G Model

EPSR-4765; No. of Pages 8 ARTICLE IN PRESS

Electric Power Systems Research xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Electric Power Systems Research

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

Influence of automatic control of a tap changing step-up transformer

on power capability area of generating unit
Jan Machowski a , Piotr Kacejko b,∗
a
Warsaw University of Technology, Poland
b
Lublin University of Technology, Poland

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

Article history: A synchronous generator is a source of real and reactive power, which can be conveniently regulated over
Received 4 January 2016 a wide range of values. The real and reactive power of a synchronous generator are limited by the reactive
Received in revised form 22 May 2016 power capability curve, which is determined by the following constraints: stator (armature) current,
Accepted 26 June 2016
rotor (field) current, load angle (owing to steady-state stability), the temperature in the end region of the
Available online xxx
stator magnetic circuit, turbine power, and the terminal voltage of the generator. Large generating units
operate on a high-voltage network by using step-up transformers, which can be equipped with on-load
Keywords:
tap changers. This article discusses how controlling the transformation ratio of step-up transformers
Large synchronous generating units
On-load tap changing control
can enlarge the area surrounded by the reactive power capability curve. Equations suitable for such an
Reactive power capability area analysis are derived, and a numerical example is presented. A new control algorithm for tap changing is
proposed.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction from the system (underexcitation) during the off-peak demand

hours.
Reactive power supplied by a generating unit to a transmission This article discusses the influence of the on-load tap changing
network is limited mainly by the rotor current limiter and permis- control of a step-up transformer on the power capability area of a
sible changes in generator terminal voltage. The influence of these generating unit.
limitations can be decreased by controlling the transformation ratio
of the step-up transformer. 2. Mathematical model
Some operators of transmission networks recommend using tap
changing transformers as step-up transformers for large generat- The power capability curve of synchronous generator is usually
ing units. In the grid code [1] a milder requirement is set, that the specified through a characteristic P(Q) on a plane with coordinates
use of tap changing transformers shall be dependent on the role corresponding to the real power and reactive power of a generator,
of generating units in the electric power system and whether an assuming that its terminal voltage is a parameter [2–7]. In this arti-
owner of a generating unit wants to offer his regulatory capacity to cle, a characteristic P(Q) is presented on a plane with coordinates
the operator. corresponding to the real power and reactive power of the gen-
In recent years, the need to enlarge the regulatory capacity of erating unit on the high-voltage side of the step-up transformer
generating units has been increasing. This results from changes (power supplied to a transmission network). The transformation
in the daily demand curve and from a growing contribution from ratio of this transformer is treated as the control variable.
intermittent sources (e.g. wind farms, solar farms) in total electric- The analysis will be carried out based on an equivalent circuit
ity generation. It is expected that large regulatory generating units diagram of a transformer presented in Fig. 1. For high voltage power
will be able to work within a broad range of changes with the ability system elements the resistance is much less than the reactance i.e.
to generate a large reactive power into the system (overexcitation) R < <X and therefore it is very common [2–6] for simplified analyses
during the peak demand hours and to absorb a large reactive power to omit the resistance and to take into account only the reactance.
It is assumed that the synchronous generator is modelled by syn-
chronous electromotive force E q behind a synchronous reactance,
i.e. Y g = 1/Z g ∼
= 1/jXd . The step-up transformer is modelled by a
∗ Corresponding author. Tel.: +48 81 5384100; fax: +48 81 5384657. series branch Z T ∼= jXT and an ideal transformation ratio ϑ inserted
E-mail address: [email protected] (P. Kacejko). on the secondary side of a transformer model (Fig. 1a).

http://dx.doi.org/10.1016/j.epsr.2016.06.044
0378-7796/© 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: J. Machowski, P. Kacejko, Influence of automatic control of a tap changing step-up transformer on
power capability area of generating unit, Electr. Power Syst. Res. (2016), http://dx.doi.org/10.1016/j.epsr.2016.06.044
G Model
EPSR-4765; No. of Pages 8 ARTICLE IN PRESS
2 J. Machowski, P. Kacejko / Electric Power Systems Research xxx (2016) xxx–xxx

Fig. 1. Equivalent two-port circuits of two-winding transformers: (a) with an ideal

transformer, (b) equivalent model for ϑ > 1, and (c) equivalent model for ϑ < 1

It was proven in [2] that a transformer of impedance Z T and

ratio ϑ may be replaced with a two-port network of ␲ type, where
the admittances of a series branch and shunt branches depend on
the transformation ratio (Fig. 1b and Fig. 1c). Such an equivalent
model works similarly to a series resonance circuit consisting of
inductive and capacitive branches. For ϑ > 1, the shunt equivalent
branch at node s is capacitive and inductive at node g (Fig. 1b).
Capacitive current flowing from voltage source V g through series
reactance ϑXT to the shunt capacitive branch increases the voltage.
Fig. 2. Equivalent circuit diagrams of generating unit.
For ϑ < 1, the shunt equivalent branch at node s is inductive and
capacitive at node g (Fig. 1c). Inductive current flowing from volt-
age source V g through series reactance ϑXT to the shunt inductive where E, V are voltages across ports, and  = arg(EV ) is the differ-
branch decreases the voltage. It appears from the resonance circuit ence of the arguments of such voltages. It is further assumed that
that the change in the step-up transformer ratio has a significant the voltage arguments are calculated with reference to the net-
influence on reactive power. work voltage: V s = Vs ∠0 and E q = Eq ∠ı, i.e. the difference between
A generating unit is analysed where the transformer is equipped arguments of both voltages amounts to ı.
with an on-load tap changer (Fig. 2a) and the transformation ratio Eq. (1) may be applied in the diagram from Fig. 2e or the circuit
can be controlled. Further analysis is conducted on per-unit val- diagram from Fig. 2d; however, when calculating reactive power
ues. The voltage on the secondary side of the step-up transformer for Fig. 2d, the power of a shunt branch must be taken into account.
is divided by the rated voltage of the transmission network. Volt- In both cases, the following is obtained:
age on the primary side of the step-up transformer is divided by
Eq Vs
the rated voltage of the generator. An ideal ratio in the transformer Ps = sin ı (2)
model (ϑ) is defined (Fig. 2b) as the ratio of the secondary volt- ϑ(Xd + XT )
age to the primary voltage. Using the known formulae [2] of the
Eq Vs Vs2
star-delta transformation, the circuit diagram in Fig. 2c may be Qs = cos ı − 2 (3)
ϑ(Xd + XT ) ϑ (Xd + XT )
transformed into the circuit diagram in Fig. 2d. It is noteworthy that
in the obtained equivalent two-port network of a generating unit, Both powers are nonlinear functions of angle ı. This angle is
all of the branches depend on sum (Xd + XT ), which is multiplied referred to as the load angle.
by factors that depend on the ratio in the same manner as in the Real power is inversely proportional to the transformation ratio;
two-port network that replaces a transformer (Fig. 2 and Fig. 2c). and for reactive power, first element is inversely proportional to the
The circuit diagram presented in Fig. 2e is obtained from the ini- transformation ratio and the second element is inversely propor-
tial circuit diagram (Fig. 2b) by bringing the voltage and impedance tional to the square of the transformation ratio. For given voltages,
to the secondary side of the transformer, i.e. the network side. any change in the transformation ratio results in changes in real
All of the circuit diagrams presented in Fig. 2 are equivalent and and reactive power.
may be used depending on the values that are calculated and the
relationships that are searched among them.
3. Capability area Q(P)
For a two-port network consisting of a series reactance X, it can
be proven [2] that the output real power and reactive power are
The capability area of the generating unit, considering control
given in the following formulae:
of the transformation ratio, results from the mathematical descrip-
EV EV V2 tion of the equivalent circuit diagram in Fig. 2 and the conditions
P= sin ; Q = cos  − (1)
X X X discussed below.

Please cite this article in press as: J. Machowski, P. Kacejko, Influence of automatic control of a tap changing step-up transformer on
power capability area of generating unit, Electr. Power Syst. Res. (2016), http://dx.doi.org/10.1016/j.epsr.2016.06.044
G Model
EPSR-4765; No. of Pages 8 ARTICLE IN PRESS
J. Machowski, P. Kacejko / Electric Power Systems Research xxx (2016) xxx–xxx 3

For Eq = Eq max , Eq. (7) takes the following form:

 2 E 2
Vs2 q max Vs
Ps2 + Qs + 2 = , (8)
ϑ (Xd + XT ) ϑ(Xd + XT )

which corresponds to the circle of the radius and the centre shift
determined by the following formulae:
Eq max Vs Vs2
rf = ; qf = − (9)
ϑ(Xd + XT ) ϑ2 (Xd + XT )
Hence, the circle centre shift is inversely proportional to the
square of the ratio, while the radius is inversely proportional to the
ratio. The circle is crossed by a vertical axis (reactive power for real
power equals zero) at the point:
Vs
 Vs

rf + qf = Eq max − (10)
ϑ(Xd + XT ) ϑ
For given real power, bus voltage and transformation ratio the
reactive power resulting from the field heating limit can be com-
puted form the following formula resulting from above Eq. (8):

Fig. 3. Capability area of a generation unit for a given voltage at bus bars. E 2
q max Vs Vs2
Qs = − Ps2 − , (11)
ϑ(Xd + XT ) ϑ2 (Xd + XT )
Condition C1 assumes that the load current cannot be higher The synchronous electromotive force of a generator is approxi-
than the permissible current of the stator winding (armature heat- mately twice as high as the rated voltage, and it can be assumed that
ing limit). On the plane (P, Q), this condition corresponds to a circle Eq max » Vs . In this case, according to formula (10), the tap changing
whose radius and location of centre result from the following rea- control after increasing the ratio ϑ moves down the arc FG resulting
soning. Real and reactive power in per-unit values is determined from the excitation current limiter (Fig. 3), and therefore reduces
by the following formulae: Ps = Vs Is cosϕ and Qs = Vs Is sinϕ. Consid- the area surrounded by the capability curve. The capability area
ering that Is = Ig /ϑ, and after squaring both formulae on both sides becomes smaller in the upper part.
2
and summing up, the dependence Ps2 + Qs2 = (Vs Ig /ϑ) is obtained The synchronous electromotive force Eq that occurs in the above
because sin2 ϕ + cos2 ϕ = 1. On the plane (P, Q), the above equation equations depends on the excitation current. The maximum value
corresponds to the circle of radius Vs Ig /ϑ and the centre at the of Eq max corresponds to the permissible current of a rotor winding.
beginning of the coordinate system. For the given network volt- In general, manufacturers do not allow excitation current values
age Vs and the load of the generator with the maximum current above rated values. Assuming that the generator rotor winding is
Ig = Ig max , the following equation is obtained: not overloaded, i.e. permissible current corresponds to rated cur-
 2 rent If max = Irf and Eq max = Erq , the rated value of electromotive force
Ps2 + Qs2 = Vs Ig max /ϑ , (4) Erq can be calculated as the multiplicity of the rated voltage of a gen-
erator with the multiplication factor depending on synchronous
which corresponds to the circle of radius rg = Vs Ig max /ϑ. reactance and the power factor. This assumption results from the
In Fig. 3, the fragment of this circle has been dotted. Points corre- reasoning presented below. The following formulae analogous to
sponding to the powers for which the condition Ig < Ig max is fulfilled Eq. (1) can be written for the equivalent circuit diagrams of the
lie inside the circle. The circle radius is inversely proportional to generator (the left side in Fig. 2b):
the ratio ϑ, which means that if a step-up transformer controller
increases the ratio, the radius of the circle gets smaller. If the con- Eq Vg Vg2 Eq Vg
Pg = sin ıge ; Qg + = cos ıge (12)
troller of a step-up transformer reduces the ratio, the radius of the Xd Xd Xd
circle gets larger. The maximum value of the stator current is to where ıge is the difference between the arguments of the generator
be adopted according to the manufacturer’s recommendations, e.g. electromotive force and the generator terminal voltage. By squaring
Ig max = IrG or Ig max = 1, 05 IrG , where IrG is the rated current. the equations on both sides and summing them up, the following
Condition C2 assumes that the excitation current cannot be is obtained:
higher than the permissible current of the rotor winding (field heat-  2  E V 2
ing limit). On the plane (P, Q), this condition corresponds to a circle Vg2 q g
Pg2 + Qg + = (13)
whose radius and location of the centre result from the following Xd Xd
reasoning. Eqs. (2) and (3) can be written as follows:  2  2 
Eq Vs Qg Pg
sin ı = Ps (5) Eq2 = Vg2 1 + Xd + Xd (14)
ϑ(Xd + XT ) Vg2 Vg2

Eq Vs Vs2 Eq. (14) is correct for any power and, in particular, for rated pow-
cos ı = Qs + 2 (6) ers PrG = SrG cosϕrG , QrG = SrG sinϕrG , and rated voltage VrG , where
ϑ(Xd + XT ) ϑ (Xd + XT )
cosϕrG is a rated power factor. It is known that synchronous reac-
After squaring both the formulae on both sides and summing tance Xd = Xd pu VrG 2 /S , where X
rG d pu is the reactance in per-unit
up, the following equation is obtained: values. After substituting these dependences in Eq. (14), the fol-
 2 lowing is obtained:
 2 
Ps2 + Qs + 2
Vs2
=
Eq Vs
(7) 2 2
2  2 
ϑ (Xd + XT ) ϑ(Xd + XT ) Erq = VrG 1 + Xd pu sin ϕrG + Xd pu cos ϕrG (15)

Please cite this article in press as: J. Machowski, P. Kacejko, Influence of automatic control of a tap changing step-up transformer on
power capability area of generating unit, Electr. Power Syst. Res. (2016), http://dx.doi.org/10.1016/j.epsr.2016.06.044
G Model
EPSR-4765; No. of Pages 8 ARTICLE IN PRESS
4 J. Machowski, P. Kacejko / Electric Power Systems Research xxx (2016) xxx–xxx

Therefore, after simple transformations, the following formulae where ıT is the difference between voltage arguments on both sides
are obtained for the synchronous electromotive force of a generator of the transformer impedance. After squaring Eq. (21) on both sides
with a rated load: and summing up, the following equation is obtained:
  2  V V 2
Erq = krf VrG , krf = 1 + Xd pu (Xd pu + 2 · sin ϕrG ) (16) V2 s g
Ps2 + Qs + 2s = (22)
ϑ XT ϑXT
The rated electromotive force of a generator is a multiplication
The equation corresponds to a circle whose radius and centre
of the rated voltage.
shift are determined by the following formulae:
Condition C3 concerns the limitation of a load angle (under-
excitation limit also referred to as stability limit). On the plane (P, Vg Vs Vs2
Q), this condition corresponds to a straight line whose location and rT = ; qT = − (23)
ϑXT ϑ2 XT
inclination result from the reasoning presented below. By dividing
Eqs. (5) and (6) on both sides, the following is obtained: By solving Eq. (22) with regard to reactive power, the following
formula is obtained:

Vs2  V V 2
Qs = tan−1 ı · Ps − (17) s g Vs2
ϑ (Xd + XT )
2 Qs = − Ps2 − (24)
ϑXT ϑ2 XT
Therefore, after substituting ı = ımax , the following equation is This formula determines reactive power as a function of real
obtained: power and both voltages.
The characteristics Q(P) given in formula (24) are arcs of a circle
Vs2 determined by Eq. (22). As XT << Xd , both the shift and radius of
Qs = mPs + c; m = tan−1 ımax ; c=− (18)
ϑ2 (Xd + XT ) this circle are many times larger than the shift and radius of the
circle determined by Eq. (8), i.e. rT >> rf and |qT | >> |qf |. There-
On the plane (P, Q), Eq. (18) delineates a straight line that for fore, arcs Vg max and Vg min corresponding to the circles (22) shown
Ps = 0 crosses the reactive power axis Q at an angle ımax at point: in Fig. 3 look almost like straight lines.
Eq. (22) can be also transformed to the following bi-quadratic
Vs2 equation:
qf = c = − (19)
ϑ2 (Xd + XT )
aϑ4 + bϑ2 + c = 0 (25)
It is noteworthy that the point corresponds to the circle centre where:
shift determined by Eq. (9). In Fig. 3, the straight line corresponding   2
to Eq. (18) is a dash-dotted line. Vs2 Vg2 Vs2
a = Ps2 + Qs2 ; b= 2Qs − ; c= (26)
According to Eq. (9), the circle centre shift, and thus the straight XT XT XT
line (18), is inversely proportional to the squared ratio. Therefore,
it can be stated that with the transformation ratio reduced by the For given values of voltages on both sides of step-up transformer
tap changing control, the straight line (18) gets lower, and thus the Vg , Vs and real and reactive power Ps , Qs injected to transmission
limitation of reactive power (section AB in Fig. 3) moves down, i.e. network Eq. (25) enables to compute the value of the transforma-
the capability area gets bigger in the lower part. tion ratio ϑ.
Condition C4, concerning the limitation of temperature in the If a step-up transformer has a constant ratio ϑ = ϑ0 = const,
end region of the stator magnetic circuit, cannot be expressed in for the given values of network voltages, the limitations Vg max
simple mathematical dependencies (end region heating limit). A and Vg min resulting from condition (20) may lie within the area
relevant curve must be determined by the generator manufacturer. ABCDEFG (Fig. 3). In this case, the capability area is limited only
In Fig. 3, this curve is BC. to the area HIEDKJ. As a result, the decisive limitation in reactive
Condition C5, concerning the mechanical power of a prime power generation may be the limitation of generator voltage rather
mover (turbine), depends on the turbine type and its adaptation than the limitation of exciting current. A similar observation was
to the operation across a broad range of real power changes (prime mentioned in [7].
mover limit). In Fig. 3, the relevant limitations have been marked The location of the abovementioned arcs Vg max and Vg min
with straight vertical lines corresponding to Pmax and Pmin . depends on network voltage Vs and is not constant. Fortunately,
Condition C6 concerns the following limitation of the generator the impact of network voltage changes is favourable and compli-
terminal voltage (generator voltage limit): ant with the needs of the power system. The arcs cross a vertical
axis (reactive power for zero real power) for the following values:
Vg max ≤ Vg ≤ Vg (20) Vs
 Vs

min
Q (Vg max ) = rT + qT = Vg max − (27)
ϑXT ϑ
For currently used generators, permissible changes in terminal
Vs
 Vs

voltage and frequency are small [8]. For a rated frequency, the per- Q (Vg min ) = rT + qT = Vg min − (28)
missible range of the generator terminal voltage changes is ±5% of ϑXT ϑ
the rated value VrG . For positive reactive power (lagging side in Fig. 3), the differ-
The formula of characteristic Q(P) for the given value Vg can be ence between voltages on the right side of formula (27) is positive
derived in the simplest way by using the circuit diagram presented and increases when network voltage Vs drops owing to the deficit
in Fig. 2a, or, strictly speaking, with the part of the circuit diagram of reactive power in the system. As a consequence, arc Vg max that
that concerns the step-up transformer. Likewise, for Eqs. (2) and (3) limits reactive power rises, making it possible to return the grow-
for the middle part of the circuit diagram from Fig. 2b, the following ing volume of reactive power. The increase in ratio ϑ has a similar
can be written: influence, i.e. results in raising the arc Vg max .
For negative reactive power (leading side in Fig. 3), the dif-
Vg Vs Vs2 Vg Vs ference between voltages on the right side of the formula (28) is
Ps = sin ıT ; Qs + = cos ıT (21)
XT ϑ ϑ2 XT XT ϑ negative, and its absolute value increases when network voltage

Please cite this article in press as: J. Machowski, P. Kacejko, Influence of automatic control of a tap changing step-up transformer on
power capability area of generating unit, Electr. Power Syst. Res. (2016), http://dx.doi.org/10.1016/j.epsr.2016.06.044
G Model
EPSR-4765; No. of Pages 8 ARTICLE IN PRESS
J. Machowski, P. Kacejko / Electric Power Systems Research xxx (2016) xxx–xxx 5

Vs rises owing to the excess of reactive power in the system. As a

result, arc Vg min that limits reactive power falls, making it possible
to absorb the growing volume of reactive power. The reduction in
ratio ϑ has a similar impact, i.e. causes the lowering of arc Vg min .
It is noteworthy that all the above mentioned electrical condi-
tions C1, C2, and C3 depend on voltage Vs at bus bars to which a
generating unit is connected. The value of this voltage affects radius
rg of circle (4), radius rf and shift qf of circle (8), as well as shift qf of
the straight line (18). The higher the value of voltage Vs , the wider
the area ABCDEFG. Therefore, characteristic Q(P) presented in Fig. 3
is always determined for a given value of voltage Vs .

4. Example

A generating unit is connected to a transmission network of

400 kV and has the following data: SrG = 1042 MVA, VrG = 27 kV,
Xd = 223%, cosϕrG = 0, 85, permissible voltage range Vg = VrG ± 5%;
maximum stator current Ig max = 1, 05 InG , SrT = 1050 MVA, XT ∼ = 17,
71%; VrT primary = 27 kV; VrT secondary = 407kVVrT secondary = 407 kV.
For comparison, this analysis has also been conducted for an
option that assumes an installed step-up transformer of constant
transformation ratio and rated voltage VrT secondary = 420 kV.
The following base units have been assumed: [S] = 1000 MVA,
[Vs ] = 400 kV, [Vg ] = 27kV. After calculating the rated data for per-
unit values, the following is obtained: SrG = 1, 042, VrG = 1, SrT = 1,
050, VrT g = 1 and VrT s = 1, 0175. The reactance in per-unit values was
calculated according to the formula Xpu = X · Vr2 pu /Sr pu , where
Vr pu and Sr pu are rated voltage and rated apparent power in per-
unit values.
The relative transformation ratio for the central tap is
ϑ0 = (407/400)/(27/27) = 1, 0175. The ratio control range is ϑ = 1,
0174 ± 7 ×0, 0125, i.e. 0, 930 ≤ ϑ ≤ 1, 105.
For an load angle limiter, it was assumed that ımax = 90o .
For the option with a transformer that has a constant trans-
formation ratio, the relative transformation ratio is obtained as
ϑ0 = (420/400)/(27/27) = 1, 05.
Fig. 4a and b are prepared for the case where the network voltage
Vs = 1, 05 Vr . Fig. 4a concerns the option where the transformer
has a constant transformation ratio. It is evident that a large upper
part of the capability area cannot be used without exceeding the
permissible voltage of the generator. Here, the limitation of the
terminal generator voltage significantly reduces the capability area.
In the lower part, such a reduction of the power capability area is
much smaller. If condition (20) is fulfilled, only the area delineated
with the bolded continuous line is available. The area enclosed with
the dotted line is available after exceeding the permissible value of
the generator voltage. In the discussed case, the network voltage is
at a maximum level. From the point of view of the power system,
the unavailability of the upper part of the area above Vg max does
not have to pose a major problem for the operation of the power
system, as no additional reactive power is needed in this case. In
Fig. 4a, vertical arrows show the direction in which arcs Vg max and
Vg min will shift following the relevant increase and decrease in the
transformation ratio.
Fig. 4b concerns the option with a tap changing transformer and
automatic control of the transformation ratio. In the upper part
of the power capability area, when the reactive power generation
Fig. 4. a Power capability area in case of step-up transformer with a constant trans-
increases, the tap changer controller increases the transforma- formation ratio for Vs = 1, 05 and ϑ = 1, 05 = const b. Power capability area in case of
tion ratio. As a consequence, arc Vg max rises, thus enlarging the tap changing step-up transformer for Vs = 1, 05 and 0, 930 ≤ ϑ ≤ 1, 105
available area within the reactive power. In the lower part of
the capability area, after increasing the absorption of the reactive
power by a generator, the tap changer controller decreases the delineated in Fig. 4b with a bolded continuous line is available. As
transformation ratio. As a consequence, arc Vg min decreases signifi- shown by the comparison of Fig. 4b and Fig. 4a, the area obtained
cantly (outside the underexcitation limit). Thus, the limitation (20) owing to the ratio control is much larger than the area with no
becomes irrelevant. Owing to the control of the ratio, the entire area control.

Please cite this article in press as: J. Machowski, P. Kacejko, Influence of automatic control of a tap changing step-up transformer on
power capability area of generating unit, Electr. Power Syst. Res. (2016), http://dx.doi.org/10.1016/j.epsr.2016.06.044
G Model
EPSR-4765; No. of Pages 8 ARTICLE IN PRESS
6 J. Machowski, P. Kacejko / Electric Power Systems Research xxx (2016) xxx–xxx

Fig. 5. a. Power capability area in case of step-up transformer with a constant trans-
Fig. 6. a Power capability area in case of step-up transformer with a constant trans-
formation ratio for Vs = 1, 00 and ϑ = 1, 05 = const b. Power capability area in case of
formation ratio for Vs = 0, 95 and ϑ = 1, 05 = const ϑ = 1, 05 = const b. Power capability
tap changing step-up transformer for Vs = 1, 00 and 0, 930 ≤ ϑ ≤ 1, 105
area in case of tap changing step-up transformer for Vs = 0, 95 and 0, 930 ≤ ϑ ≤ 1, 105

Fig. 5a and b have been prepared for the case where network need to be lowered. In this case, the control of the transformation
voltage Vs = 1, 00. In this case, in the option with a constant trans- ratio (Fig. 5b) moves the limitations Vg max and Vg min outside the
formation ratio, condition (20) limits the upper part of the area area resulting from the field heating limit and the underexcitation
less, and the capabilities of increasing reactive power generation limit.
are greater than in the case indicated in Fig. 4a. As a consequence, Fig. 6a and b are presented for the case where network voltage
it is possible to increase the voltage in the network. The limitation Vs = 0, 95. In this case, for the option with a constant transformation
of the lower part of the area is larger. It is not possible to absorb ratio, condition (20) limits the lower part of the area, which is not
reactive power from the transmission network, which is irrelevant significant, as with such low voltage in the network the absorption
for the voltage assumed here, as the voltage in the network does not of reactive power from transmission network by generator would

Please cite this article in press as: J. Machowski, P. Kacejko, Influence of automatic control of a tap changing step-up transformer on
power capability area of generating unit, Electr. Power Syst. Res. (2016), http://dx.doi.org/10.1016/j.epsr.2016.06.044
G Model
EPSR-4765; No. of Pages 8 ARTICLE IN PRESS
J. Machowski, P. Kacejko / Electric Power Systems Research xxx (2016) xxx–xxx 7

The on-load tap changing control (OLTC) of the step-up trans-

former (Fig. 7) must be realized as the voltage regulation according
to the following rule:

Vg > (Vg ref + V ) increase ϑ
(29)
Vg < (Vg ref − V ) decrease ϑ

where Vg is the generator terminal voltage, Vg ref is the reference

value provided by the substation control system, and V is the dead
zone of a controller selected according to the following formula

V = (0, 6 ÷ 0, 7) · Vtap (30)

where Vtap is the voltage on the tap of the transformer winding.

This selection of a dead zone guarantees stable control, because
the width of the dead zone ±V is greater than the tap voltage, i.e.
2V ≥ (1, 2 ÷ 1, 4)Vtap . The reference value of the tap changing
controller can be set by the substation controller (Fig. 7) depend-
ing on the reactive power of a generating unit according to the
following rule:
 Qs

Vg ref = 1+ · VrG (31)
SrG
where VrG , SrG are, respectively, the rated voltage and rated appar-
ent power of a generator, and  is a small positive number (gain).
Recommended value of this gain is equal to:
Vg max − VrG 1
= · (32)
VrG sin ϕrG
For Vg max = 1.05 · VrG this gain is equal to  = 0.05/sinϕrG .
For Qs > 0 (lagging current) formula (31) provides Vg ref > VrG .
Fig. 7. Functional diagram of a generation unit control AVR: automatic voltage
Particularly for Qs = QrG = SrG sinϕrG it is obtained that Vg ref = Vg max
regulator, OLTC: on-load tap changing controller
and generator operates with maximal permitted voltage. For Qs = 0,
formula (31) provides Vg ref = VrG . For Qs < 0 (leading current) it is
obtained that Vg ref < VrG . Obviously, value Vg ref computed from
not be reasonable. Also in this case, the control of ratio (Fig. 6b)
formula (31) must be passed through a limiter [Vg min ; Vg max ] deter-
moves the limitations Vg max and Vg min outside the area resulting
mining permitted generator voltage.
from the field heating limit and the underexcitation limit.
To minimize the number of tap changes made during a day, a
delay function can be used depending on the deviation in the con-
5. Control system trolled voltage [2] or on the distance between the current operating
point and the capability curve [9].
It is assumed that the control of a generating unit includes a gen- It is noteworthy that in rule (29), if voltage Vg is low, the trans-
erator and a step-up transformer. Therefore, the control concerns formation ratio ϑ is reduced, while in the case of high voltage it is
two facilities that differ in terms of speed and nature of control. Con- increased. At a first glance, it may seem that the opposite should be
trol of a synchronous generator is continuous and fast. Control of the case. However, this dependence is correct. Given the above def-
a step-up transformer is discrete (stepwise) and much slower than inition of a transformation ratio, the primary voltage is obtained by
in the case of a generator. In order to provide the proper control dividing the secondary voltage by the transformation ratio. There-
process for these two different devices (a generator and a trans- fore, in order to increase the primary voltage (generator side), the
former), their controllers should not operate according to the same transformation ratio must be reduced, and vice versa: to reduce the
criterion [9–11], i.e. maintain the values of the same magnitude primary voltage (generator side), the transformation ratio must be
(e.g. terminal voltage of a generator or bus-bars voltage of a sub- increased.
station). In order to provide proper process of control, a transformer Inside the control area, both controllers (AVR and OLTC) shown
controller should maintain the reference value of the voltage on the in Fig. 5 cooperate in the following way.
primary side of the step-up transformer (i.e. terminal voltage of a When network voltage falls, in the upper side of the power capa-
generator), while a generator controller should maintain the refer- bility area (owing to the shortage of reactive power in the system),
ence value on the secondary side of the step-up transformer (i.e. on generator controller AVR increases the excitation current, which
substation bus bars). increases the generator terminal voltage and, according to rule
A functional diagram of the considered control system is shown (29), transformer controller OLTC increases the ratio. As a result,
in Fig. 7. When circuit breakers are closed, then Vs = Vb . Obviously, arc Vg max rises as well. This enlarges the actual power capability
voltage control at the substation bus bar Vs is to be done with the area of a generating unit (Fig. 3).
use of current compensation to obtain a small droop (bias) in the When the network voltage rises, at the lower side of power
voltage characteristic [2]. Current compensation must be carried capability area (owing to excessive reactive power in the system),
out towards a step-up transformer. As a result, the generating unit generator controller AVR lowers the excitation current, which low-
is ‘seen’ from the side of the transmission network as a source ers the generator terminal voltage and, according to rule (29),
of voltage behind a small reactance that corresponds to the small transformer controller OLTC lowers the ratio. As a result, arc Vg min
impedance of the current compensation applied in the automatic decreases, which results in the enlargement of the actual power
voltage regulator (AVR). capability area of a generating unit (Fig. 3).

Please cite this article in press as: J. Machowski, P. Kacejko, Influence of automatic control of a tap changing step-up transformer on
power capability area of generating unit, Electr. Power Syst. Res. (2016), http://dx.doi.org/10.1016/j.epsr.2016.06.044
G Model
EPSR-4765; No. of Pages 8 ARTICLE IN PRESS
8 J. Machowski, P. Kacejko / Electric Power Systems Research xxx (2016) xxx–xxx

6. Optimal range of transformation ratio it is assumed that Qs = 0 for PV node and Vg = VrG for generator
terminals.
From above consideration it can be concluded that optimal
range of changes of the step-up transformer ratio should satisfy 8. Conclusions
two conditions:
Because of the growing number of intermittent sources of
(a) For lagging current (upper part of power capability area) maxi- energy in power systems, conventional large generating units are
mal transformation ratio ϑmax should guarantee that generator expected to have greater regulation capabilities with regard to both
voltage limit is shifted behind the field heating limit. real and reactive power.
(b) For leading current (lower part of power capability area) mini- The main limitation in the abilities of a generating unit in the
mal transformation ratio ϑmin should guarantee that generator generation or absorption of reactive power is the low permissi-
voltage limit is shifted below the underexcitation limit. ble range of generator voltage changes (generator voltage limit).
The application of step-up transformers with an on-load controlled
Condition (a) is satisfied when for Vs = Vs max and Vg = Vg max the ratio has the positive influence on the power capability area shif-
value given by (27) is larger than the value given by (10). Solving ting the generator voltage limit behind the field heating limit (for
such inequality leads to the following value: lagging current) or behind the underexcitation limit (for leading
x · Vs max current). The transformation ratio of a step-up transformer may be
ϑmax ≥ (33) controlled by a local and/or central control system or by an auto-
(x + 1) · Vg max − Eq max
matic controller that operates in voltage control mode with a dead
where x = Xd /XT . zone and time delay. The controlled voltage should be the voltage
Condition (b) is satisfied when for Vs = Vs min and Vg = Vg min the on the primary side of a step-up transformer, i.e. on the generator
value given by (28) is smaller than the value given by (19). Solving side.
such inequality leads to the following value: The formulae derived in this article make it possible to deter-
Vs min x mine the capability area of a generator-transformer unit, taking into
ϑmin ≤ (34) account the value of the step-up transformer ratio. The presented
Vg min x + 1
numerical example shows the impact of ratio changes on the shift
For generating unit with data described in Section 4 it is obtained of relevant capability curves that limit the power capability area.
ϑmin ≤ 0.927 and ϑmax ≥ 1.159. It means that range 0, 930 ≤ ϑ ≤ 1,
105 assumed in data described in Section 4 is too small (especially References
for ϑmax ). This fact is also evident from Fig. 4b, where the generator
voltage limit is still a constrain for reactive power in the upper part [1] The Grid Code, Issue 4, rev 2, 22.03.2010, National Grid Electricity Transmission
of the power capability area. plc.
[2] J. Machowski, J.W. Bialek, J.R. Bumby, Power System Dynamics: Stability and
Control, John Wiley & Sons, Chichester, New York, 2008, ISBN 978-0-470-
7. Adjustment of OLTC + AVR in load flow programme 72558-0.
[3] P. Kundur, Power System Stability and Control, McGraw-Hill, New York, 1994.
[4] C.W. Taylor, Power System Voltage Stability, McGraw-Hill, New York, 1994.
In load flow computer programmes the nodes are classified as [5] T. Van Custem, C. Vournas:, Voltage stability of electric power systems, in: The
[2,6]: (a) load nodes PQ with constant power or voltage charac- Kluwer International Series In Engineering And Computer Science, Springer
teristics of loads, (b) voltage controlled PV nodes with Q limits Science+Business Media Dordrecht, 1998.
[6] J.J. Grainger, W.D. Stevenson Jr., Power System Analysis, McGraw-Hill, Inc, 1994.
determined by power capability curve. Selected node or some PV [7] D. Reimert, Protective Relaying for Power Generation Systems, CRS Press Tay-
nodes are treated as the slack nodes. OLTC of transformers oper- lor&Francis Group, 2006.
ating in transmission network is adjusted to iterative algorithms [8] IEEE C50.13-2005 IEEE Standard for Cylindrical-Rotor 50 Hz and 60 Hz Syn-
chronous Generators Rated 10 MVA and Above.
solving network equations as described, e.g. in Refs. [6,12].
[9] R. Malkowski, Z. Szczerba, A. Zbronski, The analysis of a possible coordina-
For step-up transformers of generating units situation is differ- tion of operating algorithms for step-up transformer controller and generator
ent, because OLTC of such transformers must cooperate with AVR controller, Acta Energetica Nr 4 (2011) 35–42.
[10] J.J. Paserba, M. Shimomura, S. Tanaka, D.J. Shoup, R.T. Hellested, Enhanced Gen-
of generator as described in Section 5. The simplest way is to solve
erator Controls for the Improvement of Power System Voltage Stability, in:
the network equations using voltage controlled PV nodes with Q Symposium of Specialists in Electric Operational and Expansion Planning (VIII
limits computed in each iteration form equations describing the SEPOPE) Brasilia, Brazil, May 19-23, 2002.
power capability area. In each iteration (taking into account reac- [11] H. Kitamura, M. Shimomura, J.J. Paserba, Improvement of Voltage Stability by
the Advanced High Side Voltage Control Regulator”, in: Proceedings of the 2000
tive power and voltages form previous step) the relevant generator IEEE PES Summer Meeting, Seattle, WA, July, 2000, pp. 278–283.
voltage Vg = Vg ref must be computed from Eq. (31) and transforma- [12] D. Nedic, Tap Adjustment in AC Load Flow, UMIST, September, 2002,
tion ratio ϑ from Eq. (25). Iterations are performed until presumed https://www.researchgate.net/publication/265631174 Tap Adjustment in
AC Load Flow.
accuracy is reached. For staring point (data to first iteration)

Please cite this article in press as: J. Machowski, P. Kacejko, Influence of automatic control of a tap changing step-up transformer on
power capability area of generating unit, Electr. Power Syst. Res. (2016), http://dx.doi.org/10.1016/j.epsr.2016.06.044