understanding power transformer_factory_test_data

Understanding

Power Transformer

Factory Test Data

Mark F. Lachman

Doble Engineering Company

©Doble Engineering Company

OVERVIEW OF PRODUCTION TESTS

Core/coil: ratio,

Iex, core-to-gnd

Core/coil after VP:

Iex, core-to-gnd

SU: ratio, Rdc, Iex,

no-load/load loss,

sound, core-to-gnd

Tanking: ratio, core-to-gnd,

in-tank CTs - polarity, ratio,

saturation

CTs on cover: polarity,

ratio, saturation PA: loss, sound,

core-to-gnd


Class I includes power transformers with

high-voltage windings of 69 kV and below.

Class II includes power transformers with

high-voltage windings from 115 kV through

765 kV.

SYSTEM VOLTAGE CLASSIFICATION


Routine tests shall be made on every

transformer to verify that the product

meets the design specifications.

Design tests shall be made on a

transformer of new design to determine

its adequacy.

Other tests may be specified by the

purchaser in addition to routine tests.

GENERAL CLASSIFICATION OF TESTS


TEST TYPE PERFORMANCE DIELECTRIC MECHANICAL

Routine

Winding resistance Winding insulation resistance

(Other) Leak

Ratio/polarity/phase

relation

Core insulation resistance

(Other)

No-load losses and

excitation current

Insulation PF/C

(Other)

Load losses and

Impedance voltage

Dielectric withstand of control

and CT sec. circuits (Other)

Operation of all

devices

Lightning impulse

(Design and Other)

Control and cooling

losses (Other)

Switching impulse

345 kV (Other)

Zero-phase sequence

impedance (Design)

Low frequency test

(Applied and Induced/Partial

Discharge)

DGA (Other)

Class I in red if

different from Class II

Class II < 345 kV

is also Other

OVERVIEW OF TESTS

PD is Other for

Class I only


TEST TYPE PERFORMANCE DIELECTRIC MECHANICAL

Design/

Other

Temperature rise

Audible sound level

Other

Short-circuit

capability

Single-phase

excitation current

Front-of-wave

impulse

Design

Lifting and

moving

Pressure

OVERVIEW OF TESTS (cont.)


TEST REFERENCE

DGA

Ratio/polarity/phase relation IEEE C57.12.90-2010 clauses 6, 7

IEEE C57.12.00-2010 clauses 8.2, 8.3.1, 9.1

Winding resistance IEEE C57.12.90-2010 clause 5

IEEE C57.12.00-2010 clause 8.2

No-load losses and excitation

current

IEEE C57.12.90-2010 clause 8 IEEE C57.12.00-2010 clauses 5.9, 8.2, 9.3, 9.4

Switching impulse IEEE C57.12.90-2010 clauses 10.1, 10.2 IEEE C57.12.00-2010 clauses 5.10, 8.2 IEEE C57.12.98-1993; IEEE Std. 4-1995

Lightning impulse IEEE C57.12.90-2010 clauses 10.1, 10.3 IEEE C57.12.00-2010 clauses 5.10, 8.2 IEEE C57.12.98-1993; IEEE Std. 4-1995

Applied voltage IEEE C57.12.90-2010 clause 10.5, 10.6 IEEE C57.12.00-2010 clauses 5.10, 8.2

Induced voltage/PD IEEE C57.12.90-2010 clause 10.7, 10.8, 10.9

IEEE C57.12.00-2010 clauses 5.10, 8.2 IEEE C57.113-2010; IEEE C84.1

No-load losses and excitation

current

IEEE C57.12.90-2010 clause 8 IEEE C57.12.00-2010 clauses 5.9, 8.2, 9.3, 9.4

SEQUENCE OF TESTS


TEST REFERENCE

DGA

Load losses and

impedance voltage

IEEE C57.12.90-2010 clauses 9.1-9.4, Annex B2 IEEE C57.12.00-2010 clause 5.8, 5.9, 8.2, 8.3.2,

9.2-9.4

ONAN temperature rise IEEE C57.12.90-2010 clause 11 IEEE C57.12.00-2010 clause 8.2

IEEE C57.91-1995 Table 8 (with 2002 corrections)

DGA IEEE PC57.130/D17

ONAF temperature rise IEEE C57.12.90-2010 clause 11 IEEE C57.12.00-2010 clause 8.2

IEEE C57.91-1995 Table 8 (with 2002 corrections)

DGA IEEE PC57.130/D17

Zero-phase sequence

impedance

IEEE C57.12.90-2010 clause 9.5 IEEE C57.12.00-2010 clause 8.2

Audible sound level

IEEE C57.12.90-2010 clause 13, Annex B5 IEEE C57.12.00-2010 clause 8.2

NEMA TR1-1993

Core demagnetization

DGA

SEQUENCE OF TESTS (cont.)


TEST* REFERENCE

Insulation PF/C and

resistance

IEEE C57.12.90-2010 clauses 10.10, 10.11 IEEE C57.12.00-2010 clause 8.2

Single-phase exciting

current

IEEE C57.12.00-2010 clause 8.2 Lachman, M. F. “Application of Equivalent-Circuit Parameters to Off-Line Diagnostics of Power Transformers,” Proc. of the Sixty-Sixth Annual Intern. Confer. of Doble Clients, 1999, Sec. 8-10.

Sweep frequency response

analysis IEEE PC57.149™/D8, November 2009

Dielectric withstand of control

and CT secondary circuits IEEE C57.12.00-2010 clause 8.2

CT polarity/ratio/saturation IEEE C57.13.1-2006

Control and cooling losses IEEE C57.12.00-2010 clauses 5.9, 8.2

Operation of all devices IEEE C57.12.00-2010 clause 8.2

Core-to-ground insulation

resistance

IEEE C57.12.90-2010 clause 10.11 IEEE C57.12.00-2010 clause 8.2

SEQUENCE OF TESTS (cont.)

*Discussion of tests listed on this slide and DGA is not included in this presentation.


Tests to be discussed:

Ratio/polarity/phase relation

Winding DC resistance

No load losses and excitation current

Dielectric tests

Load losses and impedance voltage

Temperature rise

Zero-phase sequence impedance

Audible sound level

DISCUSSION OUTLINE


For each test discussion includes:

Definition and objective

Physics

Setup and test methodology

Acceptance criteria*

Abnormal data

Recourse if data abnormal

Comparison with field data (if relevant)

DISCUSSION OUTLINE (cont.)

*This discussion is based on requirements of referenced standards. If customer test specification contains requirements different from those in standards, more stringent requirements prevail.


RATIO, POLARITY, PHASE

RELATION (Routine) ©Doble Engineering Company

Definition: The turns ratio of a transformer is the ratio of

the number of turns in the high-voltage winding to that in

the low voltage winding.

Objective: The turns ratio polarity and phase relation test

verifies the proper number of turns and internal

transformer connections (e.g., between coils, to LTC, to

various switches, to PA, series auto- or series

transformer) and serves as benchmark for later

assessment of possible damage in service.

The transformer nameplate voltages should reflect the

actual system requirements. Therefore, it is important

that the nameplate drawing is approved by the customer

at the design stage.

RATIO, POLARITY, PHASE RELATION:

DEFINITION AND OBJECTIVE


3T 2T 2V

VR = 3V/2V = 1.5

TR = 3T/2T = 1.5

In ideal transformer:

TR = VR

3T 2T 1.96V

Volts per turn = 3V/3T = 1V/T

F

3V

Volts per turn = 2.95V/3T = 0.98V/T

2.95V

F VR = 3V/1.96V = 1.53

TR = 3T/2T = 1.5

= 100(1.5 – 1.53)/1.5 = –2%

0.05V

3V


PHYSICS

3V

In actual transformer Turns ratio Voltage ratio due to accuracy of the measurement and the voltage drop in the high-voltage winding. ©Doble Engineering Company

Ratio = N1/N2 = R1/R2

Polarity is determined via phase angle between two measured waveforms.

Phase relation is confirmed by testing the corresponding pairs of windings.

Tests shall be made 1. at all positions of DETC

with LTC on the rated voltage position

2. at all positions of LTC with DETC on the rated voltage position

3. on every pair of windings

R2

Balance

indicator

N2 N1

R1

H1 X0

H2

X2

Transformer in test


SETUP AND TEST METHODOLOGY



ACCEPTANCE CRITERIA

X1

H1

H2

H3

X2

X3

X0

Voltage ratio =

VH2-H1/VX2-X0 =

138/(13.2/3) = 18.108

13.2

138

With the transformer at no load and with rated voltage on

the winding with the least number of turns, the voltages of

all other windings and all tap connections shall be within

0.5% of the nameplate voltages.

For three-phase Y-connected windings, this tolerance

applies to the phase-to-neutral voltage. When the phase-to-

neutral voltage is not explicitly marked on the nameplate,

the rated phase-to-neutral voltage shall be calculated by

dividing the phase-to-phase voltage markings by 3. ©Doble Engineering Company


ABNORMAL DATA

To appreciate significance of 0.5% limit, it is instructive to

recognize the inherent errors this limit accommodates.

Actual turns RATIOTURN

Nameplate voltages

RATIONP

Deviation

100(RATIONP - RATIOTURN)/RATIONP =

Rounding off

NP voltages

creates error

Measurement RATIOMEAS

Deviation

100(RATIONP - RATIOMEAS)/RATIONP 0.5%

Measurement

introduces

error

NP voltages need to be

selected to keep well

within 0.5% (e.g., 0.2-

0.4). This assures that

measurement error

keeps RATIOmeas within

0.5% of RATIONP.

RATIOTURN

RATIONP

RATIOMEAS



RECOURSE IF DATA ABNORMAL

If deviation exceeds 0.5% for any of the measurements the

result is not acceptable.

The following steps should be considered:

Check if V/T exceeds 0.5% of nameplate voltage. If yes,

under these conditions the standard allows for deviation

from the NP voltage ratio to exceed the 0.5% limit.

Check if transformer is a duplicate of a legacy unit.

Review design data to determine if the NP voltages selected by designer create a ratio that is too far (b is

too high) from true turns ratio. Discuss possibility of

changing nameplate voltages for relevant tap positions.

Review results of production ratio tests and, if applicable,

consider retesting with analog instrument.

Exciting current reported by turns ratio instrument is a

useful diagnostic indicator.



COMPARISON WITH FIELD DATA

In verifying compliance with 0.5% deviation from the NP

voltages, the following should be recognized:

Older analog instruments produce results much closer to

the actual turns ratio than modern digital instruments.

Even within 8-200 V range, the results vary somewhat

with voltage and between different instruments.

Initial field test should be performed at the same test

voltage as the factory test with results compared with the

NP voltages and for all subsequent tests the comparison

should be made with the initial test.

The objective of the high-voltage (e.g., 10 kV) test with

external capacitor is to stress turn-to-turn insulation of both

windings for diagnostic purposes and not necessarily to

verify the 0.5% limit. In some cases, the latter could be

exceeded due to the loading effect of the test capacitor.


WINDING DC RESISTANCE (Routine)


WINDING DC RESISTANCE:


Definition: Winding DC resistance is always defined as the

DC resistance of a winding in Ohms.

Objective: The measurement of winding resistance

provides the data for:

Calculation of the I2R component of conductor losses.

Calculation of winding temperatures at the end of a

temperature rise test.

Quality control of design and manufacturing processes.

Benchmark used in field for detection of open circuits,

broken strands, deteriorated brazed and crimped

connections, problems with terminations and tap

changer contacts.



PHYSICS

i R

Domain

External

field



PHYSICS (cont.)

dy/dt

F = y/N

vmeas = iR + dy/dt R=vmeas / i

dy/dt

dy/dt

dy/dt

dy/dt

dy/dt

dy/dt



PHYSICS (cont.)

Time to stabilize resistance reading: On some units with closed

loops (e.g., GSU with two LV deltas or units with parallel

windings), it may take a long time for the reading to stabilize*; it

reduces with intermediate stability levels. This phenomenon is

not related to core saturation, which is saturating in a

reasonable time. However, as the core is being magnetized the

changing flux in the core induces voltage and sets up

circulating currents in closed loops. After the core is saturated,

there is no more induced voltage to sustain them, and the

currents begin to subside. This process, however, is associated

with LC oscillations with long time constant and may take up to

45 min to dissipate the energy. The flow of these currents

continues creating a changing flux in the core, inducing voltage

in the tested winding and thus changing the measured

resistance reading. Opening these loops, when possible,

reduces the time to stability.

* Personal communications with Bertrand Poulin, ABB, Quebec, Canada.




Data must be taken only when reading is stable. The time to stabilize the reading depends on the unit, varying from seconds to minutes.

Standard requires measurements of all windings on the rated voltage tap and at the tap extremes of the first unit of a new design.

The measured data is reported at Tave_rated_rise + 20C, e.g., 65+20= 85C and as total of 3 phases.

Transformer in test

H2

H1

H3

H0

Idc Vdc

+

+

Current

output

Voltage

input ©Doble Engineering Company


ACCEPTANCE CRITERIA

Standards give no acceptance criteria; however, a deviation

from average of three phases of 0.5% for HV and 5% for LV

could serve as practical guideline.

As important as deviation is the assurance that test data is

credible:

No excitation with no pumps - 3h and with pumps - 1h,

TTO variation 2C for 1h, and TTO-TBO 5C. This assures

that oil T represents conductor T; without reference T

resistance data has a limited value.

Test current 10% of maximum rated load current.

Voltage test leads must be placed as close as possible to

winding terminals.

Test data should be recorded only when reading is stable.

Measuring system accuracy +/-0.5% of reading with

sufficient current output to stabilize the flux.



ACCEPTANCE CRITERIA (cont.)

T stability: Experience* in the industry suggests that

relying on the T stability requirements given in the IEEE

standard does not produce a needed thermal equilibrium

and, consequently, an accurate measurement of the

winding dc resistance. To have a reliable data, the unit

should be subjected to no excitation for 2-3 days. Hence, if

the time to begin testing is of essence, it is not

unreasonable to agree to using resistance data available at

that time (assuming the IEEE T requirements have been

met), but request that resistance is re-measured later

(including cold resistance for heatrun), when the T is

stable. Obviously, the load loss and the heatrun results

should be then recalculated with the latest T.




ABNORMAL DATA

DETC H1-H3 H2-H1 H3-H2

1 3.7350 3.7352 3.7378

2 3.6470 3.6468 3.65023 3.5590 3.5580 3.5622

4 3.4714 3.4698 3.5746

5 3.3838 3.3814 3.3870

High-voltage winding

Tested Calc. % of calc.

20.9832 21.47937 97.720.4889 20.97440 97.719.9932 20.46944 97.719.6873 19.96448 98.6

19.0065 19.45952 97.7

Average

3.7360 0.03% 0.02% -0.05%3.6480 0.03% 0.03% -0.06%3.5597 0.02% 0.05% -0.07%

3.5053 0.97% 1.01% -1.98%

3.3841 0.01% 0.08% -0.09%

Deviation from average

0.6185 0.03842 99.70.5855 21.47937 100.6

0.16519 -0.11% -0.01% 0.12%0.15637 -0.12% 0.00% 0.12%

LTC X1-X0 X2-X0 X3-X016 0.16537 0.16521 0.16499N 0.1566 0.1564 0.1562

Low-voltage winding

Comparison of each measurement with the average along with design data identifies an abnormal reading in H3-H2 with DETC in 4. This potentially can be caused by a problem with DETC contacts.




If requirements associated with transformer thermal stability,

dc test current, influence of series unit or stability of the

reading are not met, a retest under different conditions

should be requested.

If acceptance criteria is exceeded, a justification from the

manufacturer should be requested. Potential problems may

include: bad crimping or brazing, incorrect conductor cross

section, loose connection, wrong design calculations.




Typically, a deviation of <5% from the factory value is

considered acceptable.

A factory value is often reported as a sum of three phase

readings at rated T. For field comparison, the per-phase

values at corresponding DETC/LTC positions should be

requested from the factory.

Comparison should be performed for readings referred to the

same T.

The field measurement should be performed at the same test

current as the factory one.

Field tests are the subject to the same thermal stability

requirements as the factory test (note that at the factory T is

measured via thermocouples and in the field the T gauge is

frequently the best option).


NO-LOAD LOSSES AND

EXCITATION CURRENT (Routine) ©Doble Engineering Company

NO-LOAD LOSSES AND EXCITATION CURRENT:


Definition: No-load losses include core loss, dielectric

loss, and conductor loss due exciting current, including

current circulating in parallel windings. Excitation current

is flowing in any winding exciting the transformer with all

other windings open-circuited.

Objective: No-load losses and excitation current, measured at specified voltage and frequency, provide the

data for:

Verification of design calculations.

Demonstration of meeting the guaranteed performance

characteristics. Since these parameters have often an

economic value attached to them, the accuracy of the

measurement becomes significant.

No-load losses are used as test parameter during the

temperature rise test.



PHYSICS

I R

F

V

B

H

Hysteresis losses

Ph = f(Bmax)

Bmax = f(Vave) F

Ieddy

Eddy

losses

Pe = f(V2rms)

PNL = Pe + Ph

Domain

rotation ©Doble Engineering Company



Start with 110% on N. As unit demagnetizes, losses drop.

Test at 100% Vrated on N, max turn bridging position with inductive LTC, and 16R if LTC with series unit.

Vave gives same Bmax as Vrms when wave-shape is a perfect sin; set based on Vave average of 3 phases

Pe is corrected for rated Vrms

Voltmeters should measure same voltage as seen by xfmr.

PNL not corrected for T if TTO-TBO 5C and 10TO_ave30C

Iexc=aver. of 3 phases in % of Irated

CT

VT

Vrms

W

I

V

Transformer in test

X0

H2 X1

X2

X3

H1

H3

Vave

A

*

*

*Vrms and Vave (calibrated in rms) will show the same voltage if perfect sine wave.

3



SETUP AND TEST METHODOLOGY (cont.)

Frequency control of motor-generator sets at GE large transformer plant in Pittsfield, MA during the early 1900. Since the primary function of these generators was to provide power for no-load loss tests, they were often referred to as magnetizers.

Historical perspective

Courtesy IEEE Power & Energy Magazine



ACCEPTANCE CRITERIA

Measured no-load losses should not exceed the

guaranteed value by more than 10% and the total losses by

more than 6%.

Assurance that test data is credible:

Test voltage is set based Vave

If oil T is not within limits, correction is applied

Frequency is within +/-0.5% of rated

Distortion 5%. The 5% limit that standard allows for

distortion of the voltage waveform is too liberal.* The

limit applies to the difference between the measured kW

and kW corrected for eddy loss due to the difference

between Vrms and Vave. To monitor the quality of the

voltage waveform, one should look at the following

criteria of the applied voltage waveform: THD < 5%, 3rd

and 5th harmonics <10% and waveform should not have

any visible distortions.





Test in parallel and series configurations, if present.

If PA is present, compare the loss difference between

non-bridging and bridging positions (max turns) with

loss measured in PA out-of-tank. If SU unit is present,

compare the loss difference between N and 16R with

loss measured in SU out-of-tank.

Test system accuracy should be within +/-3% for loss,

+/-0.5% for voltage and current, and +/-1.5C for T.



ABNORMAL DATA

Potential reasons for exceeding the guaranteed values may

include:

Variability in core steel characteristics

Different core steel

Oversights in design

Production process related factors or mistakes

Problems with windings (e.g., s. c. turn)

Wrong connection of preventative autotransformer or

series transformer or series autotransformer

Example: guaranteed no-loss - 28 kW, measured – 35 kW




Failure to meet the no-load test loss tolerance should

not warrant immediate rejection but shall lead to

consultation between purchaser and manufacturer

regarding further investigation of possible causes and

the consequences of the higher losses.

The acceptance criteria of 10% does not replace the

manufacturer’s guarantee of losses for economic loss

evaluation purposes.




Factory no-load losses and excitation test is performed

at rated voltage and three-phase excitation. Since the

open-circuit magnetizing impedance of a transformer is

non-linear, i.e., it is changing with applied voltage, a

comparison of exciting current and losses test results

obtained at low-voltage (e.g., 10 kV) and single-phase

excitation with results of the factory no-load losses and

excitation test is not possible.


DIELECTRIC TESTS


DIELECTRIC TESTS:


Definition: Tests aimed to show that transformer is

designed and constructed to withstand the specified

insulation levels are referred to as dielectric tests. They

include:

high-frequency tests: lightning and switching impulses

low-frequency tests: applied and induced/PD tests

Objective: Dielectric tests demonstrate:

compliance with users specification

compliance with applicable standards

verification of design calculations

assessment of quality and reliability of material and

workmanship

Note: Unless agreed otherwise, all dielectric tests must be performed with bushings supplied with the transformer.


HIGH-FREQUENCY:

LIGHTNING IMPULSE (Class I - design or other,

Class II - routine) ©Doble Engineering Company

HIGH-FREQUENCY - LIGHTNING IMPULSE:

OBJECTIVE

Demonstrate performance under transient high-frequency

conditions caused by lightning.

Surge of energy, from lightning striking transmission line, travels to substation and enters a transformer - full wave.

kV

s

Surge of energy, from lightning striking transmission line, travels to substation and, after reaching the crest of the surge, causes arrester operation or flashover across an insulator near transformer terminals - chopped wave (a.k.a. tail-chopped).

Surge of energy, from lightning striking transmission line, travels to substation and operates gapped silicon-carbide arrester at transformer terminals - front-of-wave (a.k.a. front-chopped).



PHYSICS

length

V

Due to impulse front high frequency, the initial voltage distribution is determined by the capacitive network, with higher voltage gradients towards the impulsed end of the winding. The higher is , the steeper are the gradients at the impulsed end of the winding. As the front passes, the distribution changes as determined by the tail of the wave.

Cg/Cs

Cg Cs

V

Full wave can be simulated by discharging capacitor while chopped wave by the operation of a gap triggered to flashover at required time.



PHYSICS (cont.)

*Assumption that FOW stresses mostly the first few turns at the impulse end is not always true; it depends on winding type and configuration, e.g., when the interleaved winding (one with high series capacitance) is in series with RV, the impulse goes through the main winding and hits RV (Personal communications with Bertrand Poulin, ABB, Quebec, Canada.) **From W. McNutt 1989 Doble tutorial: Turns in this context mean the voltage that would have been present if the applied voltage was distributed according to turns ratio.

Region A* - turn-to-turn insulation at line is tested by FOW impulse, with stress >10turns**. Region B – disk-to-disk, and layer-to-layer insulation (and turn-to-turn) is tested by FW & CW impulse, with stress 5-10turns. Region C – insulation across taps is tested by FW & CW impulse, with stress 5-10turns.

LV H1

A

B

C

A

B

B

B

B

C

H1 HV

to DETC

H0



PHYSICS (cont.)

Cg CT

Rs

Rp

Charge of Cg – generator capacitors are charged from external DC source.

*CT includes preload capacitor.

Discharge into Rp – energy from xfmr is discharged into generator, reducing voltage at tested terminal.

VT FW

Cg CT

Rs

Rp

Discharge at chop – energy from xfmr is discharged into chopping gap, reducing voltage at tested terminal to zero.

VT CW

Cg CT

Rs

Rp

FOW

Discharge into C*T –

energy from generator capacitors is discharged into xfmr, raising V at tested terminal to crest level.

VT

Cg CT

Rs

Rp




Full Wave Parameters

Magnitude FW = BIL +/- 3% RFW = 50-70% BIL

T1 = 1.67T 1.2 s +/- 30% 0.84 ÷ 1.56 s

T2 50 s +/- 20%

40 ÷ 60 s 5%

Applied test waves are of negative polarity to reduce risk of erratic external flashover.

See C57.12.90-2010 when for line terminals T1 is allowed to be >1.56 s and T2<40 s. For neutral bushing T1<10 s and T2 could be <40 s.

If the T2<40 s, it should be addressed at the bidding stage.

0.9

0.3

1.0

T1

0.5

T2

t

V

Crest voltage

Half voltage

Virtual

origin

T




Increase of series (front) resistor Rs

increases the time of voltage rise - T1.

0% change from given Rs

Cg CT

Rs

Rp

Data courtesy Reto Fausch, Haefely




Increase of parallel (tail) resistor Rp increases

the time of voltage decline to half value - T2.


0% change from given Rp

Cg CT

Rs Rp





Cg CT

Rs

Rp

Increase of series (front) resistor Rs decreases

the voltage trace overshoot - .




Chopped Wave Parameters

Magnitude CW = 1.1BIL+/- 3%

T1 1.2 +/- 30% 0.84 ÷ 1.56

TC

BIL [kV] Class I Class II 30 1.0

2.0 45÷75 1.5

95 1.8 110 2.0 125 2.3 2.3

150 3.0 TC < 6.0 30% 1

See C57.12.90-2010 for instances when could be >30% and >1s. It also permits adding resistors in chopping gap circuit to limit .

All times in the table are in s.

0.9

0.3

1.0

T1 TC

t

V

1.0

0.7

0.1




Front-of-Wave Parameters

Magnitude C57.12.00-2010 Annex A TC

30%

C57.12.90-2010 permits adding resistors in chopping gap circuit to limit . With improved arrester technology, front-of-wave tests may not be necessary

and were removed as a requirement from C57.12.00. Annex A in that standard includes the last published table of front-of-wave test levels from C57.12.00-1980, for historical reference.

0.9

0.3

1.0

TC

t

V


Current shunt and

meas. circuit

i(t)

RG

LG Glaninger: T2



T2

Impulse

control &

measuring

system

Voltage divider

and measuring

circuit

v(t)

Cg

Rs

Rp

Impulse

generator

LT, CT

xfmr

Very high di/dt induces difference

of potential. Hence, it is very

important for all return and

grounding leads to be made as

short as possible, with a minimum

R and L.

Chopping gap

and preload

capacitor Chopping gap should not be connected in series

with voltage divider no matter how convenient it is

for the test department to have a permanent setup.




Line terminal in Y

i(t)

Line terminal in

i(t)

Neutral terminal in Y

i(t)

i(t)

LV line terminal in Auto

HV line terminal in Auto

i(t) i(t)

Neutral terminal in Auto ©Doble Engineering Company



Test sequence and trace comparison

Test is performed with minimum effective turns in the winding under test, e.g., DETC = 5, LTC = 16L.

Standard:

RFW@ 50-70% BIL

CW 1

CW 2

FW

With FOW:

RFW@ 50-70% BIL

FOW 1

FOW 2

CW 1

CW 2

FW

With non-linear

protective devices:

RFW 1

RFW 2 @ 75-100% of BIL

to demonstrate growing

sensitivity to V

FW 1

CW 1

CW 2

FW 2

RFW 3 @ RFW2 voltage

RW 4

Neutral:

RFW@ 50-70% BIL)

FW1

FW2



ACCEPTANCE CRITERIA

If test equipment and tested transformer were perfectly linear, the

traces of repeated impulses, when overlaid, would perfectly

match. However, due to noise, setup imperfections or insulation

failure, discrepancies occur. Identifying their nature is the

objective of impulse data analysis.

T1, T2, Tc, voltage magnitude, , must meet requirements.

RFW and FW voltage and current traces should compare; request

to zoom in on any areas of concern.

If available, comparison of Transfer Function (TF) for RFW and FW

is used as additional diagnostic criteria. It removes sensitivity to

wave shape variations caused by impulse generator jitter (TF

should be considered only in frequency ranges where sufficient

data is present in the time domain impulse trace*).

For chopped wave test, segments of CW1 and CW2 traces prior to

moment of chop are compared. While traces after chop may be

shift, they oscillate around zero with the same frequency.

Verify that DGA results (after dielectrics) are normal.

* IEEE PC57.98TM/D07, September 2011, Draft Guide for Transformer Impulse Tests.




450 kV BIL, RFW on

HV winding – voltage

450 kV BIL, RFW on

HV winding – current




450 kV BIL, CW1 on


450 kV BIL, CW2 on





Overlay of 450 kV BIL

CW1 and CW2 - voltage




450 kV BIL, FW on


450 kV BIL, FW on

HV winding – current





RFW and FW - voltage





RFW and FW - current

High-frequency oscillations at

the beginning of current trace

are acceptable deviations,

reflecting the test setup.




450 kV BIL, FOW1 on


450 kV BIL, FOW2 on





Influence of non-linear protective

device on overlay of RFW and FW

350 kV BIL voltage traces

illustrates the need for comparing

traces of the same voltage level.




Influence of non-linear protective

device on overlay of RFW and FW

350 kV BIL current traces

illustrates the need for comparing

traces of the same voltage level.



ABNORMAL DATA

In general, whenever discrepancies occur the normal test procedure

need to be stopped and investigation performed. If the cause is found to

be external to the transformer, the corrections are made before the test

can continue.

If there is any doubt as to the cause of the discrepancies, additional

impulses need to be applied, including several FW. If the deviation

increases in magnitude, it indicates progressive dielectric failure in the

transformer.

Unusual sounds, emanating from inside the tank, should be noted; these

sounds may be helpful in locating general location of the fault.

Removing manhole covers and observing presence of gas bubbles

and/or carbon, serves as confirmation of failure and provides some

indication of the fault location.

Occasionally, the damage caused but not detected by impulse is only

detected by tests that follow: applied or induced/PD voltage tests, DGA.



ABNORMAL DATA (cont.)

Overlay of 550 kV BIL RFW and FW

voltage traces – turn-to-turn failure





current traces – turn-to-turn failure





traces – lead-to-lead failure

between RV and main LV windings

RFW voltage

FW voltage

FW current

RFW current

Voltage drop to ground

indicates one of the leads

was at ground potential

Fault to ground diverts

current around winding,

reducing measured current.


HIGH-FREQUENCY:

SWITCHING IMPULSE (Class I – other,

Class II <345 kV – other,

Class II 345 kV - routine) ©Doble Engineering Company

HIGH-FREQUENCY - SWITCHING IMPULSE:

OBJECTIVE

kV

s

FOW

CW

FW

SW

Demonstrate performance under transient high-frequency conditions

created by switching operations or network disturbance.

Surge of energy from equipment switched on or disturbance on the power system. The time to reach the crest amplitude and the total time duration of switching impulses are much longer than those of lightning impulses.



PHYSICS

length

V

Comparing to lightning impulse, the switching impulse has a much longer duration and lower frequency, resulting in voltage approaching a uniform distribution of the low-frequency steady-state voltages, i.e., voltage distributes as per turns ratio.

V

Switching impulse test consists of applying or inducing a SW between each HV line terminal and ground. Similar to a lightning wave, the switching wave can be simulated by discharging a capacitor.



PHYSICS (cont.)

*From W. McNutt 1989 Doble tutorial: Turns in this context mean the voltage that would have been present if the applied voltage was distributed according to turns.

LV D H1 HV

Region D – phase-to-ground and phase-to-phase insulation is stressed the most; stress imposed by SW is 1turns*. Charging and discharging processes are similar to those described for lightning impulse.

H1

H0

D D

D

To another

phase




Full Wave Parameters

Magnitude SW = 0.83BIL +/- 3% RSW=(50-70%)0.83BIL

Tp >100 s Td 200 s T0 1000 s

LV windings shall be designed to withstand stresses from SW applied to HV side.

Applied test waves are of negative polarity to reduce risk of erratic external flashover.

0.9

1.0

Tp Td

t

V

Crest voltage

First zero crossing T0

Virtual

origin

>90% of crest




Xfmr

Impulse

control &

measuring

system

Voltage divider

and measuring

circuit

v(t)

Cg

Rs

Rp

Impulse

generator

Note: The shown setup is for SW being applied to the HV winding. The test can also be performed with SW being induced.




ELV/2

Note: The choice of tap connections for all windings is made by the manufacturer.

Test sequence and trace comparison:

RSW@ 50-70% SW

(+) RSW - bias

SW1

(+) RSW - bias

SW2

RFW@ 50-70% BIL

CW 1

CW 2

FW

E

E/2

ELV E

E/2 -ELV/2

ELV

E

-ELV/2

ELV

-E/2




SW can saturate the core, creating an air-core conditions, i.e.,

drastically reducing impedance faced by impulse. This rapidly

decays the tail of the voltage waveform to zero, making T0<1000

s. To extend the time to saturation, prior to start of each test,

the core is magnetized in opposite direction by applying RSW (or

small dc current) of opposite polarity .

t

V

When core saturates, the

voltage collapses drastically

reducing time to zero crossing.

Bias in the core in

direction opposite

to that created by

test SW extends

time to saturation

and T0.



ACCEPTANCE CRITERIA

Tp, Td, T0, and voltage magnitude must meet requirements.

Failure detection is done primarily by scrutinizing voltage

traces for recognizable indications of failure. The test is

successful if there is no sudden collapse of voltage as

indicated on the trace.

Although overlaying RSW and SW traces in totality may

not be practical, the traces should match until the point

where the difference in the core magnetic state becomes

obvious. Normally, these differences can be easily

distinguished from drastic voltage reduction caused by a

failure.

Verify that DGA results (after dielectrics) are normal.




650 kV BIL, RSW on

HV winding – voltage 650 kV BIL, SW1 on


650 kV BIL, SW2 on


Typical reduced and full switching impulse voltage traces as measured on the HV winding; for 650 kV BIL, the BSL, i.e., the required test voltage, is 540 kV.





RSW and SW - voltage

Beginning of traces deviating

due to the difference in core

magnetic state. This is

typically more pronounced in

the overlay of reduced and full

switching waveforms





SW1 and SW2 - voltage

Slight deviation due to the

difference in core magnetic

state.



ABNORMAL DATA

In general, whenever discrepancies occur the normal test

procedure need to be stopped and investigation performed. If

the cause is found to be external to the transformer, the

corrections are made before the test can continue.

If there is any doubt as to the cause of the discrepancies,

additional impulses may be applied.

Removing manhole covers and observing presence of gas

bubbles and/or carbon, serves as confirmation of failure and

provides some indication of the fault location. ©Doble Engineering Company

HIGH-FREQUENCY–LIGHTNING AND SWITCHING IMPULSE:


If visual confirmation (e.g., carbon, bubbles) is obtained

or the data convincingly reveals a failure, the oil is

drained and internal inspection is performed.

If necessary, the unit is un-tanked. This is followed by a

thorough and well-documented investigation.

The user’s involvement in this process enhances the

quality of the investigation and that of the final product.


LOW-FREQUENCY:

APPLIED VOLTAGE (Routine) ©Doble Engineering Company

LOW-FREQUENCY – APPLIED VOLTAGE:

OBJECTIVE

The applied voltage test is a simple overvoltage test. The

early transformer engineers apparently took cues from

mechanical engineers. This is how a mechanical structure

would be tested, by applying stress that demonstrates a

safety factor of two. The applied voltage test has a 1 min

duration, with the expectation to demonstrate a long-term

capability to operate at the rated voltage.

The high-frequency tests (lightning and switching

impulse) always precede the low-frequency tests (applied

and induced voltage). This sequence is rooted in the fact

that due to a longer duration, the low-frequency tests

serve to stress further and to detect the damage caused

by the high-frequency tests.



PHYSICS

Region D – major winding -to-ground and winding-to-winding insulation are stressed the most. LV

D

HV

D

HV

LV

Shorting

lead

D




Applied Voltage Parameters

Magnitude C57.12.00-2010 Duration 1 min

Test is performed at low frequency (<500 Hz), normally, power frequency.

All terminals of tested winding are connected together; all other terminals (including all cores, buried windings with one terminal brought-out and the tank) are grounded.

A sphere-gap, set for 10% above test voltage, may be connected for protection.

Test voltage (1-phase) is determined by terminal with the lowest BIL (e.g., Neutral).

The voltage is raised from 25% or less, held for 1 min and reduced gradually.

Each winding or set of windings (e.g., in auto) is tested.

Note: On grounded-wye transformers with reduced Neutral BIL the test has a limited significance; it inly tests insulation in the vicinity of the Neutral.

E

1.1E

v



ACCEPTANCE CRITERIA

The test is a pass/fail test and is considered

passed if during the time the voltage is applied no

evidence of possible failure is observed.

The indications to monitor include unusual sound

such as thump, sudden increase in the test circuit

current and collapse in the test voltage.



ABNORMAL DATA

If unusual sound, sudden increase in the test

circuit current or circuit tripping occur, these

events should be carefully investigated by:

• observation, e.g., presence of carbon

and/or bubbles in the oil

• repeating the test

• other tests

to determine whether the failure has occurred.

Due to a significant energy being released during

applied voltage test, the test is repeated (if at all) to

confirm the failure a limited number of times (1, 2

max). The energy released is usually sufficient to

mark the location making it possible to find the

failure after un-tanking.





and/or repeating of the test and/or other tests reveal the

failure, the oil is drained and internal inspection is

performed.


LOW-FREQUENCY:

INDUCED VOLTAGE/PD

Induced:

7200 cycles

Induced:

1 hour + PD

Class I

Class II

Routine

Routine


LOW-FREQUENCY – INDUCED VOLTAGE/PD:

OBJECTIVE

The induced voltage test demonstrates the strength of

internal insulation in all windings as well as between

windings and to ground. A combination of prolonged stress

and a very sensitive PD measurement makes it a very severe

and searching test. It must be the last dielectric test to be

performed.


LOW-FREQUENCY - INDUCED VOLTAGE/PD:

PHYSICS

To stress turn-to-turn insulation to the required level, the winding needs to be excited to a level approaching twice rated voltage. At power frequency, this would overexcite the core.

Therefore, the test is performed as a higher frequency, which allows to obtain the needed volts/turn at a lower flux magnitude (v/t = dF/dt).

At higher frequency, transformers become capacitive with dangers of M-G set overexciting. This is addressed by using a variable reactor. The latter provides an additional benefit of reducing the load on MG set.

R C M G

L

Lv

xfmr

Variable reactor Lv is adjusted to reduce output from generator.

IG

VT

ILv

IT

VT

IG

IT

ILv


LOW-FREQUENCY - INDUCED VOLTAGE/PD:

PHYSICS (cont.)

Region E – with voltage distributing per turns ratio, the most stress is present in the turn-to-turn insulation of each winding as well as in winding-to-winding and winding-to-ground insulation.

LV

E

HV

E

E

E

E

E



PHYSICS (cont.)

From the physics point of view, self-sustaining electron avalanches

may occur only in gases. Hence, discharges in dielectrics may

only be ignited in gas-filled cavities, such as voids or cracks in

solid materials and gas bubbles or water vapor in liquids.

Discharges are generally ignited if the electrical field strength

inside the inclusion exceeds the intrinsic field strength of the gas.

They can appear as pulses having a duration of << 1s.

Partial discharges are defined as localized

electrical discharges that only partially bridge

the insulation between conductors and may or

may not occur adjacent to a conductor. In

insulation, the PD events are the consequence

of local field enhancements due to dielectric

imperfections.

Gaseous

inclusion

Conductor

Dielectric



PHYSICS (cont.)

To model the PD process, capacitance of the active void CC can be

viewed as part of a larger capacitive network. In that, CB is the

remaining capacitance of the immediate region in series with CC

and CA is the rest of the dielectric connected in parallel. Two

requirements must be fulfilled to initiate PD: 1) local field stress

exceeds the void’s breakdown voltage Vbd and 2) free electrons are

available.

CC

CB

CA

CC

CB CA

Vbd



PHYSICS (cont.)

Strike: As Vcc>Vbd, breakdown

occurs, charges move across

shorting the void, Vcc= 0 and

discharge stops. To make up

for imbalance, charges come

out of adjacent insulation.

CC

CB

CA

V

CC

CB

CA

Q

Q

Q Q Vcc= 0

V

Q Q Q

Q

Vcc= 0

CC

CB

CA

Q

Q Q

Q

Vcc

V

Buildup: As V , charges

move to and collect on

the surface of the void,

building potential stress

Vcc across the void.

CC

CB

CA Q Q Q

V

Vcc

CC

CB

CA

Q

Q

Vcc

V

Relaxation: Charges continue

to flow at a decreasing rate

with balance restoring. Vcc

as charges collect back on

the void’s surface.

CC

CB

CA Q

V

Vcc

Q



PHYSICS (cont.)

We cannot measure the real charge. However, as the void discharges, the

charge redistribution creates a dip* in the terminal voltage. This minute voltage

drop causes a high-frequency current to flow through a coupling capacitor

connected to a measuring system. Putting it differently, the charge movements

appear, in part, in C1 connected in parallel with CT. The integration of these high-

frequency current pulses over time produces the reported apparent charge.

Terminal

voltage

Voltage

across void PD current

Dip in terminal

voltage

C1

C2 Z

M

CT

*The detectable voltage dip is in the mV

range, while that at the void may be in

the kV range.



PHYSICS (cont.)

Measurement of partial discharge is

like trying to weigh a butterfly that

alights momentarily on scales

designed for an elephant (sometimes

during an earthquake). by Karl Haubner, Doble Australia




M G

Lv

Xfmr in

test

X0

H2 X1

X2

X3

H1

H3

C1

C2

C1

C2

C1

C2 M

Step-up

xfmr V

pC

and/or

V

Before test commences, several important steps take place:

Transformer is connected for open-circuit conditions.

Voltage is raised to verify that variable (Lv) setting allows

to reach the required test voltage.

Measuring system (M) is calibrated for PD, RIV and

voltage.




Induced Voltage/PD

Parameters

Voltage magnitude

C57.12.00-2010 clause 5.10

C84.1

Timing C57.12.90-2010

clauses 10.7, 10.8

PD/RIV criteria

C57.12.90-2010 clause 10.8/

Annex A

Voltage is gradually raised, recording pC, V and kV.

For Class I units, the test includes applying to HV winding 2.0nominal

voltage for 7200 cycles with no PD (RIV) recordings. For class II units rated

115 ÷ 500 kV, the test includes applying to HV winding 1.8nominal voltage

for 7200 cycles and 1.58nominal voltage for 1 h, recording PD (RIV) data.

For windings other than HV, when possible, taps should be selected so that

voltages on other windings are as per ANSI C84.1 and C57.12.90 clause

10.8.1 (e.g., for 115÷345 kV units , the voltage on other windings should be

1.5 times their maximum operating voltage).

Ambient Ambient

100%

Enhanced level

7200 cycles 1h level, 5 min

recordings

100%

t

V

Hold as needed

until stable (min

60 sec)

1h




PD (pC) measurements are performed using 100 ÷ 300 kHz and RIV

(V) using 0.85 ÷ 1.15 MHz frequency ranges.

For units with windings that have multiple connections (e.g., series-

parallel or delta-wye) with each connection having system voltage

>25 kV, two induced tests are performed, one in each connection. If

more than one winding has such multiple connection, then the

connections in each winding shall change between tests. In all

cases, the last test shall be for connection with highest test voltage.

To minimize the effects of external factors and stray capacitances,

the following steps are often relied on:

- filters on the power supply line

- shielding all sharp edges including those at ground potential

as well as the energized and grounded bushings

- turning off solid state power supplies, cranes and other factory

machinery

- removing air bubbles from bushing gas space

- applying pressure to suppress bubbles in the main tank.



ACCEPTANCE CRITERIA

Results are acceptable if:

Nothing unusual associated with sound, current, or voltage

is observed (see abnormal data for details).

The PD (RIV) results during 1h test period have shown:

- Magnitude 500 pC ( 100 V).

- Increase during 1 h 150 pC ( 30 V).

- No steadily rising trends during 1 h

- No sudden sustained increase during the last 20 min.

Judgment should be used on the automatically recorded 5-min

readings so that momentary excursions caused by cranes or

other ambient sources are not recorded. Also, the test may be

extended or repeated until acceptable results are obtained. DGA results (after dielectrics) are normal.




V1 PD1 RIV1 Time1 V2 PD2 RIV2 Time2 V3 PD3 RIV3 Time3

1 0.2 kV 15.4 pC 4.5 µV 00:00:03 0.3 kV 14.6 pC 5.3 µV 00:00:11 0.2 kV 138. pC 4.5 µV 00:00:20 Ambient

2 30.4 kV 24.7 pC 4.5 µV 00:00:49 30.6 kV 26.3 pC 5.3 µV 00:00:58 30.5 kV 27.2 pC 4.2 µV 00:01:07 100%


4 42.4 kV 36.7 pC 5.2 µV 00:03:33 42.0 kV 29.2 pC 7.1 µV 00:03:42 42.1 kV 30.7 pC 4.7 µV 00:03:51 1hr level

5 55.5 kV 31.9 pC 4.9 µV 00:04:03 54.5 kV 33.5 pC 13.5 µV 00:04:12 54.7 kV 33.9 pC 6.2 µV 00:04:21 Enhanced

6 42.3 kV 27.1 pC 4.6 µV 00:00:03 41.9 kV 29.3 pC 5.6 µV 00:00:35 42.1 kV 29.5 pC 4.9 µV 00:01:07 1 hr level

7 42.2 kV 27.3 pC 4.6 µV 00:05:03 41.9 kV 28.0 pC 6.1 µV 00:05:35 41.9 kV 29.8 pC 4.8 µV 00:06:10











18 41.8 kV 28.0 pC 4.6 µV 01:00:03 41.6 kV 29.1 pC 5.1 µV 01:00:35 41.6 kV 30.3 pC 4.8 µV 01:01:07 1 hr level



21 0.3 kV 18.2 pC 4.6 µV 01:04:38 0.3 kV 11.6 pC 5.2 µV 01:04:47 0.3 kV 12.0 pC 4.9 µV 01:04:56 Ambient



ABNORMAL DATA

Results are not acceptable if the pC (or V) data exceeds any

of the required criteria, and no reasonable/acceptable

justification for the source/cause is provided.

Other tests, e.g., acoustic PD, DGA, can provide confirmation

that a source of excessive partial discharge is present.

The presence of smoke and bubbles rising in the oil, audible

sounds such as thump, sudden increase in test current or

voltage collapse may all serve as a confirmation that

abnormal PD results are associated with a failure. ©Doble Engineering Company



If pC (or V) data exceeds the limits, and all the attempts

to identify and eliminate external PD sources are not

successful, a longer standing time, long duration PD

test, degassing of oil, refilling transformer under

vacuum or a heatrun test (if one is specified) are often

successfully bring the PD data within limits.

A failure to meet the partial discharge acceptance

criterion shall not warrant immediate rejection, but it

shall lead to consultation between purchaser and

manufacturer about further investigations.


and/or repeating of the test and/or other tests reveal the

failure, the oil is drained and internal inspection is

performed.


NO-LOAD LOSSES AND

EXCITATION CURRENT,

after dielectrics (Routine*)

*The test is not required by standards and no test type is

assigned to it; however, it is a wildly recognized as

standard practice and performed as routine.


NO-LOAD LOSSES AND EXCITATION CURRENT after dielectrics:

OBJECTIVE

Objective: No-load loss and excitation current, measured

at 100% and 110% of the specified voltage and frequency

after all dielectric tests are completed, provide additional

confirmation that no damage, created by dielectric tests, is

present in the transformer. If this is the last power test to

be performed, it also serves to demagnetize the core for

subsequent low-voltage tests, e.g., 10-kV exciting current

and sfra.


NO-LOAD LOSSES AND EXCITATION CURRENT after dielectrics:

ACCEPTANCE CRITERIA AND RECOURSE IF DATA ABNORMAL

No-load losses measured after dielectric tests are

compared with the results obtained before dielectric tests.

The 5% difference is often used as an acceptable criteria. Difference between the before and after data could be due

to:

Changes in the inter-laminar insulation

Temperature

Sometimes the change after initially exceeding 5% goes

away with time.

Failure to meet before and after dielectrics comparison

criteria should not warrant immediate rejection but shall

lead to consultation between purchaser and manufacturer

regarding further investigation of possible causes and

consequences.


LOAD LOSSES AND

IMPEDANCE VOLTAGE (Routine) ©Doble Engineering Company

LOAD LOSSES AND IMPEDANCE VOLTAGE:


Definition: The load losses of a transformer are losses

associated with a specified load and include:

windings I2R losses due to load current

stray losses due to eddy currents induced by leakage flux in

the windings, core clamps, magnetic shields, tank walls, and

other conducting parts. Stray losses may also be caused by

currents circulating in parallel windings or strands.

Load losses do not include control and cooling losses.

The impedance voltage of a transformer is the voltage required

to circulate rated current through two specified windings with

one winding short-circuited.



DEFINITION AND OBJECTIVE (cont.)

Objective: The impedance and load losses test provides the

data for:



characteristics. Since these parameters have often an

economic value attached to them, the accuracy of the

measurement becomes significant.

Maximum load losses are used as test parameter during

the temperature rise test.

Impedance voltage is an essential input parameter in power

system studies (e.g., load flow, transformer parallel

operation, short-circuit calculations).



PHYSICS

To create conditions when losses are limited to I2R and stray

losses, and applied voltage is equal to the voltage drop across a

loaded transformer, one winding is short-circuited and voltage is

raised until rated current is reached. The flux path is then

dominated by the leakage channel where the eddy losses in

various conducting components in the FL path are induced.

R

LV HV

FM

Vrated

Iexc R

LV HV

Irated

I2R lossesT

Eddy currents

creating losses1/T

FL

Vsc

Note: Resistance R and short

circuit of LV is not shown.



PHYSICS (cont.)

Rm Xm

XHV RHV XLV RLV

VSC

XL RL

Irated

VR_L VX_L

Irated VR_L

VX_L

VSC

Corresponds

to load loss

Corresponds

to leakage-flux

linkages of the

windings

Measured

ZSC

For most power transformers, VX_L >> VR_L.

Angle is close

to 90, requiring

high accuracy

test systems.

Compensating variable

capacitor Cc is adjusted

to reduce the input

current. VSC Iinput

IC

Irated

IC

Iinput

CC




Applied voltage is adjusted until rated current is present in the excited winding.

After data is recorded, if necessary, correction for losses in external circuit is made.

If three line currents are not balanced the average RMS value should correspond to the desired value.

The duration of the test should be kept to a minimum to avoid heating up winding conductors.

Transformer

in test CT

VT

W

I

V

A

X0

H2

X1

X2

X3

H1

H3

V

If taps are present, the following

combinations of voltage ratings are tested:

3

DETC rated rated rated max max max min min min

LTC N max min N max min N max min




For 3-wdg units, three sets of measurements are performed using three pairs of windings, producing Z12, Z13, Z23 and P12, P13, P23. Solving shown equations, determines Zi and Pi of each branch.

For test, the current is set based on capacity of the winding with lowest MVA in the pair.

When results are converted to %, all data is given based on MVA of HV winding.

1

2

3

1 2

3

Z1

Z2

Z3

Z12 = Z1 + Z2

Z13 = Z1 + Z3

Z23 = Z2 + Z3

Z1 = (Z12 + Z13 – Z23)/2

Z2 = (Z12 + Z23 – Z13)/2

Z3 = (Z13 + Z23 – Z12)/2




Since stray and I2R losses have different dependencies on T, each need to be obtained from measured losses, individually converted from test T to rated T before combined again in reported load losses. V is also converted to rated T.

Convert Rdc from

TR_test TLL_test

Measure A, V, W, T

Correct W and V

from measured

amps to rated

Calculate I2R losses

at TLL_test

Calculate stray

losses at TLL_test

(W - I2R)

Convert I2R losses

from TLL_test Trated

Convert stray

losses from

TLL_test Trated

Calculate total

losses at Trated

(stray + I2R)

Calculate %Vsc

(V / Vrated)100 = %Zsc

Correct V from

TLL_test Trated



ACCEPTANCE CRITERIA

The total losses (no-load + load) should not exceed the

guaranteed value by more than 6%.

For 2-wdg units, if Zsc>2.5%, the tolerance for measured

impedance is +/-7.5% of the guaranteed value, otherwise, it is +/-

10%. The tolerance for comparison of duplicates units produced

at the same time is +/-7.5%.

For 3-wdg units, autotransformers or units having a zigzag

winding, tolerance for measured impedance is +/-10% of the

guaranteed value. The tolerance for comparison of duplicates

units produced at the same time is +/-10%.


Thermal stability prior to test: TTO-TBO 5C.

Average of T readings (Tave_oil) before and after the test

should be used as test T. Their difference must be 5C.

Frequency is within +/-0.5% of rated.

Test system accuracy should be within +/-3% for loss, +/-0.5%

for voltage, current and RDC, and +/-1.5C for T.



ABNORMAL DATA


include:


Production process related factors or mistakes

Influence of temperature was not properly accounted for

Accuracy of measurements

Example: guaranteed load loss - 94 kW, measured – 110 kW




Failure to meet the load losses and impedance test

criteria should not warrant immediate rejection but shall

lead to consultation between purchaser and

manufacturer regarding further investigation of possible

causes and consequences.

The acceptance criteria of 6% for total losses does not

replace the manufacturer’s guarantee of losses for

economic loss evaluation purposes.




Factory and field results

cannot be compared

Factory losses are measured

under 3-phase excitation, at

rated current and reported as

sum of three phases I2R and

stray losses.

Field losses are measured

under 1-phase excitation, at

current much lower than rated

and reported as per-phase I2R

and stray losses.



COMPARISON WITH FIELD DATA (cont.)

*Since test is confined to leakage channel (where reluctance is determined by air/oil) the leakage inductance (L=/I), remains the same regardless of the current level.

Experience shows that a combined influence of different instrumentation

and test setups, difference in flux distribution under 3- and 1-phase

excitation, presence of the resistive component and averaging of factory

data can result in differences ranging from nearly perfect (<1%) to up to

6% (of the measured value).

However, the differences between factory and field test conditions

notwithstanding, the ZNP can serve as a useful guideline for evaluating

the initial value measured in the field. If, during initial test, the field per-

phase tests deviate from average (of three readings) by <3% of the

measured value, results normally are considered acceptable. The initial

per-phase test should serve as a benchmark for future testing with

acceptable difference from the initial field test being <2%.

Factory short-circuit impedance is

reported as average of three

phases, obtained at rated current*

under 3-phase excitation.

Field leakage reactance is reported

as per-phase reactive component of

the short-circuit impedance,

obtained at current* much lower than

rated under 1-phase excitation.


TEMPERATURE RISE (Design and other)


TEMPERATURE RISE:


Definition: The temperature rise is a test that verifies

transformer thermal performance through determination

of winding and oil temperature rises over ambient.

Objective: The temperature rise test provides the top-oil

rise, winding average rise and winding hot-spot rise over

ambient for:



characteristics.

Provides data for calculation of potential MVA margin.

Setup of various temperature monitoring instruments

and cooling control.


TEMPERATURE RISE:

PHYSICS

Measured:

Tto, Tt_rad, Tb_rad, Ta.

Tto Tt_rad

Tb_rad

Rad

Tb_rad

Oil

Winding

Tt_rad

Tto-a

Tw_ave*

Ta

Tw_ave-a

Tto

LV HV

Ta

T

To_ave

GRAD

Main tank

Core

Tw_hs

Ta

Ta

height

Calculated: Tto-a, Tw_ave-a, Ths-a, GRAD

*The term “winding average T rise”, Tw_ave-a, is not the T at any given point in a winding

nor is it an arithmetic average of results determined from different terminal pairs. It refers

to the value determined by measurement on a given pair of winding terminals.

Needs NL+LL

losses Need rated

current

Ths-a

Located at 3 locations around xfmr at mid-height level.


TEMPERATURE RISE:


Total losses (NL+LL) and winding cold resistance data should be available.

Test is performed for min and max MVA, and in a combination of DETC/LTC positions, producing highest load losses.

10Tamb40C and measured in containers with liquid, having a time constant as per C57.12.90-2010.

Test contains 3 key segments: - total loss run (to include 3 hr of thermal stability) - rated current run (1 hr) - hot resistance measurement (e.g., 10-20 min after shutdown)

Transformer

in test CT

VT

W

I

V

A

X0

H2

X1

X2

X3

H1

H3

V ©Doble Engineering Company

TEMPERATURE RISE:


Tto, Tt_ rad

Tb_ rad

Ta_ ave

Tto-a

t [h]

T[C]

Preceding ONAN

Cutback

ONAF shutdown

Xfmr

energized

for ONAF

ONAN

shutdown

Rhot

measurement

begins

Rated current run

Steady-state oil

T rise (change of Tto-a in 3h

1C or 2.5%

whichever is greater)

Itest

Ptotal

Irated

Measurement before

cutback determines *Tto-a

*Tto-a is corrected for difference between required and

actually used total losses (it must be 20%) and for altitude.

Total loss run

1h


TEMPERATURE RISE:


Rhot

t [min]

Rhot as function of time in

presence of decreasing

temperature is recorded

t 4 min

Tw_hot = Rhot/Rcold(234.5 + Tw_cold) – 234.5

Rhot calculated at t = 0

*If two windings are tested simultaneously in series,

the Idc is selected based on the lowest rated current.

*Instrument

connected

Instrument output

current reached

pre-selected level Flux

stabilized

t = 0

Voltage

removed

Objective: resistance of

winding at the time when

load current is still present


TEMPERATURE RISE:


Tw_hot

**To_ave_cb-a is corrected for difference between required and actually used total losses (it must be 20%)

and altitude.

Ta

**To_ave_cb-a

GRAD

Tw_ave-a

Comparison with guaranteed values,

e.g., Tto-a and Tw_ave-a 65C; Ths-a 80C

*GRAD To_ave_sd

*GRAD is corrected for difference between required and actually used load current (it must be 15%).

Tto-a

GRAD

GRAD correction for

localized hot spot

eddy currents ***Ths-a

Value used for

setting winding

T monitors

***This a simplified representation of Ths_a determination; actual design calculation is more involved.


TEMPERATURE RISE:


Series autoxfmr

windings with LTC in N LV winding

idc

vLV

idc

vLV X2 X0

i2

i1

During shutdown at the time of the first Rdc reading, the flux must be

stabilized so that resistance change is caused only by reduction in

temperature. It’s true in most cases, unless series autoxfmr is present.

X2 X0 idc

i1

i2

idc

idc

Main unit

core

Series autoxfmr

core

F F1 F2


TEMPERATURE RISE:


0.00462

0.00463

0.00464

0.00465

0.00466

0.00467

0.00468

0.00469

0.0047

0:00:00 0:00:43 0:01:26 0:02:10 0:02:53 0:03:36 0:04:19 0:05:02 0:05:46

Time [min]

LV

cir

cu

it R

dc

[o

hm

]

0.25

0.26

0.27

0.28

0.29

0.3

0.31

0.32

HV

circ

uit R

dc

[Oh

m]

X 0-X 2

H1-H2

Setup

1.5 min

Voltage

removed

t = 0

Time remaining

for stabilization

= 2.5 min

At t = 4 min, data

collection begins

with flux in main

core stable while

flux in series core

still changing.

Series autoxfmr should be

excluded from both cold

and hot Rdc measurements.


TEMPERATURE RISE:

ACCEPTANCE CRITERIA

The winding average T rise over ambient for all tested windings

should not exceed the guaranteed value, e.g., 65 or 55C.

The top-oil T rise over ambient should not exceed the guaranteed

value, e.g., 65 or 55C.

The winding hot spot T rise over ambient for all tested windings

should not exceed the guaranteed value, e.g., 80C for 65C rise

units and 65C for 55C rise units.

If shutdown is performed on each phase, results of winding average

rises should be comparable (rule of thumb: 4C difference,

presently, there is no limit in the standard).

DGA results (after heatrun) should be normal.

It is always useful to perform and review thermal scanning of all

tank walls and the cover in search for excessive overheating

(100C rise). Request image files to be provided with the certified

test report and have software to view them.

If agreed with manufacturer, the heatrun is a good time to check the

performance of temperature controllers (using a preliminary winding

T gradient) and turns ratio of CTs.


TEMPERATURE RISE:


To assure test data is credible, verify that:

T and current requirements for measuring winding cold Rdc were

complied with.

Test is performed using maximum load loss and in corresponding

DETC/LTC positions.

Test instrument type and setup used for cold and hot resistance

was the same, e.g., if two-channel measurement is used it must

be used for both hot and cold resistance tests.

If series auto is present, unless it is shown that RDC can be

measured within shutdown time constrains, the auto is excluded

from the resistance measurement*.

During shutdown, fans are turned off right after transformer is de-

energized.

The first value of winding hot Rdc was recorded not later than 4

min after shutdown.

*Lachman, M. F., et al “Impact of Series Unit on Transformer Winding DC Resistance Measurement During Heatrun”, Proc. of the Seventy-Sixth Annual Intern. Confer. of Doble Clients, 2009, Sec. T-4.


TEMPERATURE RISE:


To assure test data is credible, verify that:

Winding hot Rdc fits reasonably into the cooling curve.

Final T rises are properly corrected: GRAD for actual test

currents, Tto-a and To_ave_cb-a for actual total losses and altitude.

Test system accuracy should be within +/-3% for loss, +/-0.5% for

voltage, current and winding resistance, and +/-1.5C for

temperature.

If the test could not be done at rated frequency, the results are

converted from tested to rated frequency (see C57.12.90-2010,

Annex B). However, the fans/pumps should be operated at the power

frequency to be used when unit is in service. ©Doble Engineering Company

TEMPERATURE RISE:

ABNORMAL DATA


include:


Testing/setup mistakes

Presence of series auto-transformer

Example: guaranteed Tw_ave-a – 65C, measured – 67C

TLV_hot=[(234.5+30)4.534/4.081]-234.5=59.4ºC

y = 8.152E-07x2 - 3.235E-05x + 5.362E-03

0.005

0.00505

0.0051

0.00515

0.0052

0.00525

0.0053

0.00535

0:00

0:30

1:00

1:30

2:00

2:30

3:00

3:30

4:00

4:30

5:00

5:30

6:00

6:30

7:00

7:30

8:00

8:30

9:00

9:30

10:0

0

Time [min]

Re

sis

tan

ce

[o

hm

s]

TLV_hot=[(234.5+30)5.362/4.621]-234.5=72.5ºC

With series auto-xfmr

y = 6.962E-08x2 - 5.974E-06x + 4.534E-03

0.00442

0.00444

0.00446

0.00448

0.0045

0.00452

0.00454

0:00

0:30

1:00

1:30

2:00

2:30

3:00

3:30

4:00

4:30

5:00

5:30

6:00

6:30

7:00

7:30

8:00

8:30

9:00

9:30

10:0

0

Time [min]

Re

sis

tan

ce

[o

hm

s]

Without series auto-xfmr

Note: The example shows a quadratic function, the suitability of which was confirmed via direct fiberoptic measurements and other methods, e.g., Blume. Different functions may be used if they fit the winding behavior.


TEMPERATURE RISE:


Failure to meet the temperature rise test criteria should

not warrant immediate rejection but shall lead to

consultation between purchaser and manufacturer

regarding an investigation of possible causes and

solutions to address the problem.


ZERO-PHASE SEQUENCE

IMPEDANCE (Class I - design

Class II - routine) ©Doble Engineering Company

ZERO-PHASE SEQUENCE IMPEDANCE:


Definition: The zero-phase sequence impedance is

impedance to the single-phase current simultaneously

present all three phases. It is measured from a wye or a

zig-zag connected winding between three phase

terminals connected together and the neutral terminal.

Objective: The zero-phase sequence impedance serves

as input in analysis of unbalanced three-phase system

using symmetrical components method.



PHYSICS

In symmetrically loaded 3-phase

system, only one phase needs to be analyzed since in other phases values have the same magnitudes and only have to be shifted by 120.

In unbalanced 3-phase system, impedances in each phase are different and each phase needs to be analyzed separately.

Method of symmetrical components converts any unbalanced system into 3 balanced systems, namely positive, negative and zero-phase sequence systems.

After these are defined, the voltages and currents in the original unbalanced system are reconstructed.

Ia

Ib

Ic

Unbalanced

Ia

Ib Ic

Balanced

Ia1

Ib1

Ic1

Positive

Ib2 Ia2

Ic2 Negative

Iao

Ibo

Ico

Zero




For xfmr, the Z1 = Z2 = Zsc is known from impedance/load losses test. In zero-phase sequence system, the phase currents are in-phase with each other and flow through the xfmr only if there is a path to return to the grounded source or to circulate while satisfying the Kirchhoff’s current law.. Therefore, this test applies only to transformers with one or more windings with a physical neutral brought out for external connection.

Z0

1 2

N

1 2




Z1Ns, Z1No, Z2No are used to calculate Z1, Z2 and Z3.

If delta winding is not present, the currents shown in delta are circulating in the tank.

Z1Ns

Z1 Z2

Z3

1 2

N

Z1No

Z2No




If no delta winding is present, applied voltage should be 30% of rated Vphase_gnd and measured current Irated.

If delta winding is present, the applied voltage should be such that current in delta winding Irated.

For Y/ or /Y impedance in % is determined as: Z0 = 300(Vmeas / V r) (Ir / Imeas)

For Y/Y and autoxfmr with or without tertiary , the elements of the equivalent circuit are further determined as:

Z1 = Z1No - Z3

Z2 = Z2No - Z3

Z3 = Z2No ( Z1No - Z1Ns)

Transformer in

test

CT

VT

W

I

V

A V

1



ACCEPTANCE CRITERIA

The standard does not provide an acceptance criteria for the zero-

phase sequence values. However, the following general guidelines can

be useful (typical for 230 kV, 200 MVA core type units):

For /Y, Z0 Zsc or slightly less. Example: 50 MVA, 161/69GndY kV, Zsc = 21.9%, Z0 = 21.8%

For Y/ units, Z0 (0.8-1.0)Zsc. Example: 48 MVA, 235.75GndY/13.8 kV, Zsc = 9.9%, Z0 = 8.5%

For Y/Y/ or autoxfmrs with delta, Z1 (0.7-1.0)Zsc; with Z2

typically <1.0% or sometimes <0. Example: Auto, 18 MVA, 230GndY/60GndY/21 kV, Zsc = 4.9%, Z1 = 3.6%,

Z2 = 0.84%, Z3 = 10%

50 MVA, 69GndY/34.5GndY/13.2 kV, Zsc = 7.8%, Z1 = 6.7%,

Z2 = -0.16%, Z3 = 4.6%

For Y/Y and autoxfmrs without delta (rare occasion), magnetic

flux has a strong coupling to the tank, making, in general, the

relationship between voltage and current non-linear and the

above observations not relevant. Example: Auto, 75 MVA, 115GndY/34.5GndY kV, Zsc = 12.7%, Z1 = -9.9%,

Z2 = 27.4%, Z3 = 205.6%



ABNORMAL DATA

If unusual zero-phase sequence impedance data is

obtained the test process should be reviewed (paying

particular attention to voltages and currents used) along

with comparing the measured data with the calculated

design values.




Unusual zero-phase sequence impedance data does not

warrant a unit rejection but should lead to a consultation

between purchaser and manufacturer to understand the

possible causes and consequences.


AUDIBLE SOUND LEVEL (Design and other)


AUDIBLE SOUND LEVEL:


Definition: The audible sound level test is the measurement

of the sound pressure level around a fully assembled

transformer under the rated no-load conditions with cooling

equipment operating as appropriate for the power rating

being tested.

Objective: To protect the population from noise

inconveniences transformers are required to operate within

specified noise limits. The audible sound level test provides

the sound pressure level data for:



characteristics.

The test also serves as a quality control tool as the sound,

driven by the vibratory motion of the core, is transmitted to

the tank through direct mechanical coupling as well as is

produced by pumps and fans of the cooling system.



PHYSICS

Most of xfmr sound is generated by the core. When the core steel magnetized/demagnetized twice each cycle, the steel elongates and shortens due to a property called magnetostriction.

This produces a vibratory motion in the core transmitted to the tank through the core mechanical support and the pressure waves in the dielectric fluid. At the tank this motion radiates as an airborne sound. The vibration magnitude depends on the flux density and magnetic property of the steel.

The frequency spectrum of the sound contains mainly the even harmonics of the power frequency, i.e., 120, 240, 360, etc. The audible sound also includes a contribution emitted by pumps and fans, containing a broadband spectrum of frequencies.

Core

Dielectric

fluid

Tank

Direction of

dimensional

change Magnetostriction

caused by

domain rotation

F




Xfmr is energized with no load, at rated

(for the tap used) voltage and frequency, with tap changer on principal tap and pumps/fans operated as appropriate for the tested rating.

On certain tap positions, xfmr may produce sound levels greater than at the principal tap, e.g., engaging PA and/or series autoxfmr. Test will be performed in these positions if specified by customer.

The voltage should be set as during no-load loss test, based on Vave.

At least one test should be performed at the cooling stage for the min rating and one test at the cooling stage for max rating.

Measurements begin when xfrm reaches steady-state conditions, i.e., to allow magnetic bias to decay.

VT

Vave

Transformer

in test

X0

H2 X1

X2

X3

H1

H3

3




Microphones are located on the measurement surface at shown distance from reference sound-producing surface.

Xfmr is placed so that no acoustically reflecting surface is within 10 ft of the microphone.

If transformer H<7.9ft, measurements are made at H/2; if H7.9 ft, at H/3 and 2H/3.

First measurement is made at drain valve proceeding clockwise.

Reference sound-producing

surface is a vertical surface

following the contour of a taut

string stretched around xfmr

periphery.

Tank

LTC Drain valve

Radiator

6 ft 3 ft

1 ft

#1

Fan cooled surface

Measurement

surface

Microphone

location




The sound power rating of a transformer is determined

using one of the following three measurement methods:

A-weighted sound pressure level (most frequent)

One-third octave sound pressure level (when

specified)

Narrowband sound pressure level (when specified) ©Doble Engineering Company



Human ear can hear sounds in 20÷20000 Hz range. However, it detects some frequencies much easier than others. This uneven frequency response needs to be considered when the annoyance of unwanted sounds is to be evaluated.

To account for human’s greater sensitivity to noise at some frequencies relative to other, the measured data is passed through a weighting filter. A-

weighting is most commonly used to allow for a broad peak between 1÷6 kHz but very strongly discriminating against low frequencies.

As a result, when the average sound pressure

level is calculated, the influence of frequencies not impacting the human hearing perception is minimized.

Hz 63 125 250 500 1000 2000 4000 8000

A-filter -26 -16 -19 -3 0 1 1 -1

dB (measured) 67 76 73 70 65 66 62 52

dB (A-weighted) 41 60 64 67 65 67 63 51

A-weighted sound pressure level




The following two methods are used when a more detailed

investigation into the sources of noise is required:

In one-third octave sound pressure level measurement,

each octave band in the spectrum (i.e., 63, 125, 250, 500,

1000, 2000 and 4000 Hz) is split into three, with each “1/3

sub-band” (e.g., 63, 80, 100, 125, 160, 200, 250 Hz, etc.)

being evaluated individually.

The narrowband sound pressure level measurement is

performed at the power frequency (e.g., 60 Hz) and at

least at each of the next six even harmonics (120 Hz, 240

Hz, 360 Hz, 480 Hz, 600 Hz, and 720 Hz). Once again, each

frequency is evaluated individually.




The sound power rating is determined using the following steps:

Measure ambient sound pressure levels. This is established as an

average of measurements at a min of four locations immediately

preceding and immediately following the sound measurements

with the unit energized.

Measure combined transformer and ambient sound pressure level.

Measurements are made if ambient level is at least 5 dB or more

below the combined transformer and ambient sound pressure

level.

Compute ambient-corrected sound pressure levels. For

corrections see Table 7 in C57.12.90-2010.

Compute average sound pressure levels [in dB(A)]:

𝑳𝒑 = 𝟏𝟎𝒍𝒐𝒈𝟏𝟎𝟏

𝑵 𝟏𝟎

𝑳𝒊𝟏𝟎

𝑵

𝒊=𝟏

Li is the sound pressure level measured at ith location by one of the 3

measuring methods. Sound power levels are calculated when requested.



ACCEPTANCE CRITERIA

Computed average sound pressure level should not exceed the

audible sound levels as listed in NEMA TR1-1993, Tables 0-2 and 0-3

or as requested in customer test specification. Rectifier, railway,

furnace, grounding, and mobile transformers are not covered by

these tables.


The sound pressure measuring instrument should meet the

requirements of ANSI S1.4 for Type 1 meters.

The sound pressure measuring instrument should be calibrated

before and after each set of measurements. If calibration change

>1dB, sound measurements shall be declared invalid, and the

test repeated.

Verify that microphones were positioned at required

distances/heights, pumps/fans were operated as required for

tested power rating and voltage set based on Vave.

Verify that the ambient level was at least 5 dB or more below the

combined transformer and ambient sound pressure level.

If rated frequency is not used, 50/60 Hz conversion is applied.



ABNORMAL DATA


include:

Problems with measurement, e.g., ambient noise,

positions of microphones, sound instrument calibration,

voltage adjustment, surrounding reflecting surfaces,

etc.

Variability in core steel characteristics

Different core steel


Assembly related factors or mistakes

Example: guaranteed sound pressure level per NEMA –

75/77/78 dB(A), measured – 77/78/79 dB(A)




Failure to meet the audible sound test criteria should not

warrant immediate rejection but shall lead to consultation

between purchaser and manufacturer regarding an

investigation of possible causes and solutions to address

the problem.


CORE DEMAGNETIZATION (Routine*)

*This procedure is not required by standards but is a wildly

recognized as standard practice and performed as routine.


CORE DEMAGNETIZATION:


Definition: The core demagnetization is the process of

removing the magnetic bias in the core through a series

of steps, with each subsequent step creating magnetic

field of opposite direction and lower intensity. The first

step must bring the core to the main hysteresis loop with

the last step, upon removal, leaving no residual

magnetism in the core.

Objective: The core demagnetization creates conditions

for obtaining the low-voltage exciting current and loss

test as well as sfra benchmark data not affected by

residual magnetism .



PHYSICS

If in the presence of residual

magnetism Br, the voltage is increased

from zero, the flux varies around

minor hysteresis loops. The negative

tip of these loops lies on the main

loop. The greater the voltage, the

smaller is the offset of the minor loop

along the B axis. The bias is removed

when the main loop, symmetrical

around the origin, is reached.

Main

loop

H

B

If after reaching the main hysteresis

loop, the voltage is gradually reduced,

each minor loop will lie inside the

previous larger loop. Reduction of

voltage to zero brings working point

to the center of these loops resulting

in a demagnetized transformer.

Br = 0

Br

H

B




The core demagnetization can be performed by one of the

following:

Applying rated 3-phase voltage (holding for 5-10 min)

and reducing gradually to zero.

Applying DC voltage (e.g., 12 V), waiting until current

stabilizes, then switching voltage polarity and holding

until current reaches a lower value; this process

continues until current level is zero

Without ammeter, the above approach can be applied

but a lower level of current is reached by applying

alternate polarities of DC voltage for progressively

shorter periods of time.

If no-load losses or sound level tests are the last power

tests to be performed, they serve the function of the

core demagnetization process.



RELATIONSHIP WITH LV DIAGNOSTIC DATA

When xfmr is de-energized, the

core is constantly looking for a

state of lower energy, i.e., it

relaxes, changing its magnetic

state and moving away from the

condition immediately following

demagnetization*. This is obvious

in the low-frequency range of the

sfra trace but not in the low-voltage

excitation current data. These sfra

changes are normal and

diagnostically insignificant.

Factory

Field

Field

Factory

Data movement

with no excitation

applied between

measurements

Factory Field

mA W mA W

20.5 128 20.5 126

9.3 61 9.6 58

20.7 131 21.6 131

72 hr

24 hr

9 hr 6 hr

3 hr

1 hr

30 min

dm_init

Controlled

experiments showing

data movement*

*Lachman, M. F., et al “Frequency Response Analysis of Transformers and Influence of Magnetic Viscosity”, Proc. of the Seventy-Seventh Annual Intern. Confer. of Doble Clients, 2010, Sec. TX-11.


THE END

LAST SLIDE


UNDERSTANDING THETRANSFORMER TEST DATA

Barry M. Mirzaei – P.Eng.Hydro One

September 2012 – Chicago


2Understanding The Transformer Test DataSeptember 2012


3

No Load TestTest object is supplied from one side of the transformer(L.V.), the other side (H.V.) is left open circuit. Test voltageto be adjusted to the pre‐determined value(s)

Typical test voltage is 90% ‐ 100% and 110% of the ratedvoltage

Characteristics of the No Load Test:“Low Current – High Voltage”

Understanding The Transformer Test DataSeptember 2012


4

Induced Voltage Test

Test object is supplied from one side of thetransformer (L.V.), the other side (H.V.) is left opencircuit. Test voltage to be adjusted to the pre‐determined value(s)

Twice the rated voltage is applied for 7200 cycles for transformers with uniformly insulated windings

Characteristics of the Induced Voltage Test:

“Low Current – High Voltage and Frequency > 60”



5

Load Loss TestTest object is supplied from one side (H.V.), theother side (L.V.) is short‐circuited. Test voltage isadjusted to apply the rated current to the testobject

Load Loss:‐Resistive losses or R‐Eddy current losses in the windings‐Stray losses in leads, core plates and tank

Characteristics of the Load Loss Test:“High Current – Low Voltage”



7

Hysteresis Loss

Proportional to the frequencyand dependent on the area ofthe hysteresis loop, which, inturn, is a characteristic of thematerial and a function of thepeak flux density



8

Eddy Current Loss

Dependent on the square of frequency but is also directly proportional to the square of the thickness of the material



9

4.44 (a)

(b)

Voltage

PNL NoLoadLosses= HysteresisLoss= EddyCurrentLoss, = Coefficients

= = Exponentwithinduction




Minimizing hysteresis loss thus depends onthe development of a material having aminimum area of hysteresis loop.

Minimizing eddy current loss is achieved bybuilding up the core from a stack of thinlaminations and increasing resistivity of thematerial in order to make it less easy foreddy currents to flow.


STATEMENT OF THE ISSUE:



12

Deflection in the readings on meteringdevices (watt meters, …) were reportedwith the noise.

During the No Load test of a rebuilt 3phase 135 kV transformer in the factory,loud noises inside the tank were reported.

Not a Hydro One Asset

The noises were described as similar to“release of large amounts of air bubblesinside the oil”, started at around the 25%of the test voltage.



13

Solutions?What test data are available?

What those test data really mean?



14

Criteria & constraints for addressing the issue

Un‐necessary activities to beavoided, delivery date was critical

Un‐tanking the transformer is costly and should be avoided if there is no clear understanding about the issue

Insulation tests should not be repeated, if there is no need to do so



15

Investigation Procedure

1Oil Sample

OK 2Repeat TTR & DC Resistance

OK

3Observe The No Load TestPROBLEM

4Apply Load

TestOK

5Insulation Test?Not

Convinced to apply

6Apply reduced “Induced Voltage”




The Induced Voltage Teststresses all parts of theinsulation system, includingturn to turn, phase to phaseand winding to ground.©Doble Engineering Company

17

4.44 (a)

Concept of customized Induced test:

By applying induced voltage up to rated voltage, basically the no loadtest is being repeated with reducedinduction in the core




.

.x x x x

.

.x x x x

x x x x



Investigation Procedure

1Oil Sample

OK2

Repeat TTR & DC Resistance

OK

3Observe The No Load TestPROBLEM

4

Apply Load TestOK

5Insulation Test?

Not Convinced to apply

6Apply reduced

“Induced Voltage”

Load Loss:‐Resistive losses or R‐Eddy current losses in the windings‐Stray losses in leads, core plates and tank

Eddy Current Loss

Dependent on the square of frequency but is also directly proportional to the square of the thickness of the material

Hysteresis Loss

Proportional to the frequency and dependent onthe area of the hysteresis loop, which, in turn, is acharacteristic of the material and a function of thepeak flux density



During “Core Stacking Process” –Holes built for Core Bolts, used for proper core stacking



Core Plates Core Bolts

Photo belongs to another transformer



Core Plate

Core Bolt

Weld



In this case, the low impedance path formed by thebolts and the core clamping plates causes a local shortcircuit path which produces intense local eddy currents.The amount of heat generated by this phenomenon issufficient to considerably damage the adjacent areas.

The problem was noticeable in No‐Load test since there was higher induction to create higher current in the through bolts when compared to reduced induced test.

Increase in the Load Loss increased the probability of “Core Plates” related issues.



This picture shows the correct insulation of the core boltsPhoto belongs to another transformer .

Insulation Material©Doble Engineering Company

Understanding Transformer

Factory Testing

September 30, 2012


Understanding Transformer Factory Testing 2

On some occasions additional methods must be employed to determine the suitability of tested transformer.

These techniques may include calculated corrections or multiple tests at different loading conditions, etc.

Lets look at two actual factory cases: Case 1: Good Test Results – Bad Data Case 2: Bad Test Results – Good Transformer

Transformer Temperature Tests



Core

Coils

Oil

AVERAGE WINDING

TEMP. WINDING HOTTEST

SPOT

TOP OIL

TEMP.

Cooling

Dis

tanc

e

Temperature

AverageOil

BottomOil

Avg. Wdg. Temp.

Hot Spot

Top Oil Temp.

Gradient

TopOil

Ambient

Gradient . x H.S.F

TEMPERATURE DISTRIBUTION

Transformer Temperature Tests



UAT 39/52/65 MVA; 230 - 6.9 (XV) & 4.16 (YV) kV 60Hz (+15-5% LTC for YV)

• Heat run test was performed according to ANSI/IEEE Standards and the clients technical specification.

• Temperature results were well below Standard limits and according to client’s specification.

• Very Clean DGA Results.

• Test Results did not match design data?

Case 1: Good Results – Bad Data



Short-Circuit Method – Three Phase, 3-Winding:

Measurement System Unit Under Test

YV

HV

XVL3

L2

L1

~

SHO

RT

CIR

CU

ITSH

OR

T C

IRC

UIT

Case 1: Good Results – Bad Data



Case 1: Loading Cycle



Test Corrections Limit Total Losses (kW ) 172.600 228.464Tap Position 1R 1R Average Oil Rise 42.9 51.2Top Oil Rise 49.6 59.2 65.0Winding Gradient, YV 10.2 10.2Winding Gradient, XV 2.3 2.3Winding Gradient, HV 3.3 3.2HS over TOR, YV 11.2 11.2HS over TOR, XV 2.5 2.5HS over TOR, HV 3.6 3.5Hot Spot Factor, YV 1.10 1.10Hot Spot Factor, XV 1.10 1.10Hot Spot Factor, HV 1.10 1.10Average Winding Rise, YV 53.05 61.3Average Winding Rise, XV 45.13 53.4Average Winding Rise, HV 46.18 54.4Hot Spot Rise, YV 60.8 70.4Hot Spot Rise, XV 52.1 61.7Hot Spot Rise, HV 53.2 62.7

65.0

80.0 n: 0.63m: 0.80

Exponents

Case 1: Heat Runs Results

Winding Test (A) Rated (A) RatioHV 99.5 97.9 0.984XV 1757.0 1757.0 1.000YV 2483.0 2483.0 1.000

Winding Current for Individual Gradient Runs


Winding Current for Oil Rise Run



MVACooling Mode

Tested DesignLosses (kW) 228.682 228.464 Guar.Top Oil Rise 59.2 51.6 65.0Average Oil Rise 51.2 39.7Bottom Oil Rise 39.8 27.9

Gradient 10.2 12.8Average Winding Rise 61.4 52.5 65.0Hot Spot Gradient 11.2 14.1Hot Spot Rise 70.4 65.7 80.0



Heat Run Result vs Design Data

YV Winding

HV Winding

39.0ONAN

XV Winding Why so far off?

Case 1: Temp Rise not Expected



∆T (3 Hr) > 2 ºC

Time CH 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 0:00 1:00 2:00 3:00 5:00Measured kW 171.0 173.2 172.0 171.1 173.0 172.9 173.6 172.0 172.0 172.3 172.5 172.6 Take Take TakeMeasured Amps 101.6 102.0 103.0 101.5 102.0 102.0 100.0 99.0 99.1 99.4 99.6 99.5 HV XV YVUpper Radiator 1 2 77.43 78.66 79.56 80.10 81.55 82.66 83.83 84.18 84.57 84.73 84.92 85.09 84.98 82.27 83.61Upper Radiator 2 7 77.58 78.78 79.78 80.30 81.88 82.68 84.65 85.03 85.42 85.51 85.58 86.10 85.70 82.91 84.36

Average Upper Rads 77.51 78.72 79.67 80.20 81.72 82.67 84.24 84.61 85.00 85.12 85.25 85.60 85.34 82.59 83.99Lower Radiator 1 1 67.51 68.24 69.22 70.25 71.47 72.57 73.53 74.16 73.77 72.81 73.61 73.88 74.69 70.88 71.83Lower Radiator 2 3 63.81 64.85 65.87 66.88 67.85 68.83 69.83 69.93 70.26 69.60 70.42 70.41 69.66 66.67 68.78

Average Lower Rads 65.66 66.55 67.55 68.57 69.66 70.70 71.68 72.05 72.02 71.21 72.02 72.15 72.18 68.78 70.31Ambient # 1 4 37.86 37.96 38.00 40.00 40.75 41.27 41.64 41.94 41.70 41.35 40.92 40.53 0.00 0.00 0.00Ambient # 2 11 38.79 38.95 39.08 38.90 39.20 39.56 39.72 39.52 38.11 37.85 37.25 37.40 0.00 0.00 0.00Ambient # 3 12 35.78 35.88 36.88 36.20 37.88 38.95 39.81 39.78 39.33 39.06 38.81 38.41 0.00 0.00 0.00Average Ambient 37.48 37.60 37.99 38.37 39.28 39.93 40.39 40.41 39.71 39.42 38.99 38.78 0.00 0.00 0.00Top Oil Temp 6 81.21 81.46 82.00 83.25 84.52 85.88 87.95 89.16 89.98 90.47 91.07 91.36 91.12 88.60 89.87Top Oil Temp DV 78.00 80.00 81.00 82.80 83.00 83.50 84.00 85.00 84.00 83.95 85.00 85.40Average of Top Oil 79.61 80.73 81.50 83.03 83.76 84.69 85.98 87.08 86.99 87.21 88.04 88.38 91.12 88.60 89.87Averge Oil Rise @ 3300' 36.21 37.05 37.45 38.84 38.46 38.78 39.31 40.39 40.79 40.83 42.42 42.88 84.54 81.69 83.03Top Oil Rise @ 3300' 42.13 43.13 43.51 44.66 44.48 44.76 45.59 46.67 47.28 47.79 49.04 49.60 91.12 88.60 89.87Bottom Oil Rise @ 3300' 28.18 28.95 29.56 30.20 30.38 30.77 31.29 31.63 32.30 31.79 33.02 33.37 72.18 68.78 70.31

Case 1: Heat Runs Temp. Log



Test Corrections Limit Total Losses (kW ) 172.600 228.464Tap Position 1R 1R Average Oil Rise 42.9 51.2Top Oil Rise 49.6 59.2 65.0Winding Gradient, YV 10.2 10.2Winding Gradient, XV 2.3 2.3Winding Gradient, HV 3.3 3.2HS over TOR, YV 11.2 11.2HS over TOR, XV 2.5 2.5HS over TOR, HV 3.6 3.5Hot Spot Factor, YV 1.10 1.10Hot Spot Factor, XV 1.10 1.10Hot Spot Factor, HV 1.10 1.10Average Winding Rise, YV 53.05 61.3Average Winding Rise, XV 45.13 53.4Average Winding Rise, HV 46.18 54.4Hot Spot Rise, YV 60.8 70.4Hot Spot Rise, XV 52.1 61.7Hot Spot Rise, HV 53.2 62.7

65.0

80.0 n: 0.63m: 0.80

Exponents

Case 1: Heat Runs Results


Winding Current for Individual Gradient Runs


Winding Current for Oil Rise Run



• Stable oil temperatures must be met to achieve reasonably accurate winding gradient measurements.

• Accurate cold resistance temperature measurements are critical in determining the winding rises. YV hot resistance on Tap 1R – Cold resistance not

measured, used Tap 1N Not valid.

• Simulated load losses should be close to the expected load losses for the transformer during operation.Actual winding currents not measured during

simultaneous loading.

Case 1: Summary



GSU 820 MVA 362 / 25 kV DETC(±5%) 60Hz

• Heat run test was performed according to IEEE/ANSI Standards and the clients technical specification.

• Temperature results were below limits and according to clients requirements.

• DGA performed after heat run test found gas generation above client and the manufacturer’s acceptance limits.

Case 2: Good Unit – Bad DGA



Case 2: Bad DGA Results

Outside Lab Change Client LimitsSample # 1 2 3 - -

DescriptionBefore Heat Run [ppm]

4 hours after Heat Run [ppm]

4 hours after Heat Run [ppm]

Gas Evolution [ppm]

Gas Evolution [ppm]

H2 - Hydrogen 4 22 17 13 10O2 - Oxygen 3989 2314 300 - -N2 - Nitrogen 11045 12181 9050 - -CO - Carbon Monoxide 10 62 50 40 25CO2 - Carbon Dioxide 82 392 250 168 200CH4 - Methane 0 3 14.4 14.4 5C2H4 - Ethylene 0 5 4 4 2C2H6 - Ethane 0 27 22 22 2C2H2 - Acetylene 0 0 0 0 0

In House Lab

All measured temperature rises within calculated tolerances.



5.TX Hot Spot

4.Stray

Gassing

3. Improper Testing

2.Pump

Problem

1.Bad DGA Sample

Why ?

Case 2: Possible Causes



1. Bad DGA Data?• DGA results of the outside lab matched the results

obtained at factory.

2. Bad Pump (s) ?The most probable cause of pump overheating is the pump running backwards.

• Running ratings matched Nameplate• No Noise• Thermal Scan normal for pumps and oil flow





3. Improper Testing Method ?

• Loading per IEEE/Expedited Heating• Fiber Optic Sensors in Coils – No high temperatures• Thermocouples on structural metal parts – Normal

heating• Ambient temperature was below 40 ºC• Total heat load (kW) matched cooler rating• Maximum current was only 112% of rated/ Less

than 7 percent of allowable continuous overload current.



Only two possible causes left:

4. An oil problem due to “thermal stray gassing”.

Or5. An abnormal transformer hot spot.


An experiment is needed!



How is the gassing influenced ?

• Load dependent

• Oil temperature dependent

Case 2: Experimental Loading



Step A: Transformer Rated Conditions

1. Test Floor Open and Ventilated2. All Pumps & Fans On3. Full Rating Conditions of the transformer




Step B: Simulate Stray Gassing

1. Test Floor Closed2. Reduced Current 3. All Pumps on/ Fans adjusted to keep oil at ~ 90 ºC




Case 2: Experimental Method

Test Number

Load Condition

Oil Temperature Comments

Test #1 Overload High Oil Temperature

This is the intial heat run result. The source of gassing is indeterminate.

Test #2 Rated Load Normal Oil Temperature

Gassing under this condition is most likely not from the oil.

Test #3 Reduced Load High Oil Temperature

Gassing under this condition is most likely not from the transformer.

Test #4 Overload Normal Oil Temperature Gassing is most likely from the transformer.



Case 2: Experimental Method

Test Number

Duration [Hours] Criteria Loading

Load [%]

Top Oil Temp.

Coil Oil Temp.

Ambient [ºC]

8.0 Per IEEE Total Heat Load 108.0 73.0 90.3 37.6

1.0 Per IEEE Rated Current 100.0 70.0 92.3 38.4

Test #2 8.0 - Rated Current 100.0 53.5 78.8 24.0

Test #3 8.0 - Reduced Curr. 80.0 83.5 92.6 35.2

7.5 Per IEEE Total Heat Load 108.0 62.5 87.8 26.0

1.0 Per IEEE Rated Current 100.0 59.7 81.0 26.0

Test #1

Test #4©Doble Engineering Company


Case 2: Experiment Results

Test # 1 2 3 4Gas

Evolution [ppm]

Gas Evolution

[ppm]

Gas Evolution

[ppm]

Gas Evolution

[ppm]H2 - Hydrogen 13 0 12 3 10O2 - Oxygen - - - - -N2 - Nitrogen - - - - -CO - Carbon 40 10 36 12 25CO2 - Carbon Dioxide 168 66 272 105 200CH4 - Methane 14.4 1.8 8 4.2 5C2H4 - Ethylene 4 0.6 1.7 1.3 2C2H6 - Ethane 22 0 13.6 0 2C2H2 - Acetylene 0 0 0 0 0CO2/CO Ratio 4.2 6.6 7.6 8.8 < 3

Criteria



• Most of the gas concentrations exceed the customer limits.

• Dominant gasses are Methane, Ethane and Hydrogen (low temperature gasses or thermal stray gassing).

• No cellulose decomposition.

Case 2: Test #1 Results

The Source of the Excessive Gassing is Indeterminate.



• Gassing, all gasses within acceptance limits. • No dominant gasses.

• No cellulose decomposition.


If there was excessive gassing it would likely be from the transformer active parts.



• A gasses exceeding limits except Ethylene. • Dominant gasses are Methane, Ethane and Hydrogen. • No cellulose decomposition. • In this test the load is reduced no gassing can be

correlated with the transformer. • Dominant gasses are Methane, Ethane and Hydrogen,

this is an indication of possible thermal stray gassing. • The gassing results are similar Test #1.


This excessive gassing is likely from the oil.



• All gasses within acceptance limits.• Dominant gasses are Methane, Ethane and Ethylene.

Typical gasses for a heat run test without additional stray gassing.

• No cellulose decomposition. • The absence of gasses confirm that gas generation is

not related to the load or a transformer condition.


If there was excessive gassing it would likely be from the transformer active parts.



• The transformer successfully passed the heat run test according to ANSI/IEEE Standards.

• Test #3 results closely match with Test #1 and are indicative that the source of gasses during heat run test is thermal stray gassing of the oil.

• Gasses generated during heat run test performed are produced by thermal stray gassing of the oil used for FAT.

• The Doble Oil Lab confirmed the stray gassing tendency of the oil used for the factory heat run.

Case 2: Summary



CONCLUSION
