Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Heinrich-Hertz-Str. 9
72810 - Gomaringen
Germany
http://www.digsilent.de
info@digsilent.de
r1010
Copyright 2011, DIgSILENT GmbH. Copyright of this document belongs to DIgSILENT GmbH.
No part of this document may be reproduced, copied, or transmitted in any form, by any means
electronic or mechanical, without the prior written permission of DIgSILENT GmbH.
Contents
1 General Description 4
2.2.1 Yd-transformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5 References 35
List of Figures 36
List of Tables 37
1 General Description
The two-winding transformer model is a very detailed model for various kinds of three-phase,
two-winding transformers in power systems. It can represent e.g. network transformers, block
transformers, phase shifters or MV-voltage regulators. The model makes special consideration
for auto-transformers.
This first section describes the general model and is valid for all PowerFactory calculation func-
tions.
Particular aspects, such as saturation or capacitive effects, which are only relevant for some
calculation functions are described in the following sections.
Section 2. provides useful hints for special applications of the 2-winding transformer model.
The detailed positive-sequence model with absolute impedances (in Ohm) is shown in Fig-
ure 1.1. It contains the leakage reactances and the winding resistances of the HV and LV side
and the magnetization reactance and the iron loss admittance close to the ideal transformer.
The model with relative impedances (in p.u.) is shown in Figure 1.2. The ideal transformer of
the per-unitized model has a complex winding ratio with a magnitude of 1:1 and models the
phase shift representing the vector groups of the two windings
Figure 1.1: Positive sequence model of the 2-winding transformer (in Ohms)
Figure 1.2: Positive sequence model of the 2-winding transformer (in p.u.)
The relation between the mathematical parameters in the model and the parameters in the type
and element dialogs are described as follows:
2
Ur,HV
Zr,HV = (1)
Sr
2
Ur,LV
Zr,LV = (2)
Sr
zsc = Usc /100 (3)
PCu /1000
rsc = (4)
Sr
p
xsc 2 r2
= zsc (5)
sc
rCu,HV = R,HV,1 rsc (6)
rCu,LV = (1 R,LV,1 ) rsc (7)
x,HV = X,HV,1 xsc (8)
x,LV = (1 X,LV,1 ) xsc (9)
1
ZM = (10)
i0 /100
Sr
rF e = (11)
PF e /1000
1
xM = q (12)
1
z2
r21
M Fe
where,
The tap changer is represented by an additional, ideal transformer connected to either the HV
or the LV side (see Figure 1.3 and Figure 1.4). In most application, the winding ratio of this
transformer is real and is defined by the actual tap position (in number of steps) times the
additional voltage per steps.
Phase shifters are modelled by a complex ratio using a complex value of dutap according to
Figure 1.5.
There are two possibilities of specifying a phase shifting transformer. Either by entering magni-
tude and angle (dutap and tap ) of the additional voltage per tap step or by defining magnitude
and angle at each individual tap-step (|U + dutap |, u ). The latter is supported by the measure-
ment report in the transformer element (see also section 1.2.3).
More transformer models for further configurations are shown in section 2.2.
(a)
(b)
Figure 1.6: Yd transformer (a) in the zero-sequence system with HV side tap changer in detailed
(a) and simplified representation (b)
The load flow ComLdf calculation uses the detailed model for the transformer, that is all shunt
and branch impedances for positive- and zero-sequence system.
A component that is of special interest for load flow calculations is the tap changer. In the type
data section it is modelled using its constructive properties, in the element data section it is
defined in its control behaviour for steady-state simulation.
The basic data of the tap changer are listed in the following Table 1.1.
The parameter section for the tap-dependent impedance appears when this option is activated
(see Figure 1.7). Parameters that can be considered to be tap-dependent are the short circuit
impedances and copper losses (short circuit resistance) in the positive- and zero-sequence
systems.
For tap positions between min. and neutral and between neutral and max. tap dependent
parameters are interpolated using splines.
A very precise method tap-changer description is the so-called measurement report. Here, all
tap-dependent parameters can be entered per tap step.
Automatic tap changer control is activated by setting the corresponding option on the load flow
page of the transformer element. Additionally, automatic tap adjustment can be globally enabled
or disabled by the load flow command. The information required for tap changer control is shown
in Figure 1.9 and described in Table 1.3.
Table 1.3: Dialog fields for the automatic tap changer control
Parameter Description
According to Instead of the type data for the tap-dependent transformer
Measurement values the element-specific measurement report is used
report
Tap position Tap position used during the load flow calculation. If Automatic
Tap Changing is activated this value corresponds to the initial
tap position.
Automatic tap Activating automatic tap adjustment in load flow analysis.
changing
continuous An idealized, continuous tap changer is
Tap assumed. As a result, the tap controller can ideally comply with
changer the specified control condition
This option is useful for voltage regulators in distribution
systems having a very large number of tap steps or for thyristor
controlled tap changers.
discrete Standard option. Only integer tap positions are
considered.
HV Tap controls the HV-side.
Controlled
node LV Tap controls the LV-side
EXT Slave mode. The tap changer just follows the tap
position of the selected Master -transformer.
Only for V control mode:
Setpoint local the voltage setpoint and voltage range settings
(max./min. voltage) must be enter in the transformer dialog
bus target voltage the voltage setpoint and voltage range
settings (max./min. voltage) are taken from the controlled
busbar (topological search)
V Voltage control. For unbalanced load flow analysis, the
Control controlled phase needs to be defined additionally.
mode
Q Reactive power control (see also Figure 1.10)
P Active power control (only applicable to phase shifters, see
also Figure 1.10)
Parameter Description
Set Point V-/Q-/P- reference (depending on selected control mode)
Lower/Upper Lower and upper boundary of the controlled variable. In case of
bound discrete tap changers, the tap control can drive the controlled
variable just into a permitted band. In case of continuous tap
changers the tap controller can ideally regulate to the reference
point.
Remote Control Allows for the selection of a bus bar different from the
transformer terminals (V-control). In case of P-or Q-control the
flow through any cubicle can be controlled.
Voltage control includes optional line drop compensation. This function controls the voltage at
a remote busbar without measuring the voltage at that bus-bar. Instead, the actual value is
estimated by measuring the voltage at the HV or LV side of the transformer and simulating the
voltage drop across the line.
The principle of the line drop compensation is shown in Figure 1.11, the corresponding param-
eters are explained in Table 1.5.
Generally, there is more than just one possible solution to a load flow problem considering
automatic tap changer control. Especially in meshed networks, several transformers can control
the voltage in certain areas. In case of parallel transformers, the problem can usually be solved
by operating the two parallel transformers in a master slave mode.
In a general configuration however, especially when parallel transformer have different short
circuit impedances or different tap steps, the steady state network solution cannot be obtained
that easily.
PowerFactory addresses the mentioned problem by allowing the user to enter a controller time
constant, specifying the speed of control actions and hence the participation of several trans-
formers regulating the voltage of the same bus bar.
The approach is based on controller block diagrams according to Figure 1.12. In case of flow-
controllers (P-/Q-control) the controller sensitivity translating a power mismatch into an equiva-
lent turns-ratio percentage can be entered additionally.
In the actual load flow algorithm, which just looks at steady state conditions, controller time
constants and sensitivities are translated into equivalent participation factors.
(a) (b)
Short-Circuit calculations according to IEC assume that the shunt impedances in positive- and
negative-sequence (magnetizing reactance, iron losses) are neglected. The shunt impedances
in the zero-sequence system however must be considered. These parameters are shown in the
dialog of IEC S/C calculation.
Another detail specific to IEC calculation is the distinction between no-load and on-load tap
changers. Different impedance correction factors apply for each group. The property of on-load
variation of the tap changer therefore can be enabled in the IEC S/C calculation dialog.
This page contains additional information which is used to calculate the impedance correction
factor of the transformer.
The first criterion defines whether the transformer is a unit transformer or a network transformer.
In case of unit transformers, one common correction factor is applied to transformer and gener-
ator. Network transformers are individually.
Two different calculation procedures can be applied. The first is a general correction indepen-
dent of the actual operating conditions of a selected transformer. The second is more specific
and may lead to more precise calculation results. The selection of the correction method along
with the additional data required are shown on the S/C page, as can be seen in Figure 1.13.
The model used by the RMS simulation is identical to the load flow model. However, the tap
controller definitions are not considered here. For the simulation of tap controllers, a separate
dynamic model needs to be defined that can be interfaced with the transformer using the input
variable nntapin (tap-input).
For accurately modelling high frequency effects of transformers, additional capacitances need
to be considered, as shown in Figure 1.14.
These capacitances are equivalent capacitances of the model and not the actual winding ca-
pacitances. For obtaining equivalent capacitances from winding capacitances, the winding con-
nection (D/Y) must be considered additionally.
The high frequency model according to Figure 1.14 provides an accurate frequency response
with respect to voltages and currents at the transformer terminals. However, it is not possible to
simulate effects internal to the transformer, such as internal voltage stress.
(a)
(b)
Figure 1.14: HF Model for the external capacitances in positive sequence system (a) and zero-
sequence system (b)
For simulating nonlinear, electromagnetic transient such as transformer inrush currents or ferro-
resonance, core saturation needs to be included into the transformer model. Furthermore,
depending on the frequencies involved in the transient simulation, the transformer model has to
account for the stray capacitances between windings and winding to ground.
Figure 1.15 shows the equivalent model of 2 winding 3-phase transformer for the positive se-
quence. For simplicity, the tap changer has been left aside in the figure; however it is considered
in the model according to Figure 1.3, Figure 1.4 and Figure 1.5 as described in previous chap-
ters.
The exciting current of a transformer (no-load test) consist of an imaginary part, which is the
magnetizing current flowing through the non-linear reactance XM in Figure 1.15, and a smaller
real part flowing through the resistance RF e , which accounts for the excitation losses.
The non-linear magnetizing reactance XM represents the saturation characteristic of the trans-
former and it is defined in the transformer type (TypTr2\EMT simulation page). The model
supports the following options:
Figure 1.15: Equivalent circuit of the 2 winding 3-phase transformer for the positive sequence
The position of the magnetizing branch in the equivalent model of Figure 1.15 is defined in terms
of the distribution of the leakage reactance and resistance (TypTr2\EMT-Simulation page). De-
fault value is 0.5 which means that the total leakage impedance of the transformer (short-circuit
impedance) equally distributes between the HV and the LV winding. The user can modify the
position of the magnetizing branch in the transformer model by modifying these factors.
Figure 16 shows the magnetizing current-flux curves for the two slope and polynomial charac-
teristics. The input parameters of both curves are the same except for the saturation exponent,
which only applies to the polynomial characteristic. The input parameters are listed in Table 7.
Table 1.7: Basic data of the two-slope and polynomial saturation characteristics
The reciprocal of the p.u. unsaturated reactance is equal to the the p.u. magnetizing current
(i.e. the imaginary part of the exciting current). Therefore, the program automatically adjusts
the unsaturated reactance based on the no-load current and no-load losses entered in the load
flow page (TypTr2\Load Flow) and vice-versa:
s 2 2
1 IM Pexc
= (13)
XM Irated Srated
where,
IR , SR : Are the rated current and apparent power of the transformer respectively
The saturated reactance is also referred as the air-core reactance; it is fairly low compared with
the unsaturated reactance. Typical values for two-winding transformers are 1 to 2 times the
short-circuit inductance and 3 to 4 times for autotransformes [1].
The polynomial characteristic uses expression 14 to fit the curve asymptotically into the piece-
wise linear definition. The higher the exponent, the sharper the saturation curve:
!
M ksat
M
iM = 1 + (14)
LM 0
Where,
This polynomial characteristic is always inside the corresponding linear representation. In full
saturation the polynomial characteristic is extended linearly. Compared to the two-slope curve, it
does not contain a singular point at the knee flux and therefore its derivate (magnetizing voltage)
is continuously defined.
The p.u. values used for the definition of the saturation characteristic of the positive sequence
model are referred to the following bases quantities:
Ubase [kV]: nominal voltage of the (energizing) winding, i.e. the winding used for the no
load test
Current-Flux values
The user can also define the saturation curve in terms of measured current-flux values and
select between a piecewise linear or spline interpolation.
The current-flux values in the table are peak values in p.u.. In a power transformer with im-
pressed voltage, the magnetizing flux in p.u. is equal to the magnetizing voltage in p.u., thus
flux and voltage are interchangeable and the p.u. current-flux curve represents a p.u. current-
voltage curve as well. Furthermore, it can be assumed that the applied voltage remains fairly
linear during the non-load
tests and hence the ration between RMS and peak values of the
voltage is given by 2.
On the contrary, the magnetizing current is distorted (non-sinusoidal) because of the saturation
curve. As a consequence
of that, the ratio between the RMS and peak value of the magnetizing
current is not longer 2 and the user has to enter truly peak values in the table.
The base quantities of the p.u. values in the current-flux table are also referred to the peak
values of the corresponding nominal variables:
Sbase [M V A]
Ibase [A] = 2 1000
3 Ubase [kV ]
Ubase [KV ]/ 3
base [V s] = 2 1000
2f [kHz]
The zero sequence magnetizing current strongly depends on the construction characteristic
of the transformer core (three-legged, five-legged, shell-type, etc.) and its vector group. Fig-
ure 1.17 shows the equivalent circuit for the zero sequence.
Figure 1.17: Equivalent circuit of the 2 winding 3-phase transformer for the zero-sequence
If the transformer has delta-connected windings, then any zero sequence excitation approx-
imates a zero-sequence short-circuit, as the delta-connected winding short-circuits the zero-
sequence current. In that cases there is no need to represent zero sequence saturation.
If the transformer type does not have delta-connected windings, then the zero-sequence excita-
tion current results generally higher than the positive-sequence excitation current and strongly
depends on the core type.
To account for the higher zero-sequence linear exciting current when no delta-connected wind-
ing is available, PowerFactory allows for the definition of a linear (unsaturated) zero-sequence
magnetizing impedance. This zero-sequence magnetizing impedance and its R/X ratio is de-
fined in the load flow page (TypTr2\Load flow); the parameters are made available depending
on the vector group (i.e. hidden in case of delta-connected winding).
To account for the core type dependency of the the zero-sequence saturation characteristic, the
transformer model supports the following two options in the EMT-simulation page (TypTrf ):
3 Limbs core: use this option for three-legged core designs. In this core type, the fluxes are
roughly equal in the three legs and must therefore return outside the core through the air-
gap and the tank. Because of the fact that the air-gap and the tanks are no-magnetic, the
zero-sequence magnetizing current is nearly linear and therefore the model uses the linear
zero-sequence magnetizing impedance defined in the load flow page. In other words, it
does not consider zero-sequence saturation effects.
5 Limbs core: use this option for five-legged and shell-type cores. As the zero-sequence
fluxes return inside the core, the model uses the saturation characteristic (of the positive
sequence) in the zero-sequence magnetizing reactance as well.
The residual flux is the magnetizing flux which remains in the core after the transformer has
been switched off. A residual flux, other than a remanent 1 flux, implies then the circulation of a
magnetizing current (M = LM IM ).
Once the transformer has been switched off, this magnetizing current circulates through the no-
load losses resistance Rm and de-magnetizes the core. The flux decays then exponentially with
a time constant Lm /Rm with Lm the linear magnetizing inductance. To simulate the decaying
magnetizing current and hence the decaying residual flux it is necessary to define the no-load
losses. Otherwise, if Rm =0, the magnetizing current cannot circulate and PowerFactory will
automatically set the residual flux to 0 as soon as the transformer has been switched off.
The user can also define the residual flux in the EMT simulation by a parameter event. For
simplicity, the residual flux is entered in dq0-components using the following signals:
The dq0-transformation relates the dq0-fluxes with the abc-fluxes (phase or natural compo-
nents) as follows:
2 1 1
d 3 3 3 a
1 1
0
q = b
3 3
0 1 1 1 c
3 3 3
1 0 1
a d
1 3
1
b = 2
2
q
1 3
c 0
1
2 2
The calculation parameters c:psim c, c:psim b and c:psim c give the resulting flux (simulation
result) in natural components for the phases a, b and c respectively.
It is in general quite difficult to predict the residual flux of a transformer in a reliably way. How-
ever as the residual flux has a major impact on the amplitude of inrush currents, it has to be
considered in the model. If it is not known, typical maximum values between 0.8 and 0.9 p.u.
can be assumed for worst-case conditions.
The stray capacitances of a transformer do not only depend on its construction characteristics
of the transformer (like for instance length of the windings, insulating material, core dimensions,
etc.) but also on its installation characteristics as well (indoor or outdoor transformer, proximity
to other grounded components, walls, etc.). For that reason, the stay capacitances are not part
of the transformer type data but defined in the element (ElmTr2).
On the EMT-Simulation page of the element (ElmTr2\EMT-Simulation) the user can enable the
stray capacitances in the model by ticking the Consider Capacitances option. The model ac-
count for the following capacitances:
All transformer parameters entered in p.u. or % are referred to the transformer ratings. Trans-
former rated voltages different from nominal bus bar voltages are correctly considered.
2.2.1 Yd-transformer
This model is described in detail in section 1.1.3 as a general example for the zero-sequence
system modelling. Please refer to that section for further explanation.
If no accurate data are available from the manufacturer, the following estimations can be used
for the zero-sequence impedance voltages as seen from the grounded side:
Concerning the model for the magnetic flux saturation characteristics the transformer types
with 3 or 4/5 limbs behave differently in general. In the 3-limb design, the zero-sequence flux
defined by 15 is not guided via the transformer limbs but uses parallel paths (e.g. through the
transformer vessel, oil, ) and thus can be modelled as linear without saturation effects.
1
0 = (A + B + C ) (15)
3
The zero sequence equivalent circuit diagram of the YNyn transformers is depicted in Figure 2.1.
The equivalent circuit diagram of star connected transformers with isolated star point can be
derived from this equivalent circuit by assuming infinite grounding impedances at the respective
side.
The zero-sequence magnetizing impedance ratio depends strongly on the construction of the
magnetic circuit of the transformers. Typical ranges are:
zM 0
Core-type transformer (3-limb) zsc,0 = 3 . . . 10
zM 0
Shell-type transformer (4/5-limb) zsc,0 = 10 . . . 100 (or bank of 3 single phase units)
An internal tertiary delta winding can be considered either using the PowerFactory three-winding
model or, in a simplified way, by considering that the short circuit impedance of the internal delta
winding can be modeled by an impedance parallel to the zero sequence magnetizing impedance
of Figure 19. Hence, an internal delta winding can be modeled by simply assuming a very low
zero-sequence magnetizing reactance.
zM 0
= 1..2.4
zsc,0
The short circuit resistance of the delta-tertiary winding can be entered as R/X ratio in the Mag.
R/X field.
A zig-zag winding completely uncouples primary and secondary side of the zero sequence
system, as shown in Figure 2.3.
Figure 2.3: YNzn transformer (zero-sequence system) with HV side tap changer in detailed
representation
The PowerFactory model for the auto-transformer is a special case of the 2-winding star/star
(YY)-Transformer.
As soon as an auto-transformer symbol is entered, the option Connected Star Points (Autotrans-
former) can be checked on the Basic Data page of the element (see Figure 21). This activates
the interpretation as an autotransformer. This option only is shown when the type selected for
the transformer is of vector group YY.
The effect of this connection can be seen in Figure 22. Besides the additional connection
between the star points, only one grounding impedance can be entered.
For the YY autotransformer the currents of HV side and LV side both flow through the same
grounding impedance ZE = RE + jXE . The voltage over this grounding impedance ZE thus
affects the zero-sequence system voltages on both sides. This makes it necessary to consider
the absolute value of the impedances, currents and voltages and not the p.u.-values.
Very often, an additional delta tertiary winding is used to reduce the zero-sequence impedance
of auto-transformers. The approach for modeling this is equivalent to the internal delta tertiary
winding modeling of Yy-transformers.
Figure 3.1: Input/Output Definition of 2-winding transformer model for RMS and EMT simulation
Table 3.3: Additional parameters and signals of EMT transformer model (calculation parameter)
5 References
[1] Guidelines for representation of network elements when calculating transients. Technical
report, Cigre Working Group 33.02, 1990.
[2] Allan Greenwood. Electrical Transients in Power Systems. John Wiley & Sons, 1991.
List of Figures
1.6 Yd transformer (a) in the zero-sequence system with HV side tap changer in
detailed (a) and simplified representation (b) . . . . . . . . . . . . . . . . . . . . 7
1.14 HF Model for the external capacitances in positive sequence system (a) and zero-
sequence system (b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.15 Equivalent circuit of the 2 winding 3-phase transformer for the positive sequence 17
1.17 Equivalent circuit of the 2 winding 3-phase transformer for the zero-sequence . . 20
2.3 YNzn transformer (zero-sequence system) with HV side tap changer in detailed
representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.1 Input/Output Definition of 2-winding transformer model for RMS and EMT simulation 29
List of Tables
3.3 Additional parameters and signals of EMT transformer model (calculation param-
eter) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29