IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 41, NO.

1, JANUARY/FEBRUARY 2005 155

Transient Stability of Small Plant Generators

Connected to a Weak Utility
System—A Case Study
J. C. Das, Senior Member, IEEE

Abstract—Small generators connected through a large Instability occurs when the post fault trajectory of a power
impedance and operating in synchronism with a weak utility system does not approach steady state. The oscillations in rotor
system are more likely to pull out of step on a large perturbation, angle last typically one second or less and the transient stability
i.e., a fault in the utility’s system. Under an incident impact,
which will be distributed unevenly among various machines, every is decided in this interval. The dynamics of excitation systems,
machine will be retarded or accelerated, and undergo damped underfrequency load shedding, automatic voltage regulators and
or divergent oscillations. This paper demonstrates that the small speed governing systems are all active in this time frame [1]–[3],
machines are more vulnerable and may fall out of step. This may [26].
plunge the entire process plant into a complete power outage on Dynamic stability studies cover longer time intervals, some-
system separation. A case study of two small generating units
and the improvements in the transient stability by fast system times up to 30 s. These studies are aimed to obtain a larger
separation and load shedding is demonstrated. system response, though the attention may be focused on
a single unit or plant. The breakup of a large system can be
Index Terms—Excitation systems, generator and load modeling,
natural frequency, synchronizing forces, transient stability. studied, and much effort is concentrated on undamped or poorly
damped oscillations. Conditions for oscillations between major
subsystems may depend on many variable factors, such as load
I. INTRODUCTION levels, generator loading, and voltages, all of which are difficult
to predict with precision. The subharmonic frequencies are
T HE purpose of a transient stability study for an industrial
generating facility is to study and implement necessary
cost-effective changes so that the plant generators can ride
classed as inter-area oscillations, local machine-system oscilla-
tions, inter-unit mode oscillations, and torsional oscillations [4],
through disturbances in the power system. A power system is [5]. Power system stabilizers (PSSs) are used to eliminate and
highly nonlinear and continuously experiences disturbances. substantially reduce the incidence of spontaneous oscillations.
From a stability point of view these can be classified into two For the pulp and paper mill environment the largest size of the
categories: 1) contingency disturbances due to lightning, short generator integrated in 13.8-kV distribution is around 100 MVA
circuit, insulation breakdowns, and incorrect relay operations (mostly connected through generator reactors) No PSSs have
(these may be called large perturbations or event disturbances) been used in these installations. The dynamic characteristics of
and 2) load disturbances because of random variations in a synchronous generator change with the torque angle and the
the load demand. There are many definitions of the power regulator gain and the stability of the excitation control system
system stability in the literature, however, with respect to the is reduced. When a PSS is not applied, a fast acting or wider
fault disturbances and an initial (prefault) steady equilibrium bandwidth excitation system has a greater chance of destabilizing
point (SEP), it explores whether the post-fault trajectory will local-machine system oscillations. The National Electric Regu-
settle down to a new equilibrium point in an acceptable steady latory Council (NERC) recommends a PSS for generators that
state. The definition of the final steady state is important. The exceed 30 MVA or a group of machines that exceeds 75 MVA.
post-fault state may have periodic oscillations, which cannot be Time-domain methods of power system stability analysis
accepted. Even small fluctuations in the voltage and frequency have been in extensive use. The model for stability analysis of
are not desirable and cannot be accepted in the post fault state. a power system may consist of many differential and algebraic
Nor subsynchronous resonance, which may occur due to con- equations, depending upon the complexity of the system under
version of mechanical energy into electrical energy associated study, and the stability analysis programs are based upon
with subharmonic mode [5] and electromagnetic oscillations step-by-step numerical integration to simulate system behavior
(due to interaction between magnetic fields) can be tolerated. relative to a given disturbance. [1], [2], [6]–[8].
For a large system, the stability analysis in the time domain
means solving thousands of nonlinear differential and algebraic
Paper PID-04-26, presented at the 2004 IEEE Pulp and Paper Industry Con- equations. The direct methods in contrast integrate only the fault
ference, Portland, ME, June 21–26, and approved for publication in the IEEE
TRANSACTIONS ON INDUSTRY APPLICATIONS by the Pulp and Paper Industry on the system and determine without integrating the post-fault
Committee of the IEEE Industry Applications Society. Manuscript submitted system whether or not the pre-fault system will remain stable
for review July 1, 2004 and released for publication August 26, 2004. after the fault is cleared. This is done by comparing the system
The author is with AMEC E&C Services Inc. Tucker, GA 30084 USA
(e-mail: [email protected]). energy, when the fault is cleared to a critical energy value. There
Digital Object Identifier 10.1109/TIA.2004.840973 are several methods for determining the critical energy value,
0093-9994/$20.00 © 2005 IEEE
156 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 41, NO. 1, JANUARY/FEBRUARY 2005

i.e., the closest unstable equilibrium point (UEP) and potential

energy boundary surface method (PEBS) [9]–[14]. The Electric
Power Research Institute (EPRI) program for the direct method
is in [15].
For industrial power systems, a stability study is not gener-
ally carried out, because of the cost, unless some problems are
experienced afterwards. These are mainly due to frequent faults
or disturbances in the utility system. This means that any system
modifications to improve the stability have to be carried out
after the installations are in place. Computer programs for in-
dustrial systems are invariably in the time domain and may have
limited computational capability. Some of these commercially
available programs may have enhanced capabilities to simulate a
number of generators, excitation system, governing system, and
turbine models with user-programmable interfaces. In an indus-
trial plant the loads are well defined and their dynamic models
are important for accuracy of results. Some programs may lack
in this respect.

II. STABILITY PROBLEM OF SMALL GENERATORS

Consider an incident impact in a multimachine system. This
will be unevenly distributed among various synchronous ma-
Fig. 1. A group of synchronous machines striving for some mean retardation
chines on a large incident impact, through individual oscillations.

(1)

where is the incident impact and is the

share of the individual machines in the overall impact.
Considering that the machines are connected through negli-
gible impedances, the initial angle of impact is the same for
all the machines
(2)
where are the synchronizing power of indi-
vidual machines, and is the overall synchronizing power
brought out by the incident impact. Fig. 2. Explanation of the progress of oscillations of a synchronous machine,
Rigid machines with high synchronizing power take a large in a multimachine system on an incident impact.
share of the impact and the softer machines a smaller share of
the impact. Under this impact every machine is retarded or ac- a higher degree. The larger machines are less retarded than ini-
celerated tially and are relieved of their overload. The final torque angle
may be larger or smaller than the initial angle of impact, de-
(3) pending upon the ratio of the natural frequency of the machine
to the natural frequency of the entire network.
where is the initial retardation and is the natural frequency Fig. 2 shows that is the initial torque angle, which swings to
of the machine. Thus, the initial retardation under the impact , given by the initial impact angle and the machine produces
differs if the natural frequency differs. The natural frequency is its share of synchronizing power in the impact. All massesstrive to
given by attain same retardation, and this changes torque angle to . This
ultimate value is achieved only after some damped oscillations, in
which torque angle swings to , given by
(4)
rpm
(5)
where is the synchronizing power, is the moment of
inertia in - , and is the system frequency in hertz. where isthesystemhypotheticalnaturalfrequency,basedupon
The synchronizing forces thus brought into play cause all ma- the total synchronizing power and the moment of inertia. A gener-
chines to strive for some mean retardation, Fig. 1. Each machine ator of large synchronizing power, a rigid machine, will have high
undergoes further individual oscillations to accomplish this. natural frequency, greater than the system frequency. If it is two
After decay of oscillations the smaller machines are more re- times the system frequency, , i.e., the initial im-
tarded than they were under the initial impact and are loaded to pact angle for this machine was too large and reduces. A generator
DAS: TRANSIENT STABILITY OF SMALL PLANT GENERATORS CONNECTED TO A WEAK UTILITY SYSTEM 157

chine. A generator with higher SCR has a relatively higher field

magnetomotive force (MMF) and operates at a smaller torque
angle compared to a generator of identical rating, but of lower
SCR. Modern generators tend to have lower short-circuit ratios
compared to their predecessors, of the order of 0.5 and the in-
crease in the stability limit is achieved through high response
excitation systems. A user may have little choice with respect to
specifying higher short-circuit ratio due to manufacturing con-
straints. SCR will impact all the other generator reactances and
also its short-circuit contribution to a system. Modifications to
relaying systems, faster system separation, fault isolation, and
load shedding are some of the measures that can be considered
Fig. 3. Oscillations in a multimachine system on an incident impact and
to improve the stability limit. Existing excitation systems can be
instability of smaller softer machines. replaced, the excitation systems with voltage response time of
0.1 s or less are termed high initial response excitation systems
ofsmallsynchronizingpowerhasalownaturalfrequencyanddoes [17].
not contribute much to the system natural frequency. If it is 1/2 of
the system frequency, then . For this machine, the initial III. A CASE STUDY WITH SMALL GENERATORS
impact was too small and increases to four times its original value. Fig. 4 shows an electrical distribution system for a paper mill
Themachinemaybethrownoutofsynchronism,Fig.3.Thepower facility. The power is received at 44 kV, and the stability of
impacts in the network are most disturbing for small machines, the two small generators of 5.0 and 3.125 MVA connected to
with small synchronizing power. 4.16-kV bus is of concern, for faults in the utility system. The
The above scenario considers a power impact, i.e., a sudden mill load is approximately 28 khp of induction motors mainly
demand of power at some point in the interconnected system. connected to the low-voltage distribution system. The induc-
Under a fault condition, there will be excess power, as the tion motors and their power factor improvement capacitors are
voltage collapses and the load demand may reduce. Thus, shown lumped together on each equivalent load bus, connected
the machines will accelerate rather than de-accelerate. Also, through transformers of equivalent impedances. The implica-
the internal voltage of the machines is considered to remain tions of this lumping are further discussed.
unchanged during the power impact. Practically, the excitation
system and machine characteristics will alter this simplified
A. Statement of the Problem and Study Objective
picture of the complex phenomena of stability in a multima-
chine system. However, the vulnerability of smaller machines The utility tie represents a weak system. The three-phase fault
to fall out of step is well illustrated [16]. current and single line-to-ground fault current at the 44-kV plant
The factors affecting stability are as follows: bus is 3 kA and 2.5 kA rms symmetrical, respectively. There
• short-circuit ratio of the generator; are other consumers served from the same 44-kV bus in the
• inertia constant ; utility’s substation and a fault in the utility system causes severe
• transient reactance of the generator; voltage dips pulling out both the generators out of synchronism,
• interconnecting system impedance; plunging the whole mill into dark. A tap to other consumer from
• speed of relaying; the same 44-kV line to mill was removed, improving the situa-
• excitation system types and response; tion slightly. It is desired to maintain stability of the generators
• fast load shedding; when a system separation occurs at utility’s 44-kV breaker and
• auto-reclosing. maintain the essential loads of 5-khp supplied from the two gen-
In an industrial distribution system experiencing stability erators.
problems, some of these parameters may not be possible to be
changed easily. Auto-closing is avoided in industrial facilities, B. Data for Study
so that induction motors are not subjected to high transient The data required for the study can be divided into the fol-
torque and inrush current in case the power is switched in lowing categories.
phase-opposition to the motor residual voltage. Synchronous 1) System Data: A converged load flow study is required
motors are, generally, not suitable for auto-transfer of power. as the starting point. Data for dynamic representation of syn-
The short-circuit ratio of a generator is defined as shown by the chronous and induction machines, excitation, automatic voltage
equation at the bottom of the page. regulator (AVR), turbine/speed governing system models need
SCR can, therefore, be considered an inverse of saturated syn- to be accurately established [1], [2]. Static load data and power
chronous reactance and is a measure of the stiffness of the ma- capacitors data are required as applicable.

Per-unit excitation at normal voltage on open circuit

SCR
-
158 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 41, NO. 1, JANUARY/FEBRUARY 2005

• Referring to Fig. 4, breakers A and B at the 4.16-kV buses

are provided with time-delayed directional overcurrent re-
lays (device 67), looking in to the utility’s system. How-
ever, for a fault in the utility’s system, these relays are slow
and will not clear a ground fault in the utility’s system as
the primary windings of all utility tie transformers, T1, T2,
and T3 are delta connected.

C. Study Parameters
These parameters identify, the data to be monitored and
plotted, duration of study, and integrating time interval.

D. System Models
The accuracy of the models of excitation systems, generators,
synchronous and induction motors, and speed governing sys-
tems has a profound impact on the accuracy of the results. The
models should be based upon the manufacturer’s data, as far as
possible. For older plants all the data may not be easily avail-
able. The assumptions should have a firm base. It may be pos-
sible to derive model parameters from field test measurements
[18], [19], though it may be time consuming and sometimes dif-
ficult to implement in a continuous process plant.
1) Generator Models: A synchronous generator can be
modeled with the following: an equivalent circuit model (1),
which uses an internal voltage behind the armature resistance
and quadrature axis reactance. This is the classical simple
generator model, used for simplified stability simulations or to
represent external system connection, i.e., a utility connection.
In this model (2), saliency and transient effects and the satura-
tion effects are considered. In this model (3) in addition to the
saliency and transient/subtransient effects, the direct axis and
Fig. 4. An equivalent distribution system for study of transient stability of
small generators G1 and G2.
quadrature axis damper circuits are modeled, usually for large
machines. Further details of synchronous generator models for
the stability studies are provided in [20].
2) Disturbance Data: Initial switching status, fault type, The circuit model (2) was used for the 5- and 3.125-MVA
protective relay operating time, fault location and fault removal, generators.
clearing time of associated breakers, and load shedding data is The reactive capability curve of the generators should be con-
required. sidered and the actual operating megawatts and not the rated
3) Relaying Data: There is no incoming breaker at the point megawatt output with constraints of reactive power, lagging and
of the utility’s connection with the plant distribution and an up- leading, from the reactive power capability curve of the gen-
stream 4-kV utility breaker trips for faults in the 44-kV line and erator should be modeled. The 5-MVA generator operates at
the plant distribution system. This breaker is provided with dis- 2.7-MW active power output and the 3.125-MVA generator op-
tance relays looking into the 44-kV line and plant distribution erates at 2.0 MW.
as follows. 2) Induction Motors: The rotating load models affect the
• Approximately 80% of the 44-kV line is protected in the transient stability limits of the plant generators. Fig. 4 shows
first zone of protection. Settings for phase and ground fault that all the induction motor loads are lumped together on equiv-
relays are identical. The breaker interrupting time is 3 cy- alent transformers and an equivalent circuit model, with a single
cles and, therefore, the fault in this zone is cleared in ap- inertia constant is used. There are some errors inherent in this
proximately 4 cycles method of representation [21], as the motors will have different
• The second zone of distance protection for phase and torque–speed characteristics and, will be driving loads of dif-
ground faults extends to the remaining 20% of the 44 –kV ferent inertia and load characteristics. However, it will be im-
line and partially looks into the utility tie transformers practical to dynamically model each and every motor and its
(T1, T2, and T3 in Fig. 4) winding faults also. The second load, and an equivalent machine model is an acceptable choice.
zone faults are cleared in 18 cycles. This includes breaker If the behavior of one particular motor is required, it can be dis-
interrupting time, relaying time, and the second zone cretely modeled.
distance timer. An induction motor can be represented with a model similar
• An underfrequency relay set at 59.6 Hz with a time delay to that of a synchronous machine in the transient state, i.e., by a
of 14 cycles supplements the distance protection. Thévenin equivalent circuit of voltage behind stator resistance
DAS: TRANSIENT STABILITY OF SMALL PLANT GENERATORS CONNECTED TO A WEAK UTILITY SYSTEM 159

Fig. 5. Generator terminal voltage for a phase-to-phase fault in the utility’s system. Solid lines show fault cleared in the first zone of distance protection in 4
cycles and the dotted lines show fault cleared in 18 cycles in the second zone of distance protection.

Fig. 6. Generator terminal frequency for a phase-to-phase fault in the utility’s system. Solid lines show fault cleared in the first zone of distance protection in 4
cycles and the dotted lines show fault cleared in 18 cycles in the second zone of distance protection.

and transient reactance, calculated based upon the motor input parameters given in this standard and the manufacturer’s
parameters of stator and rotor resistances, leakage reactances, specific data.
magnetizing reactances, and slip-load characteristics. 5- and 3.125-MVA generators, Fig. 4, have dc commutator,
There are no synchronous motors in the distribution. If continuously acting, rotating exciters, which are modeled as
present these can be modeled as synchronous generators with IEEE type 1, with the following parameters:
negative active power, and generally with fixed excitation
system.
3) Excitation System: Complete modeling of excita- . Reference [19] may
tion and automatic voltage regulator systems is necessary be seen for the definition of these symbols and the control circuit
for the transient stability studies. The manufacturer’s data block diagram for the type-1 excitation system.
should be used for accuracy. References [22]–[24] provide 4) Governor and Turbine: The model should be derived
IEEE recommended types and [23] provides typical transfer based upon the manufacturer’s data. Reference [25] provides
function parameters, block circuit diagrams, Bode plots, and IEEE dynamic models. A simplified model of the steam turbine
time constants for various types of excitation systems, i.e., DC1, in terms of speed regulation (droop), lead and lag time con-
DC2, DC3, ST1, ST2, ST3, AC1, AC2, AC3, AC4, AC5A, stants, dead band, and minimum and maximum turbine power
etc. However, there can be variations in the generalized model output was used.
160 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 41, NO. 1, JANUARY/FEBRUARY 2005

Fig. 7. Active and reactive power transients in generators, phase-to-phase faults.

Fig. 8. Field current and field voltage transients on the generators.

IV. STUDY RESULTS AND INFERENCES less than the generator output. (This is necessary, so that fre-
A. Three-Phase Fault in the Utility’s System Cleared in the quency recovers.) Considering a 10-cycle delay to implement
First Zone of Distance Protection these switching actions, the stability cannot be maintained. The
generators torque angles do not recover. A replacement of the
The interrupting time of the 44-kV breaker is three cycles. existing excitation systems with static excitation systems, type
Thus, the fault is cleared in approximately 4 cycles. No other ST1 had no appreciable effect on the stability situation.
switching action takes place. The 4.16-kV generator bus suffers
a voltage dip of 83%, which improves to approximately 50%
B. Three-Phase Fault in the Utility’s System in the Second
after fault clearance and then decays fast. The induction motors
Zone of Distance Protection
rapidly slow down, heading toward a lockout condition. The
generator torque angles swing to 360 within 3 cycles of fault A fault in the second zone of distance protection is practically
occurrence and do not recover. of the same magnitude and is cleared in 18 cycles (second zone
Load shedding does not improve the situation. Referring to distance timer plus breaker interrupting time and relay operating
Fig. 4, breakers A and B are rated on the 8-cyle “TOT” basis time). Thus, the stability cannot be maintained for three-phase
(prior to 1964 basis). Consider that a fiber-optic link is estab- faults in the utility system.
lished between the utility’s system and the plant breakers A and Criteria for a three-phase fault is often used because if the sta-
B, Fig. 4, are tripped with simultaneous load shedding. Enough bility can be demonstrated for a three-phase fault, then the sta-
loads are shed so that the net load on the system is slightly bility on unsymmetrical fault need not be investigated. However,
DAS: TRANSIENT STABILITY OF SMALL PLANT GENERATORS CONNECTED TO A WEAK UTILITY SYSTEM 161

Fig. 9. Transients in the equivalent induction motor retained in service.

Fig. 10. Slip recovery of the equivalent induction motor retained in service, showing stability.

to design a system to be stable for three-phase faults may be eco- the frequency has dipped to 58 Hz at the time of fault clearance
nomically prohibitive. A bolted three-phase fault is a rare occur- and continues its downward trend after the fault clearance. The
rence. Almost 70% of the line faults are single line-to-ground lowest frequency is 55.75 Hz shortly after the generation is iso-
faults. After a three-phase fault, in order of decreasing severity lated and the load is shed at 18 cycles.
for stability, the other fault types are: When the fault is cleared in 18 cycles in the second zone of
• double line-to-ground fault; distance relaying, the frequency stays high, not dipping below
• phase-to-phase fault; 59.6 Hz., however, the voltage is 0.60 pu (40% voltage dip) after
• single line-to-ground fault. 8 cycles of fault clearance and continues its downward trend. It
The stability is obtained for these faults provided that the fol- is 0.43 per unit when the generation is isolated and the load is
lowing occur. shed after 18 cycles of fault occurrence.
• The generation is separated in 18 cycles based upon ap- Thus, the system separation and the load shedding is imple-
propriate frequency and undervoltage relay settings. mented as follows:
• All nonessential loads are shed, leaving approximately
5-khp essential loads in service, simultaneously with the • an underfrequency relay set at 59 Hz, time delay setting of
isolation of generation from the utility’s source. 2 cycles;
Figs. 5 and 6 show the voltage and frequency profiles on the • an instantaneous three-phase undervoltage relay set at
generator bus for a phase-to-phase fault cleared in the first and 60% voltage in conjunction with a timer, time delay of 4
second zones of distance protection. Fig. 5 shows that the gen- cycles.
erator terminal voltage is 0.66 pu (voltage dip of 34%) at 9 cy- The outputs of these relays are AND gated to shed load and
cles, for a fault cleared in the first zone of protection. However, simultaneously effect system separation.
162 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 41, NO. 1, JANUARY/FEBRUARY 2005

A hardwired load shedding system cannot continuously [6] Extended Transient/Mid Term Stability Program Package, ver. 3, Elect.
monitor the load/generation balance and the appropriate load Power Res. Inst., Palo Alto, CA, 1992.
[7] F. P. de Mello, “Simulating fast and slow dynamic effects in power sys-
to be shed under varying load conditions. Thus, in a hard- tems,” IEEE Comput. Appl. Power, vol. 5, no. 3, pp. 33–38, Jul. 1992.
wired system, either too much or too little load may be shed [8] M. Stubbe, A. Bihain, J. Deuse, and J. C. Baader, “STAG—A new
unified software program for the study of the dynamic behaviour of
depending upon the selected load trips. However, this is not electrical power systems,” IEEE Trans. Power Syst., vol. 4, no. 1, pp.
much of a disadvantage in the distribution system under study, 129–138, Feb. 1989.
as practically all the load is shed, except essential generator [9] A. A. Fouad and V. Vittal, Power System Transient Stability Analysis,
Using the Transient Energy Function Method. Englewood Cliffs, NJ:
auxiliaries and instrumentation loads. A power-line-controller Prentice-Hall, 1991.
(PLC)-based system can intelligently shed the load with trend [10] M. A. Pai, Energy Function Analysis for Power System Sta-
bility. Boston, MA: Kluwer, 1989.
analysis and continuous load monitoring and can supplement [11] T. Athy, R. Podmore, and S. Virmani, “A practical method for the di-
the hardwired load-shedding logic. rect analysis of transient stability,” IEEE Trans. Power App. Syst., vol.
PAS-98, no. 2, pp. 573–584, Mar./Apr. 1979.
Fig. 7 shows the active and reactive power transients and [12] P. P. Varaiya, F. F. Wu, and R. L. Chemn, “Direct methods for transient
Fig. 8 shows the field circuit transients of the generators for stability of power systems, recent results,” Proc. IEEE, vol. 12, no. 12,
phase-to-phase faults. Figs. 9 and 10 show equivalent induction pp. 1703–1715, Dec. 1985.
[13] H. D. Chiang, “The BCU method for direct stability,” IEEE Trans. Power
motor transients retained in service. Syst., vol. 9, no. 3, pp. 1194–1208, Aug. 1994.
[14] A. A. Fouad, F. Aboytes, V. F. Carvalho, S. L. Corey, K. J. Dhir, and
R. Vierra, “Dynamic security assessment practices in North America,”
V. STABILIZING MOTOR CONTACTORS IEEE Trans. Power Syst., vol. 3, no. 3, pp. 1310–1321, Aug. 1988. IEEE
Committee Report.
Severe voltage dips occur for faults in the utility’s system. En- [15] User’s Manual for DIRECT Version 4.0 (EPRI TR-105886s), Elect.
suring stability of the generators is important, however, it is also Power Res. Inst., Palo Alto, CA, 1995.
[16] J. C. Das, “Transient stability of small plant generators,” in Proc. TAPPI
important that the essential service loads do not drop out during PCE&I Conf. Symp., Williamsburg, VA, Mar. 2000, pp. 209–219.
this fault voltage dips. Motors controlled through NEMA E1 [17] IEEE Guide for Identification, Testing, and Evaluation of Dynamic Per-
and E2 starters with three-wire control may trip on voltage dips. formance of Excitation Control Systems, IEEE Std. 421.2-1990 (Revi-
sion of IEEE Std. 421A-1978).
The low-voltage contractors have even a lower tolerance for the [18] J. W. Feltes, S. Orero, B. Fardanesh, E. Uzunovic, S. Zelingher, and
voltage dips and may drop out on 20%–70% of the voltage dip in N. Abi-Sarma, “Deriving model parameters from field measurements,”
IEEE Comput. Appl. Power, vol. 15, no. 4, pp. 30–36, Oct. 2002.
the first cycle (depending upon the trapped residual magnetism). [19] G. Manchur, D. C. Lee, M. E. Coultes, J. D. A. Griffen, and W. Watson,
Stabilizing the contractors serving essential loads for 1 s is, “Generator models established by frequency response tests on a 555
MVA machine,” IEEE Trans. Power App. Syst., vol. PAS-91, no. 5, pp.
therefore, essential. This brings about other considerations. 2077–2084, Sep./Oct. 1972.
During the voltage dip the induction motors slow down. Fig. 10 [20] IEEE Guide for Synchronous Generator Modeling Practices and Appli-
shows that the equivalent induction motor slip increases from cations in Power System Stability Analyzes, IEEE Std. 1110-2002.
[21] IEEE Task Force, “Load representation for dynamic performance anal-
1% to approximately 4.5%. On voltage recovery there will be an ysis,” IEEE Trans. Power Syst., vol. 8, no. 2, pp. 472–482, May 1993. .
inrush current spike, as all motors will accelerate together (Fig. 9). [22] IEEE Working Group Report, “Computer representation of excitation
systems,” IEEE Trans. Power App. Syst., vol. PAS-87, no. 6, pp.
This simultaneous inrush current of the motors may cause further 1460–1464, Jun. 1968. .
voltage dip and stall the quick recovery of the voltage, depending [23] IEEE Committee Report, “Excitation system models for power system
upon the system impedance’s and the groups of connected motor stability studies,” IEEE Trans. Power App. Syst., vol. PAS-100, no. 2,
pp. 494–509, Feb. 1981.
loads. This is clearly shown in Fig. 9. [24] IEEE Recommended Practice for Excitation System Models for Power
System Stability Studies, IEEE Std. 421.5, 1992.
[25] IEEE Committee Report, “Dynamic models for steam and hydro
VI. CONCLUSION turbines in power system studies,” IEEE Trans. Power App. Syst., vol.
PAS-92, no. 6, pp. 1904–1915, Nov./Dec. 1973. .
Relatively small machines connected to a weak electrical [26] IEEE Committee Report, “Dynamic models for steam and hydro
system are more liable to pull out of step on a system distur- turbines in power system studies,” IEEE Trans. Power App. Syst., vol.
PAS-92, no. 6, pp. 1904–1915, Nov./Dec. 1973.
bance. This becomes of importance as much emphasis is being
placed now on embedded generation. The case study shows J. C. Das (SM’76) received the B.A. degree in math-
that the transient stability of generators was improved, with ematics and the B.E.E. degree from Panjab Univer-
least changes to the existing distribution system by fast load sity, Chandigarh, India, in 1953 and 1956, respec-
tively, and the M.S.E.E. degree form Tulsa Univer-
shedding and system separation based upon the undervoltage sity, Tulsa, OK, in 1982.
and underfrequency relays. These relays were set based upon He is currently Staff Consultant, Electrical Power
the results of the study. Systems, AMEC E&C Services Inc., Tucker, GA.
He is responsible for power system studies, in-
cluding short circuit, load flow, harmonics, stability,
REFERENCES grounding, and also, protective relaying. He conducts
courses for continuing education in power systems
[1] P. M. Anderson and A. A. Fouad, Power System Control and Sta- and has authored or coauthored more than 40 technical publications. He is
bility. New York: IEEE Press, 1994.
the author of Power System Analysis (New York: Marcel Dekker, 2002). His
[2] P. Kundar, Power System Stability and Control. New York: McGraw-
Hill, 1994. interests include power system transients, harmonics, power quality, protection,
[3] E. W. Kimbark, Power System Stability. New York: Wiley, 1956, vol. and relaying.
3. Mr. Das is a Member of the IEEE Industry Applications and IEEE Power
[4] P. M. Anderson, B. L. Agarawal, and J. E. Van Ness, Subsynchronous Engineering Societies. He is a Member of TAPPI and of CIGRE, a Fellow of
Resonance in Power Systems. New York: IEEE Press, 1990. the Institution of Engineers, U.K., a Life Fellow of the Institution of Engineers
[5] IEEE SSR Working Group, “Terms, definitions, and symbols for sub- (India), and a Member of the Federation of European Engineers (France). He
synchronous oscillations,” IEEE Trans.Power Syst., vol. 104, no. 6, pp. is a Registered Professional Engineer in the States of Georgia and Oklahoma, a
1326–1334, Jun. 1985. Chartered Engineer (C.Eng.) in the U.K., and a European Engineer (Eur. Ing.).