Network modelling of traction substation

transformers for studying unbalance effects

T.H. Chen
H.-Y. KUO

Indexing terms High-speed railway, Substation transformer, Unbalance, Single-phase connection, V-connection, Scorr connection

Abstract: The paper presents network mathematical models of traction substation transformers
for simulation of the unbalance effects, due t o
railway demands, on the power supply system.
The admittance models consider winding connections and phase-angle displacement between
primary and secondary voltage. The proposed
models are presented for the single-phase, V- and
Scott connections of transformers used in, among
other places, high-speed railway stations.

Introduction

High-speed railway demands could cause imbalances of

voltage and current on the power-supply system. The
type of high-speed railway substation transformer is the
major factor that leads to system unbalance. Simple formulas are usually used to evaluate the voltage unbalances at the point of common coupling (PCC). Although
these formulas can easily be applied to approximate the
voltage unbalances [I], the interaction of all high-speed
railway substations demands are not considered. These
interactions are significant, especially for neighbouring
substation loads. Therefore detailed models of substation
transformers are required, to enable three-phase powerflow programs to analyse the effects of large unbalanced
high-speed railway loads. The models are developed in
phase coordinates, and consider the copper loss, winding
connections and phase-angle displacement between
primary and secondary voltage. The core losses of the
transformers can easily be taken into account if related
data, that is usually obtained from the field tests, is provided. Off-nominal tapping can also be considered if
required [Z].
The copper loss of a transformer is granted if the
windings resistances are considered, that is, the conductance portion of the equivalent admittance of the transformer is not neglected. On the other hand, the core loss
of the transformer is approximated by shunt core-loss
functions on each phase of the secondary terminal of the
transformers, as used in Reference 2. It can be easily
added into the proposed equivalent transformer model
without influence on the configuration of the latter.

C IEE, 1995
Paper 1592C (P2), first received 21st April and in revised form 12th
September 1994
The authors are with the National Taiwan Institute of Technology,
Department of Electrical Engineering, 43 Keelung Road, Section 4,
Taipei, Taiwan, Republic of China
I E E Proc.-Gener. Transm. Disfrib., Vol. 142, N o . 2, March 1995

The effects on the power-supply system, due to highspeed traction loads, can be realised after detailed
analysis is done. The various ways to prevent or reduce
the degree of the voltage imbalance of the PCC can
therefore be proposed and conducted.
The derived mathematical models were confirmed by
the Electromagnetic Transients Program (EMTP) and
Personal
Computer
Simulation
Program
with
Integrated-Circuit emphasis (PSPICE). For comparing
the voltage-unbalance results, evaluated by applying the
simple formula, the results of the cases where the transformers are assumed to be ideal are presented. The
simulation results show the correctness of these derived
models.
2

Derivation of transformer models

Single-phase power is commonly used to supply electric

locomotive load. The high-speed railway between Taipei
and Kaohsiung, which is currently planned, converts
three-phase power from the Taiwan Power Company
161 kV high-voltage grid to two 25 kV single-phase
powers for high-speed electric traction loads. The phase
conversion can be achieved by various transformer connection schemes, such as single-phase, V- and Scott connections. The connection scheme of the substation
transformer and the balancing of two traction loads of
the substation will significantly affect the voltage unbalance on the PCC. These two factors are considerable,
and therefore should be taken into account in the derivation of the transformer models. The accuracy and reliability of models, of course, should be considered during
the derivation.
For simplification, the single-phase transformer model
and interconnection method are applied to compose the
proposed transformer models. Because the transformers
are approximated by their equivalent linear models, the
derived mathematical models will be constrained in
steady-state studies.
Fig. 1 shows the circuit diagrams of a single-phase
transformer, in which y, represents the equivalent admittance of the single-phase transformer obtained by a
short-circuit test.
The nodal admittance matrix of the single-phase transformer can easily be obtained by a building algorithm,

yt

Fig. 1

Circuit diagram oJa single-phase transformer

I03

and can be written as eqn. 1

%ode

the equivalent single-phase transformers in networks 1

and 2 are set k and (1 - k), respectively:

N
In a per-unit system, the turns ratio ( N 1 / N 2 )is equal to
unity. Hence, the coupling-free equivalent circuit model
of a single-phase transformer is shown in Fig. 2 [4-61.

Yr
(1 - k )
-- Yl
(1 - k )

-~Yl
(1 - k )
Y,
(1 - k )

-Yt

Yl

Yl

- Yl

-Yr

Yf

Yt

-Yt

B
a
b

(3)
Fig. 2 Coupling-free equivalent circuit model of a single-phase twowinding transformer

The circuit diagram of a single-phase transformer with

tapping in the primary side is shown in Fig. 3, in which
the k is less than 1.

The equivalent mathematical model of the single-phase

transformer, with tapping on the primary side, can therefore be obtained by combining these two equivalent networks by means of the interconnection method.
Combining eqns. 2 and 3, the nodal admittance matrix of
this transformer is shown in eqn. 4. The corresponding
equivalent circuit is figured in Fig. 5.
xode

Yb:i,

+ Yk:ie =
A

Fig. 3 Circuit diagram of a single-phase transformer with tapping on

the primary side

This transformer can be partitioned into two networks. Each network is represented by its equivalent
single-phase transformer. The primary windings of these
two single-phase transformers are in series, while the
secondary windings are in parallel. The equivalent shortcircuit admittance of the equivalent single-phase transformer should be corrected appropriately, to maintain
the original per-unit current on both sides. The two
equivalent networks of a single-phase transformer with
tapping on the primary side are shown in Fig. 4.

B
a
b
N

(4)
If node N is not tapped out, eqn. 4 can be simplified by
Kron's reduction. As expected, the result after reduction
as shown in eqn. 5 is the same as that of the general
single-phase transformer without tapping on both sides,

Fig. 4 Equivalent networks for a single-phase transformer

tapping on the primary side

with

The nodal admittance matrix of networks 1 and 2 can

be derived from eqn. 1, and are written in eqns. 2 and 3.
It is noted that the line voltages of the transformer are set
1 per unit on both sides, and the turns ratios ( N J N , ) of
104

yt

BC

Ob

Fig. 5 Coupling-free equivalent circuit model of a single-phase transformer with tapping on the primary side
IEE Proc.-Gener. Transm. Distrih., Vol. 142, No. 2, March 1995

as shown in eqn. 1 with N J N , = 1. This proves the correctness of the approach methodology:
A

Kode

from the primary side, and supplies two single-phase

loads on the secondary side. There is phase-shifting
between these two single-phase loads, so an isolation
section should be set to isolate two different phases of
power. The connection scheme is shown in Fig. 8. The
equivalent circuit model can be made available by the
interconnection method, as shown in Fig. 9.

The same approach methodology described above can be

applied to the single-phase transformer with tapping on
the secondary side, or on both the primary and secondary sides. The results are obvious and will not be presented here.

secondary

2.1 Single-phase connection

The single-phase transformer model and the linear interconnection method are fundamental to constructing the
specified transformer models. First, the single-phase connection is considered. Its connection scheme is shown in
Fig. 6 . On the basis of a single-phase transformer circuit
PCC
A

Fig. 8

V-connection

Tr f t l

Yt

I I I

IL2

2E

sL2

PCC

Fig. 6

Single-phase connection

model, the equivalent circuit model can easily be

obtained by the interconnection method, as shown in
Fig. 7. It is noteworthy that the connection of the highvoltage network does not work out phase-to-ground but
phase-to-phase. Therefore, applying the single-phase connection scheme in the traction substation, the power of
traction loads was supplied only by two phases of its
power-supply system. The more loading or the shorter
the system capacity, the more imbalance on the PCC.
Tr # I

Fig. 9

Equivalent circuit of V-connection

2.3 Scott connection

The Scott connection, which employs two single-phase
transformers, is the most frequently used transformer for
transferring three-phase power to two-phase power. The
first transformer is called the main transformer, and the
second one is called the teaser transformer [7]. Fig. 10
shows the circuit diagram of the Scott connection.

I
I

Fig. 7

Tr # 2
Equivalent circuit ofsingle-phase connection

2.2 V-connection
The V-connection of a transformer is composed of two
single-phase transformers. It draws three-phase power
IEE Proc.-Gener. Transm. Distrib., Vol. 142, No. 2, March I995

PCC

Fig. 10

.L

Scott connection
105

matrix of networks 1. 2 and 3 are written as eqns. 6, 7

and 8, respectively:

Because the size and turns ratio of these two transformers

are different, their equivalent short-circuit admittance are
therefore different. The equivalent short-circuit admittance could be represented by y M and y T for main and
teaser transformers, respectively.
The Scott transformer can be divided into three partial
networks to easily contruct equivalent circuits. Each
network is represented by a single-phase transformer.
The main transformer has a 50 percent tap on the
primary side, and is divided into two subnetworks. Their
equivalent short-circuit admittance should be corrected
as described above. The three partial networks of the
Scott connection are shown in Fig. 1 1 .

YbtL,

a1

a2

Jo

Jo

yT

ay,

-&y,
2

3
I Network

<-,

_ -1
2yM

Phase voltages of both the primary and secondary side

of the transformer are selected as base voltages. The perunit phase-to-phase voltages on both the primary and
secondary sides are therefore equal to J ( 3 ) p.u. The
nodal admittance matrices of these three subnetworks
can be obtained by eqn. 1, with little modification. The
turns ratio N J N , of networks 1 , 2 and 3 are set to
J(3)/2, 1/2 and 1/2, respectively. The nodal admittance

YT

iy7

- i y T

; r N 1 : N2

YT

iYM.

The nodal admittance of the Scott transformer can therefore be written as eqn. 9 by combining eqns. 6, 7 and 8:

2YM

2YM

-Joy,
2
JoYT
2

0
0

The nodal current inode

= Kode x

106

Eodewhere

I E E Proc.-Gener. Transm. Distrib., Vol. 142, N o . 2, March 1995

If y M = 2 y r , which is the common case, eqn. 9 can be simplified as

-Yr

-4Yr
-4Yr
-4

3 )YT
2

Jo

YT

0
0

94T

The Krons reduction can be applied to reduce eqn. IO, by eliminating internal node D, giving a more compact form, as
shown in eqn. 11. The corresponding equivalent ciruit model of the Scott transformer can therefore be configured as
shown in Fig. 12.

3YT
2

jY 7
0
0

jYT
0
0

2Yr

-2Yr

-2YT

2YT

high-speed railway substation can easily be obtained by

using the EMTP and PSPICE programs, and will be discussed here. The simulation diagram of a high-speed
railway power-supply substation is shown in Fig. 13, in
which the utility high-voltage network is represented in
utility high voltage network

Fig. 12

PCC

Equiualent circuit ofScott connection

Fig. 13 Simulation diagram oJa high-speed railway power-supply suhstation

PCC = point of common coupling

Unbalanced analysis of t h e proposed models

The large single-phase traction loads of high-speed railways are major causes of voltage unbalance on the PCC.
The different connection schemes of the substation transformer could result in different degrees of voltage unbalance. The degree of voltage unbalance on the PCC of a
I E E Proc-Gener. Transm. Distrib., Vol. 142, N o . 2, March 1995

the Thevenin equivalent circuit by short-circuit capacity

on the PCC. The proposed models of single-phase, Vand Scott connections are substituted in the diagram to
evaluate the unbalanced effect.
Voltage sources are
= 27cf

E,

1.0 sin cot (P.u.) o

E,

1.0 sin (wt - 120) (P.u.)

E,

1.0 sin (wt

+ 120)(P.u.)

,f = 60 Hz

(13)
107

The short-circuit capacity (S,) at the PCC is assumed to

be 4000 MVA, and the total demands of two single-phase
traction loads are 50 MVA. That means S, = SLl
S,, = 50 MVA. All values are represented in per-unit
values. The single-phase transformer rated power (S,,,4,)
and rated line to neutral voltage (VL,) are selected as base
power and voltage in the per-unit system. By these two
base values, the base values of the current and impedance
can be calculated. The related formulas are as follows:

compare the voltage unbalances for the three transformer

connection schemes. The results are almost the same as
those obtained by the simple formulas [l]. It should be
noted that the formulas were obtained on the basis that
the transformers are ideal. Hence, a slight difference
between these two results is reasonable. Fig. 14 corroborates the correctness of the proposed models.

Equivalent per-unit impedance of the power system:

zs(p,",)
=s
,,,,,
~ $ 3

Equivalent per-unit impedance of loads.

where Z , = equivalent impedance of the power-supply

system, Z , = equivalent impedance of loads, V,, = lineto-line voltage and V,, = line-to-neutral voltage.
Line to neutral voltage V,, is 1 p.u. in the per-unit
system. Therefore eqn. 15 can be rewritten as eqn. 16

The capacity rating of the single-phase transformer is

assumed to be 50 MVA, and the short-circuit admittance
y , is assumed to be 12.5 p.u. The resistance of the transformer is neglected in this case. The short-circuit capacity
of the substation is assumed to be 4OOO MVA, therefore
the equivalent impedance of the power-supply system,
calculated by eqn. 14, is 0.0375 p.u. The load-balancing
ratio is defined as k = S L , / S L , to simplify the discussion
of the unbalanced effect, while the balancing of the loads
on the secondary side is changed. The equivalent per-unit
impedance of two single-phase loads can be obtained by
eqn. 16. The equivalent circuit of the specified transformers is simulated by using EMTP and PSPICE. The
magnitudes and phase angles of the voltage on the PCC
are computed when the load-balancing ratio is changed.
The voltage unbalanced factor [SI is defined as d , =
I V, I I .VI I in this paper, and is calculated for all loading
conditions, where V, and V, represent the negative- and
positive-sequence voltages on the PCC.
The simulation results are shown in Fig. 14, to
sinqle-phose connection

01

0.1

0.2

0.3 0.4

0-5 0.6 0-7

0-8 0-9

Fig. 14
Comparison of voltage unbalances for the three transformer
connection schemes

108

Conclusion

Models of single-phase transformers, with and without

tapping, were introduced in this paper, followed by
models of three specified transformers that are commonly
used in high-speed railway stations. The three transformer connection schemes are single-phase, V- and Scott
connection. The building algorithm of the nodal admittance matrix and the interconnection method are used to
develop the proposed models. The proposed approach
can easily be used to construct circuit models for various
transformer connections, and be easily realised.
Commercial packages, EMTP and PSPICE, are used
to simulate the circuit models, and are also applied to
corroborate the correctness of the proposed mathematical models of the specified transformers. The simulation
results show that the proposed models can accurately
represent their corresponding transformer in the steadystate studies.
The proposed models are simple, and can be easily
implemented into a three-phase power-flow program to
undertake various studies. The effects of large unbalanced
high-speed railway demands on the power-supply system
can be simulated accurately via the three-phase powerflow programs, in which the proposed transformer
models were correctly implemented. Further studies can
also be conducted to reduce the harmful effects of the
large unbalanced loading condition.

References

1 CHEN, T.H.: 'Criteria to estimate the voltage unbalances due to

high-speed railway demands'. IEEE Power Engineering Society

Winter Meeting, 94WM 234-5 PWRS, New York, 1994
2 ARRILLAGA, J., ARNOLD, C.P., and HARKER, B.J.: 'Computer
modelling of electrical power systems' (Wiley, Chichester, New York,
1983). pp. 7-10
3 CHEN, T.H., CHEN, M.S., INOUE, T., KOTAS, P., and CHEBLI,
E.A.: 'Three-phase cogenerator and transformer models for distribution system analysis'. IEEE Trans. Power Defiu., 1991, 6, (4), pp.
1671-1681
4 LAUGHTON, M.A.: 'Analysis of unbalanced polyphase networks by
the method of phase co-ordinates. Part 1 . System representation in
phase frames of reference', Proc. IEE, 1968, 115, (X), pp. 1163-1172
5 SALEH, A.O.M., LAUGHTON, M.A., and STONE, G.T.: 'M- to
N-phase transformer models in phase co-ordinates', IEE Proc. C,
1985,132, (l), pp. 41-48
6 GORMAN, M.J., and GRAINGER, 1.1.: 'Transformer modelling for
distribution system studies. Part 11. addition of models at YBus and
ZBus', IEEE Trans. Power Deliu., 1992,7, (2), pp, 575-580
7 GONEN, T.: 'Electric power distribution system engineering'
(McGraw-Hill, New York, 1986), pp. 149-150
8 'IEEE Recommended Practice for Electric Power Distribution for
Industrial Plants'. IEEE Std 141-1976 (Wiley-Interscience, New York,
1976), pp. 58-59
IEE hoc.-Gener. Transm. Distrib., Vol. 142, No. 2, March 1995