IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. IA-13, NO.

2, MARCH/APRIL 1977 107

Stability Considerations for Industrial Power SystenIw PoNT

RICHARD H. McFADDEN, MEMBER, IEEE MAY 4 1977
TECHNI1CAL LIBRARY
Abstract-Transient stability is assuming increasing importance in
the design and operation of industrial power systems as trends toward
x LOOVERS
large synchronous process drives and in-plant generation continue.
Stability is the ability of a power system with multiple synchronous Ed E po
machines to survive disturbances such as faults, switching, and impact
loading without losing synchronism. The stability of a system under
specified disturbances can be examined by computerized transient Fig. 1. Simplified two-machine power system.
stability studies, and various methods exist to improve stability levels
if necessary. machines and the reactance of the connecting circuit. Losses
in the machines and the resistance of the line are neglected for
INTRODUCTION simplicity).
If the internal voltages of the two machines are EG and
OR YEARS, system stability has been a problem almnost
exclusively to electric utility engineers. Within the past
EM, and the phase angle between them is 0, we can easily
demonstrate [1], [2] that the real power transmitted from
decade, however, increasing numbers of industrial and com- the generator to the motor is
mercial facilities have installed local generation, large snychro-
nous motors, or both. This means that systems stability is of
concern to a growing number of industrial plant electrical engi- P EG EM ~~sinG.
neers and consultants. x
The maximum value of P obviously occurs when 0 = 900.
STABILITY FUNDAMENT-ALS Thus

What Do We Mean by Stability? EGEM

Pmax =-

Fuindanmentally, stability is a property of a power system

containing two or more synchronous inachines. The system is This is the steady-state stability limit for the simplified two-
stable, unzder a specified set of conditions, if all of its snychro- machine system. Any attempt to transmit more power than
n1Ouis nmachiines reInain in step with one another (or having Pmaxwill cause the two machines to pull out of step, with the
pulled out of step, regain synchronism soon afterwards). The given values of internal voltages.
emphasis on specified conditions in this definition is intended This model of a simple power system tells us that at least
to stress the fact that a system which is stable under one set of three electrical characteristics of the system affect stability.
conditions may be unstable under some other set of condi- They are
tions. How the system coniditions affect stability will be cov-
ered later.
I) internial voltage of the generator,
2) reactances of the machines and transmission system,
Steady-State Stability 3) internal voltage of the motor.
Although the discussion in the rest of this chapter will re- The higher the internal voltages, and the lower the system and
volve around stability under transient conditions such as machine reactances, the greater power can be transmitted
faults, switching operations, etc., we should also be aware that under steady-state conditions.
a power system may become unstable under steady-state
conditions. Transient Stability
The simplest power system to which stability considera- The preceding brief look at steady-state stability serves as a
tions apply consists of a pair of synchronous machines, one background for an examination of the more complicated prob-
acting as a generator and the other acting as a motor, con- lem of transient stability. This is true because the same three
nected together through a reactance. See Fig. 1. (In this model electrical characteristics which determine steady-state stability
the reactance is the sum of the transient reactances of the two limits, affect transient stability in the same way. However, it is
important to note that a system which is stable under steady-
Paper TOD-76-22 approved by the Industrial and Commercial Power state conditions is not necessarily stable when subjected to a
Systems Committee of the IEEE Industry Applications Society for pre- transient disturbance.
sentation at the 1972 Industrial and Commercial Power Systems Tech- By transient stability, we mean the ability of a power sys-
nical Conference, May 1-4, Milwaukee, WI. Manuscript released for tem to survive a sudden change in generation, load, or system
publication June 28. 1976.
The author is with the Westinghouse Electric Corp., Pittsburgh, PA characteristics without a prolonged loss of synchronism. To
15221. see how a disturbance affects a synchronous machine, we will

Authorized licensed use limited to: Centro Universitário Fei. Downloaded on March 08,2022 at 14:21:22 UTC from IEEE Xplore. Restrictions apply.
108 IF[l' O1'1AN.\ VI( INS (IN IN lI)iSTtN'YAl.l(ICAl ION S. NIA RC('Il/At'RI 1I977

1list look at the steady-state charac:terisiics, which are de-

scribed by the steady-state torque equLa;tionl I ] '

,7p2 AMON)
A
T = -
8
5SRFR sin 6?,

whiere
T mechanical shaft torque.
p inumber of poles of inachine.
rR air-gap flux,
FR rotor field MMF,
R mechaniical angle between rotor anid stator field lobes. I 2. Torque versuLs rotor angle relationship tor svncihronous
The air-gap flux OS stays constant as long as the voltage at
R
miachines in \teady state.
the iimachinie does not change, if the effects of saturationi of the
core iron are neglected. Therefore, if the field excitatio3n re- Of course, if tihe rotor anole overshoot does not transitorily
mains unchanged, a change in shlaft torque Twill cause a corre- exceed 180°. or if the disturbance causing the rotor swinlg is
spornding change in rotor angle a R. (This is the angle by which, promptly remiioved, the inacihine mlay remaini in synchronism
for a imiotor, the peaks of the rotating stator field lead the cor- witli the systemii. The rotor arngle theni oscillates in decreasing
responding peaks of the rotor field. For a generator. the rela- swings until it settles to its final value (less thani 900). The
tion is reversed). Fig. .2 graphically illustrates the variationl of oscillatiotns are damped by mechanical load and losses in the
rotor angale with shaft torque. With the machinie operating as a system, especially in the damper winidings of the machine.
mlotor (when rotor angle anid torque are positive), torque in- A change in the rotor angle of a machine (,enerally requiires
creases with rotor angle until reaches 90 electrical degrees.
SR a change in the speed of the rotor. For example, if we aSSum11e
Beyond 900, torque decreases with inicreasing rotor angle. As that the stator field trequenicy is conistant, it is necessary to at
a result, if we incr-ease the required torque outpuit a slny-
of least mcalomentarily slow dowtn the rotor of a synichronious
chronious motor beyond the level correspondinig to (O rotor motor to perm-nit the rotor tield to tall farther behind the stator
anigle, it will "slip a pole". Unless the load torque is reduced field anid thus inicreases aR. The rate at whichl the rotor speed
below the 900 level (the pullout torque), the motor will coni- cani change is determ'lined by the mIom1teInit of iInerttia of the
tinue slipping poles indefinitely. The problemiis that can follow rotor plus whatever is Imechalically COLIpled to it--prime
fromi extenided operationi in this out-of-step condition will be imover-, load, reductioni gears, etc. This imieans that a miachine
discussed further below. withl high inlertia will be less likely to pull out unider a disturb-
A getierator operates similarly. Increasing torque illpLtt Ulltil ance of brief durationi thlian a low-itnertia machiine, all other
the rotor angle exceeds 900 will result in pole slipping anld loss characteristics being equal.
(If synchronism with the power system. assuming constat
electrical load. Two-Machine Svsteeins
Similar relationis apply to the other parameters of the The previous discussioII of the tranlsienlt behavior oft svnI-
tor-que equation. For example, air-gap flux OSR is a fuLnctionl chrontous machines is based on a single machine conlnected to a
of the voltage at the machine. Thus if the other f-actors remiiain good approximiiationi of an infinite bus. An example is the typ-
coInstaInt, a chanige in system voltage will cause a chanige in ical industrial situationi where a synichlronious imiotor of at Imlost
rotor angle. Likewise, changing the field excitation will cause a a few thousand horsepower is connected to a ultilitV company
change in rotor angle. if constanit torque alnd voltage are systemn with a capacity of thouLsanids of megawatts. Ulnder
maintamred. these coniditionis we canl safely nieglect the eftfect of the ima-
All of the above discussion refers to rather gradual changes clhin e oni thie power systemii.
in the cohiditions affectin0c the torque angle, so that approxi- However, a systeml conIsistinlc, oinly of two imachines of coni-
mately steady-state conditionis always exist. However, the parable size coniniected throuLg a transmiiissioni link becomes
coupliing between the stator anid rotor fields of a snychronous riiore complicated, because the two macliiies can affect one
machine is somewhat elastic. This meanis that if an abrupt another's performance. The imedium through which this occuLrs
rather thani a gradual change occurs in onie miiore of the pa-
ot is the air-gap flux. This is a fuinctionI of miiachine terminial volt-
ramneters of the torque equation, the rotor angle will tetnd to age, which in tuLrn is affected by the chlaracteristics of the
overshoot the final value determiniedl by the changed conidi- trainsmiiission systei, the amounlt of power being transmlitted.
tions. This disturbanice may be severe enougl to carry the the power factor, etc.
ultinate steady-state rotor angle past 900, or the transietit hI the steady state, the rotor angles of the two mnachliines are
Iswiigi" rotor angle past 1 80. Eitlher event results in the deterinilied by the simultaneous sOlution of their- respective
slippinig of a pole. If the coniditions whlichi caused the original torqlue equationis. Under a tranisienlt disturbance, as in the one-
disturbance are not corrected, the machine will then continue machine system, the rotor angles move toward values corre-
to slip poles-in short, pull out of step with the power system sponiding to the changed system coinditions. Even if these new
to which it is connected. values are withinl the steadv-state stability limits of the systemii,

Authorized licensed use limited to: Centro Universitário Fei. Downloaded on March 08,2022 at 14:21:22 UTC from IEEE Xplore. Restrictions apply.
McFADDEN: INDUSTRIAL POWER SYSTEMS 109

an overshoot may result itn loss of synchronism. If not, both SYSTEM DISTURBANCES THAT CAN
rotors will undergo a damped oscillation and ultimately settle CAUSE INSTABILITY
to their new steady-state values. The most common types of disturbances teniding to pro-
Anl important concept here is synchronizing power. The duce instability in inidustrial power systems are (not neces-
more real power that can be transmitted over the transmission sarily in order of probability)
link between the two machinies, the more likely they are to
remaiin in synchronism in the face of a transient disturbance. 1) short circuits,
In fact, synchlroinous machinies separated by a sufficiently low 2) loss of a tie circuit to a public utility,
impedance behave in effect as one composite machine, since 3) loss of a portion of on-site generation,
they tend to remain in step with one another regardless of 4) starting a motor which is large relative to system gen-
external distuirbanices. erating capacity,
5) switching operations,
Multimtiachinze Systems 6) impact loading on motors,
7) abrupt decrease in electrical load on generators.
At first glance, it would appear that a power system incor-
porating many synchronous machines would be extremely The effect of each of these disturbances should be apparent
complex to analyze. This is true if we need a detailed, precise from the previous discussion of stability fundamentals. Items
analysis; a large digital computer and a sophisticated program 1) through 5) tend to reduce voltage levels, ultimately requir-
are required for a complete transient stability study of a multi- ing an increase in machine angles to maintain a given load.
machine system. However, many of the multimachine systems Items 6) and 7) directly increase the rotor angles of the af-
encountered in industrial practice contain only synchronous fected machines.
motors which are similar in characteristics, closely coupled
electrically, and connected to a high-capacity utility system.
Under most types of disturbance, the motors will remain syn- SOLUTIONS TO STABILITY PROBLEMS
chronous with each other, although they may all lose synchro-
nism with the utility. Thus the problem is similar to a single Generally, anything which decreases the severity or dura-
synchronous motor connected through an impedance to an tion of a transient disturbance will make the power system less
infiniite bus. The simplification should be apparent. More com- likely to become unstable under that disturbance. In addition,
plex systems, where the machines are of comparable sizes and increasing the moment of inertia per rated kVA of the sny-
are separated by substantial impedances, will usually involve a chronous machines in the system will raise stability limits by
full-scale computer stability study. resisting the changes in rotor speeds required to change rotor
angles.
System designi primarily affects the amount of synchroniz-
PROBLEMS CAUSED BY INSTABILITY ing power that can be transferred between maclhines. Two
machines connected by a low-impedance circuit such as a short
The most immediate hazards of asynchronous operation of cable or bus run will probably stay synchronized with each
a power system are the high transient mechanical torques and other under all conditions except a fault on the connecting
currents which usually occur. To prevent these transients from circuit, a loss of field excitation, or an overload. The greater
causing mechanical and thermal damage, synchronous motors the impedance between the machines, the less severe a disturb-
and generators are almost universally equipped with pullout ance will be required to drive them out of step. This meanis
protection. For motors of small and moderate sizes, this protec- that from the standpoint of maximunm stability all synchro-
tioin is usually provided by a damper protection or pullout nous machines should be closely connected to a common bus.
relay which operates on the low power factor occurring during Limitations on short circuit duties, economics, and the re-
asynchronous operation. The same function is usually pro- quirements of physical plant layout usually combine to render
vided for large motors, generators, and synchronous condens- this radical solution impractical.
ers by loss-of-field relaying. In any case, the pullout relay acts Design and selection of rotating equipnmenit can be a major
to trip the machine breaker or contactor. Whatever load is contributor to improving system stability. Most obviously, use
being served by the machine is naturally interrupted. Conse- of induction instead of synchronous motors eliminates the
quently, the primary disadvantage of a system which tends to potential stability problems associated with the latter. (Under
be unstable is the probability of frequent process interruptions. rare circumstances an induction motor/synchronous generator
Out-of-step operation also causes large oscillatory flows of real system can experience instability, in the sense that undamped
and reactive power over the circuits connecting the out-of-step rotor oscillations occur in both machines, but the possibility
machines. Impedance- or distance-type relaying protecting is too remote to be of serious concern). However, economic
these lines may falsely interpret the power surges as a line considerations often preclude this solution.
fault, tripping the line breakers and breaking up the system. Where synchronous machines are uised, stability can be en-
Although this is primarily a utility problem, large industrial hanced by increasing the inertia of the mechanical system.
systems or those where local generation operates in parallel Since the H constant (stored energy per rated kVA) is propor-
with the utility may be susceptible. tional to the square of the speed, fairly small increases in

Authorized licensed use limited to: Centro Universitário Fei. Downloaded on March 08,2022 at 14:21:22 UTC from IEEE Xplore. Restrictions apply.
llQ 110 tEF TIRANSACTIO)NS ON INI)IISTRY APPLICAT IONS. N1ARCH1/ ARIL 1977

synclhl-oInots speed canl pay siginificant dividends in higher 11ouLIS aii tlimetic capability. precisioni, anid ability to sto-re and
inertia. If carried too far, this mav becomlie self-defeating retrieve hugh amiOuIntlS of intforilationl m1lade it a niatLuiral for Sta-
because higher speed machines liave smlaller diameter r-otorS. bilitv stLidies. Transient stability programis were writteni anid
WK2 varies with the square of the rotor radius, so the increase used by miiajor electrical SupplieCS, ultility' comiipaniies, ConSul11t-
in H due to higher speed m1ay be offIset by a decrease due to inu firmiis, and Universities. and have nlow almost totally re-
the lower WK2 of a sinaller diameter rotor. placed the older- miiethods.
A furthier possibility is to LIse syn1chironious machines with
low transient reactances, thereby permiitting the maximum How Stabilits' Programns Work
flow of synchronizing power. The applicability of this solution Mathenmatical imetlhods of stability analysis depenid oni a
is limited mostly by short-circuit considerations and machinie repeated solutioni of the swing equation for each imachine,
design problems.
MVA (H) d26R
Si'stemn protection often offers the best prospects for im- Pa
proving the stability of a power system. The most severe dis- I80Qf dt2
turbance which an industrial power system is likely to experi- where
ence is a shiort circuit. To prevent loss of synchronism. as well
as to limit personnel hazards and equipmzent damage, short P0 accelerating power (input power minus outpuit
power), MW,
circuits should be isolated as rapidly as possible. A system
MVA rated MVA of machine,
which tends to be unstable should certainly be equipped with
H inertia constant of machine, MW-s/MVA,
instanitanieous overcurrent protectioni on1 all of its primar-y
feeders, which are the most exposed portion of the primary J, systemn f'requenicy, Hz,
6R rotor angle, degrees,
system. As a general rule, in fact, instantaneous relaying slhould
be used throughlout the systemii wherever- selectivity permits.
t tiime, S.
Voltage regulator and exciter characteristics affect stability The programl begins with the results of a load-flow study to es-
because, all other things being equial, higlier field excitatio-n tablish initial power and voltage levels ini all machines and
requlires a smaller rotor anigle. Consequently, stability is en- interconnecting circuits. The specified disturbance is applied at
hanced by a properly applied regulator and exciter which a time defined as zero, aned the resulting changes in power
respotid rapidly to transient effects and furnish a high deg,ree levels are calculated by a load flow routine. Using the calcu-
of field forcing. In this respect. miodern solid-state voltage lated accelerating power values, the swing equation is solved
regulators and static exciters can contribLite markedly to for a new value of 5R for each imachline at an increimental tiIme
improved stability. (On the other hand, a mismatch in exciter (e."., 0.01 s) after the disturbance. The voltage and power
and regulator characteristics cani make an existing stability levels corresponding to this tnew angular positioIn of the syn-
problem even worse.) clhreonous machines are then ulsed as base informationi for an-
other iteration. In this way, the performance of the system is
TRANSIENT STABILITY STUDIES calculated for every integration interval out to as much as
several minlulteS.
Knowing how to cure an unstable power system is not very
valuable if, in order to test our proposed reimedies, we have to Sim2ulationi of the System
stage stability tests on the actual system. This is especially true A moder-n tranisienit stability study comIpttter prograim can
if the system in question is still in the design stage. Conse- sitnulate virtually any set of power system components in suf-
quently, we need a fast, simple, and inexpensive way to Simlu- ficient detail to give accurate results. The simnulationi of rotat-
late the transient performance of a power systemii under a ing mnachililes anid related ecluipmenit is, of cour-se, of special
variety of normal and abnormal conditionis. importance in stability stuidies. The siniplest possible represen-
tation for a synichronouis milotoI- or generator involves only a
Histors' conistanit iJnterinal voltage, a constant tranisient reactance, and
The first transient stability studies responiding to this nieed the rotating inertia (H) conistant. This approximation neglects
were done by semieducated guess. When this technique proved saturation of the core iron, voltage regulator action, the
insufficiently accurate, and as rotating-machine theory was influenice of the conistructiotn of the machinie oni the transienlt
developed, simple power systems consisting of only a few ma- reactanices tor the direct anid quadrature axes, and most of the
chines were analyzed by matLial calculations or mechanlical characteristics of the prime milover or load. Nevertheless, this
models [3] . These inethods were neither precise nor applicable so-called "classical" representatiorn is often accurate enough to
to complex systems. In the thirties and forties, further devel- give reliable results, especially when the time period being
opment of modelinig techniques led to the use of the ac tnet- studied is rathier short. (Limiitingc, the study to a short period--
work analyzer in stability studies. This analog device permitted say, 1/2 secoind or less--meanis that neither the voltage regu-
the simulation of much larger systems than previously possi- lator nor the governor-, it anly, has time to exert a significanit
ble, but still suffered from the disadvantages of imprecise rep- effect). The classical represenitation is generally ulsed for the
resentation of synchroinous machines, proneniess to humlian smaller and less intluential machiines in a system. or where the
error, limited capacity, and highl cost per study runl. Finally in more detailed informiationn requtired for better simulations is
the fifties the digital comnputer caine uponI the scene. Its enor- iiot available.

Authorized licensed use limited to: Centro Universitário Fei. Downloaded on March 08,2022 at 14:21:22 UTC from IEEE Xplore. Restrictions apply.
McFADDEN: INDUSTRIAL POWER SYSTEMS III

As additional data on the machines become available, better B) For all significant transformers and autotransformers,
approximations can be used. This permits more accurate re- 1) kVA rating.*
sults which remain reliable out to longer time periods. Modern 2) Impedance.*
large-scale stability programs can simulate all of the following 3) Voltage ratio.*
characteristics of a rotating machine: 1) voltage regulator and 4) Winding connection.*
exciter, 2) steam system or other prime mover, including gov- 5) Available taps and tap in use.*
ernor, 3) mechanical load, 4) damper windings, 5) salient 6) For regulators and load tap-changing transformers:
poles, 6) saturation. Induction motors can also be simulated in regulation range, tap step size, type of tap changer
detail, together with the speed-torque characteristics of their control. *
connected loads. C) Short circuit capacity (steady-state basis) of utility sup-
In addition to rotating equipment, the stability program ply, if any.*
can include in its simulation practically any other major sys- D) kVAR of all significant capacitor banks.*
tem component, including transmission lines, transformers, E) Description of normal and alternate switching arrange-
capacitor banks, and even voltage regulating transformers and ments.*
dc transmission links in some cases. II) Load data: real and reactive electrical loads on all
significant load buses in the system.*
Simulation of Disturbances III) Rotating machine data.
The versatility of the modern stability study is also appar- A) For major synchronous machines (or groups of identical
ent in the range of system disturbances that can be repre- machines on a common bus).
sented. The most severe disturbance which can occur on a 1) Mechanical and/or electrical power ratings (kVA,
power system is usually a three-phase bolted short circuit. hp, kW, etc.).*
Consequently, this type of fault is most often used to test 2) Inertia constant H or inertia WK2 ol rotating ma-
system stability. Stability programs can simulate a three-phase chine and connected load or prime mover.*
fault at any location, with provisions for clearing the fault by 3) Speed.*
opening breakers either after a specified time delay, or by the 4) Real and reactive loading, if base-loaded generator.*
action of overcurrent, underfrequency, overpower, or imped- 5) Speed-torque curve or other description of load
ance relays. This feature permits the adequacy of proposed torque, if motor.*
protective relaying to be evaluated from the stability 6) Direct-axis subtransient,* transient,* and synchro-
standpoint. nous reactances.*
Short circuits other than the bolted three-phase fault cause 7) Quadrature-axis subtransient, transient,* and syn-
less disturbance to the power system. Although most stability chronous reactances.
programs cannot directly simulate line to line or ground faults, 8) Direct-axis and quadrature-axis subtransient and
the effects of these faults on synchronizing power flow can be transient* time constants.
duplicated by applying a three-phase fault with a properly 9) Saturation information.
chosen fault impedance. This means that the effects of any 10) Potier reactance.
type of fault on stability can be studied. 11) Damping data.
In addition to faults, the stability program can simulate 12) Excitation system type time constants, and limits.
switching of lines and generators. This is particularly valuable 13) Governor and steam system or other prime mover
in the load-shedding type of study, which will be covered type, time constants, and limits.
below. B) For minor synchronous machines (or groups of machines).
Finally, starting of large motors on relatively weak power 1) Mechanical and/or electrical power ratings.*
systems and impact loading of running machines can be 2) Inertia.*
analyzed. 3) Speed.*
4) Direct-axis synchronous reactance.*
Data Requirements for Stability Studies C) For major induction machines or groups of machines.
The data required to perform a transient stability study, 1) Mechanical and/or electrical power ratings.*
and the recommended format for organizing and presenting 2) Inertia.*
the information for most convenient use are covered in detail 3) Speed.*
in the application guides for particular stability programs. The 4) Positive-sequence equivalent circuit data (e.g., Rl.
following is a summary of the generic classes of data needed. X1, XM,R2, X2).*
Note, of course, that some of the more esoteric information 5) Load speed-torque curve.*
is not essential; omitting it merely limits the accuracy of the 6) Negative-sequence equivalent circuit data.
results, especially at times exceeding five times the duration 7) Description of reduced-voltage or other starting ar-
of the disturbance being studied. The more essential items are rangements, if used.*
marked by an asterisk (*). D) For minor induction machines: detailed dynamic repre-
I) System data. sentation not needed, represent as a static load.
A) Impedances (R + jX) of all significant transmission lines, IV) Disturbance data.
cables, reactors, and other series components.* A) General description of disturbance to be studied, includ-

Authorized licensed use limited to: Centro Universitário Fei. Downloaded on March 08,2022 at 14:21:22 UTC from IEEE Xplore. Restrictions apply.
112 112 VRANSA(7I IONS ON INDUIS F'RY AtPPLI \FIONS, MA4R(H/APRII.
~~~~~~~~~~~~IEFE 1977

I I I -~~~~~~~~~~~
IC
I 1 1 1 - I ~~~~~~~~~~~~~~~~~&Er.L ~~~~~~~~~-IOJ I CC~~~~~~~~~~
1. {. 1 (1 1 1~ ~ ~ ~ ~ ~ ~ ~ ENJ 0,
1 1II 1 ~~~~~C
1Ge~~~~~~~~~~~~~~~j. r~~~~~~Je24

I I 1 E1IiF IC

I I I ~~~~~~-I I~~I
C CI

w IEo I I ? II
-LJ I E ~C IIC -

BtIIriE'~~~I 1 I ~ I
l IiE E
EE E.I1 1E
0 I11 1 j~~~~~~~~~~r5 C I~~~~ C L-
I I I ~~~~~~~ 1b.~~~~~~~E En
II I ~~~~~~~~~~~~~~~~~~~~~~I£
IL~~~~~~~~~~~~~~~~~~~~~~~~~~~EF
I I 1 I I-Li:, IL~ ~~~~~~~~~~~~~~~~ I
- --- ----
.. -
--- . --- --- ..--- ---- --- ---

II ICr~~~~~~~~~~~tLL

-- - - - -- - - - -- - .-.------
. .4 ,~~~~~~~~~~~~~~~~~~~~4

TIME,t SECONDS
F'i L. 3. Compuiter printoLit of swing curves for Case I Pault oni systemi
in Fig. 4.

inig (as applicable) initial switchinig status, fault type, simiulationi. Normially a printing initerval of 0.01 or 0.021 s is

location and duiration; switching operations and timngn:- uised: loniger initervals reduice the c miputer costs slightly, but

m-anufactur-er, type, and settings of protective relays and increase the risk of missing fast swings. ot' rotor anigle. The

clearingp timne of associated breakers.* COmIpuIter timie cost is niearly pr-oportional to thie total study
B) Limiits on acceptable voltage, currenit, or power swinos.* time, so this para meter shiouild be closely contr-olled for the

V) Study parameters. sake of economiy.

A) Durationi of study.* This is especially imiportanit if -the systemi and miachinies

B) Integratino interval.* have beeni represen ted approximnately or incomipletely, becauise

C.) Outpuit printing interval.* the errors will accumiulate anid r-ender- the r-esults meianinguless
D) Data output required.* after somie poinit. A timie limiit of' five tim-ies the duiration of thle
iajo r diStuirbanice being stuidied is generally long enoughi to

Stability, Programn Outpuit show whlethier the systemi is stable or- not, while keep)ing costs

to a r-easoniable level.
Most stability programs give the user a wide chioice of re-

suits to be printed out. The program can calculate and print tinterpretinig Results--Swinig CurvTes
any of thie following iniformation funictioni of time:
as a
Thle reSuLlts of a comiputer
stability stuidy are fairly transient

easy to uiser learnis thie basic p)rinciples

understand onice the
rotor angles. torques, anld speeds of synichronious
underlying stability problemis. The miost direct way to deter-
mnac hines,
m-ine fromi thl Stuidy results whiethier a systemn is stable is to
2) real and reactive power tlows throuIghout the systemn,
3) voltages and voltage angles at all buses,
look at a set Of SWingl- CUrves, for the mnachinies in the systemi.
4) system- frequenicy,
Swing cur-ves are simplyI plots of rotor angles aoainist timie:, if

the curves of all the miachinies plotte-d

involved are on common
5) torques and slips of all induction machines.
axes, we cani easily see whethier thiey divergre (inidicating( inista-
The com-binationi of these r2-sults selected by the ulser can be bility) or settle to new steady-state valuies.
printed out for each printing interval (also user-selected) dur- For exam-ple.. Fig. 3 shiows swinay curves for a system whiich
ing, the course of the study period. is stable unrder- thie disturbance applied. This is a reproductioni
The value of the study is stronigly affected by thie selection of' an actual comipujter printout1. A simiplified siniale-linie dia-

of the proper printing interval and the total durationi of thie gram of thie system- appear-s as Fig. 4. Note that whlile the

Authorized licensed use limited to: Centro Universitário Fei. Downloaded on March 08,2022 at 14:21:22 UTC from IEEE Xplore. Restrictions apply.
 McFADDEN: LNI)NJSTRIAL POWER SYSTEMS 113

U -IIL i-Tr 'S,T.rf

E. F- ',+ '

Mr/,k

Fig. 4. Single-line diagram of system whose swing curves appear in

Figs. 3 and 5.

37 * . -------0-------C-C-C-C-
-* . .--------.----------.---.--. ---------

I I D U) D I IC I I I cc
I 0 I I C II U C I I I I I C CI
I ILU IEEC I £r1 IC I I I C
I DC0 EE EF C EI I J I IC I
I uI E I C A E E A A 0 A A A C I I I I I I
Du I
I E CA A A I E. Uui A 1 I I C DO
00 ) I L B b, Ct jB 8 8B r E I C AA I I C 0 00
I EE btA C I I E UOIC .1 I I u )DO L
C CC C ECCCC I I ' *)UU
U 1 0) 0 D I BEE
16 b-B-E-E-AA *---------.----------
. -------
EE I I I I E b1 I E 1 I 8 E E E I
I II It E dB~;~E
. i c EEEE E I
I I I I I I EL F.. E E E E E E E E C IT
I I Ic A c I r I
IxI I I I I IC I A CI I I
LLI I I I I 1 I I C C C I I
LUi
1 1 1 I i I I COC C A I I I
I At I G.A. lO.II I 1* 1 1I A I I I
3
II g IA
1 A I I

0 I C J.G3
IG I 1 I 1 CI A
a
LUJ I t) I Sh *H 1 1 1 1 1 I
0 I (F. I GES. SJ 5 I Ii I I I A I
I I I II I I£ II ~ ~ ~I ~ I
~~~ ~ IA~ ~~T
Co I £ I I I 1 I A A
I I.a
II II1 1 1 1 I AlII
0 I I I IiI I 1 1 I 1
0
.I I I i'1 T IT A A
£31, 24 . _______ !..------..+----
. -- 4.-----+-------+ -~-+_--__-_____ ..
...........................____
1 IX i E r r IX IA
$ I I 1 1 1 1 I I I
+ II I 1 1 1 I I 1 A
tzn 0 1 I I i 1 1 I I II
U I I I I I1I I A 1t

z-*~
l | I 1 1 1 IT T I I
|- r i I I --
I
-A
________~~~ I__
~45
I__ _I__ __I__
'') a .) -1 *i ;'? ) . , a.41j0; , "I ,7 ,R n .9 1 *s 0
TIME, SECONDS

Fig. 5. Computer printout of swing curves for Case 11 fault on system

shown in Fig. 4.

three-phase bolted fault on a synchronizing bus feeder (Case I), tween the synchronizing bus and one of the generator buses
cleared by instantaneous tripping of the feeder breaker, causes and is cleared by tripping the tie circuit breaker. The swing
all five generators to experience swings of varying magnitude, curves for this condition are shown in Fig. 5. Generator No. 1
the oscillations in the rotor angles are obviously damped and is disconnected from the system and suffers a severe overload,
can be expected to die out. causing it to decelerate, as shown by a unidirectional negative
By contrast, in Case II the fault is applied to the tie be- change in rotor angle. The other machines stay in synchronism.

Authorized licensed use limited to: Centro Universitário Fei. Downloaded on March 08,2022 at 14:21:22 UTC from IEEE Xplore. Restrictions apply.
114 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, MARCH/APRIL 1977

Fig. 6. Single-line diagram of a typical large industrial power system

with on-site generation.

STABILITY STUDIES ON A TYPICAL SYSTEM 138-kV utility line ahead of the plant transformers, on a
Probably the best way to examine some of the typical ap- feeder from each of the 13.8-kV buses, and on a feeder from
plications of stability analysis to industrial power systems is to each 4160-V bus. Of course, the simulation would include
look at the stability studies which would go into the design of clearing of the fault via the proposed relaying. If any of these
a' typical large industrial system including 20 MVA of local studies show an unstable condition, further stability studies
generation and 40 MVA of purchased power capacity. The might be required to test the effectiveness of various proposed
stability studies which might be applied to this system are dis- solutions.
cussed below. 3) In the system shown in Fig. 6, some considerations
1) The basic layout of the primary system can be affected should certainly be given to automatic load shedding. If the
by stability considerations. For example, an initial design power company suffers an outage on the 138-kV lines while
choice might be to connect the generated and purchased the plant is running at nearly full load, the 20 MVA of localr
power buses through only one tie circuit. However, stability generation will abruptly be subjected to an overload approach-
studies could show that inadequate synchronizing power is ing 300 percent. This overload will promptly cause the gener-
available to prevent the generators and the utility from losing ators to trip off, leaving the plant with no power at all, even
synchronism during primary system faults unless two ties are though 20 MVA of perfectly sound generation is available to
provided. The same sort of considerations might dictate that maintain service to the most critical loads. Obviously, a
the 4160V bus ties be operated closed, to provide the lowest method of automatically interrupting noncritical loads com-
possible impedance between the synchronous motors and the mensurately with the loss of system capacity would be
power sources to enhance stability. valuable.
2) Related to the design of the basic layout is the problem One such possibility would be to trip noncritical feeders
of protective relaying. The system can be designed for maxi- whenever the utility tie breakers are opened. However, this
mum inherent stability by closely coupling all of the machines. wired-in scheme lacks flexibility. To permit shedding only the
Or the same objective can often be obtained by designing the amount of load required to prevent system collapse, many
protective scheme for the fastest possible clearing of faults. industrial plants with local generation use underfrequency
Since the former choice may involve economic sacrifices in the relaying. This scheme depends on the fact that an overloaded
form of higher capacity switchgear, often the latter choice generator slows down, dropping the system frequency.
represents the best solution. However, extra-fast relaying can A two-stage load shedding scheme might operate as follows.
conflict with the requirements of a selectively coordinated The first-stage relay operates at 59 Hz and a time delay of 6
system, unless expensive zone protection schemes (bus differ- cycles, tripping 10 or 15 MVA of noncritical load; the second
ential, pilot wire, etc.) are used. Balancing all of these factors, stage operates at 58 Hz and a delay of 30 cycles, tripping an
probably the best procedure is to design the system layout additional 20 MVA of somewhat more critical load.
around process requirements, provide the fastest relaying pos- In designing the load-shedding scheme for the system in
sible within the constraints of selectivity and economics, and Fig. 6, we would first run a stability study to calculate the
then check the proposed layout and relaying by a series of sta- decay of system frequency when the utility tie is lost, without
bility studies simulating the more probable fault conditions. In any load shedding. Using this frequency decrement curve, we
the Fig. 6 system, three-phase faults might be applied on one can estimate the amounts of load to be shed and the frequency

Authorized licensed use limited to: Centro Universitário Fei. Downloaded on March 08,2022 at 14:21:22 UTC from IEEE Xplore. Restrictions apply.
IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. IA-13, NO. 2, MARCH/APRIL 1977 115

and time delay settings for the underfrequency relays. Then REFERENCES
these data can be used in the stability study program to calcu- [1] Electrical Transmission and Distribution Reference Book, Westing-
late the system frequency versus time curve with the proposed house Electric Corporation, East Pittsburgh, PA., 1964, chap. 13.
load shedding. If sufficient load is shed fast enough to prevent (2] E. W. Kimbark, Power System Stability, vol. 1. New York:
Wiley, 1948.
system collapse, the validity of the proposed relay scheme and [3] A. E. Fitzgerald and Charles Kingsley, Jr. Electric Machinery.
settings is confirmed. Usually several runs are made with dif- New York: McGraw-Hill, 1961, chap. 5.
ferent system conditions in each load shedding analysis.
4) In the system shown in Fig. 6, the effect of starting one
of the large synchronous motors could be substantial, espe-
cially under abnormal conditions when one or more power Richard H. McFadden (S'67-M'67) was born in
sources are out of service. This effect can be evaluated by a Philadelphia,
PA in 1946. He graduated from
Lehigh University, Bethlehem, PA, in 1967
series of stability studies simulating motor starting under vari- with the B.S.E.E. degree, and received the J.D.
ous conditions of system capacity, prestart motor terminal degree from Duquesne University, Pittsburgh,
PA, in 1976.
voltage, etc. The study results yield motor accelerating times, After briefly working as an electrical engi-
real and reactive power flows, and bus voltages at all critical neer for Sun Pipe Line Co., he joined Westing-
house Electric Corp. in 1967 as a District
points in the system. Engineer. Since then, he has held various appli-
As the above discussion should indicate, transient stability cation engineering positions in Westinghouse,
analysis should be an integral part of the design or expansion and is now Project Manager in charge of engineering support for major
of any industrial power system containing several large syn- industrial projects in the Eastern U.S.
Mr. McFadden is a Registered Professional Engineer in Pennsylvania,
chronous machines, and should be considered even if only one a member of NSPE and ABA, and Vice Chairman of the Power Systems
machine is being applied. Support Committee of IEEE IAS Industrial Power Systems Department.

Cable Tray Installation for Process Plants

RALPH D. CHAMLEE, SENIOR MEMBER, IEEE, AND GEORGE W. SWANSON

Abstract-The function of a tray system in a process plant is to pro- turers and designers alike are innovating to supply solutions to
vide an efficient wiring method for power, lighting, instrumentation, problems as they arise.
communications, and grounding. This paper reviews tray system design, The results are gratifying. The components go together
material selection, installation, and problem areas. The presentation is
based on experience in the design and installation of complete tray relatively fast and make a flexible, accessible system, readily
systems for oil refineries, chemical plants, and power plants. repaired, changed, or expanded. With proper selection of
materials, the components should be expected to have a
INTRODUCTION relatively long life.
The routing of cables is generally less rigorous than the
E LECTRICAL installations in outdoor process plants have routing of conduits and therefore requires less engineering
Echanged appearance in recent years. More and more time to develop layouts. It appears that a given circuit-foot
projects are being designed and constructed using cables of cable in tray requires less skill and possibly less cost to
supported by a system of trays for the various electrical install in comparison to an equivalent circuit-foot of wire in
systems rather than using the more conventional installation conduit.
consisting solely of wire in rigid conduit.
Code changes in both the 1971 and 1975 National Electrical FUNCTION
Code (NEC) have done much to pave the way for the increased
use of cable tray systems. As an industry we are feeling our The function of a cable tray system in a process plant is to
way along, developing designs and installation methods to provide a complete and efficient installation of electrical
capitalize on the benefits of the cable tray system. Manufac- conductors to serve the requirements for power, lighting,
grounding, communications, and instrumentation. Since the
Paper IOD-75-42, approved by the Petroleum and Chemical Indus- cables are not self-supporting, in the sense that a conduit is,
try Committee of the IEEE Industry Applications Society for presen- a complete tray system is used to give this support.
tation at the 1975 IEEE Petroleum and Chemical Industry Technical Continuous runs of trays supported at regular intervals on
Conference, Milwaukee, WI, September 15-17. Manuscript released for
publication January 9, 1976. structural supports, which are provided basically for the piping
The authors are with C F Braun & Co., Alhambra, CA. system, make up the basic raceway system. By its nature, the
Authorized licensed use limited to: Centro Universitário Fei. Downloaded on March 08,2022 at 14:21:22 UTC from IEEE Xplore. Restrictions apply.