You are on page 1of 6

Effect of Terminal Connections on SFRA Results of

Three-Winding Power Transformers


Esam Al Murawwi, Redy Mardiana, Qi Su

Abstract- This paper presents some case studies on the effects
of terminal connections that affect Sweep Frequency Response
Analysis (SFRA) results. Wrong or different terminal
connections might lead to confusing SFRA results which could be
misinterpreted in the assessment of transformer winding
conditions. Effect of grounding the tertiary winding, opening the
tertiary winding, and connecting the neutral point in the terminal
connections are highlighted in this work.
I. INTRODUCTION
Many problems can happen to power transformers in their life-
time starting from the moment they are shipped to site until
the moment of achieving their end-life (ageing or failures) [1,
2]. Also, many competitions are happening in the world and
every industry and manufacture wants to make the transformer
as less costly as possible with as much high quality as
possible. Because of those reasons, there is a need for
evaluating and testing the transformer to assess its condition
either before connecting it to the power network during
commissioning stage or during its maintenance phase.
Sweep Frequency Response Analysis (SFRA) is a global
technique used by many industries and manufactures to assess
the mechanical integrity of the core, winding, and clamping
structure of power transformers. This is usually done by
injecting a low voltage signal (2-15V) with the sweeping
frequency from 20Hz to 2MHz into a transformer terminal and
measuring the response at another terminal as shown in Fig.1
[1]. The amplitude and phase transfer function of SFRA are
determined by Fourier analysis.
Fig. 1: Basic principle of SFRA
There are some standards as well as ongoing work to publish a
standard related to the SFRA. The existing standards are the
China national standard since 2005 [3], and the CIGRE
standard since 2008 [4]. The active projects to publish a
standard are the IEC [5] and the IEEE [6]. As we can see from
the above standards and dates, SFRA is a new area and has
been investigated by many researchers [713].
SFRA test is carried out in many conditions of transformer.
First, it is used to test a new transformer at factory to check its
quality and to create the reference fingerprint. Second, it is
used during the commissioning stage to check its condition
after transportation. Third, it is used during the service stage to
check its healthiness. Finally, it is used after special cases such
as earthquake, lightning, and faults to check its condition.
According to IEC draft standard 60076-18 [5], each frequency
response can be divided into three regions as depicted in Fig.2.
The information for each region is given as follow:
The lower frequency region dominated by the core.
The middle frequency region dominated by the
interactions among the windings.
The higher frequency region controlled by the individual
winding structure internal connections.
The highest frequencies region (usually above 1MHz)
dominated by the measurement set-up and connection
leads.
Note that the starting and ending frequencies of the different
regions can vary from transformer to transformer.
The method to analyze the transformer conditions is
commonly based on the waveform comparison [1]. There are
three methods of waveform comparison:
1. Time-based (a comparison between current SFRA results
with previous results of the same unit).
2. Type-based (a comparison between SFRA of one
transformer with another sister unit).
3. Design-based (a comparison between the three phases of
the same transformer).
Fig. 2: General relationship between frequency response, transformer
structure, and machine set-up
The test engineer can follow the same order as mentioned
above to test the transformer conditions. In this case, he
should start with the time-based comparison if previous SFRA
results are available. Note that the terminal connections have
to be the same with the previous measurement to achieve a
good comparison. To avoid different connections and wrong
result interpretation, it is necessary to take photos of the
connections for future tests. Fig. 3 shows a time-based
comparison of the 40 MVA transformer after shifting it to
another substation. The results indicate that the waveforms are
matching except for the core region. This is because the
winding resistance test was carried out before shifting which
magnetized the core (core remanence).
Fig. 3: a time-based comparison of the same transformer after shifting.
If the previous results are not available, a type-based
comparison is recommended. Fig. 4 shows a type-based
comparison of four 40MVA transformers.
Fig. 4: A type-based comparison between four sister transformers.
If there is neither previous result nor sister unit to compare
with, a design-based should be the solution. Fig. 5 shows a
design-based comparison of the three phases of 40 MVA
transformers. In the low frequency region (core region), the
middle phase is different than the other two phases which is a
normal condition. Also, the variations above 1 MHz
(connections region) is normal as well. It is worth to mention
that the test engineer can combine two methods or more for
better condition assessment such as a combination between
type-based and design-based comparisons.
Fig. 5: A design-based comparison between the three phases of the same
transformers.
Deviations of SFRA traces due to terminal connections can be
as big as the one caused by a fault. This could lead to
misinterpretation of the results (e.g. internal faults happened
whereas it is just a different or wrong connection). To avoid
misinterpretation of the results, special connection
consideration must be taken into account while doing the
SFRA test.
Before shifting (without DC injection)
After shifting (with DC injection)
II. SFRA CONNECTIONS- CASE STUDIES
Commonly, there are four SFRA tests on the three-winding
transformer. The tests include:
1. Injecting into HV winding while LV winding is open, and
tertiary winding is closed [HV_LV (OC)].
2. Injecting into HV winding, LV winding is shorted
(without including the neutral), and tertiary winding is
closed [HV_LV (SC)].
3. Injecting into LV winding, HV winding is open, and
tertiary winding is closed [LV_HV (OC)].
4. Injecting into Tertiary winding while HV winding and LV
winding are open [TV_LV and HV (OC)].
Figs. 6 to 9 show each of the cases. The difference between
the above tests is the terminal connection and each case has its
own graph different than the others. One of the problems that
occur while doing the SFRA tests is that the test engineer may
make different or wrong connections. This problem may affect
the result interpretations. For example, keeping the tertiary
winding grounded or opened instead of closing it (in normal
practice, the tertiary winding must be closed), or including the
neutral when shorting the LV side in the second case of the
testing as shown in Fig. 7 (in normal practice, neutral is not
included in the shorting)
Fig. 6: Injecting in HV winding, LV winding is open, and tertiary winding
is closed [HV_LV (OC)].
Fig. 7: Injecting in HV winding, LV winding is shorted (without neutral), and
tertiary winding is closed [HV_LV (SC)].
Fig. 8: Injecting in LV winding, HV winding is open, and tertiary winding is
closed [LV_HV (OC)].
Fig. 9: Injecting in Tertiary winding, HV winding and LV winding are open
[TV_LV and HV (OC)].
III. EFFECT OF GROUNDING THR TERTAIRY
WINDING
The normal practice while doing the SFRA test is to close the
tertiary winding. However, if the tertiary winding is grounded,
which is a T-bar connecting the tertiary busing to the
transformer tank as shown below in Fig. 10, then the result
may vary.
Fig. 10: tertiary winding grounded
Figs. 11 to 13 illustrate the different cases of SFRA testing
with the tertiary winding closed and grounded:
A. Injecting in HV, LV open circuit, (Tertiary closed vs.
Tertiary Earthed) as shown in Fig. 11.
B. Injecting in HV, LV Short circuit, (Tertiary closed vs.
Tertiary Earthed) as shown in Fig. 12.
C. Injecting in LV, HV open circuit, (Tertiary closed vs.
Tertiary Earthed) as shown in Fig. 13.
Source
Reference
Response
Ground Connection
Tertiary
closed
Fig. 11: Injecting in HV, LV open circuit:
a. Tertiary closed, b. Tertiary closed and earthed.
Fig. 12: Injecting in HV, LV short circuit:
a. Tertiary closed, b. Tertiary closed earthed
Fig. 13: Injecting in LV, HV open circuit:
a. Tertiary closed, b. Tertiary closed and earthed
It is obvious, from Figs. 11 to 13, that the grounding of the
tertiary winding has more effect when the voltage source is
injected in the low voltage side as shown in Fig.13. The reason
is that the low voltage winding is close to the tertiary winding.
IV. EFFECT OF KEEPING THE TERTAIRY WINDING
OPENED
Apart from keeping the tertiary grounded, keeping it opened
also has an effect on the result. The following two cases were
taken for 40MVA transformers:
A. Injecting in HV, LV open circuit, (Tertiary closed vs.
Tertiary opened) as shown in Fig. 14.
B. Injecting in LV, HV open circuit, (Tertiary closed vs.
Tertiary opened) as shown in Fig. 15.







Fig. 14: Injecting in HV, LV- OC, (Tertiary closed Vs. Tertiary opened)
1U
IV
1W
1U
IV
1W
a.
b.
1U
IV
1W
1U
IV
1W
a.
b.
2u
2v
2w
a.
2u
2v
2w
b.

Fig. 15: Injecting in LV, HV- OC, (Tertiary closed Vs. Tertiary opened)
Apart from the variations on the high frequency region as
shown in Fig. 15, in the middle region (indicated by the
circle), one can see a double peak for the tertiary opened,
while a one peak is present in the case for the tertiary closed.
This is consistent for the three phases of the transformer when
each phase is tested separately.
V. EFFECT OF INCLUDING THE NEUTRAL IN THE
TERMINAL CONNECTION
In normal practice, the terminals are short circuited but not
connected to the neutral point when the voltage source is
injected in the HV side and LV side is short circuited. Fig.16
shows the difference between including and excluding the
neutral while shorting the LV side.
It is worth noting that if the neutral is included in the shorting
in previous tests, then it should be included also while
repeating the test in the future. In this case, a fair comparison
and right conclusions about the healthiness of the transformer
can be achieved. For this reason, it is recommended to take
photos of the terminal connections to avoid such problems.
Fig. 16: Injecting in HV, LV short circuit: with and without neutral
VI. SPECIAL SFRA CONSIDERATIONS TO ACHIEVE
RELIABILITY
The following procedures are helpful to obtain better SFRA
results, prevent SFRA future interpretations problems, and get
reliable results [1, 14, 15]:
The SFRA test has to be done with the same transformer
conditions (e.g., assembled, oil filled) to get reliable
results.
The transformer should be isolated. Meaning that all
connections from the transformer (except the tank) and
auxiliary connections should be removed.
Cleaning of the bushings contacts and tightening the
connections clamps should be done to assure reliable
contact.
Three shielded high frequency cables should be used
(source, reference, and measurement).
The test leads should be as short as possible.
Ground extension of the test leads has to be of low
inductance.
Ground extensions of the measuring cable shields should
not have electrical contacts with the terminal contacts.
Noises should be eliminated and grounding to be checked
before SFRA test since noisy curves may result from poor
grounding.
SFRA test should be the last test before transportation and
the first test after the transportation.
The status of the transformer should be noted (e.g.,
transformer age, number of failures, type of failures, etc).
If reference data is available, the results should be
available hardcopy and softcopy in case it needs to be
uploaded for easy comparison during future tests.
Tap changers position, core being magnetized or not
(e.g., after winding resistance test), whether the oil is
presented or not, and whether the oil is recently drained or
not should be documented and available during the SFRA
test.
Photos of the SFRA terminal connections should be taken
and available for any future SFRA tests. The same
connections should be used for other repeated SFRA tests.
It is better to use the same instrument while repeating the
test to achieve good repeatability. Otherwise, using an
instrument with the same frequency range, accuracy
range, settling time, and applied voltage is recommended.

VII. CONCLUSION
The effect of terminal connections on SFRA results of three-
winding transformer has been presented and discussed. The
terminal connections include grounding the tertiary winding,
opening the tertiary winding, and connecting the neutral point
in the connections. Wrong or different terminal connections
might lead to different waveform of SFRA and
Without Neutral
With Neutral
misinterpretation of the results. One of the main conditions to
achieve good waveform comparisons and avoid
misinterpretation is to have the same terminal connections for
the previous, current, and future tests.
REFERENCES
[1] Kraetge, M. Kruger, M.Viljoen, and A. Dierks, Aspects of Practical
Applications of Sweep Frequency Response Analysis (SFRA) on Power
Transformers. 6
th
Southern Africa Regional Conference, CIGRE 2009.
[2] M. de Nigris, R. Passglia, R. Berti, L. Bergonzi, and R. Maggi,
Application of Modern Technologies for the Condition Assessment of
Power Transformers. CIGRE 2004.
[3] The Electric Power Industry Standard of Peoples Republic of China,
Frequency Response Analysis on Winding Deformation of Power
Transformers, DL Standard, DL/T911-2004, ICS 27.100, F24,
Document No. 15182-2005, June 2005.
[4] CIGRE Working Group A2/26, Mechanical condition Assessment of
Transformer Windings using Frequency Response Analysis (FRA).
Paris, April 2008.
[5] IEC 60076-18 Ed.1: Power Transformers-Part 18: Measurement of
Frequency Response. 2009 (not yet published).
[6] IEEE PC57.149TM/D4 Draft, Trial-Use Guide for the Application and
Interpretation of Frequency Response Analysis for Oil Immersed
Transformers, 2007 draft version (not yet published).
[7] L. Satish, and A. Saravanakumar, Identification of terminal connection
and system function for sensitive frequency response measurement on
transformers, IEEE Trans. Power Deliv., vol.23, no.2, April 2008.
[8] N.Abeywickrama, Y.V.Serdyuk, and S.M.Gubanki, High-frequency
Modeling of Power Transformers for use in Frequency Response
Analysis (FRA), IEEE Trans. Power Deliv., Vol. 23.No.4, October
2008.
[9] R. Ragavan and L. Satish, An Effective method to compute transfer
function of a Transformer from its equivalent circuit. IEEE Trans.
Power Deliv., vol.20, no2, pp.780-788, April 2005.
[10] M. Florkowski and J. Furgal, Detection of transformer winding
deformations based on the transfer function Measurements and
simulations. Meas. Sci. Technol., vol.14, pp. 1986-92, Sep. 2003.
[11] M. Wang, A. J. Vandermaar, and K. D. Srivastava, Improved Detection
of Power Transformer Winding Movement by Extending the FRA High
Frequency Range, IEEE Trans. Power Deliv., Vol. 20 .No.3, July 2007.
[12] A. Shintemirov, W. H. Tang, Q. H. Wu, A Hybrid Winding Model of
Disc-Type Power Transformers for Frequency Response Analysis,
IEEE Trans. Power Deliv., Vol. 24.No.2, April 2009.
[13] K.G.N.B. Abeywickrama, Y.V. Serdyuk, and S.M. Gubanski,
Exploring possibilities for characterization of power transformer
insulation by frequency response analysis (FRA). IEEE Trans. Power
Deliv., vol.22, no.3, pp.1375-1382, July 2006.
[14] M. Wang, John Vandermaar, and K.D. Srivastava, Transformer
Winding Movement Monitoring in Service- Key Factors Affecting FRA
Measurements. IEEE Electrical Insulation Magazine Vol. 20, No. 5.
September/October 2004.
[15] M. Karlstorm, P. Werelius, M. Ohlen, L. Adeen, and E. Brynjebo,
Considerations to Ensure Measurement Repeatability when using
SFRA on Transformers, 2008.

You might also like