IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 21, NO.

1, FEBRUARY 2006 301

Comparative Analysis Between Synchronous

and Induction Machines for Distributed
Generation Applications
Walmir Freitas, Member, IEEE, Jose C. M. Vieira, Student Member, IEEE, Andre Morelato, Member, IEEE,
Luiz C. P. da Silva, Vivaldo F. da Costa, and Flavio A. B. Lemos

Abstract—This paper presents a detailed comparative analysis limit the amount of distributed generation in a given system.
between synchronous and induction machines for distributed The results can be a useful technical guide for utility engineers,
generation applications. The impacts of these generators on the and energy producers decide which machine is more suitable,
distribution network performance are determined and compared
by using computational simulations. The technical factors an- taking into account the main characteristics of their network.
alyzed are steady-state voltage profile, electrical power losses, This paper is organized as follows. Section II describes the
voltage stability, transient stability, voltage sags during unbal- network component models employed in this paper. The im-
anced faults, and short-circuit currents. The results showed that pacts on the steady-state voltage profile are addressed in Sec-
the best technical choice depends on the network characteristics, tion III. Section IV discusses the electrical power losses. The
i.e., the main factors that may limit the penetration level of dis-
tributed generation. impacts on the system voltage stability margin are determined
in Section V. Section VI presents the results considering tran-
Index Terms—Distributed generation, induction generator,
sient stability. Voltage sags during unbalanced faults are ana-
short-circuit currents, steady-state voltage profile, synchronous
generator, transient stability, voltage sag, voltage stability. lyzed in Section VII. The short-circuit currents supplied by the
generators during faults are investigated in Section VIII. Finally,
Section IX summarizes the main conclusions.
I. INTRODUCTION
II. NETWORK COMPONENT MODELS
R ECENTLY, the interest in distributed generation has
considerably increased due to market deregulation, tech-
nological advances, governmental incentives, and environment
In this paper, all network components were represented by
three-phase models. In the studies about steady-state voltage
impact concerns [1]–[3]. At present, most distributed genera- profiles, power losses, and stability, the network variables were
tion installations employ induction and synchronous machines, represented by phasors. Such analyses were conducted by using
which can be used in thermal, hydro, and wind generation a load flow and a transient stability program. On the other hand,
plants [3]. Although such technologies are well known, there is in the studies about voltage sags and short-circuit currents, the
no consensus on what is the best choice under a wide technical network variables were represented by instantaneous values.
perspective. These cases were analyzed by using an electromagnetic tran-
Based on these facts, it is important to understand the dif- sient program. The simulation package adopted was the Sim-
ferent impacts provoked by this choice on several technical PowerSystems for use with Matlab/Simulink, version 2.3 [4].
factors. This paper presents research results considering dis- In all cases, the distribution network feeders were rep-
tinct scenarios and technical factors. The factors analyzed are resented by a series RL impedance, because they can be
steady-state voltage profile, electrical power losses, voltage sta- considered short-lines, and the transformers were modeled by
bility, transient stability, voltage sags during unbalanced faults, employing the circuit. In the steady-state studies, the loads
and short-circuit currents. These factors were investigated by were represented by constant power models, as is usual in load
using different kinds of power system analysis programs, e.g., flow programs, whereas in the dynamic studies, active power
load flow programs, transient stability programs, and electro- loads were represented by constant current models and reactive
magnetic transient programs. Simulation results showed that power loads were represented by constant impedance models,
the choice should be done considering the main factors that may as recommended in [5].

Manuscript received March 7, 2005; revised July 22, 2005. This work was A. Induction Generators
supported by FAPESP and CNPq, Brazil. Paper no. TPWRS-00119-2005.
W. Freitas, J. C. M Vieira, A. Morelato, L. C. P. da Silva, and V. F. da Costa Although most induction generators in operation are em-
are with the Department of Electrical Energy Systems, State University of ployed in wind power plants [3], [6], such machines have
Campinas, C.P. 6101 13083-852, Brazil (e-mail: [email protected]; jcarlos@
dsee.fee.unicamp.br; [email protected]; [email protected]. also been used in medium-size hydro and thermal plants [3],
br; [email protected]). [7]–[9]. Therefore, in order to keep the results as generic as
F. A. B. Lemos is with the Department of Electrical Engineering, Catholic possible, the mechanical torque was considered constant, i.e.,
University of Rio Grande do Sul, C.P. 1429 90619-900, Brazil (e-mail:
[email protected]). the regulator and prime mover dynamics were neglected. The
Digital Object Identifier 10.1109/TPWRS.2005.860931 squirrel-cage rotor induction generator was represented by
0885-8950/$20.00 © 2006 IEEE
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
302 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 21, NO. 1, FEBRUARY 2006

a sixth-order model in the electromagnetic transient studies,

which was reduced for a fourth-order model in the transient
stability simulations [10]. In all cases simulated, part of the
reactive power consumed by the generator was locally supplied
by capacitors installed at the terminal of the machine, whose
compensation capacity was adopted equal to of the machine
capability, as is usual in this case [3].

B. Synchronous Generators Fig. 1. Single-line diagram of system 1.

At present, most distributed generation systems employ syn-

chronous generators, which can be used in thermal, hydro, or
wind power plants. In the electromagnetic transient simulations,
the synchronous generators were represented by an eight-order
model, which was reduced to a sixth-order model in the transient
stability simulations [10]. Usually, synchronous generators con-
nected to distribution networks are operated as constant active
power sources, so that they do no take part in the system fre-
quency control. Therefore, in this paper, the mechanical power
was considered constant, i.e., the regulator and prime mover dy-
namics were neglected. Similar to the induction generator case,
this option also leads to results that are more generic. In addi-
tion, typically, there are two different modes of controlling the
excitation system of distributed synchronous generators. One
aims to maintain constant the terminal voltage (voltage control
mode), and the other one aims to maintain constant the power
factor (power factor control mode) [3], [11]. Power factor con-
trol mode is usually adopted by independent producers to max-
imize the active power production [3]. In consequence, unitary
power factor operation is adopted. Thus, both forms of control
are employed in this paper. In the voltage regulator cases, the
controller set point was fixed at 1 p.u., whereas in the power
factor regulator cases, the controller set point was fixed at 1 (uni-
tary power factor). A functional description of excitation sys-
tems acting as a voltage or power factor regulator is provided
in [11].

III. STEADY-STATE VOLTAGE PROFILE

Voltage violations due to presence of distributed generators
can considerably limit the amount of power supplied by these
generators in distribution networks [3], [12]. Before installing
(or allowing the installation of) a distributed generator, utility
engineers must analyze the worst operating scenarios to guar-
antee that the network voltages will not be adversely affected
by the generators. These scenarios are characterized by [12] the
following: Fig. 2. Steady-state voltage profile for different generators. (a) Minimum
• no generation and maximum demand; demand (10%). (b) Maximum demand (100%).
• maximum generation and maximum demand;
• maximum generation and minimum demand. substation transformer tap was adjusted to maintain the nodal
In this paper, it was considered that the minimum demand voltage in all buses within the allowable range for minimum
corresponds to 10% of the maximum demand. Moreover, the and maximum demand considering the case without generator
allowable steady-state voltage variation was adopted equal to tap % . Assume that an independent producer wishes
%( p.u.). The single-line diagram of the system to install six 5-MW distributed generators at bus 8 through
used in this section is shown in Fig. 1. Such network consists dedicated transformers. For this situation, simulation studies
of a 132-kV, 60-Hz subtransmission system with short-circuit were carried out to verify what kind of generator allows such
level of 1000 MVA, represented by a Thevenin equivalent installation under a steady-state voltage profile viewpoint.
(Sub), which feeds a 33-kV distribution system through one The network voltage profile considering maximum and min-
132/33-kV, /Yg transformer. The feeder ratio is 4.3. The imum demand and different generators is presented in Fig. 2.
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
FREITAS et al.: COMPARATIVE ANALYSIS BETWEEN SYNCHRONOUS AND INDUCTION MACHINES 303

TABLE I TABLE III

GENERATOR TERMINAL VOLTAGES STEADY-STATE VOLTAGE VARIATION DUE TO GENERATOR
DISCONNECTION (V )

TABLE II TABLE IV
MAXIMUM ALLOWABLE NUMBER OF AC GENERATORS CONSIDERING VOLTAGE REGULATION (V )
STEADY-STATE VOLTAGE PROFILE VIOLATIONS

can be utilized to quantify the impact provoked by generator

disconnections:

(1)
In this figure, the allowable values of nodal voltage (
p.u.) are represented by horizontal dotted lines. In this case,
it was considered that the six generators were injecting nom- where is the total number of buses, is the magnitude
inal active power (5 MW) into the network. It can be seen that of the nodal voltage of bus in the presence of distributed gen-
some nodal voltages will violate the superior limit during min- erators, and is the magnitude of the nodal voltage of bus
imum demand if a constant power factor synchronous gener- without distributed generators.
ator is adopted, whereas, if induction generators are chosen, The results are summarized in Table III considering that the
then some nodal voltages will be below the inferior limit during six generators are tripped off during maximum and minimum
maximum demand. On the other hand, if constant voltage syn- demand. It can be observed that the cases with induction gen-
chronous generators are employed, then the nodal voltages will erators or constant power factor synchronous generators lead
remain within the allowable range for both demand cases. In to the smallest variations of the voltage. In the case of con-
this case, the power factor of the synchronous generator varied stant power factor synchronous generators, the generators do not
from 0.986 inductive to 0.990 capacitive for the minimum and supply or consume reactive power. Therefore, the difference of
maximum demand values, respectively. The generator terminal the distribution of reactive current between this case and the case
voltages are shown in Table I. without generators is small. Thus, when the generators are dis-
In order to determine the maximum number of ac genera- connected, the steady-state operating point do not change con-
tors that can be installed without steady-state voltage violations, siderably. Similarly, in the case of induction generators, the gen-
the nodal voltages were calculated for each generator added in erators practically do not inject or consume reactive power as
a one-by-one basis (from one to six generators). The results well, remembering that part of the reactive power consumed by
are presented in Table II. If the constant voltage synchronous the induction generator is locally supplied by capacitors. On the
generator is selected, it is possible to install the six generators other hand, in the case of constant voltage synchronous genera-
without steady-state voltage violations. On the other hand, in the tors, the voltage variation is larger. In this situation, the amount
other cases, there will be voltage violations. The third column of reactive power injected/supplied by the generators is signif-
in Table II shows what will be the problem if a new generator icant. Thus, the steady-state operating point changes consider-
is installed. The most restrictive case is related to the constant ably after the disconnection of the generators.
power factor synchronous generator.
B. Steady-State Voltage Regulation
A. Steady-State Voltage Variation Due to Another important issue related to steady-state voltage is
Generator Disconnection the regulation characteristic of the network, i.e., how much
One important issue related to steady-state voltage profile is the nodal voltages change between maximum and minimum
to determine how much the nodal voltages will change when demand cases. It is desirable that the nodal voltages change as
the distributed generators are suddenly disconnected, because little as possible during load variations. The following global
the actuation time of voltage controllers in distribution systems, index can be employed to analyze this question:
e.g., under load tap change transformers, is slow [3]. Thus,
network operators want such variations to be as small as pos- (2)
sible. In order to analyze this issue, the following global index
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
304 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 21, NO. 1, FEBRUARY 2006

TABLE V
ACTIVE POWER LOSSES (IN MEGAWATTS)

Fig. 3. Single-line diagram of system 2.

improvement saturation point has been reached. This character-

istic has already been reported in other works [15]. The adoption
of constant voltage synchronous generators leads to the largest
reduction of the losses because this generator supplies the ac-
tive and reactive loads locally, decreasing the magnitude of the
current in the feeders. On the other hand, the usage of induction
generators does not cause a great reduction in the active power
losses; indeed, when the six generators are operating, the system
losses increase. In this case, the generators consume reactive
power from the network, rising the magnitude of the currents
circulating in the feeders. The losses behavior in the presence
of constant power factor synchronous generators is situated be-
tween the other two cases, because these generators supply lo-
cally active power but do not provide or consume reactive power.
where is the magnitude of the nodal voltage of bus
2) Minimum Demand: In this case, typically, the presence of
during maximum demand, and is the magnitude of the
the generators increases the active power losses, independent of
nodal voltage of bus during minimum demand.
the generator employed. In this situation, a large amount of ac-
The results are shown in Table IV. It can be seen that the
tive power generated is exported to the subtransmission system,
usage of constant voltage synchronous generators leads to the
influencing adversely the distribution system losses. The usage
best characteristic of voltage regulation (minimal variation).
of the constant voltage synchronous generators can be related to
Such a fact occurs because the generators’ reactive power
the worst case under losses viewpoint because, in order to keep
output changes according to the load variation, resulting in the terminal voltage at 1 p.u., the generator consumes a large
a good voltage regulation. On the other hand, the usage of amount of reactive power.
constant power factor synchronous generators or induction
generators implies the worst voltage regulations (maximal vari-
V. VOLTAGE STABILITY
ation), because these machines are not voltage self-regulated.
Normally, it is expected that the installation of generators
IV. ACTIVE POWER LOSSES close to the loads leads to a gain in the system voltage sta-
bility margin. However, the impact on the margin depends on
Although active power losses are not a technical factor that the reactive power exchanged between the generator and the net-
can limit the amount of distributed generation, it is an impor- work, which is different from distinct technologies. Therefore,
tant economical factor [13], [14]. Therefore, in this section, the in this section, the PV curves of the system shown in Fig. 3,
electrical losses of the system shown in Fig. 1 are determined which is derived from [3], are analyzed. Such network com-
for different scenarios of generation and load. The losses were prises a 132-kV, 60-Hz subtransmission system with short-cir-
calculated by using the following equation: cuit level of 1500 MVA, represented by a Thevenin equivalent
(Sub), which feeds a 33-kV distribution system through two
(3) 132/33-kV, /Yg transformers. An ac generator with capacity
of 30 MVA is connected at bus 6, which is connected to the net-
where is the total active power losses of the system, work through a 33/0.69-kV, /Yg transformer. This machine
is the active power supplied by the substation, can represent one generator in a thermal generation plant as well
is the active power supplied by the generators, and as an equivalent of various generators in a wind or small-hydro
is the total active power consumed by the loads. generation plant. In some cases, such a machine was simulated
The results are presented in Table V. The values of the active as an induction generator and in other ones as a synchronous
power losses for the case without distributed generators are generator. The PV curves were obtained by varying the active
also shown. The following facts can be observed. and reactive loads and keeping the active power injected by the
1) Maximum Demand: It can be verified that during heavy ac generator at the nominal level (30 MW). The active power
load, typically, the installation of distributed generators leads to supplied by the generator was kept constant because, usually,
a decrease of the electrical losses. Initially, each generator added such generators are not rescheduled by the system operator. In
implies a reduction of the losses. However, after the third gener- addition, simulation results showed that the extreme cases (i.e.,
ator is installed, then, if a new generator is connected to the net- the smallest/largest stability margin) are obtained when the gen-
work, the losses start to increase. This fact indicates that the loss erator is supplying nominal power.
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
FREITAS et al.: COMPARATIVE ANALYSIS BETWEEN SYNCHRONOUS AND INDUCTION MACHINES 305

Fig. 4. PV curves of bus 4.

The PV curve of bus 4 is shown in Fig. 4. It can be verified

that the presence of the synchronous generators augments the
system stability margin, independent of the excitation system
control mode. In addition, the usage of the constant voltage syn-
chronous generator leads to the largest gains, because this ma-
chine provides active and reactive power to local loads. On the
other hand, in the case with an induction generator, the system
stability margin is reduced. In this case, the operating point of
the system goes from point to on the curve after only
one step in the load increment. It can be verified, by using dy-
namic simulation, that the point on the PV curve represents
the steady-state stability limit of the induction generator. If the
load augments further, the machine rotor speed increases mono-
tonically. This fact occurs because when the loads increase, the
generator terminal voltage decreases. As the electrical torque is
proportional to the terminal voltage, it also decreases; in con-
sequence, the rotor speed increases to compensate the electrical
torque reduction. From a determined point, the rotor speed in- Fig. 5. Dynamic responses of the generators for a three-phase short circuit.
(a) Terminal voltage of the induction generator. (b) Rotor angle of the
creases unlimitedly, leading the system to a voltage collapse. Of synchronous generator.
course, at this point, the generator would be disconnected by the
protection system and the system would return to the PV curve
also accelerate during short circuits, and as a result, the reactive
for the case without generators, if there are no dynamic loads.
power consumed by the generators increase considerably, which
may lead the system to a voltage collapse. Thus, in this case, the
VI. TRANSIENT STABILITY stability phenomena can be verified by analyzing the dynamic
Typically, the actuation time of the protection system of dis- response of the rotor speed or the terminal voltage [16].
tribution network is rather slow [3]. In addition, the value of the Fig. 5 presents the dynamic responses of the different gener-
inertia constant of ac-distributed generators is low; usually it is ators for a three-phase-to-ground short circuit applied at bus 4
smaller than 2 s. Therefore, transient stability issues can limit at s, which is eliminated at 15 cycles by tripping branch
the amount of active power exported by distributed generators 2–4, when the generator is injecting 25 MW into the network.
to the system. Thus, in this section, the dynamic behavior of the It can be seen that only the case with the constant voltage syn-
generators during three-phase-to-ground faults is analyzed. The chronous generator is stable.
system employed for this investigation is the same as presented The different behavior of each generator can be explained
in Fig. 3. by analyzing the response of the reactive power exchanged be-
During short circuits, usually, synchronous generators accel- tween the generator and the network for each situation, which
erate, so that they may become unstable due to loss of synchro- is shown in Fig. 6. In the case of the induction generator, the
nism. The stability of synchronous generators can be determined reactive power exchanged takes into account the reactive power
by analyzing the dynamic response of the rotor angle [10]. Al- supplied by the capacitors. It can be verified that the reactive
ternatively, in the case of induction generators, these generators power injected by the constant voltage synchronous generator
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
306 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 21, NO. 1, FEBRUARY 2006

TABLE VI
CRITICAL POWER INJECTED BY THE GENERATORS FOR DIFFERENT
FAULT CLEARANCE TIME (IN MEGAWATTS)

Table VI. It can be verified that the usage of the constant voltage
synchronous generator permits that the critical power assumes
the highest values. On the other hand, in the case of induction
generator or constant power factor synchronous generator, the
values of critical power are smaller.

VII. VOLTAGE SAGS

The incidence of unbalanced short-circuits in distribution net-
works is relatively frequent. During such short circuits, voltage
sags may occur in the system buses. The presence of ac gen-
erators may influence the magnitude and the duration of these
voltage sags. It will depend on the impact of these generators on
the system short-circuit level and the dynamic behavior of the
reactive power exchanged between the generator and the net-
work. Thus, this section presents an analysis about voltage sags
due to unbalanced faults by using electromagnetic transient sim-
ulations. The network employed is the same shown in Fig. 3.
Fig. 7(a) and (b) presents the dynamic responses of the nodal
Fig. 6. Dynamic behavior of the reactive power exchanged between the
generator and the network during a three-phase short circuit. (a) Constant voltages of buses 4 and 5, respectively, for a 400-ms phase-A-to-
voltage synchronous generator. (b) Constant power factor synchronous ground short circuit applied at bus 4 at ms. In this paper,
generator. (c) Induction generator. voltage sag magnitude refers to the remaining voltage, as recom-
mended in [17]. Analyzing the voltage of bus 4, one can verify
increases during and after the fault. This fact has a positive im- that the voltage sag magnitude (minimum value of voltage) is
pact on the transient stability response of the generator. In the smaller in the presence of the generators, i.e., the voltage sag
case of the constant power factor synchronous generators, it can problem is aggravated by the installation of the generators. It
be noted that the reactive power injected by the generator in- occurs because the generators increase the system short-circuit
creases during the fault due to the delayed response of the exci- level. On the other hand, analyzing the voltage of bus 5, one can
tation system. However, soon after the fault clearance, the exci- see that in the presence of the constant voltage synchronous gen-
tation system acts to keep unitary power factor operation. This erator, the voltage sag magnitude is larger, i.e., the voltage sag
fact reduces the reactive power injected, affecting adversely the problem is improved by the generator. In the case of the constant
transient stability performance of the system. On the other hand, power factor synchronous generators, there is practically no dif-
in the case of the induction generator, although during a fault the ference between the situation with and without generator con-
generator injects reactive power into the network due to self-ex- sidering the voltage sag magnitude of bus 5, whereas the voltage
citation phenomenon [3], soon after the fault clearance, the gen- sag of bus 5 is adversely affected by the induction generator. In
erator consumes a large amount of reactive power, which can all cases, the ac generator and its transformer are installed at bus
lead the system to a voltage collapse if the generator was not 5, so that the factor that predominantly influences the voltage
disconnected quickly. behavior of this bus is the response of the reactive power ex-
One important aspect related to the transient stability issue changed between the generator and the network.
is to determine the critical active power, i.e., the maximum ac- In order to obtain a better understanding of the influence of
tive power that the generator can inject exhibiting a stable re- each type of generation on voltage sags, many repeated simula-
sponse for a determined actuation time of the protection system. tions were carried out considering different clearance times of
In order to evaluate this question, repeated transient stability the fault described previously. Voltage sags can be characterized
simulations were conducted for different fault clearance times. by their magnitude (minimal value of voltage) and duration (pe-
The same contingency previously described was simulated, i.e., riod that the voltage remains below a determined value) [17]. In
a three-phase-to-ground short circuit at bus 4, which is elim- this paper, the value adopted to calculate the duration voltage
inated by tripping branch 2–4. The results are summarized in sag was 0.85 p.u. The results are summarized in Table VII,
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
FREITAS et al.: COMPARATIVE ANALYSIS BETWEEN SYNCHRONOUS AND INDUCTION MACHINES 307

TABLE VII
VOLTAGE SAGS DUE TO A PHASE-TO-GROUND SHORT CIRCUIT

Fig. 7. Response of the nodal voltage of buses 4 and 5 for a phase-A-to-ground Therefore, it can be verified that, typically, the installation of
short circuit applied at bus 4. (a) Nodal voltage of bus 4. (b) Nodal voltage of
bus 5. a generator in an industry plant can reduce the voltage sag at
this bus. However, the voltage supplied to the other consumers
where the voltage sags of buses 4 (where the fault is applied) may be adversely affected by this installation.
and 5 (where the generator is installed) are shown.
Analyzing the behavior of bus 5 voltage, one can confirm VIII. SHORT-CIRCUIT CURRENTS
that the usage of the constant voltage synchronous generator im- The installation of ac generators may elevate the values of
proves the voltage performance under sag magnitude viewpoint. the short-circuit currents, becoming mandatory to update the
In the case of constant power factor synchronous generator, for protection and/or the network devices. Moreover, the relay set-
some situations, the voltage sag magnitude is improved. How- tings need to be readjusted to detect faults properly. Thus, in this
ever, when the fault clearance time increases, the presence of section, the short-circuit currents supplied by the ac generators
the generator affects negatively the response of bus 5 voltage. during balanced and unbalanced faults are determined by using
On the other hand, in the case of the induction generator, inde- electromagnetic transient simulations. The fault and ground re-
pendent of the fault clearance time, both the magnitude and the sistances were set equal to 0.001 ohm.
duration of the voltage sag are adversely affected when com- Fig. 8 presents the dynamic behavior of the currents sup-
pared with the case without generators. Such differences can be plied by the generators (stator current) during a three-phase-to-
explained through the dynamic behavior of the reactive power ground short circuit applied at bus 5 at ms. The system
exchanged between the generators and the network, as previ- employed is the same as presented in Fig. 3. It can be seen
ously discussed. In the case of bus 4 voltage, independent of the that the current response is different from each generator. In the
type of the generator employed, in all cases, the voltage sag is case of the induction generator, although initially the magnitude
aggravated by the installation of the generators due to the in- of the currents is high, they decrease quickly because this ma-
crease in the system short-circuit level. chine has no capacity to provide sustained short-circuit currents
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
308 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 21, NO. 1, FEBRUARY 2006

Fig. 9. Stator current during a phase-A-to-ground short circuit. (a) Constant

voltage synchronous generator. (b) Constant power factor synchronous
generator. (c) Induction generator.

Fig. 8. Stator current during a three-phase-to-ground short circuit. (a) Con-

stant voltage synchronous generator. (b) Constant power factor synchronous during three-phase faults. In this situation, the network three-
generator. (c) Induction generator. phase voltages drop to zero and the capacitor bank becomes
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
FREITAS et al.: COMPARATIVE ANALYSIS BETWEEN SYNCHRONOUS AND INDUCTION MACHINES 309

TABLE VIII
SHORT-CIRCUIT CURRENTS SUPPLIED BY THE AC GENERATOR (IN P.U.)

TABLE IX cited by the network. On the other hand, in the case of the syn-
PRE-FAULT TERMINAL VOLTAGE (IN P.U.) chronous generators, all currents present a sustained response
due to the presence of the excitation system.
In order to obtain a better understanding of the short-circuit
currents supplied by ac generators, many repeated simulations
were conducted for different faults. The results are summarized
in Table VIII. In this table, the second, sixth, and tenth columns
unloaded. Consequently, there is no external excitation source show the maximum (peak) value of the stator current after the
for the generator, and it becomes unable to produce voltage. fault for A, B, and C phases, respectively. The other columns
Theoretically, this fact could become the detection of faults by show the rms value of the stator currents at different instants
protection systems based on over-current relays more difficult. after the fault application.
However, in this case, voltage-based relays could be used. In the The previous discussion can be confirmed by analyzing these
case of synchronous generators, it can be observed that the usage results. In addition, it can be verified that the largest peak values
of the excitation system as a voltage regulator permits that the of currents are related to the synchronous generator cases, and
generator supplies sustained short-circuit current. Nevertheless, the smallest values can be related to the induction generator
if the excitation system is used as a power factor regulator, this cases. This fact can be partially explained by analyzing the
capability is decreased. pre-fault magnitude of the terminal voltage, which is shown in
Fig. 9 shows the currents supplied by the generators during Table IX. The larger the pre-fault magnitude of the terminal
a phase-A-to-ground short circuit applied at bus 5 at voltage, the larger the peak of current. In addition, the results
ms. In this case, the induction generator can supply sustained show that, usually, the peak of current supplied by the induction
short-circuit currents during the fault. However, it is interesting generator is lower than the current supplied by the synchronous
to observe that the current of the faulted phase (phase-A) de- generator. This characteristic can be useful if it is desirable to
creases quickly to zero. Only phase B and C currents present a expand the distributed generation in networks with constraints
sustained response. It occurs because these phases remain ex- related to elevation of the short-circuit level.
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
310 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 21, NO. 1, FEBRUARY 2006

IX. CONCLUSION Induction generator (5 MVA):

This paper presented an extensive study about the impacts p.u. p.u.
provoked by the connection of induction and synchronous gen-
erators to distribution networks. The objective was to determine p.u. p.u.
the main technical differences between these generators. In the p.u.
case of synchronous generators, we analyzed the usage of the
excitation system as a voltage or a power factor regulator.
B. Data of System 2
To sum up, it was verified that from the viewpoint of a steady-
state voltage profile, voltage stability, and transient stability, the Substation transformer 1 (100 MVA):
usage of constant voltage synchronous generators is advanta-
geous and permits to increase the allowable penetration level of p.u. p.u.
distributed generation. The usage of induction generators may p.u.
be interesting in networks suffering from constraints related to
the increase in the short-circuit levels. In the case of voltage Substation transformer 2 (100 MVA):
sags, it was observed that the usage of constant voltage syn-
p.u. p.u.
chronous generators can improve the dynamic performance of
the voltage at the installation point. However, other consumers p.u.
may be adversely affected due to more intense voltage sags. Ac-
Generator transformer (30.5 MVA):
cording solely to the technical factors analyzed in this paper,
the usage of constant power factor synchronous generators may p.u. p.u.
be considered the worst option. However, other factors must be
p.u.
considered to decide what is the best option in global terms, for
example, economical and political aspects. Impedances of the feeders:

APPENDIX
In this section, the systems data are presented. All symbols
used are defined in [4].

A. Data of System 1
Substation transformer (100 MVA): Synchronous generator (30 MVA):

p.u. p.u. p.u. p.u.

p.u. p.u. p.u.
p.u. p.u.
Generator transformers (5.1 MVA):
p.u. p.u.
p.u. p.u.
p.u.

Impedances of the feeders: Induction generator (30 MVA):

p.u. p.u.
p.u. p.u.
p.u.

REFERENCES
Synchronous generator (5 MVA): [1] Impact of Increasing Contribution of Dispersed Generation on the
Power System, 1999. CIGRÉ Working Group 37.23 CIGRÉ, Tech.
p.u. p.u. Rep..
[2] Dispersed Generation, 1999. CIRED Working Group 4, CIRED, Tech.
p.u. p.u. Rep..
p.u. p.u. [3] N. Jenkins, R. Allan, P. Crossley, D. Kirschen, and G. Strbac, Embedded
Generation, 1st ed. London, U.K.: Inst. Elect. Eng., 2000.
p.u. p.u. [4] SimPowerSystems User’s Guide, 2002. TransÉnergie Technologies Inc..
[5] IEEE Task Force, “Load representation for dynamic performance anal-
ysis,” IEEE Trans. Power Del., vol. 8, no. 1, pp. 472–482, Jan. 1993.
[6] V. Akhmatov, H. Knudsen, A. H. Nielsen, J. K. Pedersen, and N. K.
Poulsen, “Modeling and transient stability of large wind farms,” Int. J.
Elect. Power Energy Syst., vol. 25, no. 1, pp. 123–144, 2003.
Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.
FREITAS et al.: COMPARATIVE ANALYSIS BETWEEN SYNCHRONOUS AND INDUCTION MACHINES 311

[7] R. Belhomme, M. Plamondon, H. Nakra, D. Desrosiers, and C. Gagnon, Andre Morelato (M’89) graduated in electronics engineering from Instituto
“Case study on the integration of a nonutility induction generator to the Tecnológico Aeronáutica in 1970 and received the Ph.D. degree in 1982 from
Hydro-Quebec distribution network,” IEEE Trans. Power Del., vol. 10, the State University of Campinas, Campinas, Brazil.
no. 3, pp. 1677–1684, Jul. 1995. He is currently a Full Professor of electrical engineering at the State Univer-
[8] N. P. McQuin, P. N. Willians, and S. Williamson, “Transient electrical sity of Campinas. During late 1991 to 1992, he worked at Hitachi Research Lab-
and mechanical behavior of large induction generator installations,” in oratory, Hitachi, Ltd., in Japan. His general areas of research interest are power
Proc. 4th Int. Conf. Electrical Machines Drives, Sep. 1989, pp. 251–255. system control and stability, distributed generation, and parallel processing ap-
[9] J. R. Parsons, “Cogeneration application of induction generators,” IEEE plications.
Trans. Ind. Appl., vol. IA-20, 1984.
[10] P. Kundur, Power System Stability and Control. New York: McGraw-
Hill, 1994.
[11] J. D. Hurley, L. N. Bize, and C. R. Mummert, “The adverse effects of
excitation system var and power factor controller,” IEEE Trans. Energy
Convers., vol. 14, no. 4, pp. 1636–1641, Dec. 1999.
[12] C. L. Masters, “Voltage rise: The big issue when connecting embedded
generation to long 11 kV overhead lines,” Power Eng. J., vol. 16, no. 1, Luiz Carlos Pereira da Silva graduated in electrical engineering in Federal
pp. 5–12, 2002. University of Goias, Goias, Brazil, in 1995 and received the M.Sc. and Ph.D.
[13] S. K. Salman, “The impact of embedded generation on voltage regula- degrees in electrical engineering from the State University of Campinas, Camp-
tion and losses of distribution networks,” Proc. IEE Colloq. Impact Em- inas, Brazil, in 1997 and 2001, respectively. From 1999 to 2000, he was visiting
bedded Generation Distribution Networks, 1996. Ph.D. student at the University of Alberta, Edmonton, AB, Canada.
[14] C. L. T. Borges and D. M. Falcão, “Impact of distributed generation At present, he is an Assistant Professor at the State University of Campinas.
allocation and sizing on reliability, losses and voltage profile,” in Proc. His research interests are power system stability and control and distributed gen-
IEEE/Power Engineering Society Power Tech Conf., 2003. eration.
[15] J. Mutale, G. Strbac, S. Curcic, and N. Jenkins, “Allocation of losses
in distribution systems with embedded generation,” in Proc. Inst. Elect.
Eng., Gener., Transm., Distrib., vol. 147, Jan. 2000, pp. 7–14.
[16] O. Samuelsson and S. Lindahl, “On speed stability,” IEEE Trans. Power
Syst., vol. 20, no. 2, pp. 1179–1180, May 2005.
[17] IEEE Recommended Practice for Monitoring Electric Power Quality,
IEEE Std 1159-1995, Jun. 1995.
Vivaldo Fernando da Costa received the Ph.D. degree in electrical engineering
from the State University of Campinas, Campinas, Brazil, in 1992.
From 1993 to 1994, he was a Visiting Professor at the University of British
Walmir Freitas (M’01) received the Ph.D. degree in electrical engineering from
the State University of Campinas, Campinas, Brazil, in 2001. From 2002 to Columbia, Vancouver, BC, Canada. At present, he is an Associate Professor
2003, he was a Postdoctoral Fellow at the University of Alberta, Edmonton, at the State University of Campinas. His research interests are power system
stability and control and distributed generation.
AB, Canada.
At present, he is an Assistant Professor at the State University of Campinas.
His areas of research interest are power system stability and control, distributed
generation, and power electronic applications.

Jose C. M. Vieira (S’98) graduated in electric engineering from the Federal

University of Uberlandia, Uberlandia, Brazil, in 1996 and received the M.Sc. Flavio A. B. Lemos received the Ph.D. degree in electrical engineering from
degree from the State University of Campinas, Campinas, Brazil, in 1999. He is the Federal University of Santa Catarina, Santa Catarina, Brazil, in 2000.
presently working toward the Ph.D. degree at the State University of Campinas. From 1996 to 1997, he was a Visiting Researcher at Brunel University,
From 1999 to 2003, he was with FIGENER as a Consulting Engineer. His Uxbridge, U.K. At present, he is an Assistant Professor at the Catholic Uni-
research interests are distributed generation, power system control and dynamic, versity of Rio Grande do Sul, Brazil. His research interests are power system
optimal power flow, and energy market. stability and control and distributed generation.

Authorized licensed use limited to: Petrobras. Downloaded on November 28,2024 at 01:17:19 UTC from IEEE Xplore. Restrictions apply.