Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Synchronous Machine
ElmSym
DIgSILENT GmbH
Heinrich-Hertz-Str. 9
72810 - Gomaringen
Germany
http://www.digsilent.de
info@digsilent.de
r1017
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
1.5.2 Equations with stator and rotor flux state variables in stator-side p.u.-system 17
1.5.3 Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.5.4 Equations with stator currents and rotor flux variables as used in the Pow-
erFactory model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.5.6 Saturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.5.8 Saturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2 EMT-Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3 References 31
List of Figures 32
List of Tables 33
1 General Description
This document describes the PowerFactory synchronous machine models, as used for the vari-
ous steady state and dynamic power system analysis functions supported by PowerFactory , as
there are:
This document describes the model equations that are implemented in PowerFactory . A list of
input and output variables can be found in 2.
For steady state load flow calculations a synchronous machine can be modelled by an equiv-
alent voltage source behind the synchronous reactance. However, in actual load flow calcula-
tions, the controlled operation of a synchronous generator is typically modelled.
Figure 1.1 shows the basic concept of a controlled synchronous machine modelled for load flow
analysis
• Voltage control
Large synchronous generators at large power stations typically operate in voltage control mode
(“PV” mode).
Smaller synchronous generators, e.g. embedded in distribution grids typically keep the power
factor constant (”PQ”-mode).
When enabling the Voltage control option of the generator element, the generator will control the
voltage directly at its terminals. For more complex control schemes, e.g. controlling the voltage
at a remote bus bar or controlling the voltage at one bus bar using more than one generator, a
Station Controller model needs to be defined.
In this case, the Station Controller adds an offset to the reactive power operating point specified
in the synchronous generator element:
Q = Q0 + K · ∆QSCO (1)
For more details, please refer to the Technical Reference Document of the Station Controller.
Generally, reactive power limits are only considered, if the synchronous machine is in voltage
control and the load flow option Consider reactive power limits is enabled. If this option of the
Load Flow command is disabled and the specified reactive power limits are exceeded, Power-
Factory just generates a warning message but doesn’t apply any actual limit to the generator’s
reactive power output.
In the case that it is difficult to achieve a well balanced load flow state, an additional scaling factor
can be applied to the reactive power limits. This scaling factor is more for “debugging reasons”
and doesn’t have any physical interpretation. The reactive scaling factor is only considered if
the load flow option Consider Reactive Power Limits Scaling Factor is enabled.
Fix reactive power limits. Fix reactive power limits can either be specified at Element or Type
level of the synchronous machine. Type limits are used when the option Use Limits Spec-
ified in Type is enabled, otherwise, the model takes the Element limits that can either be
defined on a p.u.-basis or using actual units (Mvar).
User-defined Capability Curve. The user-defined capability curve allows specifying a com-
plete, active power and voltage dependent capability diagram (see Figure 1.2). User-
defined capability diagrams are defined using the object IntQlim, which is stored in the
Operational Library.
For more information, please refer to the Technical Reference of the Capability Diagram
For assigning a capability diagram to a Synchronous Machine Element, the correspond-
ing reference (pQlim) must be set (see Figure 1.3). If this pointer is assigned, all other
attributes relating to reactive power limits are hidden and the local capability diagram of
the Synchronous Machine Element displays the reactive power limits defined by the Ca-
pability Curve object (IntQlim) at nominal voltage.
Besides these local settings, the corresponding options on the “Active Power Control”-tab of
the Load Flow Command either activate or deactivate the influence of Secondary Controller or
Primary Frequency Bias.
Reference Machine
The option Reference Machine has two consequences:
• Machine balances active power if the Load Flow option As Dispatched is selected and the
Balancing-Option by Reference Machine is enabled, or the Load Flow Option Secondary
Control is selected and no Secondary Controller is specified in the network.
In this case, PowerFactory considers in all isolated grids a common frequency deviation df and
establishes an active power balance through this variable and the primary frequency bias of the
individual generators:
P = P0 + Kpf · ∆F
With:
The Primary Controlled Load Flow represents that state of a power system following an active
power disturbance, in which the primary governors have settled and the system finds a “quasi
steady-state” before the secondary controlled power plants take over the active power balancing
task.
During the “primary frequency controlled” state, there is a deviation from nominal frequency.
For simulating the “Secondary Controlled” state, which is an (artificial) steady state following the
settling of the secondary control system:
P = P0 + K · ∆PSCO (2)
with:
For more information related to the Power Frequency Controller object, please refer to the cor-
responding Technical Reference Manual.
This situation can be modelled by enabling the Load Flow Option According to Inertias. In this
case, the variable dF represents an equivalent frequency rate of change and active power will be
balanced according to the inertia of all generators (defined by the Acceleration Time Constant,
to be found on the RMS-/EMT-Simulation page).
When considering active power limits in a load flow calculation, PowerFactory makes reference
to the Operational Limits (Min. and Max.).
This option decides whether a synchronous machine can be used for driving an island-network.
Typical applications are:
In case of contingencies that split the system two of more isolated areas, PowerFactory requires
at least one synchronous generator with this option being enabled for assuming that the cor-
responding island can continue operating after having been islanded. Otherwise, the load flow
calculation will assume a complete black out in the corresponding island (all loads and genera-
tors unsupplied).
• Subtransient equivalent
• Transient equivalent
• Synchronous equivalent
The distinction of the time dependence is due to the effect of increased stator currents on the
induced currents in the damper windings, rotor mass and field winding. In the case of a fault
near to a generator the stator current can increase so that the resulting magnetic field weakens
the rotor field considerably. In steady state short circuit analysis, this field-weakening effect is
represented by the corresponding equivalent source voltage and reactance. The associated
positive sequence model of a synchronous machine is shown in Figure 1.4. The delayed effect
of the stator field on the excitation and damping field is modelled by switching between the
source voltage E”, E’ and E depending on the time frame of the calculation.
Figure 1.4: Single-phase equivalent circuit diagram of a generator for short-circuit current cal-
culations which include the modelling of the field attenuation
In the “complete short circuit method”, the internal voltage source is initialized by a preceding
load flow calculation.
The “complete short circuit method” calculates subtransient and transient fault currents using
subtransient and transient voltage sources and impedances.
Based on the calculated subtransient and transient (AC-) currents, PowerFactory derives other
relevant short-circuit indices, such as peak short circuit current, peak-break current, AC-break
current, equivalent thermal short circuit current by applying the relevant methods according to
IEC60909 (see next section).
The IEC 60909 (equivalent to VDE 102/103) series of standards only calculates the subtransient
time phase. Short circuit currents of longer time phases are assessed based on empirical
methods by multiplying the subtransient fault current with corresponding factors.
Figure 1.5 shows the basic IEC 60909 short circuit model of a synchronous machine.
When calculating initial symmetrical short-circuit currents in systems fed directly from genera-
tors without unit transformers, for example in industrial networks or in low-voltage networks, the
following impedances have to be used
Z S1 = RS + jd00 (3)
Normally it is assumed that X2 = Xd00 . If Xd00 and Xq00 differ significantly the following can be
used:
1 00
X200 = X2 = · Xd + Xq00
(5)
2
00
ZS0 = RS0 + jXS0 (6)
For the subtransient reactance, the saturated value has to be used leading to highest possible
fault currents.
IEC 60909 makes no provision of the pre-fault state. It always considers a voltage factor cmax of
1,1 (or 1,05 in LV-networks). Because this approach would lead to overestimated fault currents,
the impedance is corrected by a correction factor KG :
Un Cmax
KG = (7)
UrG 1 + χ00d sin ϕrG
with Cmax : Voltage factor, see IEC 60909-0, item 2.3.2, page 41, Table 1.1.
All other short circuit indices are calculated precisely according to the IEC60909 (VDE 102/103)-
standard.
Besides IEC60909, PowerFactory supports short circuit calculation according to ANSI C-37.
Similar to short circuit calculations according to IEC60909, only subtransient fault currents are
actually calculated.
For further details related to ANSI C-37, please refer to the original ANSI C-37 standard and
corresponding literature.
The IEC 61363 standard describes procedures for calculating short-circuits currents in three-
phase ac radial electrical installations on ships and on mobile and fixed offshore units.
The calculation of the short-circuit current for a synchronous machine is based on evaluating the
envelope of the maximum values of the machine’s actual time-dependent short-circuit current.
The resulting envelope is a function of the basic machine parameters (power, impedance, etc.)
and the active voltages (E”, E’, E) behind the machine’s subtransient, transient and steady-state
impedance. The impedance are dependent upon the machine operating conditions immediately
prior to the occurrence of the short-circuit condition.
When calculating the short-circuit current, only the highest values of the current are considered.
The highest values vary as a function of time along the top envelope of the complex time-
dependent function. The current defined by this top envelope is calculated from the equation:
√
iK (t) = 2Iac (t) + idc (t) (8)
00 0
00 0
Iac (t) = (Ikd − Ikd )e−t/Td + (Ikd
0
− Ikd )e−t/Td + Ikd (9)
The subtransient, transient and steady-state currents are evaluated using equations:
00
Ikd 00
= Eq0 /Zd00 with Z 00d = (Ra + jXd00 ) (10)
and
0
Ikd 0
= Eq0 /Zd0 with Z 0d = (Ra + jXd0 ) (11)
Internal voltages considering terminal voltage and pre-load conditions are calculated using
equations:
√
00
Eq0 = U 0 / 3 + I 0 ∗ Z 00d (12)
√
0
Eq0 = U 0 / 3 + I 0 ∗ Z 0d (13)
√ 00
Idc (t) = 2(Ikd − I0 sin φ0 )e−t/Tdc (14)
The OPF (Optimal Power Flow) function in PowerFactory allows the user to calculate optimal
operational conditions, e.g. the minimization of losses or production costs by adjusting the
active and reactive power dispatch of the generators.
To consider the synchronous machine in the OPF calculation the following options have to be
assigned on the “Optimization” tab of the synchronous machine element.
It is possible to enable and disable the active and reactive power optimization of the machine.
The active power flag allows the active power dispatch of the machine to be optimized in the
OPF calculation. On the other hand, the reactive power flag allows the voltage reference of the
machine to be adapted according to the OPF optimization function.
When these options are disabled, the synchronous machine is treated as in a conventional load
flow calculation during the execution of the OPF.
For every machine a minimum and maximum active and reactive power limit can be defined.
For the reactive power limits there is also the possibility to use the limits which are specified in
the synchronous machine type (enable the flag Use limits specified in type).
The active and reactive power limits will be considered in the OPF if and only if the individual
constraint flag is checked in the synchronous machine element and the corresponding global
flag is enabled in the OPF dialogue.
The table Operating Costs specifies the costs ($/h) for the produced active power (MW) of the
units. The representation of the data is shown automatically on the diagram below the table for
checking purposes. The cost curve of a synchronous machine is calculated as the interpolation
of the predefined cost points.
The equivalent circuits of the synchronous machine model for harmonics are shown in Figure
1.6.
Figure 1.6: Synchronous machine models for positive, negative and zero sequences
The average inductance experienced by harmonic currents, which involve both the direct axis
and quadrature axis reactances, is approximated by
L00d + L00q
L00 = (15)
2
b
R = k(f ) · r with k(f ) = (1 − a) + a · (f /fnom ) (16)
When enabling the option Consider Transient Parameters the harmonic inductance is calculated
from xd”, xd’ and xd, as entered for the RMS-simulation or EMT-simulation functions. Only in
a very narrow band around nominal frequency, the effect of the transient and synchronous
reactance is visible (see also Figure 1.7).
Because of the highly accurate representation around nominal frequency this model can in-
crease the accuracy of subsynchronous resonance studies based on frequency domain analy-
sis.
Figure 1.7: Frequency domain representation of synchronous machine (Consider Transient Pa-
rameters)
Figure 1.8 to Figure 1.10 show the equivalent circuit diagrams of the PowerFactory synchronous
machine models, which are represented in a rotor-oriented reference system (Park coordinates,
dq-reference frame).
The d-axis is always modelled by 2 rotor loops representing the damping and the excitation
winding.
For the q-axis, PowerFactory supports two models, a model with one rotor loop (for Salient Pole
machines) and a model with two rotor loops (for Round Rotor Machines).
Figure 1.8: d-axis equivalent circuit for the synchronous machine representation
Based on the equivalent circuit diagrams according to Figure 1.8 to Figure 1.10, the following
differential equations can be derived describing the PowerFactory synchronous machine model
for time domain simulations.
Figure 1.9: q-axis equivalent circuit for the synchronous machine representation (round rotor)
Figure 1.10: q-axis equivalent circuit for the synchronous machine representation (salient rotor)
1.5.2 Equations with stator and rotor flux state variables in stator-side p.u.-system
Using stator and rotor flux as state variables for the description of the synchronous machine
model the following set of equations is resulting1 .
1 dψd
ud = rs id + − nψq
ωn dt
1 dψq
uq = rs iq + + nψd (17)
ωn dt
1 dψ0
u0 = rs i0 +
ωn dt
dψe
ue = re ie +
ωn dt
(18)
dψD
0 = rD iD +
ωn dt
of currents in Figure 1.8 to Figure 1.10 showing the actual orientation of currents of the PowerFactory model
dψk
0 = rk ik +
COn dt
(19)
dψQ
0 = rQ iQ +
COn dt
dψQ
0 = rQ iQ + (20)
ωn dt
For completing the model, the flux linkage equations are required:
d-axis:
q-axis, round-rotor:
te = ψd iq − ψq id (24)
1.5.3 Mechanics
The accelerating torque is the difference between the input torque (mechanical torque) tm and
the output torque (electromechanic torque) te of the generator.
Jtot ωn2 dn dn
= Ta,tot = tm + te
p2z Pr dt dt
(25)
dϑ
= ωn n
dt
The inertia of the generator and the turbine, plus the inertia of the mechanical load, can then be
expressed in a normalized per unit form as the inertia time constant Htot in [s], with
The inertia time constant H can be given based on the rated apparent generator power, as
shown in the equation above, or based on the rated active generator power. The mechanical
starting time or acceleration time constant Ta,tot in [s] is then
1.5.4 Equations with stator currents and rotor flux variables as used in the PowerFac-
tory model
For obtaining maximum effectiveness with regard to the numerical accuracy and robustness of
the model, the multiple time-scale properties of the equation system shall be used by partionning
the equations into “fast” and “slow” equations.
PowerFactory uses rotor flux and stator currents as state variables because this choice of state
variables leads to the best possible multiple time-scale separation and hence to the best nu-
merical properties.
ψd00 = ke ψe + kD ψD
(28)
ψq00 = kx ψx + kQ ψQ
xmd xlD
ke =
xd2
xmd xle
kD =
xd2
xmq xlQ (29)
kk =
xq2
xmq xlx
kQ =
xq2
and:
ψd = x00d id + ψd00
(31)
ψq = x00q iq + ψq00
1 dψd00
u00d = − nψq00
ωn dt
(32)
1 dψq00
u00q = + nψd00
ωn dt
x00d did
ud = rs id + − nx00q iq + u00d
ωn dt
x00q diq
uq = rs iq + − nx00d id + u00d (33)
ωn dt
x0 di0
u0 = rs i0 +
ωn dt
The parameters according to Table 1.1 that have been used in the equation systems and the
equivalent circuit diagrams are typically not available for synchronous machines.
The classical input parameters of a synchronous machine, as they can be entered directly into
the PowerFactory synchronous machine model are depicted in Table 1.2.
For converting the set of input parameters according to Table 1.2 into the set of internal parame-
ters according to Table 1.1, there are several methods described in the literature. Some of them
are more accurate, some of them are highly simplified but easier to realize.
PowerFactory applies a highly accurate parameter conversion method, as described in [1]. This
method consists of th following formulas for the d-axis: Auxiliary variables:
x1 = xd − x1 + xrl
(xd − xl )2
x2 = x1 −
xd (34)
x1 x00
x2 − xd
d
x3 = x00
1 − xdd
xd 0 xd xd
T1 = 0 Td + 1 − 0 + 00 Td00
xd xd xd
T2 = Td0 + Td00 (35)
q
T3 = Td0 Td00
x2 T1 − x1 T2
a=
x1 − x2 (36)
x3
b= T2
x3 − x2 3
r
−a a2
Tle = + −b
2 4
r (37)
−a a2
TlD = + −b
2 4
Tle − TlD
xle = T1 −T2 TlD
x1 −x2 + x3
TlD − Tle
xlD = T1 −T2 Tle
x1 −x2 + x3 (38)
xle
re =
ωn Tle
xlD
rD =
ωn TlD
The q-axis parameters can be calculated analogously to the d-axsis parameters in case of a
round rotor machine (2 rotor-loops).
For a salient pole machine (1 rotor loop), the internal parameters can be calculated as follows:
(xq − xl )(x00q − xl )
xlQ =
xq − x00q
(39)
x00q xq − xl + xlQ
rQ =
xq ωn Tq00
1.5.6 Saturation
The model described in the previous section was a purely linear model not considering any
saturation effects.
Generally, there exists saturation for all reactances of the synchronous machine model. How-
ever, for the purpose of system analysis, only main flux saturation has to be considered in the
model by considering saturation of the magnetizing reactiances xm d and xm q.
whereas the level of saturation depends on the magnitude of the magnetizing flux:
q
ψm = (ψd + xl id )2 + (ψq + xl iq )2 (41)
PowerFactory supports different approximations for saturation. In case of the saturation model
1, based on the two parameters SG10/SG12, a quadratic approximation is applied:
If ψm > Ag :
Bg (ψm − Ag )2
csat = (42)
ψm
else:
csat = 0 (43)
In case of a tabular input, csat is calculated based on a spline approximation of the sampled
values.
Saturation in d-axis can be measured by no-load field tests. However, the saturation of the
mutual reactance xmq in the q- axis cannot be measured easily and therefore assumptions
have to be taken for q-axis saturation:
• In case of a round rotor machine, it is assumed that saturation in q-axis is equal to d-axis
saturation.
• In case of a salient rotor machine the saturation characteristic in q-axis is weighted by the
ratio xq /xd .
1
ksatd =
1 + csat
1 (44)
ksatq = xmq0
1 + xmd0 csat
Saturated magnetizing reactances apply to all formulas (21), (25), (23) and (28), (29), (30).
Saturation in subtransient reactances is not considered, which represents a valid approximation
because the subtransient reactance of a generator is only very weakly influenced by main flux
saturation.
The saturation of stator leakage reactances is a current-dependent saturation, i.e. high currents
after short-circuits will lead to a saturation effect of the leakage reactances. Because the use
of unsaturated subtransient reactances would therefore lead to underestimated maximum short
circuit currents, it is recommended to use saturated values for xd” and xq” (“saturated” refers
here to current saturation).
For all other parameters (transient and synchronous reactance), unsaturated values shall be
entered. The influence of main flux saturation is considered by the model as described above.
For RMS-simulations, stator flux transients are generally not considered. Neglecting stator flux
transients in (33) leads to the following simplified stator voltage equations for RMS-simulations:
uq = rs id − x00q iq + u00d
(45)
ud = rs iq − x00d id + u00q
u00d = −nψq00
(46)
u00q = nψd00
Assumption that magnetizing voltage is approx. equal to magnetizing flux (for saturation) leads
to the following approximation:
q
ψm ≈ um = (ud + rs id − xl iq )2 + (uq + rs iq + xl id )2 (47)
1.5.8 Saturation
Figure 1.11 shows the definition of the main flux saturation curve. The linear line represents the
air-gap line indicating the excitation current required overcoming the reluctance of the air-gap.
The degree of saturation is the deviation of the open loop characteristic from the air-gap line.
The characteristic is given by specifying the excitation current I1.0pu and I1.2pu needed to obtain
1 p.u respectively 1.2 p.u. of the rated generator voltage under no-load conditions. With these
values the parameters sg1.0 (=csat (1.0pu) ) and sg1.2 (=csat (1.2pu) ) can be calculated.
ie (1.0p.u)
sg1.0 = −1
i0
(48)
ie (1.2p.u)
sg1.2 = −1
1.2i0
q
s
1.2 − 1.2 sg1.2
g1.0
Ag = q
sg1.2
1 − 1.2 sg1.0 (49)
sg1.0
Bg =
(1 − Ag )2
Alternatively, a sampled excitation current vs. voltage curve can be entered into the PowerFac-
tory model.
Rotor current and rotor flux of the PowerFactory model is not expressed in a stator per-unit
system as it has been used in section 1.5. PowerFactory uses the following p.u. definitions,
which are also known as “no load p.u.-system”:
Rotor currents:
eie = xmd0 ie
eiD = xmd0 iD
(50)
eix = xmq0 ix
eiQ = xmq0 iQ
Rotor-flux:
xmd0
ψee = ψe
xe0
xmd0
ψeD = ψD
xD0
xmq0 (51)
ψex = ψx
xx0
xmq0
ψeQ = ψQ
xQ0
With
dψee
u
ee = eie + Te0
dt (53)
dψeD
0 = eiD + TD0
dt
dψex
0 = eix + Tx0
dt (54)
dψ Q
0 = eiQ + TQ0
dt
dpsiQ
0 = eiQ + TQ0 (55)
dt
with
xe0
Te0 =
re ωn
xD0
TD0 =
rD ωn
xx0 (56)
Tx0 =
rx ωn
xQ0
TQ0 =
rQ ωn
In the p.u.-system used for rotor variables of the PowerFactory model, there will be 1 p.u. stator
voltage in case of no load conditions and 1 p.u. excitation voltage (and no saturation).
PowerFactory defines several rotor angles based on different references. The rotor angle is
defined as the position of the d-axis. The following variables are available:
• fipol / [deg]: Rotor angle with reference to the local bus voltage of the generator (terminal
voltage)
• firot / [deg]: Rotor angle with reference to the reference voltage of the network (slack bus
voltage)
• firel / [deg]: Rotor angle with reference to the reference machine rotor angle (slack gener-
ator)
• dfrot / [deg]: identical to firel
• phi / [rad]: Rotor angle of the q-axis with reference to the reference voltage of the network
(=firot-90◦ )
Figure 2.1: Input/Output Definition of the synchronous machine model for stability analysis
(RMS-simulation)
2.2 EMT-Model
Figure 2.2: Input/Output Definition of the synchronous machine model for stability analysis
(EMT-simulation)
3 References
[1] B. Oswald. Netzberechnung 2: Berechnung transienter Vorgnge in
Elektroenergieversorgungs-netzen. VDE-Verlag, 1 edition, 1996.
List of Figures
1.6 Synchronous machine models for positive, negative and zero sequences . . . . 15
1.9 q-axis equivalent circuit for the synchronous machine representation (round rotor) 17
1.10 q-axis equivalent circuit for the synchronous machine representation (salient rotor) 17
2.1 Input/Output Definition of the synchronous machine model for stability analysis
(RMS-simulation) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.2 Input/Output Definition of the synchronous machine model for stability analysis
(EMT-simulation) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
List of Tables