OIL AND CELLULOSE

Uncertain area

Conventional Method Cons

• Transformer needs to reach Thermal Equilibrium

• Oil Moisture changes with temperature

• Not accurate in lower temperatures

• Aged oil resolved higher Water% than New oil

63

DFR found to be the most accurate method to estimate the insulation moisture.

DFR & Moisture %

64

Test Case :500 MVA 230kV GSU,1976

• CH and CL increased 100%!

• DGA had gone up in the past but was constant lately.

• The unit was taken out.

• One reason for increasing P.F. is moisture. Should the unit be dry-out?

• Just shipping to the closest maintenance cost $500K.

• Kinectrics offered DFR to diagnose.65

Measurement Year 2000 Year 2009CHL 0.16 0.2CH 0.29 0.59CL 0.48 0.99

Conventional Power Factor :

Test Case

• CHL is normal showing Water content 0.5-1%

• CL does not match with any moisture model curve, So it is not moisture.

• The whole curves shifted up. Insulation conductivity has changed.

• Conclusion: Contamination, no dry-out needed. Shipped out to replace the windings.

66

After opening tank: carbon deposit found on LV and HV winding due to

arcing of tank shipping bolt to the core.

DFR Result

CHCL

CHL

DFR Is Able to:Estimate the moisture content of power

transformers accurately. Identify unsatisfactory conditions during routine

testing. Detect contamination on insulating system. Characterize transformers to avoid potential

catastrophic failures.

67

Drawbacks:

Needs basic transformer design data Needs expertise to interpret results

FRA (SFRA)

Two methods:- Applying LV Impulse (Obsolete) :Time domain- AC , Sweep Frequency (SFRA)

68

Test techniques

Freq response analysis (FRA)• Impedance/admittance/transfer

measurements• Typically 1kHz – 1MHz• Network analyzer or equivalent• Detects deformations/displacements• Compare against other phases , previous

measurement, or sister unit• The lead lengths need to be as short as

possible, and the test configuration must remain constant for repeated tests.

69

Application

70

• Detecting Faults which involves:Winding deformations Core movementsFaulty core groundsPartial winding collapseHoop bucklingBroken clamping structuresShorted turns and open windings

• Generate a base-line data for future comparison

• To Confirm Smooth Transportation

FRA

71

FRA

72

FRA

Repeatability is Key in FRA Measurement

73

Core NOT grounded

Core grounded

FRA-Typical Curves

74

FRA-Interpretation

• Low frequencies

– Core problems

– shorted/open windings

• Medium frequencies

– Winding deformations

• High frequencies

– Tap connections

– Other winding connection problems

75

FRA-Applied voltage

It is usually 10V.

76

2.8 VOmicron

10 VFRAX, Doble and others

FRA-Proper Grounding

77

Good grounding practice

Poor grounding practice

FRA-Test Case -180 MVA off-shore• Acetylene up

• Change in winding resistance

• LV distorted at 5 kHz – 500 kHz 78

Shipped for repair

79

Time for More Questions!

80