Electric Power Systems Research 145 (2017) 149–156

Contents lists available at ScienceDirect

Electric Power Systems Research

journal homepage: www.elsevier.com/locate/epsr

Influence of tertiary stabilizing windings on zero-sequence

performance of three-phase three-legged YNynd transformers. Part II:
Tank overheating hazard and short-circuit duty
Angel Ramos a,∗ , Juan Carlos Burgos b
a
Electrical Distribution Division, Gas Natural Fenosa, Avda. San Luis 77, Madrid 28033, Spain
b
Universidad Carlos III de Madrid, Leganés, Madrid, Spain

a r t i c l e i n f o a b s t r a c t

Article history: This paper analyzes the influence of stabilizing windings in three-phase three-legged YNynd transformers
Received 4 October 2015 on tank overheating hazard and on the short-circuit performance of such transformers in the event
Received in revised form of unbalanced conditions involving zero-sequence currents as asymmetrical faults. It complements a
20 December 2016
previous companion paper that presented experimental zero-sequence measurements and equivalent
Accepted 2 January 2017
circuit calculations for three power transformers. Those previously calculated zero-sequence equivalent
Available online 10 January 2017
circuit parameters are used here to evaluate the comparative performance of YNynd power transformers
when the delta stabilizing winding is closed and when it is open. Additionally, it analyzes the effects of
Keywords:
Three-winding transformers
other variables, such as relative position of the stabilizing winding, presence of magnetic shunts in the
Tertiary stabilizing windings tank walls, and system grounding.
Transformer zero-sequence performance © 2017 Elsevier B.V. All rights reserved.
Short-circuit duty
Tank overheating hazard

1. Introduction ditions, or short-circuit problems, which have been extensively

discussed in technical literature [2–11], tank heating hazard in
Tank overheating hazard and short-circuit duty are proba- such transformers is rarely discussed in detail. In Ref. [12], Garin
bly the most important issues to evaluate when questioning the mentioned the possibility of core-form transformer tank overheat-
need for stabilizing windings in three-phase three-legged YNynd ing due to zero-sequence flux. The advisability of using stabilizing
transformers. In both situations, the transformer’s zero-sequence windings to avoid overheating hazard is reported in Ref. [13].
performance under different transformer and grid configurations Information about zero-sequence impedances and tank heating
in the event of fault conditions should be carefully reviewed. modeling was reported in Ref. [14] and, more recently, Ref. [15]
Circulation of zero-sequence flux through the tank walls and developed a detailed analysis of tank overheating hazard in a
cover and other metallic structural elements occurs in three- three-phase three-legged YNynd transformer with and without
phase three-legged transformers when zero-sequence unbalanced stabilizing winding. Technical standards [16,17] also point out
conditions arise. In the event of severe asymmetrical faults, the pro- some issues related to tank heating as a consequence of zero-
tection system ordinarily eliminates such conditions in less than sequence flux. Nonetheless, published information about tank
one second, so that no hazardous tank overheating takes place. overheating hazard is somewhat contradictory and it is probably
Nevertheless, in cases of highly resistive short-circuits or open cir- insufficient to reach a conclusion about the necessity of utilizing
cuit faults, circulation of zero-sequence currents and fluxes could stabilizing windings to mitigate this hazard under various possible
remain undetected causing significant tank overheating [1]. electrical grid and transformer configurations.
Unlike other relevant topics related to zero-sequence perfor- As regards short-circuit duty, the need for appropriate ter-
mance of wye–wye transformers, such as third-harmonic current tiary stabilizing winding mechanical design, as well as the greater
and voltages, temporary overvoltages due to unbalanced con- weakness of these windings when compared with primary and
secondary windings, was identified from the very beginning of
their utilization, as referred to in Blume’s patent [18]. Nonetheless,
∗ Corresponding author. extensive use of stabilizing windings was accompanied by lower
E-mail addresses: [email protected], [email protected] rated power design in such windings to reduce the cost to power
(A. Ramos).

http://dx.doi.org/10.1016/j.epsr.2017.01.003
0378-7796/© 2017 Elsevier B.V. All rights reserved.
150 A. Ramos, J.C. Burgos / Electric Power Systems Research 145 (2017) 149–156

Table 1 Table 2
Set of zero-sequence measurements taken. Characteristics of transformers subject to zero-sequence measurements.

Test number Test code HV winding LV winding Stabilizing winding Data Transformer

1 HOO Energized Open circuit Delta open #1 #2 #3

2 HOC Energized Open circuit Delta close
3 HSO Energized Short circuit Delta open Rated power (MVA) 25/25/8.33 75/75/25 150/150/50
4 HSC Energized Short circuit Delta close Rated voltage (kV) 45/16.05/10 220/71/10 230/71/20
5 LOO Open circuit Energized Delta open Vector group YNyn0 + d11 YNyn0 + d11 YNyn0 + d11
6 LOC Open circuit Energized Delta close Short-circuit impedance 10.8% 14.0% 14.1%
7 LSO Short circuit Energized Delta open Inner winding Stabilizing Stabilizing Low voltage
8 LSC Short circuit Energized Delta close Intermediate winding Low-voltage Low-voltage High-voltage
Outer winding High-voltage High-voltage Stabilizing
Magnetic shunts No Yes No

utilities [6]. Although there were frequent references to their weak-

ness and to the need for appropriate stabilizing winding mechanical Table 3
design, 25% or 35% rules (in relation to power rating) became very Ampere-turn balance measurements in transformer #1.
popular [19]. Far from considering that practice outdated, many Test code (N·I)1 (%) (N·I)2 (%) (N·I)3 (%) MMF0 + (N·I)tank (%)
technical specifications still feature the 35% rule [20].
HOO 100 – – 100
A century after the beginnings of their use, problems related to HOC 100 – 78.7 21.8
stabilizing windings’ short-circuit duty have been not been solved HSO 100 87.2 – 12.9
effectively. As reported in Ref. [21], based on information from 14 HSC 100 113.5 25.1 13.1
high-power laboratories around the world, one in four transformers LOO – 100 – 100
LOC – 100 90.9 9.1
fails in short-circuit withstand tests. Moreover, the failure rate rises
LSO 99.1 100 – 1.0
to over 40% in the case of transformers above 100 MVA. Based on LSC 38.3 100 61.3 0.8
this information, the authors suggest several guidelines for enhanc-
ing short-circuit performance, one of them consisting of avoiding
Table 4
tertiary stabilizing windings. This practice has been applied suc-
Ampere-turn balance measurements in transformer #2.
cessfully in Italy and Turkey.
The purpose of this paper is to contribute to clearer understand- Test code (N·I)1 (%) (N·I)2 (%) (N·I)3 (%) MMF0 + (N·I)tank (%)
ing of the effect of stabilizing windings on tank overheating hazard HOO 100 – – 100
and short-circuit duty in three-phase three-legged YNynd trans- HOC 100 – 99.6 1.2
formers. It analyzes the advantages and disadvantages of stabilizing HSO 100 97.1 – 2.9
HSC 100 118.2 20.0 0.3
windings’ contribution in relation to these problems, taking into
LOO – 100 – 100
account aspects such as relative position of the windings, presence LOC – 100 99.2 1.5
or absence of magnetic shields, and neutral grounding of high- LSO 90.3 100 – 9.7
voltage and low-voltage windings. LSC 23.2 100 73.4 3.6

2. Tank overheating hazard in YNynd transformers Table 5

Ampere-turn balance measurements in transformer #3.
The aim of this section is to advance in the understanding of Test code (N·I)1 (%) (N·I)2 (%) (N·I)3 (%) MMF0 + (N·I)tank (%)
the influence of stabilizing windings on tank overheating hazard in
HOO 100 – – 100
core-form YNynd power transformers. Special attention is paid to
HOC 100 – 97.1 3.1
the relative position of stabilizing windings and to the effect of mag- HSO 100 87.3 – 13.0
netic shields on tank walls, as these aspects are scarcely discussed HSC 100 54.3 44.2 1.6
in technical papers. LOO – 100 – 100
First, tank overheating hazard is identified qualitatively by LOC – 100 90.3 9.7
LSO 97.5 100 – 2.6
means of ampere-turn balance analysis, taking advantage of zero- LSC 89.5 100 4.0 6.6
sequence measurements carried out in the companion paper [22].
Next, zero-sequence temperature rise tests in different
transformers and under different operating configurations are ampere-turns in non-energized windings are expressed as a per-
summarized. centage of ampere-turns in energized windings (i.e. ampere-turns
of the primary winding are the reference in the denominator for
2.1. Ampere-turn balance from measurements tests HOO, HOC, HSO and HSC, and ampere-turns of the secondary
winding are the reference in the denominator for tests LOO, LOC,
As demonstrated in Part I of this paper [22], a complete repre- LSO and LSC). The sum of the winding ampere-turns and the magne-
sentation of zero-sequence behavior can be obtained from a set of tomotive forces created by the currents circulating in the tank walls
onsite low-voltage measurements (see Table 1). Those tests were (with the proper references for tank current) are the magnetomo-
performed on transformers #1, #2 and #3 (Table 2) to calculate tive force required to create zero-sequence flux MMF0 , as indicated
zero-sequence impedances in each test scenario, from which var- in Eq. (1). An example of calculation of Tables 3–5 is reported in
ious zero-sequence equivalent circuit parameters were developed Appendix A.
and validated [22].
MMF0 + (N · I)tank = (N · I)1 + (N · I)2 + (N · I)3 (1)
Current measurements for the HV winding (I1 ), LV winding (I2 )
and stabilizing winding (I3 ) presented in Part I [22], together with The information shown in Tables 3–5 identifies the operating
internal design data (number of turns), are used in this section to conditions of transformers with higher tank overheating hazard. A
determine ampere-turn balances in each transformer and for each greater [MMF0 + (N·I)tank ] percentage supposes a higher tank over-
test configuration. Tables 3–5 present the results of the ampere- heating hazard, as it indicates that the circulating currents in the
turn balance. For more comprehensive analysis of these results, tank walls are high (MMF0 is usually low).
 A. Ramos, J.C. Burgos / Electric Power Systems Research 145 (2017) 149–156 151

The most hazardous conditions for all transformers are obvi- was indicated in Section 2.1, the ampere-turn winding imbal-
ously in the HOO and LOO configurations, which represent ances offer a qualitative indication about tank heating hazard.
operating conditions when asymmetric faults or imbalances occur 3) Perform zero-sequence temperature rise tests as acceptance
in the primary (in the case of HOO) or secondary side (in the case tests in the transformer manufacturer laboratory. The number
of LOO) with transformer neutral ungrounded in the secondary and configuration of the tests must be determined according to
side (or primary in the case of LOO) and with the stabilizing wind- item 2 and they must be done with a neutral current equal (at
ing open (i.e. YN-y). In these cases (HOO and LOO), zero-sequence least) to item 1. A more detailed procedure is reported below.
ampere-turns in the energized winding are not compensated by the 4) Complementarily to zero-sequence temperature rise tests (or
other windings and so zero-sequence magnetic flux is potentially alternatively in clearly justified situations), a detailed and rig-
high, producing potentially hazardous tank currents and heating. orous finite element analysis could be carried out in order to
In other cases, under other operating conditions, tank heat- prevent tank overheating or internal hot spots in structural
ing hazard seems to be moderate or low. In authors’ opinion, parts. Desirably, this analysis should be done in early stages of
the ampere-turn winding imbalances offer a qualitative indica- transformer engineering and construction, so that a more ade-
tion about this hazard but, in order to evaluate the necessity of quate internal design or material selection could be derived.
stabilizing winding to avoid tank heating hazard, a more accurate
analysis should be complemented with zero-sequence temperature The procedure for specifying and performing zero-sequence
rise tests, as shown in the next section. temperature rise tests should be as follows:
Nonetheless, some interesting results may be derived from
Tables 3–5: - Apply a single-phase voltage source at rated frequency to HV
winding between the shorted line terminals and its neutral. LV
- In the case of transformer #1, despite what is commonly thought, winding may be open-circuited (HOO test) or shorted (HSO test).
the absence of a stabilizing winding does not eliminate ampere- In case of LOO or LSO tests, the voltage source is applied to LV
turn winding imbalances (for example, in tests HOC and HSC the winding with HV winding open-circuited or shorted, respectively.
imbalance is higher than in test HSO). - The excitation must be progressively increased up to the neu-
- Comparing Tables 3 and 4, it appears that the presence of tral current reach the maximum permitted value determined
magnetic shunts significantly reduces ampere-turn winding (see step 1 in the methodology above) and must remain constant
imbalances. throughout the test.
- The test may be ended when temperatures are stabilized (rate of
change below 1 ◦ C per hour).
2.2. Zero-sequence temperature rise tests - Temperature probes must be installed in each transformer tank
wall and cover (medium height and at the height correspond-
In the previous section, zero-sequence ampere-turn balances ing to windings top or ending of magnetic shunts). Additional
were shown in three transformers under different operating con- probes may be installed in the transformer oil outlet and inlet
figurations, identifying the situations in which tank overheating pipes. An ambient temperature measurement must be obtained.
hazard was higher. In practice, even in such situations, the mag- The exact number and position of temperature probes should
nitude and duration of zero-sequence flux and current circulation be agreed between user and manufacturer, depending on trans-
through the tank walls is limited by grid zero-sequence impedance former design, thermal analysis, etc. Temperatures must be
and transformer protection settings. Zero-sequence currents below recorded each 5 min.
this limit value could remain undetected, causing permanent zero- - If possible, fiber optic sensors may be installed to determine tem-
sequence flux and current circulation and subsequent tank (and perature rise of internal structural parts (i.e. clamping structures,
structural part) heating. bolts, etc.).
In order to detect overheating hazards due to zero-sequence flux - Just before the end of the test, a thermographic map of tank walls
circulation in case of omission of stabilizing windings in wye–wye should be obtained (with a thermographic camera). Coherence
three-legged transformers, the authors propose to include zero- between measurements of the thermographic camera and tem-
sequence temperature rise tests as mandatory type tests in the perature probes should be checked.
technical specification for the transformer purchasing process. The - Before and after each temperature rise test, transformer oil sam-
complete methodology proposed by the authors to evaluate over- ples must be extracted for dissolved gas-in-oil analysis (H2 , CH4 ,
heating hazards in wye–wye three-legged transformers without C2 H6 , C2 H4 , C2 H2 , CO, CO2 , O2 , N2 ) in order to detect internal hot
stabilizing windings include the following steps: spots during the tests.

1) Determine the maximum permitted value for neutral current in With the objective of implementing the methodology described
the transformer. This value must be determined depending on above, several transformer manufacturers were asked to evaluate
selectivity and sensitivity of the transformer protection system. the influence of permanent zero-sequence excitation under dif-
Neutral currents above this value must be tripped out for the ferent operating conditions on three-phase three-legged YNynd
transformer protection settings (an ordinary maximum permit- transformers. In the absence of a contractual obligation, as temper-
ted value for neutral current is about 0.2–0.5 p.u.). ature rise tests require high-voltage laboratories (with a relevant
2) Identify working conditions of the transformer that may gener- cost and duration), some limitations to the complete set of tests
ate a relevant circulation of neutral current. This aspect is very were agreed. Despite these restrictions, several zero-sequence tem-
dependent on primary and secondary neutral grounding of the perature rise tests were performed with very significant results.
transformer. For example, if a transformer has its primary neu- As it was indicated in the Part I of this paper, there are three
tral directly grounded and the secondary neutral is ungrounded, internal designs that cover the vast majority construction types of
then only primary zero-sequence excitation tests would be nec- three-phase three-legged YNynd power transformers. These types
essary (i.e. HOO condition). In the case that both transformer are represented for the analyzed transformers numbered with #1,
neutrals are grounded, then different working conditions would #2 and #3. As it was shown in the previous section, the ampere-turn
be occur (i.e. HSO and LSO conditions). In this regard, it can be balance analysis offered a qualitative indication that tank overheat-
very useful to have a zero-sequence ampere-turn analysis. As it ing hazard was higher in the transformer typology #1 (i.e. with
152 A. Ramos, J.C. Burgos / Electric Power Systems Research 145 (2017) 149–156

Table 6 - As expected, in the case of utilization of stabilizing windings, tank

Characteristics of transformer subject to zero-sequence temperature rise tests.
overheating hazard is not a concern.
Data Transformer #1B - In the case of omission of stabilizing windings, when both pri-
Rated power (MVA) 25/25/8.33 mary and secondary neutrals remain grounded, tank overheating
Rated voltage (kV) 45/16.05/10 hazards is neither a concern.
Vector group YNyn0 + d11 - In the absence of stabilizing winding, when one of the trans-
Short-circuit impedance 10.7% former neutrals is ungrounded detailed attention should be paid
Inner winding Stabilizing
to tank overheating hazard. In those cases, rigorous calculations,
Intermediate winding Low-voltage
Outer winding High-voltage simulations and tests should be discussed between manufacturer
Magnetic shunts No and user to ensure that tank temperature design limits are not
exceeded.

Table 7
Zero-sequence temperature rise test results in transformer #1B. 3. Short-circuit duty of YNynd transformers
Test configuration Neutral current (p.u.) Maximum tank
temperature rise (K)
Short-circuit duty is one of the key aspects when designing and
operating power transformers, as during these regimes windings
HOO 0.3 60.2
are subject to high currents and forces. In fact, short-circuit forces
HOC 1.0 40.8
HSO 1.2 33.4 largely determine the transformer’s mechanical design. As indi-
HSC 1.2 30.3 cated in the introduction, in the particular case of three-winding
YNynd transformers, stabilizing winding design is particularly crit-
ical to short-circuit forces, as these windings are typically thermally
inner stabilizing winding and no magnetic shunts) and, especially, designed for approximately one third of the transformer’s rated
in case of HV winding zero-sequence excitation. power.
The results of the zero-sequence temperature rise tests carried The aim of this section is to analyze the magnitude of the trans-
out on a transformer of this internal design characteristics (Table 6), former currents in the case of primary or secondary single-phase
numbered #1B (similar to transformer #1), are shown below. line-to-ground faults (1LG), and to compare the short-circuit duty
Following the test procedure indicated, four zero-sequence tem- of the different transformer types under study in several grid con-
perature rise tests were performed on transformer #1B, all of them figurations.
applying the voltage source to the HV winding. In the case of the
HOO test (most hazardous configuration), the neutral current was 3.1. Transformers, scenarios and equivalent circuits considered in
limited to 0.3 p.u., but in the rest of the tests neutral currents equal simulations
or above the rated current of the transformer were applied. In all the
tests, zero-sequence excitation was maintained for several hours to For the purpose of performing quantitative assessment of short-
achieve a thermal steady state. circuit duty in three-phase three-legged YNynd transformers,
Summarized results of the tests are shown in Table 7, where neu- different simulations with various real scenarios were carried out
tral current in the energized winding is expressed in per-unit values considering the variables involved in the wye–wye transformers’
in relation to the transformer rated current, and where maximum short-circuit duty behavior.
tank temperature rise is the highest measurement above ambient Table 8 shows the variables and options considered in the sim-
temperature obtained at various points of the tank walls and cover. ulations analyzed in this paper.
The main conclusions drawn from the experimental results of Fig. 1 shows the equivalent circuits used to calculate single-
transformer #1B are as follows: phase-to-ground fault currents.
Information about the transformers’ impedance parameters is
- The results of the temperature rise tests are consistent with the reported in Part I of this paper [22], considering that the equivalent
qualitative ampere-turn balance approach (transformers #1 and circuit model used for the simulations were model B. On the other
#1 B are twin units). hand, calculation of grid impedances was based on the grid short-

Table 8
Scenarios considered in simulations.

Variables Number of options Options considered Observations

Transformer design types 3 - 45/15 kV 25 MVA (Design type These 3 transformers represent the
#1) main three-legged transformers’
- 220/66 kV 75 MVA (Design type design types (Fig. 1 of Part I [22])
#2)
- 220/66 kV 150 MVA (Design type
#3)

Operation of stabilizing winding 2 - Closed

- Open

Neutral grounding of primary and 4 - Grounded/grounded

secondary windings - Grounded/ungrounded
- Ungrounded/grounded
- Ungrounded/ungrounded

Types of fault considered 3 - Secondary 3LG

- Primary 1LG
- Secondary 1LG
 A. Ramos, J.C. Burgos / Electric Power Systems Research 145 (2017) 149–156 153

design (i.e. winding arrangement and use of magnetic shunts) and

neutral grounding.
Tables 10–12 show short-circuit currents for transformers #1,
#2 and #3, respectively, depending on fault type, neutral grounding
and operation of the stabilizing winding. In these tables, the results
consider the case of average short-circuit power of the primary and
secondary grids (medium/medium). I1 , I2 and I0M are expressed in
per-unit values referred to the rated power/current of the trans-
former and I3 is expressed in per-unit values referred to the rated
power/current of the stabilizing winding (1/3 of the transformer
rated power).
The influence of the presence of the stabilizing winding, of neu-
tral grounding and of the transformer’s internal design can be
clearly deduced from the results shown in Tables 10–12.
When the stabilizing winding is closed, single-phase fault cur-
rents increase significantly in the case of neutral grounding on at
least the faulted side.
As regards core-form transformers’ internal design, when com-
paring results in #1 and #2 (and in the rest of the transformers
of these types), the presence of magnetic shunts is identified as
not very significant to short-circuit duty. Conversely, the relative
position of the stabilizing winding is of great significance:

- When the stabilizing winding is the innermost winding, short-

circuit currents increase highly. In secondary 1LG faults (the most
frequent fault), both I2 and I3 are extraordinarily high, and con-
sequently the electrodynamic forces in the transformer are very
high.
- When the stabilizing winding is the outermost winding, short-
circuit currents increase moderately/highly. Special attention
should be paid to high stabilizing winding currents in the case
of primary 1LG faults (less frequent faults) and, to a lesser extent,
to I1 . However, short-circuit stresses are easier to manage in this
transformer design than in the previous one.

What is the reason behind this behavior? The explanation can

be found on the zero-sequence impedances calculated in the Part
I of this paper [22]. For example, in case of secondary 1LG faults,
GND/GND transformer neutral grounding configuration and stabi-
lizing winding closed, the short-circuit currents of the transformer
are mainly determined by the value of the impedance Z0 in the
LSC test. The values of this impedance are 3.29% in transformer #1,
3.24% in transformer #2 and 13.88% in transformer #3, which cause
very high currents I2 and I3 in transformers #1 and #2.
Fig. 1. Equivalent circuits for calculating line-to-ground fault currents when the 3-
impedance zero-sequence model is used (for primary L-G fault calculations, only the Although presenting results with all grid short-circuit power
left-hand voltage source (solid line) is applied; for secondary L-G fault calculations, scenarios is too cumbersome to include in this paper, it could
only the right-hand voltage source (dotted line) is applied). be indicated that the analysis with other scenarios different
than medium/medium showed a monotonically influence on
Table 9 the absolute values of short-circuit currents. Additionally, some
Grid short-circuit power considered in simulations. complementary simulations were performed considering other
Grid short-circuit power (MVA) Grid voltage transformers, other grid Z0 /Z1 ratios, a six-impedance zero-
sequence equivalent circuit (i.e. model C in Part I of this paper
220 kV 132 kV 66 kV 45 kV 15 kV
[22]) and different on-load tap changer positions. In all the cases,
High 20,000 8000 4000 2500 650 the above conclusions keep unaltered. Detailed information about
Medium 10,000 2000 2000 500 250
Low 2500 500 500 100 25
these complementary simulations can be found in Ref. [23].

3.3. Temporary overvoltages results and conclusions

circuit power information shown in Table 9 and the hypothesis
(usual practice) that grid Z0 /Z1 ratio is equal to 3. Complementing the analysis of short-circuit currents in Section
3.2, Tables 13–15 show short-circuit voltages for transformers #1,
3.2. Short-circuit current results and conclusions #2 and #3, respectively, depending on fault type, neutral grounding
and operation of the stabilizing winding. As in the previous sec-
The main objective of the Sections 3.2 and 3.3 is to summarize tion, the results consider the case of average short-circuit power of
the results that were obtained in the simulations that were run the primary and secondary grids (medium/medium). U1 and U2 ,
combining the variables described in the Section 3.1, highlighting respectively for the primary and secondary windings, the high-
the influence of the stabilizing winding depending on transformer est rms line-to-ground power-frequency voltage on a sound phase
154 A. Ramos, J.C. Burgos / Electric Power Systems Research 145 (2017) 149–156

Table 10
Short-circuit current results for transformer #1.

Fault type Transformer neutral grounding Stabilizing winding closed Stabilizing winding open

I1 (p.u.) I2 (p.u.) I3 (p.u.) I1 (p.u.) I2 (p.u.) I0M (p.u.)

3LG – 6.33 6.33 – 6.33 6.33 –

Primary 1LG GND/GND 5.75 3.05 8.11 4.18 3.51 0.67
GND/UG 5.55 2.46 9.28 2.65 1.88 0.77
UG/GND 1.68 1.68 – 1.68 1.68 –
UG/UG 1.68 1.68 – 1.68 1.68 –
Secondary 1LG GND/GND 5.87 10.24 13.10 4.94 5.32 0.39
GND/UG 1.83 1.83 – 1.83 1.83 –
UG/GND 4.57 9.55 14.93 2.24 2.98 0.74
UG/UG 1.83 1.83 – 1.83 1.83 –

Table 11
Short-circuit current results for transformer #2.

Fault type Transformer neutral grounding Stabilizing winding closed Stabilizing winding open

I1 (p.u.) I2 (p.u.) I3 (p.u.) I1 (p.u.) I2 (p.u.) I0M (p.u.)

3LG – 6.66 6.66 – 6.66 6.66 –

Primary 1LG GND/GND 5.59 3.45 6.43 4.89 4.54 0.35
GND/UG 5.27 2.31 8.89 2.63 2.19 0.44
UG/GND 2.17 2.17 – 2.17 2.17 –
UG/UG 2.17 2.17 – 2.17 2.17 –
Secondary 1LG GND/GND 6.81 13.33 19.55 6.29 6.46 0.17
GND/UG 2.32 2.32 – 2.32 2.32 –
UG/GND 4.23 11.57 22.02 2.43 2.89 0.45
UG/UG 2.32 2.32 – 2.32 2.32 –

Table 12
Short-circuit current results for transformer #3.

Fault type Transformer neutral grounding Stabilizing winding closed Stabilizing winding open

I1 (p.u.) I2 (p.u.) I3 (p.u.) I1 (p.u.) I2 (p.u.) I0M (p.u.)

3LG – 6.61 6.41 – 6.41 6.41 –

Primary 1LG GND/GND 7.06 3.77 9.89 4.20 3.54 0.66
GND/UG 5.35 2.04 9.93 2.52 1.83 0.68
UG/GND 1.78 1.78 – 1.78 1.78 –
UG/UG 1.78 1.78 – 1.78 1.78 –
Secondary 1LG GND/GND 5.59 5.96 1.12 5.80 5.91 0.12
GND/UG 1.99 1.99 – 1.99 1.99 –
UG/GND 2.74 4.42 5.03 2.27 2.89 0.62
UG/UG 1.99 1.99 – 1.99 1.99 –

Table 13
Short-circuit voltage results for transformer #1.

Fault type Transformer neutral grounding Stabilizing winding closed Stabilizing winding open

U1 (p.u.) U2 (p.u.) U1 (p.u.) U2 (p.u.)

Primary 1LG GND/GND 1.13 1.02 1.19 1.17

GND/UG 1.13 1.03 1.26 1.34
UG/GND 0.93 0.96 0.75 0.88
UG/UG 0.93 0.96 0.93 0.96
Secondary 1LG GND/GND 1.02 0.94 1.12 1.16
GND/UG 0.98 0.94 0.98 0.94
UG/GND 1.10 0.96 1.34 1.30
UG/UG 0.98 0.94 0.98 0.94

Table 14
Short-circuit voltage results for transformer #2.

Fault type Transformer neutral grounding Stabilizing winding closed Stabilizing winding open

U1 (p.u.) U2 (p.u.) U1 (p.u.) U2 (p.u.)

Primary 1LG GND/GND 1.17 1.02 1.22 1.19

GND/UG 1.17 1.02 1.27 1.32
UG/GND 0.92 0.86 0.92 0.96
UG/UG 0.92 0.96 0.92 0.96
Secondary 1LG GND/GND 1.01 0.92 1.09 1.14
GND/UG 0.99 0.94 0.99 0.94
UG/GND 1.11 0.94 1.42 1.35
UG/UG 0.99 0.94 0.99 0.94
 A. Ramos, J.C. Burgos / Electric Power Systems Research 145 (2017) 149–156 155

Table 15
Short-circuit voltage results for transformer #3.

Fault type Transformer neutral grounding Stabilizing winding closed Stabilizing winding open

U1 (p.u.) U2 (p.u.) U1 (p.u.) U2 (p.u.)

Primary 1LG GND/GND 1.18 1.16 1.22 1.17

GND/UG 1.21 1.35 1.25 1.35
UG/GND 0.92 0.97 0.92 0.97
UG/UG 0.92 0.97 0.92 0.97
Secondary 1LG GND/GND 1.04 1.14 1.05 1.15
GND/UG 0.99 0.93 0.99 0.93
UG/GND 1.17 1.22 1.35 1.30
UG/UG 0.99 0.93 0.99 0.93

during a single-phase ground fault, expressed in per-unit values currents when there the neutral is grounded on at least the faulted
referred to the rated voltage of the winding. side, as the zero-sequence impedance of the transformer is lower
As one might expect, when the stabilizing winding is open, an when the stabilizing winding is closed. Moreover, the relative posi-
increase of the temporary overvoltages occurs in the non-faulted tion of the stabilizing winding was demonstrated to have great
phases, in the case of neutral grounding at the faulted side. This importance: when the stabilizing winding is the innermost wind-
increase affects both the faulted side and the non-faulted side. ing, the increase in short-circuit currents is much higher than when
When non-faulted side is ungrounded, the increase is slightly the stabilizing winding is the outermost winding.
higher. It can be concluded that the stabilizing windings in three-phase
Nonetheless, the omission of the stabilizing winding produces three-legged YNynd transformers may be omitted when primary
temporary overvoltages that can be easily withstood for the insu- and secondary neutral are both grounded (the most frequent sit-
lation of the analyzed transformers. uation), thus reducing very significantly the severe mechanical
requirements to which the transformer is subject during asym-
4. Conclusions metric short-circuits in case of YNynd transformers, and without
tank overheating hazard. When the primary or secondary neutral
Clear understanding of the influence of stabilizing windings remains ungrounded, in the absence of the stabilizing winding tank
on tank overheating hazard and short-circuit duty in three-phase overheating hazard should be critically reviewed by manufacturer
three-legged YNynd transformers is a key factor when evaluating and user, taking protection settings into account. In such cases, uti-
the need for stabilizing windings in such transformers. lization of magnetic shunts in the tank walls should be considered
This paper presented experimental data and simulations for as a possible alternative to the stabilizing winding.
zero-sequence ampere-turn balancing, zero-sequence tempera-
ture rise tests and short-circuit duty in three-phase three-legged Acknowledgment
YNynd transformers. Special attention was paid to features such
as the relative position of the stabilizing winding, the presence of The authors would like to thank Samuel Fernandez, from Uni-
magnetic shunts and the effect of primary and secondary neutral versidad Carlos III de Madrid, for his contribution in short-circuit
grounding, which are usually underestimated or scarcely analyzed. calculations.
In a qualitative approach, ampere-turn balance analysis per-
formed on three transformers revealed that most operating Appendix A.
configurations in three-phase three-legged YNynd transformers
present moderate or low tank overheating hazard due to circulation In this section, an example of calculation of the ampere-turn
currents in tank walls. That is true with the obvious exception of the balance shown in Tables 3–5 is presented.
situation in which only the primary or secondary neutrals remain Example of calculation of test configuration HOC in transformer
ungrounded and the stabilizing winding is open, as no compensa- #1:
tion of the zero-sequence ampere-turns is achieved in this case. As reported in Table 3 of Part I of the paper [22], the measured
Additionally, measurements show that the presence of magnetic currents in the zero-sequence impedance tests were: I1 = 1.581%;
shunts reduces ampere-turn winding imbalances, as they provide a I2 = 0; I3 = 1.243%, expressed in percentage from HV rated current
circulation path for the zero-sequence flux, so no currents circulate and SW rated current respectively.
through the tank walls. Considering module (phase): I1 = 5.07 A (−83.44◦ ); I2 = 0 A;
To complete the evaluation of tank overheating hazard in the I3 = 10.361 A (+93.59◦ ).
case of omission of stabilizing windings, a complete methodology HV winding has a number of turns of N1 = 426, HV winding has a
and procedure for specifying and performing zero-sequence tem- number of turns of N2 = 152 and SW winding has a number of turns
perature rise tests have been proposed. These tests were applied of N3 = 164.
to a transformer (designed with innermost stabilizing winding and Therefore, HV Ampere-turns = 2159.8 AT (−83.44◦ ); LV Ampere-
without magnetic shunts) concluding that tank overheating hazard turns = 0 AT; SW Ampere-turns = 1699.2 AT (93.59◦ ).
is not a concern unless, in case of absence of stabilizing windings, From Eq. (1): MMF0 + (N·I)tank = 471.2 AT (−72.7◦ ).
only one of the transformer neutrals is ungrounded. The results In order to achieve a more comprehensive analysis of these
of the temperature rise tests were consistent with the qualitative results, ampere-turns in non-energized windings are expressed as
ampere-turn balance approach in a twin transformer. a percentage of ampere-turns in energized windings (i.e. ampere-
For short-circuit duty, different fault scenarios were simulated turns of the primary winding are the reference in the denominator
to evaluate the influence of the presence of the stabilizing wind- for tests HOO, HOC, HSO and HSC, and ampere-turns of the sec-
ing, internal design characteristics (stabilizing winding position ondary winding are the reference in the denominator for tests LOO,
and presence of magnetic shunts) and neutral grounding schemes. LOC, LSO and LSC). In the example:
Through such simulations, it was quantified that the presence of a (N·I)1 = 2159.8/2159.8 = 100%
closed stabilizing winding significantly increases single-phase fault (N·I)2 = 0%
156 A. Ramos, J.C. Burgos / Electric Power Systems Research 145 (2017) 149–156

(N·I)3 = 1699.2/2159.8 = 78.7% [12] A.N. Garin, Zero-phase sequence characteristics of transformers. Part II:
MMF0 + (N·I)tank = 471.2/2159.8 = 21.8% equivalent circuits for transformers, Gen. Electr. Rev. 43 (April (4)) (1940)
78–82.
[13] F. Descans, Transformateurs à couplage étoile-étoile. Revue ACEC-Charleroi,
References 1969.
[14] T. Ngnegueu, M. Mailhot, A. Munar, Zero-phase impedance and tank heating
[1] M.A. Tsili, S.A. Papathanassiou, Analysis of a three-limb core power model for three-phase three-leg core type transformers coupling magnetic
transformer under earth fault, in: Proceedings of 16th International field and electric circuit equations in a finite element software, IEEE Trans.
Conference on Electrical Machines (ICEM), Cracow, Poland, September, 2004. Magn. 31 (May (3)) (1995) 2068–2071.
[2] L.F. Blume, Influence of transformers connections on operation, AIEE Trans. 33 [15] P. Penabad-Duran, C. Alvarez-Marino, X.M. Lopez-Fernandez, Transformer
(May) (1914) 753–770. tertiary stabilizing windings. Part II: overheating hazard on tank walls, XXth
[3] J.F. Peters, Harmonics in transformer magnetizing currents, AIEE Trans. 34 International Conference on Electrical Machines (ICEM) 31 (May (3)) (2012)
(September) (1915) 2157–2182. 2369–2374.
[4] L.N. Robinson, Phenomena accompanying transmission with some types of [16] IEC 60076-8, Power Transformers. Part 8: Application Guide, 1997.
star transformer connections, AIEE Trans. 34 (September) (1915) 2183–2195. [17] IEEE Std. C57.105, Guide for Application of Transformer Connections in
[5] J.F. Peters, M.E. Skinner, Transformers for interconnecting high-voltage Three-Phase Distribution Systems, 1978 (Reaffirmed 2008).
transmission systems for feeding synchronous condensers from a tertiary [18] U.S. Patent 1,173, 094 by L.F. Blume, Tertiary Winding, United States Patent
winding, AIEE Trans. 40 (June) (1921) 1181–1199. Office, February 1916.
[6] J. Mini Jr., L.J. Moore, R. Wilkins, Performance of auto transformers with [19] R.M. Charley, Recent progress in large transformers, J. Inst. Electr. Eng. 69
tertiaries under short-circuit conditions, AIEE Trans. 42 (October) (1923) (October (418)) (1931) 1189–1207.
1060–1068. [20] P. Penabad-Duran, X.M. Lopez-Fernandez, C. Alvarez-Marino, Transformer
[7] K.E. Gould, Instability in transformer banks, AIEE Trans. 46 (May) (1927) Tertiary Stabilizing Windings. Part I: Apparent Power Rating, XXth
676–682. International Conference on Electrical Machines (ICEM) (2012) 2369–2374.
[8] A. Boyajian, Inversion currents and voltages in auto-transformers, AIEE Trans. [21] G. Bertagnolli, L. Bergonzi, G. Cannavale, G. Caprio, F. Iliceto, B. Dilli, O.
49 (January) (1930) 810–818. Gülyyesil, Power Transmission Reliability. Technical and Economic Issues
[9] A. Boyajian, B.A. Cogbill, The Whys of the Wyes. The Behavior of Transformer Relating to the Short-Circuit Performance of Power Transformers. Paper no.
Y Connections, Publication General Electric Company GEA-6605/GET-3388B, 12-207, CIGRE Session, 2000.
1957. [22] A. Ramos, J.C. Burgos, Influence of stabilizing windings on zero-sequence
[10] B.A. Cogbill, Are stabilizing windings necessary in all Y-connected performance of three-phase three-legged YNynd transformers Part I:
transformers? AIEE Trans. Power Appar. Syst. 78 (October (3)) (1959) equivalent circuit models, Electr. Power Syst. Res. 144 (2017) 32–40.
963–970. [23] A. Ramos, Consideraciones Acerca de la Utilización de Arrollamientos de
[11] O.T. Farry, Tertiary windings in autotransformers, AIEE Trans. Power Appar. Estabilización en Transformadores Estrella-Estrella (Pd. D. Dissertation),
Syst. 80 (April (3)) (1961) 78–82. Universidad Carlos III de Madrid, 2016, January (in Spanish).