CHAPTER 6

MACHINE CHARACTERISTICS
Original Author: Revised by:
C. F. Wagner C. F. Wagner

EFORE, the growth of the public utilities into their

B present enormous proportions with large generating

stations and connecting tie lines, machine perform-
ance was largely judged in terms of the steady-state char-
acteristics. The emergence of the stability problem gave
rise to the analysis of the transient characteristics of
machines and was largely responsible for our present
knowledge of machine theory. A further contributing urge
was the need for more accurate determination of short-
circuit currents for the application of relays and circuit
breakers.
The variable character of the air gap of the conventional
salient-pole synchronous generator, motor, and condenser
with its concentrated field windings requires that their
analysis follow a different line from that for machines such
as induction ‘motors, which have a uniform air gap and
distributed windings. Blonde1 originally attacked this
problem by resolving the armature mmf's and fluxes into
two components, one in line with the axis of the poles and Fig. 2—Cut-away view of conventional waterwheel generator.
the other in quadrature thereto. When the study of the
transients associated with system stability was undertaken this conception was quickly recognized as an invaluable
too11,2. Since that time the method has been extended by
subsequent investigators,3-g notably Doherty and Sickle,
who introduced into the industry several new constants,
such as transient reactance and subtransient reactance to
describe machine performance under transient conditions.
This chapter treats of the characteristics of synchronous
and induction machines in the light of the development of
the past twenty-five years. It will consider steady-state
and transient conditions for both salient pole and cylin-
drical rotor machines under both balanced and unbalanced
conditions. There follows a discussion of the character-
istics of induction motors under such transient conditions
as might contribute to the short-circuit current of a system
and might influence the choice of a circuit breaker.

I. STEADY-STATE CHARACTERISTICS OF
SYNCHRONOUS MACHINES
The two general types of synchronous machines are the
cylindrical rotor machine or turbine generator which has
an essentially uniform air gap and the salient-pole gener-
ator. Figs. 1 to 5 illustrate the outward appearances and
cross-sectional views of typical modern machines.
Typical saturation curves for a hydrogen-cooled turbine
generator, a waterwheel-generator and a synchronous con-
denser are shown in Figs. 6, 7, and 8 respectively.
Because of the necessity of matching the speed of water-
Fig. l—Cut-away view of umbrella-type waterwheel wheel-generators to the requirements of the waterwheels
generator. it is difficult to standardize units of this type. However,
145
146 Machine Characteristics Chapter 6

Fig. 5—Hydrogen-cooled frequency changer set installed on

the system of the City of Los Angeles, 60 000 kva; 600 rpm;
50 cycle-11 500 volts; 60 cycles-13 200 volts.

rent. Depending upon the application, either the field

current for rated voltage in the air gap or the actual field
current for rated voltage, including saturation, is used.

1. Unsaturated Cylindrical-Rotor Machine Under

Steady-State Conditions
Fig. 3—Steam turbine generator installed at the Acme Sta- The vector diagram of Fig. 9 is the well-known diagram
tion of the Toledo Edison Company, 90 000 kw, 85-percent
power factor, 85-percent SCR., 13 800 volt, a-phase, 60-cycle.
of a cylindrical-rotor machine. Consistent with the policy
of this book, familiarity with this diagram is assumed. Let
great strides have been made with large 3600-rpm con- it suffice merely to indicate the significance of the quan-
densing steam turbine-gencrators. These find their great- tities. The vectors et and i represent the terminal voltage
est application in the electric utility industry. Table 1 of to neutral and armature current, respectively. Upon add-
Chap. 1 gives some of the specifications20 for these ing the armature resistance drop, ri, and armature leakage
machines. reactance drop, xii to etr the vector el is obtained, which
The concept of per-unit quantity is valuable in compar- represents the voltage developed by the air-gap flux <p,
ing the characteristics of machines of different capacities which leads el by 90 degrees. This flux represents the net
and voltages. However, care must be exercised in the case flus in the air gap. To produce this flux a field current, It,
of generators to use the same reference value for field cur- is required. The current I, can be taken from the no-load

Fig. 4—Cut-away view of hydro-

gen-cooled turbine generator.
Chapter 6 Machine Characteristics 147

Fig. 6—Saturation curves for typical hydrogen-cooled turbine Fig. 8—Saturation curves for typical hydrogen-cooled con-
generator. denser.

field current If; it is the voltage taken from the air-gap line
of the no-load saturation curve for the abscissa corre-
sponding to It. The side ,-IB of the triangle, since it is pro-
portional to -1 i and consequently proportional to the arma-
ture current, can be viewed as a fictitious reactance drop.
It is called the drop of armature reactance and is desig-
nated z,i. The reactance drops .rli and .w’ can be com-
bined into a single term called the synchronous reactance
drop and there results
xd =x1+% (1)

It follows from the foregoing that the internal voltage. ei,

is equal to the vector sum of et, ri and j xdi. The field
current, If, can be determined for any condition of loading
(neglecting saturation, of course) by merely calculating
ei and taking It from the air-gap line of Fig. 10.
At no load the axis of the field winding, the line OC, leads
the terminal voltage by 90 degrees. At zero power-factor,
the vector diagram reduces to that shown in Fig. 11, which
shows that, except for the effect of the resistance drop, the
foregoing statement would still be true. As ri is only about
Fig. 7—Saturation curves for typical waterwheel generator.
one or two percent in practical machines, the statement

saturation curve of Fig. 10 as being the current required

to produce el. But, the armature current produces an mmf
by its so-called armature reaction, which is in time phase
with it and in terms of the field can be expressed as Ai. To
produce the net mmf represented by the current, It, the
field current must be of such magnitude and the field
structure must adjust itself to such position as to equal If.
In other words, If has now such position and magnitude
that If and Ai added in vectorial sense equals I,. The
triangle O.JB, formed by drawing AB perpendicular to i or
Ai and OB perpendicular to OC, is similar to the triangle
ODC; OB has the same proportionality to OC and -4B to
AZ’ as el has to I,. Neglecting saturation, OB, designated
.
as ei, is thus the open-circuit voltage corresponding to the Fig. 9—Vector diagram of cylindrical-rotor machine.
 Machine Characteristics Chapter 6

done, it can be seen that x,i can likewise be thought of as

arising from the two components of i in the form of x,&i
and xaqiq, respectively, in leading quadrature to id and i,.
In the case of a cylindrical rotor machine, &d and x,, are
both equal to xa but a case will soon be developed for which
they are not equal.
The synchronous reactance, xd, can be obtained most
conveniently from the no-load curve and the full-load zero
power-factor curve. In Fig. 10 OA is the field current
required to circulate full-load current under short-circuit
conditions, the terminal voltage being zero. In this case
all of the internal voltage (the ri drop can be neglected
justifiably) must be consumed as synchronous reactance
drop (%-Ji)within the machine. If there were no saturation,
the internal voltage can be determined by simply reading
the terminal voltage when the short-circuit is removed,
maintaining the field current constant meanwhile. This
voltage would in Fig. 10 be equal to AB. Thus the un-
saturated synchronous reactance per phase is equal to the
phase-to-neutral voltage AB divided by the rated current.
When the saturation curve is expressed in per unit or per-
cent it is equal to AB; but where expressed in generator-
I .
terminal voltage and field amperes, it is equal to F(lOO)
fg
10—No-load and full-load zero power-factor character- Ifsi
istics of a generator. in percent or - in per unit.
IfI3
2. Unsaturated Salient-Pole Machine Under Steady-
State Conditions
Given the proper constants, the performance of an un-
saturated salient-pole machine at zero power-factor is the
same as for a uniform air-gap machine. For other power-
factors, conditions are different. A vector diagram for such
machines is shown in Fig. 12. As before et and i are the

Fig. 11—Vet tor diagram of cylindrical-rotor generator at zero

power-factor. Fig. 12—Vector diagram of salient-pole machine.

can be accepted as true for all practical purposes. How- terminal voltage to neutral and the armature current,
ever, as the real load is applied to the machine the angle 6 respectively, and ei is the “voltage behind the leakage
increases from zero and the lead of OC ahead of et increases reactance drop.” The flux + is required to produce ei. This
from 90 degrees to 90 degrees plus 6. The angle 6 is a real flux can be resolved into two components +d and Cp,. The
angle; it can be measured without much difficulty. flux <Pd is produced by If and AZ& the direct-axis com-
It is convenient for some purposes to resolve the reac- ponent of Ai, and +, is produced by Ai,, the quadrature-
tions within the machine into two components, one along axis component of Ai. Here the similarity ceases. Because
the axis of the field winding and the other in quadrature of the saliency effect, the proportionality between the
thereto. In Fig. 9, the armature current is divided into mmf’s and their resultant fluxes is not the same in the two
the two components, id, and, i,, in which the subscripts are axes. When saturation effects are neglected ad can be re-
significant of their respective components. When this is garded as made up of a component produced by If acting
Chapter 6 Machine Characteristics 149

the projection of BF upon OG is equal to x&d so that OG

becomes equal to ed, the fictitious internal voltage, which
is proportional to If.
The armature resistance is usually negligible in deter-
mining either the angle 6 or the excitation and for this case
et sin 6 = x,i, = x,i cod++@ (2)
Upon expanding the last term and solving for 6
x,i cos 4
tan 6= (3)
Fig. G--Flux resulting from a sinusoidal mmf in e,+x,i sin $
(a) direct axis, From Fig. 14, the internal voltage
(b) quadrature axis.
ed=et cos 8+x& sin (4-i-6) (4)

alone and a component produced by Ibid. The component The unsaturated synchronous reactance, xd, can be de-
produced by 1f can be regarded as producing the internal termined from the no-load and full-load zero power-factor
voltage ed. The mmf produced by -Aid has a general sinu- curves just as for the machine with uniform air gap. The
soidal distribution in the direct axis as shown by Fig. 13(a). quadrature-axis synchronous reactance is not obtained so
The resultant flux because of the variable reluctance of the
air gap has the general shape indicated. It is the sinusoidal
component of this flux that is effective in producing the
L&did drop shown in Fig. 12. In the quadrature axis, the
component of mmf is likewise sinusoidal in nature as shown
in Fig. 13b, and gives rise to the distorted flux form. In
proportion to the mmf the sinusoidal component of flux is
much less than for the direct axis. The effect of this com-
ponent is reflected in the x aq i, drop of Fig. 12. In general
X aq is much smaller than &,d.
The armature resistance and leakage reactance drops
can also be resolved into its two components in the two
axes much as x,$ of Fig. 9 was resolved. When this is done
the internal voltage ed can be obtained by merely adding
rid and ri, and then j x, i, and j xd id to the terminal voltage
et. The notation ed is used to differentiate the internal
voltage in this development from that used with the
cylindrical rotor machine theory.
Another form of the vector diagram of the machine is
presented in Fig. 14, which shows much better the relation
Fig. 15—Determination of internal angle, 6, and excitation
between those quantities that are most useful for calcula-
of a saturated salient pole machine when loading is known.
tion purposes.. If from B the line BP of length x,i is drawn
perpendicular to i, then since angle CBP is equal to $+S,
easily but fortunately there is not as much need for this
the distance BC is equal to x$ cos (++6), or x&. By
quantity. It can be determined from a test involving the
comparing this line with the corresponding line in Fig. 12,
determination of the angular displacement of the rotor as
it can be seen that the point P determines the angle 6. This
real load is applied to the machine and the use of Eq. (2),
relation provides an easy construction for the determina-
which gives
tion of the angle 6 having given the terminal voltage, the
et sin 6
armature current, and the power-factor angle, 6. Further, (5)
xqzi cos (++6)
or it can be determined by means of a slip test. The slip
test is described in the A.I.E.E. Test Code for Synchronous
Machines10 of 1943 for a determination of xd. The test for
the determination of x, is identical except that the mini-
mum ratio of armature voltage to armature current is used.

3. Saturation in Steady-State Conditions

Short-circuit ratio is a term used to give a measure of the
relative strengths of the field and armature ampere turns.
It is defined as the ratio of the field current required to
produce rated armature voltage at no load to the field
Fig.14—Determination of internal angle, 6, and excitation of current required to circulate rated armature current with
an unsaturated salient-pole machine when loading is known. the armature short-circuited. In Fig. 10 the SCR is equal
150 Machine Characteristics Chapter 6

I becomes determinable from Eq. (1) or from the triangle

to F. When no saturation is present it is simply the just discussed.
fsi

reciprocal of the synchronous impedance, xd. The foregoing analysis is not strictly correct, as it neg-
It is impossible to specify the best specific SCR for a lects certain changes in saturation in the pole structure.
given system. In the past it has been the practice in Eu- The leakage from pole to pole varies approximately pro-
rope to use somewhat smaller SCR’s than was the practice portional to the field current and the point T was deter-
in this country. In recent years, however, the trend in this mined upon the basis that this leakage was proportional
country has been toward smaller values. The Preferred to the field current MS. The increased field leakage at the
Standards for Large 3600-rpm Condensing Steam Turbine- higher excitation produces greater saturation in the field
Generators20 specifies SCR of 0.8. poles and this in turn increases the mmf required to force
The desire for smaller SCR’s springs from the fact that the flux through the pole. The net effect is to increase the
the cost is smaller with smaller SCR. On the other hand, field current over that determined by the method just
static stability is not as good with smaller SCR. Regula- discussed causing the two curves to separate more at the
tion is also worse but both of these effects are alleviated in higher voltages.
part by automatic voltage regulators. For most economical The concept of the determination of the curve of rated
design a high SCR machine usually has a lower xd’. There- current at zero-power-factor by the method just described
fore, both because of its lower xd’ and higher WR2 a high is valuable and in an attempt to retain the advantages of
SCR has a higher transient stability. This is not usually this method the concept of Potier reactance, xr,, is intro-
a significant factor particularly in condensing turbine ap- duced. The Potier reactance is the reactance that, used in
plications, because transient stability is not of great im- a triangle of the general type described, will just fit be-
portance in the systems in which they are installed. It tween the two curves at rated voltage. It can be deter-
may be quite important for hydro-generators; the Boulder mined from test curves, see Fig. 16, by drawing DE equal
Dam machines, for example, are designed for SCR’s of
2.4 and 2.74.
The effects of saturation arise primarily in the deter-
mination of regulation. Tests indicate that for practical
purposes both the cylindrical rotor and the salient-pole
machine can be treated similarly. Consideration will be
given first to the characteristics for zero-power-factor load-
ing. Fig. 11 shows that for zero power-factor, the ri drop
of the machine is in quadrature to the terminal voltage
and internal drop and can have little effect upon regulation.
It will therefore be neglected entirely.
The determination of the rated-current xero-power-
factor curve can be developed as follows. Take any ter-
minal voltage such as MN of Fig. 10. The voltage behind
leakage reactance is obtained by adding to this voltage the
leakage reactance drop, SR, which gives the line PQ. The
distance PR then gives the field current necessary for
magnetizing purposes. In addition, however, field current
is required to overcome the demagnetizing effect of the
armature current. This mmf is represented on the curve
by the distance ST, giving MT as the field current required
to produce the terminal voltage OM with rated current
in the armature. Other points on the rated-current
zero-power-factor curve can be obtained by merely Fig. 16—Zero power-factor characteristics of generator.
moving the triangle RST along the no-load saturation
curve. to OA and then EF parallel to OB. The distance FG is
Upon sliding the triangle RST down to the base line, it then the Potier reactance drop. Potier reactance is thus a
can be seen that the total field current required to circulate fictitious reactance that gives accurate results for only one
rated current at short circuit which is represented by the point, the point for which it is determined. For most ma-
point A, can be resolved into the current OD necessary to chines it is sufficiently accurate to use the one value ob-
overcome leakage reactance drop and the current DA re- tained at rated voltage and rated current. Potier reactance
quired to overcome demagnetizing effects. Neither leakage decreases with increased saturation. Sterling Beckwith19
reactance nor the field equivalent of armature current are proposed several approximations of Potier reactance, the
definite quantities in the sense that they can be measured two simplest are :
separately. They may be calculated but their values are
5, = 21+0.63tXd’-X1)
dependent upon the assumptions made for the calculations.
Synchronous reactance, xd, is a definite quantity and is and
equal to the distance A0 expressed in either per unit or 2, = 0.8 xd’.
percent. When either x1 or xa is assumed, then the other For other loads at zero-power-factor, the conventional
Chapter 6 Machine Characteristics 151

method is to divide the lines BA and FD of Fig. 16 in

proportion to the armature current. Thus for three-fourths
rated current the regulation curve would be the line HJ in
which BH and FJ are three-fourths of BA and FD,
respectively.
For power-factors other than zero, several methods are
available to determine the regulation. They all give sur-
prisingly close results, particularly at lagging power-
factors. The problem may take either of two forms; the
determination of the terminal voltage when the load cur-
rent, load power-factor, and excitation are given, or the
determination of the excitation when the load current, load
power-factor, and terminal voltage are given. The resist-
ance drop is so small that it is usually neglected.
(a) Adjusted Synchronous Reactance Method*-
This method utilizes the no-load and the rated-current
zero-power-factor curves. To obtain the excitation at any
other power-factor for rated current, an arbitrary excita-
tion is chosen such as OC of Fig. 17. The no-load voltage

Fig. 18—Determination of excitation, including the effects of

saturation.

the point F. The vertical distance OF is then the terminal

voltage for the particular excitation. Following this pro-
cedure another excitation is chosen and the
construction
repeated from which the clotted line is obtained. The inter-
section of the line with the normal voltage gives the excita-
tion for the desired power-factor at rated load. If the
machine is not operating at rated current. the zero-power-
factor curve corresponding to the particular current should
be used.
(15) General Method-For lack of a better name this
method has been called the “General Method.” It is based
upon the assumption that saturation is included by reading
the excitation requirements from the no-load saturation
curve for a voltage equal to the voltage behind the Potier
reactance drop.
The method is described in Fig. 19 with all terms ex-
pressed in per unit. The voltage, ep, is the Potier internal
voltage or the voltage behind the Potier reactance drop.

Fig. 17—Determination of regulation curves for power-factors

other than zero by the “adjusted synchronous reactance
method.”

CA is then regarded as an internal voltage and the distance

AB as an internal drop of pure reactance, which is laid off
in proper relation with the terminal voltage as indicated
by the power-factor of the load. The construction is as
follows: The adjusted synchronous reactance drop AB is
laid off to make an angle with the X-axis equal to the
power-factor angle. A line equal to the distance AC is then Fig. 19—Determination of field current for round rotor
scribed from the point A until it intercepts the Y-axis at machine with saturation included by adding a in phase
*Described as Method (c) Pars. 1.540 in Reference 10. with ep.
152 Machine Characteristics Chapter 6

The distance jk represents the synchronous reactance drop,

x&. If there were no saturation the synchronous internal
voltage would be Ok-. When using per unit quantities
throughout this is also equivalent to the field current.
This method includes the effect of saturation by simply
adding s1 the increment in field current for this voltage in
excess of that required for no saturation, to Ok in phase
with eP, giving as a result, Oh. When per unit quantities
are not used the construction is a little more complicated.
It involves the construction of eP separately SO that s can
be obtained in terms of field current. This quantity is then
added to the diagram for no saturation in terms of the field
current. In Fig. 18, first lay off from the terminal voltage,
Oa, and then the x, drop aC at an angle with the horizontal
equal to the power-factor angle. Oc then represents eP. By
scribing this back to the ordinate and reading horizontally,
the excitation corresponding to this voltage is obtained.
The effect of saturation is introduced by the distance s.
The field current required if there were no saturation is
obtained by the construction Oj and jk where Oj represents
the excitation, up, required to produce the terminal voltage
at no load and jk the excitation, ot, for the synchronous
reactance drop, read from the abscissa. These vectors
correspond to et and jk, respectively, in Fig. 19 except
that they are in terms of field current. If kh, equal to the
saturation factor, s, is added along a line parallel to OC,
the total excitation Oh is obtained.
(c) Round Rotor Potier Voltage Method*-This
Fig. 20—Reactive power capacity of steam turbine generator
method is the same as (b) except that the effect of satura-
20 000 kw, 23 529 kva, 0.85 p.f., 0.8 SCR, at 0.5 psig hydrogen.
tion s, in Fig. 18 is, for the sake of simplicity laid off along
Ok, making om the desired excitation. As can be seen,
sending voltage held at a constant value ed. As shown in
there is little difference between those two methods. This
Chaps. 9 and 10 the power circle of a line of such character-
method gives the best overall results, especially at leading
power factors. The particular name of this method was istics has its center in the negative reactive axis at z and
assigned to distinguish it from the next method.
edet
(d) Two-Reaction Potier Voltage Method-This its internal voltage, ed, must be such that its radius,
-,
method is similar to that of (c), except that the two-reac- xd
tion method of construction shown in Fig. 14 is used to passes through the point of rated real power and rated
determine the excitation before including the saturation reactive power. Actually, however, the center is usually
factor s. Fig. 15 shows the entire construction, For the located at a point equal to (SCR) times (rated kva). This
sake of comparison with other methods, the construction is to take care of saturation effects. Since, however, with
is also shown in Fig. 18. The construction is the same as 1
no saturation - is equal to SCR, it can be seen that for
(c) except that the line Oq is made to pass through the xd
point g instead of k. This arises because x, is smaller than xd. this condition both relations reduce to an equivalence.
The leading kvar capacity (underexcited) of air-cooled
4. Reactive Power Capacity condensers is usually about 50 percent of the lagging kvar
The capacity of a synchronous machine to deliver re- capacity but for hydrogen-cooled condensers about 42
active power is dependent upon the real power that it percent.
delivers. Two limitations from the heating standpoint are
recognized: (1) that due to the armature, and (2) that due II. THREE-PHASE SHORT CIRCUIT
to the field. Figure 20 shows the reactive power capability In addition to its steady-state performance, the action
of a standardized 3600-rpm steam turbine-generator. Real of a machine under short-circuit conditions is important.
power is plotted as abscissa and reactive power as ordinate. The presence of paths for flow of eddy currents as provided
All the curves are arcs of circles. The line centering about by the solid core in turbine generators and by the damper
the origin represents the limit imposed by the condition of windings in some salient-pole machines makes the treat-
constant armature current whereas the other arc by con- ment of these machines, from a practical viewpoint, less
stant field current. With regard to the latter, the generator complicated than that for salient-pole machines without
can be likened to a simple transmission line of pure react- damper windings. For this reason the three-phase short-
ance, zd, with the receiver voltage held at a constant value, circuit of these types of machines will be discussed first.
et, the terminal voltage of the generator, and with the Armature resistance will be neglected except as it influences
*Described as Method (a) Para. 1.520 in Reference 10. decrement factors.
Chapter 6 Machine Characteristics 153

5. Three-phase Short-Circuit of Machines with

Current Paths in Field Structures
Consideration will be given to a simultaneous short-
circuit on all phases while the machine is operating at
no-load normal voltage without a voltage regulator.
The general nature of the currents that appear is shown
in Fig. 21. They can be divided into two parts:
a. An alternating component in the armature and as-
sociated with it an unidirectional component in the field.
These two components decay or decrease together with the

Fig. 22—Symmetrical component of armature short-circuit

current (three-phase short circuit from no-load rated volt-
age). Values are rms.

tion is important only as it affects the field current neces-

sary to produce normal voltage at no load. The steady-
state value of short-circuit current is thus equal to the
line-to-neutral voltage read from the air-gap line for the
value of field current required to produce normal voltage
divided by the synchronous reactance in ohms.
Transient Component-If the excess of the symmet-
rical component of armature currents over the steady-state
component be plotted on semi-log paper, it can be seen
that this excess, except for the first few cycles, is an ex-
ponential function of time (the points lie in a straight line).
Extending this straight line back to zero time and adding
the steady-state component, the so-called transient com-
ponent, id’, or armature current is obtained. This compo-
nent is defined through a new reactance, called the transient
same time constants. The alternating armature component reactance by means of the expression
can be regarded as being produced by its associated uni-
erated
directional component in the field. All phase components id’ = -

of the alternating current are essentially the same except x’d

that they are displaced 120 electrical degrees.* The manner in which this quantity is related to the ex-
b. An unidirectional component in the armature and an ponential and steady-state terms is shown in Fig. 22.
alternating component in the field or in the damper wind- In discussing this component, the presence of the
ings. In this case, likewise, the alternating current in the damper-winding currents of salient-pole machines and
field winding can be regarded as produced by the uni- rotor eddy currents of turbine generators can, for the
directional component in the armature. moment, be neglected. Before short-circuit occurs the flux
associated with the field windings can be broken up into
6. Alternating Component of Armature Current two components (see Fig. 23), a component @ that crosses
This component can in turn be resolved into several the air gap and a component <PI, a leakage flux that can
components, the r.m.s. values of which are shown in Fig. be regarded as linking all of the field winding. Actually,
22. They are : of course, the leakage flux varies from the base of the pole
a. The steady-state component to the pole tip. The flux @i is so weighted that it produces
b. The transient component the same linkage with all the field turns as the actual leak-
c. The subtransient component age flux produces with the actual turns. It is approximately
proportional to the instantaneous value of the field current
Each of these components
will be discussed separately. If. The total flux linkages with the field winding are then
Steady-State Components-The steady-state com- those produced by the flux (++@I). As the field structure
ponent, as its name implies, is the current finally attained. rotates, a balanced alternating voltage and current of nor-
Because of the demagnetizing effect of the large short- mal frequency are produced in the armature. Because the
circuit current, the flux density within the machine de- armature resistance is relatively small, its circuit can be
creases below a point where saturation is present. Satura- regarded as having a power-factor of zero. The symmet-
*The machine used in this case was a salient-pole machine. As rical current thus produced develops an mmf that rotates
will be seen later, such machines also contain a second harmonic synchronously and has a purely demagnetizing, as con-
component of current. This type of machine was chosen to show trasted with cross magnetizing, effect on the field fluxes.
more clearly the presence of field and damper currents. It is a well-known fact that for the flux linkages with a
154 Machine Characteristics Chapter 6

possible to increase the exciter voltage instantaneously to

an amount that would produce this steady-state field cur-
rent, then this component of short-circuit current would
remain sustained. It is important to grasp the significance
of this truth. There is always a constant proportionality
between the alternating current in the armature and the
unidirectional (often called direct-current) component of
current in the field winding, whether the operating condi-
tion be steady-state or transitory.
The initial value of armature current, as stated, grad-
ually decreases to the steady-state and the induced current
in the field winding likewise decreases to its steady-state
magnitude. The increments of both follow an exponential
curve having the same time constant. Attention will next
be given to considerations affecting this time constant.
If a constant direct voltage is suddenly applied to the
field of a machine with the armature open-circuited, the
current builds up exponentially just as for any circuit hav-
Fig. 23—Air-gap and leakage fluxes at no load. ing resistance and inductance in series. The mathematical
expression of this relation is:
circuit to change instantly, an infinitely large voltage is
necessary and the assumption is justified that, for the tran- r,=; [&ij (6)
sition period from the no load open-circuited condition to
the short-circuited condition, the flux linkages with the field in which
winding can be regarded as constant. This is equivalent to
e, is the exciter voltage.
saying that the flux (@++I) remains constant. In order
rf is the resistance of the field winding in ohms.
that this flux remain constant in the presence of the de-
T&, is the open-circuit transient time constant of the
magnetizing effect of the armature current, it is necessary
machine or of the circuit in question in seconds.
that the field current If increase to overcome the demag-
t is time in seconds.
netizing effect of the armature current. If If increases then
Q>l, which is proportional to it, must likewise increase. It The time constant is equal to the inductance of the field
follows then-that @ must decrease. Consideration of the winding divided by its resistance. In the case of the short-
steady-state conditions has shown that the air-gap voltage, circuited machine, it was shown that at the first instant
el, is proportional to the air-gap flux a. The armature the flux linkages with the field winding remain the same
as for the open-circuit condition, but that the direct com-
current for short-circuit conditions is equal to e’. If + and
Xl
ponent of field current increases to 2 times the open-
consequently el had remained constant during the transi-
tion period, then the transient component of short-circuit circuit value before short-circuit. Since inductance is
current would be merely the no-load voltage before the defined as the flux linkages per unit current, it follows
short-circuit divided by the leakage reactance and the then that the inductance of the field circuit under short-
transient reactance would be equal to the armature leakage xd’
circuit must equal - times that for the open-circuit con-
reactance x1. However, as just shown, the air-gap flux
decreases and, therefore, the armature current is less. It dition. The short-circuit transient time constant that is
follows then that the transient reactance must be greater the time constant that determines the rate of decay of the
than the armature leakage reactance. It is a reactance transient component of current must then equal
that includes the effect of the increased field leakage
occasioned by the increase in field current. Td’=- xd’T’d,, in seconds
Under steady-state conditions with no saturation, the xd
armature current can be viewed as produced by a fictitious The component of armature current that decays with
internal voltage equal to xd id whose magnitude is picked this time constant can then be expressed by
from the air-gap line of the no-load saturation curve for
the particular field current. At the first instant of short- -7
(id’ --id)6 Td
circuit, the increased armature current, id’, can likewise be
viewed as being produced by a fictitious internal voltage When t is equal to Td’ the magnitude of the component
behind synchronous reactance, whose magnitude is x&’ or has decreased to c-l or 0.368 times its initial value. This
erated instant is indicated in Fig. 22.
xd-7, if the short-circuit be from rated voltage, no load.
Subtransient Component--In the presence of damp-
The voltage provides a means for determining the initial er windings or other paths for eddy currents as in turbine
value of the unidirectional component of field current by generators, the air-gap flux at the first instant of short-
picking off the value of 1f on the air-gap line of the no-load circuit is prevented from changing to any great extent.
saturation curve corresponding to this voltage. If it were This results both from their close proximity to the air gap
Chapter 6 Machine Characteristics 155

and from the fact that their leakage is much smaller than only one can equal zero at a time. Therefore at times of
that of the field winding. Consequently, the initial short- three-phase short-circuits, the alternating component of
circuit currents of such machines are greater. If this excess current in at least two and probably all three phases must
of the symmetrical component of armature currents over change from zero to some finite value. Since the armature
the transient component is plotted on semi-log paper, the circuits are inductive, it follows that their currents cannot
straight line thus formed can be projected back to zero change instantly from zero to a finite value. The “theorem
time. This zero-time value when added to the transient of constant flux linkages” must apply to each phase sepa-
component gives the subtransient current, id”. This sub- rately. The application of this theorem thus gives rise to
transient current is defined by the subtransient reactance in an unidirectional component of current in each phase equal
the expression and of negative value to the instantaneous values of the
alternating component at the instant of short circuit. In
this manner the armature currents are made continuous as
shown in Fig. 24. Each of the unidirectional components
The subtransient reactance approaches the armature leak-
age differing from that quantity only by the leakage of the
damper windings.
Since the excess of the armature currents represented
by the subtransient components over the transient com-
ponents are sustained only by the damper winding cur-
rents, it would be expected that their decrement would be
determined by that of the damper winding. Since the
copper section of this winding is so much smaller than
that of the field winding, it is found that the short-circuit
subtransient time constant, Td”, is very small, being
about 0.05 second instead of the order of seconds as is Fig. 24—The inclusion of a d-c component of armature
characteristic of the transient component. The component current whose existence is necessary to make the armature
of armature current that decays with this time constant is current continuous at the instant of short circuit.
(id” -id’) and can be expressed as a function of time as
in the three phases decays exponentially with a time con-
(i:’ stant T,, called the armature short-circuit time constant.
The magnitude of this time constant is dependent upon
Thus the time in seconds for this component to decrease the ratio of the inductance to resistance in the armature
to 0.368 times its initial value gives Td” as indicated in circuit. As will be shown the negative-sequence reactance,
Fig. 22. x2, of the machine is a sort of average reactance of the
Tests on machines without damper windings show that armature with the field winding short-circuited, so that it
because of saturation effects, the short-circuit current even is the reactance to use in determining T,. There exists then
in this case can be resolved into a slow transient component the relation
and a much faster subtransient component. The influence
52
of current magnitudes as reflected by saturation upon the T --in seconds (8)
‘- 27rjra
transient and subtransient reactance is discussed in more
detail under the general heading of Saturation. in which r, is the d-c resistance of the armature. The
quantity 271-jmerely converts the reactance to an induc-
7. Total Alternating Component of Armature Cur- tance.
rent The maximum magnitude which the unidirectional
The total armature current consists of the steady-state
value and the two components that decay with time con-
stants Td’ and Td”. It can be expressed by the following
equation
t t
-)I -7
&,= (i;‘--$)e Td +(iJ-id)e Td +id

The quantities are all expressed as rms values and are equal
but displaced 120 electrical degrees in the three phases.

8. Unidirectional Component of Armature Current

To this point consideration has been given to flux link-
ages with the field winding only. The requirement that
these linkages remain constant at transition periods de-
termined the alternating component of armature current.
Fig. 25—Representation of instantaneous currents of a three
Since these components in the three phases have a phase phase system. (a) Three separate vectors projected on x-axis,
displacement of 120 degrees with respect to each other, (b) Single vector projected on three axes.
156 Machine Characteristics Chapter 6

component can attain is equal to the maximum of the The short-circuit time constant is affected in a similar
alternating component. Therefore, manner
xd’ + xext
~da(max-)
=fieF (9)
Tdt=---- Tie in seconds (14)
xd xdfxext

A symmetrical three-phase set of currents can be repre- For the armature time constant, the external reactance
sented as the projection of three equal-spaced and equal must be added to the negative-sequence reactance of the
length vectors upon a stationary reference, say the real machine and the external resistance to the armature resist-
axis. They can also be represented as the projection, as it ance of the machine. The expression then becomes
rotates, of one vector upon three stationary axes, spaced
X2+x&
120 degrees. These axes can conveniently be taken as T- in seconds (15)
a- 2rj(ra+Text)
shown in Fig. 25, as the horizontal-axis and two axes hav-
ing a 120-degree relation therewith. Since the initial mag- Because of the much lower ratio of reactance to resistance
nitude of the unidirectional components are the negatives in external portions of circuits, such as transformers or
of the instantaneous values of the alternating components transmission lines, in the vast majority of cases T, for
at zero time, then the unidirectional components can be faults out in the system is so small as to justify neglecting
represented as the projection of a single vector onto the the unidirectional component of current.
three equal-spaced axes. This fact is used at times to
determine the maximum magnitude which the unidirec- 11. Short Circuit from Loaded Conditions
tional component can attain. By its use it is unnecessary The more usual case met in practice is that of a short-
to await a test in which the maximum happens to occur. circuit on machines operating under loaded conditions. AS
This method is in error, however, for machines in which before, the short-circuit current in the armature can be
x~” and xd” are radically different. divided into two components, a symmetrical alternating
component, and a unidirectional component.
9. Total RMS Armature Current Alternating Component-The alternating compo-
The rms armature current at any instant is nent in turn can be resolved into three components: (1)
steady state, (2) transient, and (3) subtransient. Each of
&izTG--
these components will be discussed individually.
The minimum current thus occurs in the phase in which The load on the machine affects the steady-state com-
the unidirectional component is zero and the maximum ponent only as it influences the field current before the
occurs when the unidirectional component is a maximum, short circuit. The field current can be determined by any
that is, when maximum dissymmetry occurs. Since the of the methods discussed under the heading of “Steady-
maximum value that the unidirectional component can State Conditions.” Saturation will be more important than
attain is 2/2 e, then for the no-load condition. The steady-state short-circuit
Xd current is then equal to the line-to-neutral voltage read
I
from the air-gap line for the field current obtained for the
(10) loaded condition divided by x&
In the discussion of the determination of the transient
Of course, a rms value as its name implies, is an average component from the no-load condition, it was stated that
quantity and is usually taken over a cycle or half cycle the quantity that remained constant during the transition
of time. The foregoing expression assumes that both the period from one circuit condition to another, is the flux
alternating and the unidirectional components do not de- linkages with the field winding. For the short-circuit from
crease, because of the natural decrement, during the first loaded conditions this same quantity can be used as a basis
cycle. In reality the decrement is usually sufficient to for analysis. Consideration will be given first to a load
make the effect noticeable. In applying circuit breakers before short circuit whose power factor is zero, lagging,
it is usual to use a factor 1.6 instead of -\/3. This factor and whose current is &L. The flux linkages before short
includes a small decrement. circuit will be determined by a superposition method,
obtaining first the linkages with the field winding for zero
10. Effect of External Impedance
armature current and any terminal voltage and then the
If the short-circuit occurs through an external imped- flux linkages with armature current, &L,, and zero terminal
ance rext+j xext, and Text is not too large, their effect can voltage. The total flux linkages is the sum of the two
be introduced by merely increasing the armature constants values so obtained.
by these amounts. Thus the components of short-circuit Let $1 be the flux linkages with the field winding at
current become no-load at rated voltage. For any other terminal voltage
idI1 =
e rated such as et, the flux linkages + will be equal to
(11)
J;dl’+xext

e rated
$+I (16)
id’=-.----- rated
W)
xd’+ %xt
By definition the transient reactance of a machine is
eairgap at no load equal to the reactance which, divided into the line-to-
id= (13)
xd+%xt neutral rated voltage, gives the transient component of
Chapter 6 Machine Characteristics 157

short-circuit current at no-load normal voltage. If this circuit current i’ can be determined by dividing these
short-circuit current is designated as &‘I, then linkages into the total flux linkages just determined. This
gives
erated
id’1 = - (17) et+Xd’idL
xd’
i’= -= + 09
xd’#l xd’
At the instant of short-circuit from no-load at rated volt-
e rated
age, the flux linkages with the field winding, $Q, remain
constant. The demagnetizing effect of the armature cur- The numerator of this quantity can be regarded as an
rent is overcome by an increase in the field current. Thus internal voltage, ed’, which is equal to the terminal voltage
the armature current idI with its associated field current plus a transient reactance drop produced by the load
which is always proportional to it, can be regarded as pro- current.
ducing the flux linkages $1 with the field winding. For any When the power factor of the loads considered is other
other armature current, id’, assuming always that the than zero lagging, the vector sense of current and terminal
armature is short-circuited, the flux linkages with the field voltage must be introduced. This can be accomplished by
computing t?d’for the operating condition in the same man-
winding are equal to -$ $1. Combined with Eq. (17),

xd’J/l
id’l can be eliminated giving $ = id’-. While this
erated

expression was derived from considerations applying only

to the instant of transition, its application is more general.
The only necessary considerations that must be satisfied
are that the armature be short-circuited and that the field
current contain a component of current to overcome the
demagnetizing effect of the armature current. But these
conditions are always satisfied even under steady-state
conditions of short circuit, so, in general, it is permissible
to replace id’ in this expression by idI,. The flux linkages
with the field winding for the steady-state short-circuit
xdjh
condition thus become &I,-.
e rated
By application of the superposition theorem, the total
flux linkages with the field winding can then be regarded
as the sum of the flux linkages produced by the terminal
voltage, namely *I,LQ and those by the armature cur-
erated

Xd’lC’l
rent with zero terminal voltage, namely &I,-. If the
erated

armature current lags the voltage by 90 degrees, then the Fig. 27—Machine internal voltage as a function of reactance.
linkages are directly additive, and there results for the Full-load rated voltage.
flux linkages with the field
ner that ed was determined in Fig. 14, except that xd should
be replaced by xd’. The voltage ed’ should then replace
et+xd’ &L in (19). However, for nearly all practical pur-
(18) poses it is sufficiently accurate to replace ed’ by the amph-
tude of a quantity ei’, which is usually referred to as the
Since the flux linkages with the field winding produced voltage behind transient reactance to distinguish it from
by a unit of current id under short-circuit conditions is similar internal voltages for which leakage, synchronous
xd’+l or subtransient reactance is used. The construction for
equal to - then the transient component of short-
erated this quantity is shown in Fig. 26 and to assist in the ready
evaluation of the amplitude the curves in Fig. 27 are
provided. The transient component of short-circuit current
is then
I
i’- ei
cm
xd’

The subtransient component of short-circuit current is

obtained in a manner similar-to the transient component
except that the subtransient reactance is used in the calcu-
Fig. 26—Construction for the determination of internal volt- lation of the internal voltage ei”. For loads of zero-power-
ages ei’ and ei”. factor lagging the subtransient reactance drop, xd”idL,
158 Machine Characteristics Chapter 6

caused by the armature current is directly additive to the . xd - xd’

2,= ---e,e-G cos (27rjl+a)+z COS (27rft+cU)
terminal voltage and for zero-power-factor leading directly xdxd’
subtractive. For other power-factors ei” can be obtained
xq -xd’ x +xd’
from Fig. 27 by using xd”. The subtransient component of ------etc-k cos (4?rft+c+*ete -+a coscx (22)
2x;xq ‘d Q
short-circuit current is then
xd - xd’
ib = -e t es+ cos (27r$!+cu-120’)
(21) xdxd’
*d

+z cos (27ft+a-- 120”)

Unidirectional Component-In the three-phase
short-circuit from no load, the unidirectional component -xd’
of current was introduced to prevent a non-continuous -- 2, cos (4Tft+a- 120”)
2x:x9 etE
transition of the instantaneous value of current from the
no-load to the short-circuit condition. The unidirectional e-k cos (a!+120°) (23)
current performs a similar role for the short-circuit from
loaded condition. Before the short-circuit the armature
current is equal to idL and has some position with reference cos (27rjt+a+ 120”)
to et such as shown in Fig. 28. The subtransient com-
+>os (2n;ft+a+ 120”)

(24)

Where
et = Terminal voltage before short-circuit.

(25)

2x,&~
T,= (26)
T(xd’+xq)’

cy= Angle which indicates point on wave at which

short-circuit occurs.
The instantaneous field current, Id, is

Fig. 28—Showing that idc for a short circuit from load is equal
(27)
to the negative of 2 times the difference between i” and i.

Where
ponent, i”, lags ei” by ninety degrees so i” and idr, will be If= Initial value of field current.
determined with respect to each other. The 2/2 times the
vector difference between these two quantities (since they III. UNBALANCED CONDITIONS
are rms magnitudes) gives the unidirectional component
necessary to produce smooth transition. The magnitude 13. Phase Currents for Unbalanced Short Circuits
of this quantity varies between this amplitude and zero As explained in the chapter relating to Symmetrical
depending upon the point in the cycle at which short- Components, the unbalanced operating conditions of a.
circuit occurs. rotating machine can for most purposes be described in
Other Considerations-Time constants are not influ- terms of three characteristic constants: the positive-se-
enced by the nature of loading preceding the short-circuit. quence impedance, the negative-sequence impedance, and
Total rms currents can be determined by the relations the zero-sequence impedance. The short-circuit currents
already given. can be resolved, as before, into the steady-state, transient,
and subtransient components. The difference between
12. Three-Phase Short Circuit of Salient-Pole Ma- these components decreases exponentially as before. The
chine without Damper Windings components of armature current and the time constants
For most applications it is sufficiently accurate to treat for the different kinds of short-circuits are given below for
the salient-pole machine without damper windings just as short-circuits at the terminals of the machine.
other machines. It must be recognized, however, that this For three-phase short-circuit:
is only an approximate solution. Among other complica-
ei” it=-
ei’ i=-
ei
tions, in reality a strong second harmonic is present in the j”EII Td’ = ‘& Td; CW
armature current. Doherty and NickleG have developed xd xd’ xd

expressions for the armature currents for a three-phase For terminal-to-terminal short circuit, the a-c components
short circuit from no load. These are given below. of the phase currents are given by
Chapter 6 Machine Characteristics 159

up by the armature current. If the axis of the short-

circuited field winding lines up with the axis of pulsating
field then the current is large and if the rotor is moved
xd’+J;Z
Td’------ T’do m-0 through 90 electrical degrees then the current is much
xd+xZ
smaller. The first position corresponds to the case of a
in which x2 is the negative-sequence impedance of the transformer in which the secondary winding is short-
machine circuited, the field winding in this case corresponding to
For terminal-to-neutral short circuit, the a-c components the secondary winding of the transformer. This is the
of the phase currents are given by . position in which the subtransient reactance, xd”, is deter-
mined. It is equal to one-half of the voltage from terminal-
3ei” ‘I 3ei’
2“f =
2= to-terminal divided by the current. For the second position
xd”+xZ+xO xd’+x2+xO
the field winding is in quadrature to the pulsating field
3ei xd’+x2+xO
and consequently no current flows in the field winding.
i= Td’ = Tie (30)
xd+xZ+xO xd+x2+xO The armature current is then determined by the magnetiz-
in which x0 is the zero-sequence impedance of the machine. ing characteristics of the air gap in the quadrature axis.
The subtransient time constant, Td”, does not change sig- The subtransient reactance, xq”, is determined when the
nificantly with different conditions and, therefore, the field is in this position and is equal to one-half the quotient
single value is used for all conditions. The unidirectional of the voltage divided by the current. The reactance for
components and the rms values are determined just as intermediate positions varies between these two amounts
described under the general subject of “Short Circuit from in accordance with the curve shown in Fig. 29.
Load. ” The above values of ei, ei’ and ei” will naturally
be those values corresponding to the particular load
condition.
The ratio of the phase currents for terminal-to-neutral
to three-phase short circuits can be obtained from Eq’s (30)
and (28). Thus, for the phase currents
Terminal-to-neutral short circuit 3x,-J”
Three-phase short circuit =xd”+x2+xO
The negative-sequence impedance, x2, is usually equal to
xd”, but for many machines x0 is less than xd”. For
these cases, the terminal-to-neutral short-circuit current
is greater than the three-phase short-circuit current. The
generator standards require that the machine be braced
only for currents equal to the three-phase values. In order
that the terminal-to-neutral current not exceed the three-
phase current a reactor should be placed in the neutral of
the machine of such value as to bring the zero-sequence
impedance of the circuit equal to xd”. Thus, the neutral
reactor, xn, should be
1
xn >=-(xd” -XI)) Fig. 29—Relation between subtransient and negative-ae-
3 quence reactance.

14. Negative-Sequence Reactance

When a set of negative-sequence currents is made to flow
The negative-sequence impedance of a machine is the through the armature with the field short-circuited and
impedance offered by that machine to the flow of negative- rotating in its normal direction, then the field winding
sequence current. A set of negative-sequence currents in takes different positions successively as the armature field
the armature creates in the air gap a magnetic field that rotates with respect to it. The nature of the impedances
rotates at synchronous speed in a direction opposite to that in the two extreme positions, that is, where the field wind-
of the normal motion of the field structure. Currents of ing lines up with the magnetic field and where it is in
double frequency are thereby established in the field, quadrature with it, should be somewhat the same as xd”
and in the damper winding if the machine has one. The and x9/‘, the only significant difference being the fact that,
imaginary component of the impedance is called the nega- in the determination of xd” and x,“, currents of normal
tive-sequence reactance and the real component the nega- frequency were induced in the field, whereas, in the nega-
tive-sequence resistance. These will be discussed sepa- tive-sequence case the currents are of twice normal fre-
rately, in the order mentioned. quency. One would expect therefore that the negative-
If a single-phase voltage is applied across two terminals sequence reactance x2 is some sort of a mean between xd”
of a salient-pole machine without dampers while its rotor and xq”, and such is the case. According to the AIEE test
is stationary, the resulting current is dependent upon the code, 10 the definition of negative-sequence reactance is
position of the rotor with respectt to the pulsating
set field
equal to “the ratio of the fundamental component of re-
160 Machine Characteristics Chapter 6

active armature voltage, due to the fundamental negative- where I equ’als the root-mean-square armature current in
sequence component of armature current, to this compo- the short-circuited phase; and E equals the root-mean-
nent of armature current at rated frequency.” A rigorous square open-circuit voltage between terminals before the
interpretation of this definition results in z2 equal to the short-circuit is applied or the no-load voltage correspond-
5, “+xd” ing to the field current at which I is read.
arithmetic mean . However, several different defi- In general, the same arguments can be applied to other
2
nitions can be given for x2. That this is possible is depend- types of machines such as turbine generators and salient-
ent largely upon the fact that when a sinusoidal set of pole machines with damper windings when the parameters
negative-sequence voltages is applied to the armature the Xd” and x9" are used. For such machines the difference
currents will not be sinusoidal. Conversely if the currents between xq” and xd” is not great. The values for x9” and
are sinusoidal the voltages will not be. xd” of a machine with copper dampers are given in Fig. 29.
In Table 1 are shown expressions 11 for x2 based upon For such machines the difference between x2 based on the
different definitions. This table is based on a machine different definitions of Table 1 will become inconsequential.
without damper windings for which xq” is equal to xq, and In addition, for turbine generators, saturation introduces
xd” is equal to xd’. In this table variables of much greater magnitude than those just con-
sidered. For these machines negative-sequence reactance
b-dg-m
can be taken equal to xd”.
-dx,+vG’ Method of Test-In addition to the method implied
For each test condition it is possible to establish definitions by the AIEE Code and the ASA whereby x2 is defined as
based on whether fundamental or root-mean-square cur- the arithmetic mean for xd” and xq”, x2 can be determined
rents are specified. For example, in the first definition if directly from test either by applying negative-sequence
the fundamental component of armature current is used voltage or by the method shown in Fig. 30.
in calculating x2 then the expression in the first column
should be used, but if the root-mean-square figure of the
resultant current is used then the expression in the second
column should be used.
In order to orient one’s self as to the relative importance
of the different expressions, figures have been inserted in
the expressions given in Table 1 for a typical machine
having the constants xd’ = 35y0, x, = 70%, and xd = 100%.
The magnitudes are tabulated in the righthand columns
of Table 1. From the standpoint of practical application,
the negative-sequence reactance that would result in the
proper root-mean-square current for method (3) would
appear to be the most important. However, the method
of test to determine this quantity involves a sudden short-
circuit and from this standpoint proves rather inconveni- EA=O EB=E EC=-E EM= g(O+dE--aE) =+
ent. On the other hand, the figure for x2 obtained from .EA* jia2-a)E
P 32/3 Es2 jE
zz=---=i
the use of the root-mean-square values in a sustained
Ea2 =.73 =
I aa 143
single-phase short-circuit current [method (4)], is nearly If 4 = cos+ 5, where P = wattmeter reading,
equal to this quantity. When the resistance is negligible
this negative-sequence reactance is equal to then, z?=z(sin 4-t-j cos $)=r~+jx2
VW
x2=------xd
A.,‘= Fig. 30—Determination of the negative-sequence impedance
(31)
I of symmetrically-wound machines.

TABLE 1—-DEFINITIONSOF NEGATIVE-SEQUENCE REACTANCE

Chapter 6 Machine Characteristics 161

With the machine driven at rated speed, and with a single- induces the voltage E, in the stator and SE, in the rotor.
phase short-circuit applied between two of its terminals In the rotor the impedance drop is
(neutral excluded) the sustained armature current and the
r&l- jsxrL (32)
voltage between the terminal of the free phase and either
of the short-circuited phases are measured. The reading since the reactance varies with the frequency of the cur-
of a single-phase wattmeter with its current coil in the rents in the rotor. The rotor current is therefore deter-
short-circuited phases and with the above mentioned volt- mined by the equation
age across its potential coil is also recorded. The negative-
sequence impedance equals the ratio of the voltage to the SE, = r,L+ jsx,L
current so measured, divided by 1.73. The negative- or
sequence reactance equals this impedance multiplied by the
ratio of power to the product of voltage and current.

15. Negative-Sequence Resistance It follows from this equation that the rotor circuit can be
The power associated with the negative-sequence cur- completely represented by placing a circuit of impedance
rent can be expressed as a resistance times the square of 2+ jxr across the voltage E,. The total power absorbed by
the current. This resistance is designated the negative-
sequence resistance. For a machine without damper wind-
: must be the sum of the rotor losses and the useful shaft
ings the only source of loss is in the armature and field
resistances, eddy currents, and iron loss. The copper loss r,
in the armature and field is small as is also the iron and power, so that, resolving ; into the resistances rr and
eddy loss in the armature, but the iron and eddy loss in l-s
the rotor may be considerable. Copper damper windings -r,, the power absorbed by rr represents the rotor cop-
S
provide a lower impedance path for the eddy currents and
hinder the penetration of flux into the pole structure. The per loss. The power absorbed by Er, represents the
S
relatively low resistance of this path results in a smaller
negative-sequence resistance than if the flux were per- useful shaft power.
mitted to penetrate into the rotor. For higher resistance Neglecting re and the real part of z,, the only real power
damper windings the negative-sequence resistance increases is that concerned in the rotor circuit. Assume that the
to a point beyond which the larger resistance diminishes induction motor drives a direct-current generator. At
the current in the rotor circuits sufficiently to decrease small slips the electrical input into the stator is equal to
the loss. the copper loss, i.e., the 12rr, of the rotor plus the shaft
Induction-Motor Diagram-The nature of the neg- load. With the rotor locked, the shaft load is zero, and the
ative-sequence resistance is best visualized by analyzing total electrical input into the stator is equal to the rotor
the phenomena occurring in induction motors. In Fig. 31 copper loss. At 200-percent slip, i.e., with the rotor turning
at synchronous speed in the reverse direction, the copper
I:r,
loss is I:r,, the electrical input into the stator is 2
1-2 - I,2r,
and the shaft load -rJ,2 or - A negative shaft
2 2 *
load signifies that the direct-current machine instead of
functioning as a generator is now a motor. Physically that
is just what would be expected, for as the slip increases
from zero the shaft power increases to a maximum and
Fig. 31—Equivalent circuit of induction motor. then decreases to zero for loo-percent slip. A further in-
crease in slip necessitates motion in the opposite direction,
is given the usual equivalent circuit of an induction motor which requires a driving torque. At 200-percent slip the
in which electrical input into the stator is equal to the mechanical
input through the shaft; half of the copper loss is supplied
ra = stator resistance.
x,= stator-leakage reactance at rated frequency. from the stator and half through the shaft. This is the
rr= rotor resistance. condition obtaining with respect to the negative-sequence
in which the rotor is rotating at a slip of 200 percent rela-
xr = rotor-leakage reactance at rated frequency.
tive to the synchronously rotating negative-sequence field
zrn = shunt impedance to include the effect of magnetiz-
in the stator. Half of the machine loss associated with the
ing current and no-load losses.
E-II- applied voltage. negative-sequence current is supplied from the stator and
I,= stator current. half by shaft torque through the rotor.
I,- rotor current. The factors of fundamental importance are the power
s = slip. supplied to the stator and the power supplied to the shaft,
which can always be determined by solving the equivalent
The justification for this diagram is shown briefly as circuit involving the stator and rotor constants and the
follows: The air-gap flux created by the currents I, and I, magnetizing-current constants. A more convenient device,
162 Machine Characteristics Chapter 6

supplied to the stator, which, in turn is equal to the loss

absorbed by the negative-sequence resistance. Therefore,
the total decelerating power is equal to the positive-
sequence power output plus the loss in the negative-
sequence resistance.
The assumption was made that the stator resistance
and the losses in the magnetizing branch were neglected.
For greater refinements, the stator resistance and the losses
in the magnetizing branch can be taken into consideration
by substituting them in the equivalent circuit and reducing
that circuit to simple series resistance and reactance,
wherein the resistance becomes the negative-sequence re-
sistance and the reactance the negative-sequence reactance.
The ratio of the negative-sequence shaft power to the loss
in the negative-sequence resistance is then equal to the
rr
ratio of the power loss in 5 for unit negative-sequence

current in the stator to r2. This ratio can be obtained easily

by test by measuring the shaft torque and the negative-
Fig. 32—Development of negative-sequence resistance and
sequence input when negative-sequence voltages only are
reactance from equivalent circuit of induction motor. applied to the stator.
(a) Negative-sequence diagram for induction motor; (b) neg- While this analysis has premised induction-motor con-
lecting armature and no load losses; Cc) simplified network- struction, the conclusions can also be applied to synchro-
negative-sequence resistance and reactance.
nous machines.
Method of Test-While r2 and x2 can be determined by
since s is constant and equal to 2 for the negative-sequence, applying negative-sequence voltage from another source
is to reduce the equivalent network to a simple series im- of supply to the armature, the following method has the
pedance as shown in Fig. 32 (c). The components of this advantage that the machine supplies its own negative-
impedance will be called the negative-sequence resistance sequence voltage. Two terminals of the machine under
r2, and the negative-sequence reactance x2. The current test are short-circuited and the machine driven at rated
flowing through the negative-sequence impedance is the frequency by means of a direct-current motor. The equiv-
current flowing through the stator of the machine, and the alent circuit and vector diagram for this connection are
power loss in r2 is equal to the loss supplied from the stator shown in Fig. 33. The positive-sequence power per phase
of the machine and the equal loss supplied through the at the terminals is equal to the product of El and 11 and
shaft. the cosine of the angle #. This power is positive. However,
The total electrical effect of the negative-sequence re- the negative-sequence power output per phase is equal to
sistance in system analysis problems is obtained by in- the product of E2, r2, and the cosine of the angle between
serting the negative-sequence resistance in the negative- E2 and 12, and since 12= -II, and El = E2, the negative-
sequence network and solving the network in the usual
manner. All three of the sequence currents are thus af-
fected to some extent by a change in the negative-sequence
resistance. The total electrical output of a generator, not
including the shaft torque developed by negative-sequence
current, is equal to the total terminal power output plus
the losses in the machine. However, the negative- and
zero-sequence power outputs are merely the negative of
their losses. In other words, their losses are supplied by
power flowing into the machine from the system. There-
fore, the contribution of the negative- and zero-sequences
to the electrical output is zero. The total electrical output
reduces then to that of the positive-sequence and to include
the positive-sequence armature-resistance loss it is neces-
sary only to use the positive-sequence internal voltage in
the calculations. Or viewed differently, since there are
no internal generated voltages of the negative- or zero- r~-3vwP+w)
-
sequence, the corresponding internal power must be zero. 212
In addition to this electrical output, which produces a In which
torque tending to decelerate the rotor, there also exists the P,= shaft input.
negative-sequence shaft power supplied through the rotor. P F+W= friction and windage loss.
It was shown that this power tending to decelerate the Fig 33— Negative-sequence resistance of a synchronous
rotor is numerically equal to the negative-sequence power machine.
Chapter 6 Machine Characteristics 163

sequence power output is the negative of the positive- end-connection fluxes. The departure from this ideal con-
sequence power output, which, of course, must follow since dition introduced by chording and the breadth of the phase
the output of the machine is zero. A negative output is belt determines the zero-sequence reactance.
equivalent to a positive input. This input is equal to r21i The zero-sequence resistance is equal to, or somewhat
per phase. Therefore, the positive-sequence terminal out- larger than, the positive-sequence resistance. In general,
put per phase is rzri, and adding to this the copper loss due however, it is neglected in most calculations.
to II, gives the total shaft power due to the positive- Method of Test-The most convenient method for
sequence as 3 (~1; + rJ:) . test of zero-sequence impedance is to connect the three
Now from Fig. 32(a), if zm be neglected the negative phases together, as shown in Fig. 34, with the field short-
sequence input per phase is equal to

from which it follows that

r2 =?+T#). (34)

As shown previously the negative-sequence shaft power

per phase is equal to %I;, which on substituting 2rr from
Rotor at synchronous speed
(34) reduces to (rZ -r$j. But since r,=rl, the expression (or blocked)
for the negative-sequence shaft power per phase can also Zero-sequence impedance;
be written (r2 -r$j. Incidentally, from this the rotor
losses are equal to 2(rz - rl)Ii. Therefore the total shaft
input into the alternating-current machine is equal to Fig. 34—Connection for measuring zero-sequence imped-
3 [rz1;+d:+ (r2 -rl)fi] and, since li’l=12, reduces to 6r& ance.
Including the effect of friction and windage, P(F+w), and
calling P, the total input into the alternating-current circuited. This connection insures equal distribution of
machine from the driving tool, current between the three phases. For this reason it is
preferable to connecting the three phases in parallel. The
&-p(F+W)
r2=
61; zero-sequence impedance is then equal to Zo=E as in-
31
I dicated in the illustration.
and, since & = - where 1 is the actual measured phase
fi
current, IV. PER UNIT SYSTEM
Ipa - p(F +W,l
r2= (36) The performance of a whole line of apparatus, regardless
212 of size, can often be expressed by a single set of constants
The foregoing neglects the effects of saturation. Tests on when those constants are expressed in percentages. By
salient pole machines with and without dampers verify this is meant that the loss will be a certain percentage of
the fact that the loss varies as the square of the negative- its kilowatt rating, its regulation a certain percentage of
sequence currents. The loss for turbine generators, on the its voltage rating, etc. The advantage of this method of
other hand, varies as the 1.8 power of current. representation extends to a better comparison of perform-
ance of machines of different rating. A 100-volt drop in a
16. Zero-Sequence Impedance transmission line has no significance until the voltage base
The zero-sequence impedance is the impedance offered is given, whereas, as a percentage drop would have much
to the flow of unit zero-sequence current, i.e., the voltage significance.
drop across any one phase (star-connected) for unit current A disadvantage of the percentage system is the confusion
in each of the phases. The machine must, of course, be that results from the multiplication of percentage quanti-
star-connected for otherwise no zero-sequence current can ties. Thus, a 20-percent current flowing through a 40-
flow. percent reactance would by simple multiplication give 800
The zero-sequence reactance of synchronous machines is which at times is erroneously considered as 800-percent
quite variable and depends largely upon pitch and breadth voltage drop, whereas, the correct answer is an 8-percent
factors. In general, however, the figures are much smaller voltage drop.
than those of positive and negative sequences. The nature The per unit system4 of designation is advanced as pos-
of the reactance is suggested by considering that, if the sessing all the advantages of the percentage system but
armature windings were infinitely distributed so that each avoids this last mentioned disadvantage. In this system
phase produced a sinusoidal distribution of the mmf, then the rating quantity is regarded as unity. Any other amount
the mmfs produced by the equal instantaneous currents of the quantity is expressed as a fraction of the rated
of the three phases cancel each other and produce zero amount. It is the same as the percentage system except
field and consequently zero reactance except for slot and that unity is used as a base instead of 100. The foregoing
164 Machine Characteristics Chapter 6

multiplication example would in the per unit system be

expressed as follows : A 0.20 per unit current flowing (37)
through a 0.40 per unit reactance produces an 0.08 per
w.
unit voltage drop, which is correct. But et=ed- x&d -j&iq
A further advantage of the percentage and per unit and k=E cos &-jE sin 8
systems lies in the elimination of troublesome coefficients.
However, this is not an unmixed blessing as a definite Upon substituting Zt and E in equation (37), there results
disadvantage of the use of the per unit system lies in the ..
i, -Jtd = -j [ed - x&d -jx,&]
loss of the dimensional check.

V. POWER EXPRESSIONS +ji[E cos &-jE sin e] (38)

It is frequently necessary to know the manner in which Equating reals
the power output of a machine varies with its excitation
and internal angle. A particular application of this knowl- x,(x.+xJi +E sin 8
i,= -
edge is the stability problem. Several simple cases will X&k Q $0
be considered. x,E sin e
t (39)
17. Machine Connected through Reactance to In- x*x~+xqx,+xqxo-
finite Bus and also Shunt Reactance across its And equating imaginaries
Terminals, Resistance of Machine Neglected
x.+xc
The schematic diagram for this case is shown in Fig. id= - [ed - xdid] -tE cos 8’
35(a), which also shows the significance of the various x.&c 0

symbols to be used in this discussion. The reactances z,, (X1+Xc)ed-XgE COS e

L (40)
z,, and the one indicated by the dotted lines represent the %&+~d~~+~dxc
branches of an equivalent 7r circuit, for which the resistance
components are neglected. For the purposes of determin- The power output, P, is equal to the sum of the products of
ing the power output of the generator the reactance shown the in-phase components of armature current and terminal
dotted can be neglected. The vector diagram which ap- voltage, namely
plies is Fig. 35(b). The total machine current is equal to i, P&&d-xdid) +id(x&)

= e& + (2, - xd) i&q

= [ed+ (xq-xd)id]& (41)

The power is then obtained by calculating i, and id from

(39) and (40) and inserting into (41). If E and ed are ex-
pressed in terms of rms volts to neutral and reactances in
ohms per phase, then the above expression gives the power
in watts per phase; but if the emf’s are expressed in terms
of the phase-to-phase volts the expression gives total power.
On the other hand, if all quantities are expressed in p.u.
then the power is also expressed in p.u. where unity is
equal to the kva rating of the machine. If ed’ rather than
ed is known then cd should be replaced by ed’ and xd by zd’
wherever they appear in Eqs. (40) and (41).
For the special case of a machine with cylindrical rotor
in which x, =xd, the expression reduces immediately to
P = edi,
X,Eed sin e
F (42)
&Jo +x,x, + z.&
Fig. 35—Machine connected to infinite bus through a re-
actance. Another interesting special case is that for which the
shunt reactance is not present or xI = 00. Then

1x0+x,
which from the internal and external currents one can
ed--E cos e E sin e
obtain* P= ed + (xq - xd)
xc+ xd

edE Sin e+ (xd - x,)E2 Sin 28

L (43)
And inserting the equivalents of iT, and & &+xd 2(xd+&)(x,+&)

And if x0= 0 and X~= 00, then

edE sin e (xd-xq)E2 sin 28
P= + (44)
*The symbol caret over a quantity indicates a phasor quantity. xd 2xdxq
Chapter 6 Machine Characteristics 165

Fig.
- 37—Vector diagram of salient-pole machine including
Fig. 36—Power-ang1e diagram of a salient-pole machine- effect of series resistance.
excitation determined to develop rated kva at 80-percent
power factor. xd=l.ls; xq=O.75.
(46)

In Fig. 36 is shown a power-angle diagram of a salient- From (45)

pole machine whose excitation is determined by loading at iq-- 5 et sin 0+rQJ (47)
full kva at 80-percent power factor. %
An expression frequently used to determine the maxi-
Substituting (47) into (46)
mum pull-out of turbine generators is the following
r2
OC et cos 8+‘et sin 8+-id+xdid-ed=o
Pull-out in kw =&rating of generator in kva) 2, xq

where OC is the field current for the particular operating from which
condition and OD is the field current for the rated-current
zero-power factor curve for zero terminal voltage (see Fig.
17). This expression is based upon the maintenance of
id-------
1
r2+xdxq [
Xqed-ret sin 8-z,& COS 8
1
..* (48)

rated terminal voltage up to the point of pull-out. At and substituting in (47)

1
pull-out the angle 6 of Fig. 15 is equal to 90 degrees. Since . 1
the extent of saturation is measured by the voltage behind a,=- red+x& sin e-ret cos 8 (49)
r2-hdxq
the Potier reactance drop, it can be seen from Fig. 15 that
for 6 equal to 90 degrees this voltage is less than rated The power output, P, is equal to the sum of the products
voltage, and that therefore little saturation is present. of the in-phase components of i and et, or
From Eq. (44) since xd = x, and 8 = 90 degrees, the pull-out P = iqet cos 8+idet sin 8 (50)
GE
is -. But ed is proportional to OC on the air-gap line and Upon substituting (48) and (49) this reduces to
xd
&I is likewise proportional
Examination
to OD on the air-gap line.
of Eq. (44) shows that even if the excita-
tion is zero (ed=O) the power-angle curve is not equal to
p=-!!L
T2 + xdxq
ed(r cos 6+xq sin 0) +yet sin 2&?%?t
1(51)
(xd - i,) E2 sin 20
zero, but equal to . This results from the The power input into the machine is equal to P plus ri2.
2xdxq
The expression for this quantity does not simplify and it
effects of saliency. Note that it disappears for uniform
is better to calculate it through the intermediate step of
air-gap machines for which xd =x,. Advantage is some-
evaluating ri2, which is equal to r(ii+ii).
times taken of this relation in the case of synchronous
The foregoing expressions apply to the steady-state con-
condensers to obtain a somewhat greater capability in the
ditions. In stability problems it is necessary to determine
leading (under-excited) kva range. With some excitation
the average power from instant to instant. In general for
systems (see Chap. 7, Excitation Systems) it is possible to
this purpose it is permissible to neglect both the unidirec-
obtain negative excitation. The excitation voltage, ed, in
tional component of currents and the subtransient com-
Eq. (44) can be somewhat negative without producing an
ponent of the alternating current, leaving only the transient
unstable power-angle diagram. By this device the leading
component. These latter are determined by the instanta-
kva range can be increased as much as 15 or 20 percent.
neous value of ed’. It follows then that the power expres-
18. Inclusion of Machine Resistance or External sions are simply those derived for the steady-state condition
Resistance with ed replaced by ed’ and xd by zd’.
If the machine is connected to an infinite bus through a
resistance and reactance circuit, the external resistance VI. EFFECT OF CHANGE IN EXCITATION
and reactance can be lumped with the internal resistance Field forcing in certain industrial applications and con-
and reactance and the following analysis used. The vector siderations of system stability require that the voltage
diagram for this case is shown in Fig. 37 for which increase in response to a sudden need. This increase is
et sin 6+rid-xqiq=o (45) brought about automatically either by means of the same
166 Machine Characteristics Chapter 6

control that produced the increase in load or through the the field circuit. The beauty of the per unit system is
use of a voltage regulator. It is necessary, therefore, to be exemplified in the analysis of this problem. In p.u. the
able to predetermine the effect of an increase in exciter differential equation for the field circuit takes the following
voltage upon the output of the synchronous machine. In form
general, significant changes in exciter voltage never require
less than about one-tenth of a second to bring about the e, = Q-f-T; dx (52)
’ dt
change. By the time this effect has been felt through the
synchronous machine, which has a time constant of about In this equation e, represents the exciter voltage or the
a second, it will be found that the result is always slow voltage across the field if there is no external field resistor
when compared to the subtransient and unidirectional in the field circuit. The unit of e, is that voltage required
components of the transients associated with the change. to circulate such field current as to produce rated voltage
In other words, variations in exciter voltage are reflected at no load on the air-gap line of the machine. The term ed
only in the transient components. As an example, suppose is the synchronous internal voltage necessary to produce
it is desired to calculate the armature current of a machine the instantaneous value of armature current for the given
for a three-phase short-circuit while it is operating at no armature circuit regardless of what it may be. Its unit is
load with a voltage regulator set for rated voltage. rated voltage. It is synonymous with field current when
Immediately after the inception of the short circuit there unit field current is that field current necessary to produce
is a slight lag in the regulator until its contacts and relays rated voltage at no load on the air-gap line. It will be seen
close. The exciter voltage (and voltage across the field of then that the use of ed is merely a convenient way of
the main machine) then rises as shown in the upper curve specifying the instantaneous field current during the tran-
of Fig. 38. The bottom curve refers to the armature cur- sient conditions; it is the field current necessary to produce
the armature current existent at that instant. As shown
previously, ed’, is proportional to the flux linkages with the
field winding. It is the quantity that, during the transition
period from one circuit condition to another, remains con-
stant. The foregoing equation has its counterpart in the
more familiar forms

e&+N(lo-8)d$ (53)

or

To familiarize the reader with (52), suppose that normal

exciter voltage is suddenly applied to the field winding at
no load. Since the armature is open-circuited ed’ and ed
are equal and the equation can be written

Fig. 38—Illustration showing relative importance of different ded .

components of armature short-circuit current and response When steady-state conditions are finally attained dt IS
of transient component to the exciter voltage.
equal to zero and ed = e,. This states that since e, = 1.0,
rent, the dotted line showing the nature of transient com- ed must also equal 1.0, that is, the excitation is equal to
ponent if there were no regulator, the exciter voltage the normal no-load voltage. It will attain this value ex-
remaining constant. The line immediately above shows ponentially with a time constant TJ,.
how the transient component changes as a result of the Another example. Suppose the synchronous machine
change in exciter voltage. To approximately the same to be short-circuited from no-load and to be operating
scale, the cross-hatched area shows the increment in cur- without a regulator. At any instant the armature cur-
rent caused by subtransient effects. The blackened area rent, i, is equal to ed’/xd’. But since ed, which can be
shows how the unidirectional component would contribute regarded as the instantaneous field current required to
its effect. This component is quite variable and for a produce i, is equal to xd i, then eliminating i between
short-circuit on the line might be entirely completed in a these equations
cycle or less. In any event regardless of its magnitude it
can be merely added to the transient and subtransient (56)
component. It is independent of the exciter voltage.

19. Fundamental Equation Then equation (52) takes the form

Being restricted to the transient component, the effect of de{

1 =$3,‘+T:,-
exciter response can then be defined entirely by effects in dt
Chapter 6 Machine Characteristics 167

or if it is to be expressed in terms of armature current name, has been called by the author “the follow-up meth-
od. ” It is a method that can be applied to any problem
1 = zi+T:s’s
d (57) involving a simple time constant.
’ ddt’ To demonstrate the method, consider a simple resist-
The sustained magnitude of i, is then ance-inductance circuit to which the voltage, E, is applied.
Let the differential equation for this circuit be
. 1
z=-
xd E = RI+Lg (59)
dt
The initial magnitude of i, since ed’ remains constant
during the transition and is initially equal to 1, is where the symbols have the customary significance.
1 Dividing through by R, there results
. I
2=7
L dI
xd g=I+E z 030)
The homogeneous equation for (57) is
,di dI .
. The coefficient of dt IS called the time constant of the
0 = &ft+ T&,Xd $
circuit and will be designated by T, giving
OP
xd’ di E dI
O=i+T:oG z (53) R=IfTdt

E-1
Thus i changes from -$ to -!- exponentially with a time
xd
dI
A=- R (61)
xd’ I dt T
constant equal to -Tdo.
xd
In this expression i is the steady-state current that I
In all problems involving a transition from one circuit
condition to another the one quantity (when subtransient approaches for the instantaneous value of E. I is the
effects are neglected and the time constant in the quadra- instantaneous magnitude of current. If the current at
ture axis is zero) that remains constant within the machine any instant is plotted by the point a (Fig. 40) and the
is the flux linkages with the field winding, which in turn E
corresponding value of -R for that instant is plotted as
is reflected in the quantity ed’. It is necessary, therefore,
to calculate ed’ for the circuit condition preceding the the point b (Fig. 40) displaced horizontally by a time T,
transition. All the discussion of the following cases as-
sumes that this point is understood and that ed’ is known
for the beginning of the transient period.
Several cases will now be discussed.

20. Three-phase Short Circuit of Unsaturated Ma-

chine
The problem is to determine the transient component
of short-circuit current in response to the exciter voltage
given in Fig. 39. This is most quickly and conveniently
found by a graphical method, which, for want of a better
Fig. 40—Construction derivation of “Follow-up Method.”

then the vertical distance between a and b gives the nu-

merator of (61) and the horizontal distance the denomina-
dI
tor. The slope of the line between a and b is equal to dt.

If an interval of time At is chosen following the instant

under consideration and E is assumed constant over the
interval then the change in I during the interval, AI, is
dI
equal to -At. The fmal value of current for the interval
dt
E
is then given by the point c. If - at an instant At later is
R
then plotted and the line drawn from c then the value for
AI for the second interval is obtained. Following such
procedure it is possible to construct the complete curve
Fig. 39—Illustration of “Follow-up Method.” for I. The construction is illustrated in Fig. 39, in which
168 Machine Characteristics Chapter 6

the curve marked E is the instantaneous magnitude of E . 1

2 =z. Starting from this value the actual magnitude of
from time t =O. Plot E displaced to the right a time T.
i is obtained as a function of time.
Let Ii be the initial value of I at t=O. Divide the time into
intervals of length At. Draw the line ab, then cd, ef, etc. 21. Unsaturated Machine Connected to Infinite Bus
The accuracy will be greater the smaller the intervals and As stated previously the subtransient and unidirectional
can be increased somewhat for a given element width by components of current are not of importance in the sta-
bility problem. For this application it is desirable to de-
using T-$f instead of T for the distance by which the
termine how ed’ varies as this influences the power output
steady-state curve which I tends to approach, is offset of the machine and in turn dictates the degree of accelera-
horizontally. tion or deceleration of the rotor. The circuit shown in
Now returning to the problem in hand. The differential Fig. 35(a) is typical of a setup that might be used for an
equation governing the case is given by (52). The exciter analytical study to determine the effect of exciter response
voltage e, is assumed given and expressed in p.u. For a in increasing stability limits. Another case of considerable
three-phase short circuit at the terminals of the machine importance is the action of a generator when a heavy load,
ed is equal to xai and ed’ =zd’i. Therefore Eq. 52 becomes such as a large induction motor, is connected suddenly
across its terminals or across the line to which is is con-
di nected. In starting the motor the line voltage may drop
~‘==X&+Xd’T&-- (62)
dt an excessive amount. The problem might be to determine
Dividing through by xd the amount to which this condition could be ameliorated
f by an appropriate excitation system. Since reactive kva
e, di
-~i+zT$~- (63)
is more important than the real power in determining
xd dt regulation, the motor can be represented as a reactor and
The construction dictated by this equation and the the circuit in Fig. 35(a) utilized. Having determined the
manner in which ed’ varies, the power in the case of the
follow-up method is shown in Fig. 41. F is plotted against stability problem and the terminal voltage (f&i’-xd’i) in
the case of the voltage problem, can be calculated easily.
Equation (52) must be used again to determine the manner
in which ed’ varies in response to changes in exciter voltage
and phase position of the rotor with respect to the infinite
bus. The instantaneous armature current can be found in
terms of the rotor angle 8 and ed’ by replacing f?dand xd of
Eq. (40) by ed’ and xd’, respectively, giving
id = (X,+X&d’-X,E cm 8
(64)
&&+~d’&+~d’~o

The synchronous internal voltage, ed, is equal at any

instant to
ed=ed’+ (xd-xd’)id (65)
and upon substituting (64)
(xCk?$ed’-x#E COS 8
ed = et’ + (xd - xd’)
~&+~d’~,+~d’&

(~8~~+~d~s+~d~c)edf-~~~(2d-~dfjE COS 8
L
&~o+~d’~~+~d’~o

Substituting this expression in (52), there results.

&&+xdxn+xd%
ex= ed’ -
%%+~d’~~+~d’&

&bd - xd’) ded’

E cos 0 + T:odt 033)
~&+xd’~,+~d’~o

which can be converted to

Td’ xrhd - xd’)
Fig. 4l—Transient component of short-circuit current, i’, as T’e”+ E COS 8=ed’+Td’~ ded’ (67)
do ~&o+~d~~+xdxo
influenced by excitation.
in which
%&o+xd’~,+~d’&
time, its zero being displaced an interval ‘$TI, from refer- Td’ = T do’ 038)
&xo+xd%+xd&

ence zero. The initial value of i is determined through ed’ The time constant Td’ is the short-circuit transient time
which was 1.0 at t = 0. This makes the initial amount of constant.
 Chapter 6 Machine Characteristics 169

If 8 were constant or if its motion as a function of time From this point the follow-up method can be used as
were known then the whole left-hand side could be plotted before. After ed’ is determined as a function of time any
(displaced by the time Td’) and treated by the follow-up other quantity such as terminal voltage can be obtained
method as the quantity that ed’ tends to approach. Un- readily.
fortunately 8 is not in general known beforehand, and it is
necessary to calculate 0 simultaneously in small increments 23. Saturation
in a simultaneous solution of cd’ and 8. The magnitude of In analyzing transient phenomenon of machines in the
8 is determined by the electromechanical considerations unsaturated condition, the theory was built around the
discussed in the chapter dealing with System Stability. In concept of the transient internal voltage, ed’, a quantity
solving for ed’ a progressive plot of the left-hand side can evaluated by using the transient reactance, xd’. In the
be made or (67) can be transformed to the following form presence of saturation it was found that for steady-state

1
ded’ 1 conditions by the introduction of the Potier reactance,
x&d - zd’)
-=-
E COS &ed’ (69) x, (see Sec. 3) the proper regulation was obtained at full
dt Td’ &&+xdx,+xdxo
load zero power-factor. The use of x, and ep also resulted in
and the increment calculated from the equation satisfactory regulation for other power-factors. In extend-
ing the analysis into the realm of transient phenomenon.
ep will continue to be used as a base from which to intro-
duce additional mmf into the field circuit to take care of
A shunt resistance-reactance load such as an induction saturation effects. The treatment will follow quite closely
motor is not much more difficult to solve numerically but the same assumptions as were used in determining the
the expressions become too involved for analytical solution. steady-state regulation according to the Two-Reaction
It is necessary only to calculate id in terms of ed’ and 8 just Potier Voltage method of Sec. 3(d).
as was done before and then follow the same steps as used With this assumption the fundamental Eq. (52) for the
for the reactance load. field circuit becomes
d&
22. Unsaturated Machine Connected to Resistance- e, = ed+ (s due to e,) + T&,z (76)
Reactance Load
A case not too laborious to carry through analytically As before ed represents, neglecting saturation, the voltage
is that for which a resistance-reactance load is suddenly behind the synchronous reactance of the machine or what
applied to a synchronous machine. Let reXt and xext be the is equivalent the field current required to produce the
external resistance and reactance. The addition of a sub- instantaneous ed’, including the demagnetizing effect of
script t to machine constants indicates the addition of reXt the instantaneous armature current. The total field cur-
or xext to the respective quantity. The equations of Sec. 17 rent is obtained by adding s to ed. In some cases it is found
then apply to this case, if et in the equations is made equal simpler to convert all of the right hand side to the single
to zero and xd replaced by xdt, etc. variable ep but in others it is simpler to retain the variable
Following the same procedure as previously, there re- in the form of ed’. Two applications of this equation will
sults from Eq. (48) when ed and xd are replaced by ed’ be discussed.
and x& and et is equal to zero. Machine Connected to Infinite Bus-The circuit
shown in Fig. 35(a) is the one under discussion and for
which Eq. (66) applies for the unsaturated condition. This
equation can be expanded to include saturation, in accord-
The field current or its equivalent, the synchronous in- ance with Eq. (76), to the following
ternal voltage, is then
x~~o+~d~~+~d~o x&d - xd’)
ex= edt - Ems8
ed = ed’ + (Xdt - x&) id ~m%+~d’~.+~d’& %%+~d’x.+xd’&

%t de:
= ed’+ (xdt -xi&) ed’ +(S due to e,)+Td!,-. (77)
r2,+xkqt dt
This can be converted to
(72)
ded e, - (s due to e,) x&d-&) E COS 8
Substituting this expression in (52) there results that dt= T;, +(x&o+ xdx. + xdxc) 27

in which Td’ is defined from Eq. (68). Before (78) can be

used it will be necessary to determine ep in terms of ed’.
The components of current, i, and id, can be determined
which can be converted to
from (39) and (40) by replacing ed, by ed’ and xd by a$.
Thus
(74) x,E sin t9
iq= (79)
in which wo+x,x,+x&i

T; =
xdb,t+T:
T’ ia= (x.+x&d’-x,E COB8
xd&,t +< do- ~&+~d’~.+~d’x, ’
170 Machine Characteristics Chapter 6

The direct-axis component of eP is equal to assumption that is usually made in stability studies that
where quick response excitation is installed, ed’ may be
epd = ed’ - (xd’ -xp)id
regarded as constant.
&,(xd’
= wc+qJ,+x,xc ed’+ Machine Connected to Resistance-Reactance
%&+xd’&+xd’& x&+dxs+&.l’xo
Load-This case is the same as that considered in Sec. 22
and the quadrature-axis component of eP is except that saturation effects are to be included. Upon
(x, - x,)i,
including the saturation term s into Eq. (74) there results
ePQ=
that
-x,)E sin 0

1
X*(X,
= (82) ded
x8x0+x,x8+x,x0 e,-((s due to e,) =ed’+Td’x
The amplitude of ep is then equal to
in which
e,=dK (83)
xd’txqt + r:
While this quantity does not simplify greatly, it does not Tdf = -- T’do. (86)
xdtxqt+T:
appear so formidable after numerical values are inserted.
eP can thus be calculated for any instantaneous value of It is well to recall again that this analysis neglects sub-
.f?d’and the .s corresponding thereto substituted in Eq. (78). transient effects and assumes that the time constant in the
Equation (73) provides a means for computing increments quadrature axis is zero. If in Eqs. (48) and (49) et is made
of change in ed’ for use in step-by-step solution. Thus equal to zero, f?d is replaced by ed’ and the corresponding
changes in reactance associated with ed’ are made, and in
ddAt
Aed’= x addition the subscripts are changed to indicate total re-
actances, Then
As s becomes small and saturation effects disappear, the
solution relapses into the same type as used when satura- id = (87)
tion is negligible (Eq. 66), for which the follow-up method
is frequently applicable. rt
i, = (88)
The relations just developed are useful in estimating the xd:xqt Cr;ed’.
extent to which ed’ varies in system stability problems. The total current is then
Fig. 42 shows the results of calculations on a system in
which a generator is connected to a large network, repre- . l/x:t+rt
sented as an infinite bus, through a reactance equal toj0.6.
2= ed’. (89)
xd’txqt i- r”t

The voltage ep is
ep = id-

ufion substituting ed’ from (90) into (83) and using (86)
also there results that
~(x~t+~~,(x&+r~)e, d(x~t+f~,(xq~+i~+s due to e >
-
P
xdtxqt + ri xdtxqt-bi

As can be seen from Fig. 43 the solution of this equation

lends itself well to the follow-up method. On the right-
hand side the assumed exciter response curve, e,, is plotted
as a function of time. Multiplying this quantity by the
coefficient of e,, the term epm is obtained. This is the value
eP tends to attain if there were no saturation effects. -4s
in the follow-up method, the zero of time from which the
Fig. 42—Effect of rate of response upon ed’ as a line-to-line instantaneous curve of eP is drawn, is displaced to the left
fault represented by the three-phase shunt load j0.2 is applied an amount Td’ minus half the interval of time chosen in
to generator which had been operating at 90 percent power- the step-by-step solution. Along the ordinate of eP a curve
factor. 20 percent of air-gap mmf required for iron at rated s1 equal to the second term is plotted in which s is obtained
voltage.
from the no-load saturation curve shown in (b). For any
instantaneous value of ep, sl is plotted downward from epm
A line-to-line fault is assumed applied to the connecting as the construction progresses. So starting from the initial
transmission line on the high tension bus at the generating value of ep, of which more will be said later, a construction
end which is equivalent to a three-phase short circuit line is drawn to a point for which s1 was the value corre-
through a reactance of j0.2 ohms. The curves justify the sponding to the initial value of eP. For the second interval
Chapter 6

Fig. 43—Graphical determination of terminal voltage as polyphase series resistances, rext and reactances x-t are suddenly
applied.

sl is taken for the value of e,, at the end of the first interval sudden disconnection of a loaded generator from the bus
or, to be slightly more accurate for an estimated average throwing its load upon the remaining units or the starting
value of eP for the second interval. And so the construction of an induction motor by direct connection to a generator.
proceeds. For the latter case, if the capacity of the induction motor
By the same reasoning whereby eP was obtained in Eq. is a significant fraction of the kva of the generator, a severe
(90) the terminal voltage et can likewise be obtained, giving drop in voltage results. Thus a 500-hp motor thrown on
a 330%kva generator produces an instantaneous drop in
et = i1/x2ext + r2ext
voltage of the order of 13 percent. The effective impedance
of the induction motor varies with slip and to be rigorous
this variation should be taken into consideration. It is
and substituting ed’ from (90) usually sufficiently accurate to use the blocked rotor re-
actance for the motor impedance up to the speed corres-
(92) ponding to maximum torque in calculating the factor which
determines ePo in terms of e,. Beyond the slip correspond-
This permits of the calculation of et from eP after the ing to maximum torque, the effective impedance varies
construction has been completed. rapidly to the running impedance. Simultaneously with
During the transition from one operating condition to the increase in impedance the lagging kva likewise drops
the next, only ed’ remains constant; eP changes. It is es- off which results in a considerable rise in voltage. This
sential therefore that ed’ be computed for the initial operat- effect is clearly shown in Fig. 44 taken from some tests
ing condition. The conventional construction shown in made by Anderson and Monteith.20 As running speed is
Fig. 43(c) can be used. This determines the initial value approached the generator voltage rises, the excitation be-
of ed’ for the new operating condition from which the initial ing too high for the particular loading. To form a better
value of eP can be computed by Eq. (90). idea of the magnitudes involved in such calculations, Fig.
Common cases for which these calculations apply are the 45 shows curves of terminal voltage as an induction motor
determination of regulation for loads suddenly applied to equal in horse power to 20 percent of the kva of a generator
a generator. Instances in which this can occur are the is suddenly thrown upon an unloaded generator for differ-
172 Machine Characteristics Chapter 6

problem generally and have plotted the results in curve

form.
The analysis has been carried out for both self-excited
and separately-excited exciters. The results for the former
are plotted in Fig. 46, and for the latter in Fig. 47. These
curves are plotted in terms of the four parameters: (1)
magnitude of load change (2) Xiaat (3) T&,, and (4) rate
of exciter response, R. The response is defined in the
chapter on Excitation Systems. It is shown by Harder
and Cheek22 that variations in x,, saturation factor of the
generator and power factor between zero and 60 percent
have little effect upon the maximum drop. The assumed
value of xd for these calculations was 120 percent. An
Fig. 44—Performance of 3333 kva, 0.6 power-factor, 3600 rpm, accurate figure for maximum voltage drop can be obtained
1.7 short-circuit ratio generator as a single 500-hp induction- for values of xd other than 120 percent by first expressing
motor pump is started. Induction-motor starting torque
reactances and the applied load on a new kva base, such
equal to full-load torque and pull-out torque equal to 2.8
full-load torque. Full lines represent operation with fixed that xd on the new base is 120 percent, and then applying
excitation and dotted lines under regulator control. the curves. For example, suppose a load of 1500 kva (ex-
pressed at full voltage) of low power factor is to be applied
to a 3000-kva generator having 30-percent transient re-
actance and 150-percent synchronous reactance. Suppose
that the generator time constant is 4.0 seconds and the
exciter has a nominal response of 1.0. To determine the
drop, express the transient reactance and the applied load
on the kilovolt-ampere base upon which xd is 120 percent.
The base in this case will be 3000 x 120/150 = 2400 kva.
On this base the transient reactance xd’ is 30 X 2400/3000 =
24 percent, and the applied load is 1500/2400 -62.5 per-
cent. If the exciter is self-excited then from the curves of
Fig. 46, the maximum voltage drop is 15 percent for 62.5-
percent load applied to a generator having 24-percent tran-
sient reactance, a time constant of 4.0 seconds, and an
exciter of 1.0 nominal response. This same maximum drop
10 would be obtained with the machine and load under
0 0.2 0.4 0.8 0.8 1.0
consideration.
TIME -SECONDS
The initial load on a generator influences the voltage
Fig. 45-Terminal voltage of a 500 kva, SO-percent power-
drop when additional load is suddenly applied. As shown
factor engine-type generator Cxd= 1.16, x, =0.59, xd’ =0.30, in Fig. 48, a static or constant-impedance initial load re-
13 percent saturation) as a 100-hp induction motor is con- duces the voltage drop caused by suddenly applied load.
nected. However, a load that draws additional current as voltage
decreases may increase the voltage drop. Such loads will
ent rates of response of the exciter. Ordinarily one is be referred to as “dynamic” loads. For example, a running
primarily interested in the minimum voltage attained dur- induction motor may drop slightly in speed during the
ing the accelerating period and so the calculations have voltage dip so that it actually draws an increased current
been carried out to only 1.0 second. The curves show con- and thereby increases the maximum voltage drop. The
ditions for constant excitation and for exciters with 0.5 and dynamic initial load curve of Fig. 48 is based on an initial
1.0 ratios, respectively. load that draws constant kilowatts and power factor as the
voltage varies.
24. Drop in Terminal Voltage with Suddenly-
Applied Loads
VII. CONSTANTS FOR USE IN STABILITY
When a relatively large motor is connected to a gener-
ator, the terminal voltage may decrease to such an extent
PROBLEMS
as to cause undervoltage release devices to operate or to The stability problem involves the study of the electro-
stall the motor. This situation arises particularly in con- mechanical oscillations inherent in power systems. A
nection with the starting of large motors on power-house fundamental factor in this problem is the manner in which
auxiliary generators. The best single criterion to describe the power output of the generator varies as the position of
this effect when the generator is equipped with a regulator its rotor changes with respect to some reference voltage.
to control the excitation is the maximum drop. The pre- The natural period of power systems is about one second.
vious section describes a method whereby this quantity Because of the series resistance external to the machine,
can be calculated. However, the problem arises so fre- the time constant of the unidirectional component of arm-
quently that Harder and Cheek22,23 have analyzed the ature current is usually so small as to be negligible in
Chapter 6 Machine Characteristics

Fig. 46—Maximum voltage drop of a synchronous machine WITH SELF-EXCITED EXCITER as affected by (a) magni-
tude of load change, (b) x’d sat, CC) T’d,, and (d) rate of exciter response. xd’ on curves refer to saturated or rated-voltage
value. Assumptions used in calculations: xd’= 1.07 xd’ aat; xd = 1.20; x,=0.75; no-load saturation curve/air gap line normal
voltage = 1.2; time lag of regulator =0.05 second; added load is constant impedance of 0.35 pf.; initial load zero.
 Machine Characteristics Chapter 6

--

Fig. G--Maximum voltage drop of a synchronous machine WITH SEPARATELY-EXCITED EXCITER as affected by
(a) magnitude of load change, (b) X’ d lat. (c) Td,,’ and cd, rate of exciter response.
Chapter 6 Machine Characteristics 175

Further consideration of this constant is given in Part

XIII of this chapter.

26. Network Calculator Studies

For most problems the synchronous machine can be
represented by its transient reactance and a voltage equal
to that behind transient reactance. For the rare case for
which salient-pole theory is required, the following pro-
cedure can be followed. It is impossible to set up the two
reactances in the two axes by a single reactor, but if the
reactance, ic,, is used and a new voltage, es& introduced as
representing the internal voltage, both position of the rotor
and the variations in ed’ can be carried through quite
simply.
Fig. 49 shows a vector diagram similar to Fig. 14 in
which e,-,dis included. This voltage is laid off along ed and
f!?d’and terminates at the point a. The reading of power
Fig. 48—Effect of type (whether dynamic or static) and initial at esd is the same as the actual output of the machine. As
load, assumed at 0.88 power factor, upon the maximum volt-
the exciter voltage changes ed’ and f?cldlikewise change.
age drop when l00-percent low-power-factor load is suddenly
applied to an a-c generator.

comparison with this natural period. The subtransient

component is likewise so small that its effects can be
neglected. There remains then only the transient com-
ponents, those components associated with the time con-
stants of the field winding, that are important.

25. Representation of Machine

The transient stability problem is primarily concerned
with the power-angle relations during system swings fol-
lowing a disturbance. Because of the dissymmetry of the
two axes, it is necessary theoretically to take this dissym-
metry into consideration. However, in most cases an
impedance is in series externally to the machine so that Fig. 49—Construction of eqd for network calculator studies.
the difference in reactances in the two axes becomes a
smaller proportion of the total reactance. The results of
calculations presented in Chap. 13 show that for most The incremental changes in esd can be obtained as follows.
practical purposes it is sufficiently accurate to represent From Fig. 49 it is evident that at any instant
the unsymmetrical machine with a symmetrical machine eqd=ed’+ (xq-xd’)id (94)
having the same zd’.
From Eq. (52)
In spite of the close agreement of salient-pole with
cylindrical-rotor results, a few cases arise for which it is ded’
dt’$-pd
necessary to use salient-pole theory. Relations for calcu-
lating the power output have been given in Sets. 16 and 17 and
and for computing the change in internal voltages in Sec.
22(a). It is shown in the latter section that if the exciter Aed’ = d$At = $-(c-eJAt (95)
do
is of the quick-response type, the voltage ed’ can, for all
practical purposes, be regarded as constant. Methods for where &d’ is the increment of ed’ in the increment of time
the inclusion of these factors into the stability calculations At. From Fig. 49 there results also that
have also been treated in Chap. 13. ed=eqd+ (xd-&d
A knowledge of the inertia constant, H, is a requisite for so that

1
the determination of the acceleration and deceleration of
the rotor. It represents the stored energy per kva and can Aed’ = $- ex-eqd- (xd-x&j At.
do
be computed from the moment of inertia and speed by the
following expression In network calculator studies of system stability, Q,
H=0.231WR2(rpm)210d e&, and id are known at any instant. From Eq. (94) it is
kva
(93) evident that the increment of epd is equal to the increment
in ed’. Thus
where H = Inertia
W@=Moment
constant
of inertia
in kw-sec. per kva.
in lb-ft*.
Aeqd=$-
do
e=-ee,d- (xd-x,&d
1 At (98)
176 Machine Characteristics Chapter 6

This method can be applied regardless of the number of of unidirectional and negative-sequence currents. The crest
machines involved in the study. value of the negative-sequence current, iz, is 1/2& and the
To obtain the initial value of eqd, calculate f?d’ from
crest value of i” is l/z II’
the steady-state conditions before the disturbance. ed’ is
the quantity which remains constant during the instant In the case of neztive-sequence currents, part of the
representing the change from one operating condition to loss is supplied by the shaft and part is supplied through
another. The proper eqd is obtained by changing the the armature. The loss associated with the circulating
magnitude of eQd until Eq. (94) is satisfied. currents in the rotor as developed in Section 15 is approx-
To include the effect of saturation, break the reactance imately equal to 2(rz - r@. Assuming for the moment that
X which represents the machine, into two components the loss varies as the square of the current and neglecting
xz’ and (x, -xP), the latter being next to the voltage es& the differences due to the frequencies in the two cases, the
The voltage at the junction of these two reactances is eP, loss for the unidirectional components of current is
the voltage behind x,. The effect of saturation will be
included by adding the saturation factor s taken from the
no-load saturation curve (see Fig. 17) for ePr to the excita-
tion obtained by neglecting saturation. This corresponds
to method (d) of Sec. 3 for steady-state conditions. Eq.
(98) then becomes Actually, however, the loss varies more nearly as the 1.8
power of the current so that the expression becomes
w-2 - r1)
Ir
(xd >

27. Armature Resistance Now considering the effect of frequencies. Since the
For most stability studies the loss associated with the depth of current penetration varies inversely as the
resistance of the armature is so small as to be negligible. square root of the frequency, the resistance varies directly
The exception to this rule is the case for which a fault as the square root of the frequency. The loss for the uni-
occurs near the terminals of a generator. directional component is then
The losses in an a-c generator during a three-phase short
circuit can be large enough to affect significantly the rate 2(r2 fib-2 -4
or (xd”)l.8 l
(101)
at which the rotor changes angular position. This is of fl(xd”) I”

particular importance for stability studies. Two of the

most important factors determining this effect are the loca- 36,
I I I I I I I I I I I I I.‘*
tion of the fault and the value of the negative-sequence
resistance. The latter is difficult of evaluation particularly
for turbo-generators-the type of machine in which the
III I I I I I I I I I I I I
effect is greatest. One must rely almost entirely upon cal- \\I 11 11 11 11 11 11
I --I
culations, which are extremely complicated. For a-c board - .I4
studies of system stability it is convenient to represent the
machine losses by means of a resistance placed in series in
the armature. The value of this resistance should be chosen .I2
so that its loss, with the reactance of the machine repre-
\I L
sented by xi, be equivalent to that of the machine under ‘-

actual conditions. An approximate evaluation of this g2.0 .I0 5

2
equivalent resistance will be developed for a turbo-
s Q
generator.
g 1.6 .08 z
Let the initial value of the subtransient component of a
short-circuit current be designated, i”. The components F
of the unidirectional current have a maximum value fii” 1.2 \I I I I I I I I I I I
.06
and are related in the three phases in a manner as discussed Ill
----. _ --- 11
in Sec. 8. The sum of the unidirectional components in all
three phases produce an essentially sinusoidal wave of mmf .8 f-yy\\I.o4
that is stationary with respect to the armature. This sta-
tionary mmf develops a flux that in turn generates currents
.4 +++-J-I.02
having a frequency of 60 cps in the rotor. This effect is
similar to that produced by negative-sequence currents in
the armature except that the latter produce a sinusoidal 0
<mm. wave that rotates at a speed corresponding to 60 cps D5 .I0 .I5 .20 .25 .30
in a direction opposite to the rotation of the shaft and TIME IN SECONDS
ultimately generates circulating currents in the rotor hav- Fig. 50—Development of req of a turbo-generator for the con-
ing a frequency of 120 cps. The magnitudes of the mmf dition of a three-phase short circuit across the terminals of
waves in the two cases are equal for the same crest values the machine for various duration of the short circuit.
Chapter 6 Machine Characteristics 177

The loss associated with this current is

rl[ (p-)A?+J.J. (104)

To form an idea of the order of magnitudes of these

losses, let
xdf’ = 0.09. T, = 0.09.
x:=0.15. Td=0.6.
xd = 1.25.
r2 = 0.035.
r1= 0.005.
The results of the calculations are shown in Fig. 50. The
upper dashed curve is the loss associated with the uni-
directional component and the lower dashed curve the load
losses. The full line represents the total losses. The cur-
rent flowing in the generator as represented on the board

is constant and equal to -$. The equivalent resistance,

ras, to be inserted in series with xd’ must be such that the

integrated loss over any interval must be the same as that
in Fig. 50. The dot-dash curve in Fig. 50 gives the values
of reel obtained by this method.
Figure 51 gives similar values of r,, for other types of
machines. The curves in Fig. 51(a) were calculated for
short circuits at the terminals of the machines, those in
Fig. 51(b) are for three-phase short circuits across the
terminals of a transformer connected in series with the
machine.

VIII. UNBALANCED SHORT CIRCUITS ON

MACHINES WITHOUT DAMPER WINDINGS
Because of the dissymmetry of salient-pole machines
without damper windings, the armature currents at times
of three-phase short-circuits, as shown in Sec. 12, contain
second-harmonic components. For unsymmetrical short-
circuits, such as from terminal-to-terminal, the wave forms
of currents and voltages become even more complex. Both
odd and even harmonics are present.
Fig. 51—Typical equivalent resistance, r eq, for different types 28. Terminal-to-Terminal Short Circuit
of machines.
In particular consider a salient-pole machine in which
(a) for three-phase short circuit across the terminals used
(b) for three-phase short circuit across the terminals of a series-
saturation is neglected and which is operating at no load
connected transformer of 10 percent impedance. to which a short-circuit is suddenly applied across two
terminals. The short-circuit current5 in these phases is
then
Since the unidirectional current decreases exponentially
with a time constant T,, the loss as a function of time is (105)
i=(59fxd’)+(xq-xd’) cos2(2~f~+~~)
u/z(r, - ?-I)E-lz in which &, indicates the phase position during the cycle
(102) at which the short-circuit occurred.
(xd”) “’ ’
It will be observed that this can be resolved into two
In addition to the losses associated with the unidirec-
components
tional current, the load losses as reflected by rl can also be
significant for a three-phase fault across the terminals. 4 Ifsin (27$++1)
First: (106)
Neglecting the sub-transient component, the a-c com- (zq+xd’) + (“q- xd’) cos 2 (27$+&3)
ponent of short-circuit for a three-phase short circuit from -\/3 If sin $0
no-load is (107)
Second: (xq+xd’) + (x,--xd’) cos 2 (27$+40)
The first component is shown in Fig. 52(a) for a typical
machine and consists of odd harmonics only. The second
178 Machine Characteristics Chapter 6

xd’+m
Thus keedyate = 6If xd+dz

sin (27$+40)
(108)
[(x,+xd’)+(xq- xd’) cos 2 (t-h+@]
With the assistance of Fig. 52 it will be seen from Eq.
(105) that the maximum amount of the odd harmonic
d3 If
component is equal The maximum
to - value of
t&d’ ’
the total current is dependent upon the instant during
the cvcle at which short-circuit
u
occurs and reaches a
43 If
maximum of V.

Assuming no decrement for either the odd or even

harmonics
~3 If sin & l+b2
(109)
.
2rms(even) = - 2 -4s
J l-b2
. If- 1
%ma (odd) (110)
xd’ + ~&-$d’ di-=??

(110

Fig. 52—Armature current and field current in a synchronous

machine when a terminal-to-terminal short circuit is sud-
denly applied.

xd’=o.30 xd’l.1 x, =0.75 (b. = 9o”

component is shown in Fig. 52(b) for &,= +90” and con-

sists of even harmonics only. The latter component is de-
pendent upon the instant during the cycle at which the
short-circuit occurs and may vary anywhere between the
values given and the negative of those values in accordance
with the coefficient, sin 40. Figure 52(c) gives the total
current, the sum of Figs. 52(a) and 52(b).
The units chosen are the p.u. in which for the machine
operating at no-load at rated circuit voltage If would be
equal to 1.0 and in this case the current i is given in terms
of crest magnitude of rated phase current.
The components of armature current shown in Figs.
52(a) and 52(b) have associated with them the field cur-
rents shown in Figs. 52(d) and 52(e), respectively, the for-
mer consisting only of even harmonics and the latter only
of odd harmonics. In Fig. 52(f) is shown the total field
current. The average magnitude of this current is equal to
Xd+4x
~- ---If.
xd’+&.t&’
The odd-harmonic component of field current and its as-
sociated even harmonic in the armature decay to zero with
time. The even harmonics of the field and their associated
odd harmonics of armature current decay to constant, Fig. 53—Wave form of voltage across terminals of a water-
steady-state amounts. Their initial values are in excess of wheel generator without damper windings for a terminal-to-
their steady-state magnitudes by the amount the average terminal short circuit from no-load. xq/xd’= 2.5.

of If is in excess of its steady-state amount, If. The steady- (a) Initial value of odd harmonic component (decays slowly);
state value of i is then equal to the initial amount of the (b) initial value of even harmonic component for sin do= 1 (decays
odd-harmonic component multiplied by rapidly). Its magnitude varies between that given and its
negative depending upon the point during the cycle at which
x/+4= short circuit occurs. It may be zero.
xd + dxqxd” (c) Total initial value for sin +0x 1
Chapter 6 Machine Characteristics 179

The rms total current is equal to the square root of the

sum of the squares of those components. It must be re-
membered that the unit of current is the crest of rated
terminal current. When expressed in terms of the rated
rms current the above figures must be multiplied by 4.
The voltage from the short-circuited terminals to the
free terminal, neglecting decrements, is equal to

ea-eb=eab= --31fK[sin(2~~~+4o)+3b sin 3(%$+&I)

+51,sin 5(27rjt+&)+. . .]
+31r sin (b[2b cos 2(27rfl+&)
+4b2 cos 4(27rj-t+~) + . . 1 (112)
in which
Fig. 55—Equivalent circuit to which Fig. 54 may be reduced.
K=Am
013) b= _ dx&d’-- 1
dXcl/Xd’ + 1
dzJxd’+ 1
and b has its previous significance.
Like the short-circuit current this voltage can likewise
be resolved into two components that together with the at times of unbalanced short circuit, that resonance occur
total voltage are plotted in Fig. 53. The maximum possible between the reactance of the machine and the load
voltage, that, which occurs when sin 40 is equal to unity, is with the possibility that dangerously high voltages might
result. Considering a purely capacitive load such as an
unloaded transmission line, the schematic diagram is shown
eab(maximum for max. flux linkages)
in Fig. 54 and the equivalent circuit in Fig. 55 for the
When sin &=O, the even harmonic component is equal condition of a terminal-to-terminal short circuit. The emf
to zero and for this case the maximum voltage is applied to the circuit is equal to the open-circuit voltage
for the same short-circuit condition. The oscillographic
results of tests made on a particular machine as terminal-
eab(maximum for minimum flux linkages) = (115)
to-terminal short circuits are applied for different amounts
of connected capacitance are shown in Fig. 56. Resonance
The corresponding line-to-neutral voltages for the termi-
nal-to-terminal short-circuit are ?,$ of the above figures. will occur near points for which the quantity Z =n 2
7
In all of these expressions the crest value of rated line-to- dxdlz,

neutral voltage has been used as a base. When the rms where n represents the integers 1, 2, 3, etc., and also the
figure is used, the above quantity must be multiplied order of the harmonic. The nature of this resonance phe-
by 4. nomenon is illustrated more clearly by the curve of Fig. 57,
For a terminal-to-neutral short circuit, neglecting de- in which is plotted the maximum voltage during short-
crements, the short-circuit current is circuit in per unit.
To orient one’s self with regard to the length of line
3lJcos (27rjl+$o) -cost&j involved in these considerations, the figure in miles which
(116)
i=~‘+z,+%l)+i xc+xq) cc% 2(2n;ft+&) appears below each oscillogram of Fig. 56 represents ap-
proximately the length of single-circuit 66- or 220-kv
29. Unsymmetrical Short Circuits Under Capa- transmission line that, with a generator having the char-
citive Loading acteristics of the one used in the test, is required to satisfy
When a salient-pole machine without damper windings the given value of x,/d*. These figures were arrived
is loaded by a highly capacitive load,12* l3 there is danger, at by assuming a generator capacity of 25 000, 75 000, and
200 000 kva for 66-, 132-, and 220-kv lines, respectively.
For smaller machines the length will decrease in proportion.
The possibility of the existence of such resonant condi-
tions can be determined for other types of loads and other
types of faults by setting up the network for the system
and replacing the machine by the reactance xw. This
circuit should be set up for the positive-, negative-, and
zero-sequence networks and the networks connected in
accordance with the rules of symmetrical components. Any
condition for which the impedance as viewed from the
machine is zero or very small should be avoided.
Since the danger of these high voltages arises from the
dissymmetry of the machines, it can be eliminated effec-
Fig. 54—Schematic diagram of a three-phase, salient-pole
alternator to which a three-phase bank of capacitors and a tively by the installation of damper windings. Fig. 58
terminal-to-terminal short circuit are applied simultane- presents oscillographic evidence of the voltages existing
ously. for machines equipped with different types of dampers as
 Machine Characteristics Chapter 6
180

Fig. 56—Effect upon the terminal voltage of varying the shunt capacitive reactance when a terminal-to-terminal short cir-
cuit is applied to a machine without damper windings.
Chapter 6 Machine Characteristics 181

Fig. 57—Experimental values of crest voltages (twelfth cycle)

from terminal a to b when switch in Fig. 54 is closed. Unit of
voltage is crest of terminal-to-terminal voltage before short
circuit. xq/xd’= 2.2. Machine without damper winding.

terminal-to terminal short circuits and capacitive react-

antes are applied simultaneously. While a continuous
or connected damper winding is most effective, a non-
If
connected damper winding having a ratio of 2, equal to

at least 1.35 will be found adequate for practically all

purposes.

IX. DAMPER WINDINGS

The addition of copper damper windings to machines
effectively simplifies the characteristics of the machines
as viewed externally in that harmonic effects are largely
eliminated. However, the addition of other possible cir-
cuits for current flow complicates the internal calculations.
The influence of dampers can in most cases be evaluated
in terms of their effect 14 upon the subtransient reactances Fig. 58-Effect of damper windings.
in the two axes. Terminal-to-terminal short circuit:
(a) No dampers.
30. Types of Damper Windings (b) Connected copper damper.
Damper windings are of several general types. (c) Connected high resistance damper.
Connected Dampers-These consist essentially of (d) Non-Connected copper damper.
windings similar to a squirrel-cage or an induction motor. Terminal-to-neutral short circuit:
They are continuous between poles as shown in Fig. 59 in (e) No damper.
which (a) shows the connection between poles for a slow- (f) Connected copper damper.
speed machine and (b) shows the additional bracing re- (g) Connected high resistance damper.
quired in the form of an end ring for higher speed ma- (h) Non-connected copper damper.
chines. In this type of damper, x~” and xd” have nearly
the same magnitudes. in Fig. 60. They are somewhat cheaper than connected
Non-connected Dampers-The dampers in each pole dampers but at the expense of no longer being able to
face are independent from those in adjacent poles, as shown make x~” and xd” equal.
182 Machine Characteristics Chapter 6

connected in series to slip rings. By connecting a variable

resistor externally to the slip rings the starting character-
istics can be varied at will.
The general characteristics of damper windings will be
discussed under the following heads.

31. Balancing Action and Elimination of Voltage

Distortion
One of the earliest needs for damper windings arose from
the use of single-phase generators and, later, phase bal-
ancers. Such machines if unequipped with damper wind-
ings have characteristics which resemble closely those of
a three-phase machine without damper windings when a
single-phase load is drawn from it. Voltage distortion
similar to that discussed under unbalanced short-circuits
occurs. In addition, if this condition persists the currents
that flow in the body of the pole pieces, produce excessive
heating. The addition of damper windings provides a low-
resistance path for the flow of these currents and prevents
both wave distortion and excessive heating. Because of
the steady character of the load, damper windings in
single-phase machines and phase balancers must be heavier
Fig. 59—Connected damper windings:
than those in three-phase machines.
(a) Slow-speed machine.
The best criteria of a polyphase machine to carry un-
(b) High-speed machine.
balanced load are its negative-sequence reactance and
resistance. The former reflects its ability to prevent un-
balancing of the voltage and the latter its ability to carry
the negative-sequence current without undue heating of
the rotor. These properties are particularly important for
such fluctuating loads as electric furnaces. Not only do
the dampers reduce voltage unbalance but also reduce
wave form distortion.

32. Negative-Sequence Reactance and Resistance

As discussed previously the negative-sequence reactance
and resistance of a machine are both affected by the damper
windings. Table 2 shows the effect of different types of
windings upon a 100-kva generatorI and Table 3 upon a
5000-kva synchronous condenser.14 Both of these tables
represent test results.

TABLE 2—CONSTANTS OF A SYNCHRONOUS GENERATOR AS

AFFECTED BY TYPE OF DAMPER WINDING (100 KVA,
2300 VOLTS, 25.2 AMPERES)

Fig. 60—Two types of non-connected damper windings.

Special Dampers-In this classification fall such

dampers as double-deck windings, which are in effect a
double winding, one of high resistance and low reactance
and the other of low resistance and high reactance. The
principal uses of this type are in motors where the com-
bination provides better starting characteristics. At low
speeds the high reactance of the low-resistance winding
forces the current to flow through the high-resistance wind-
ing, which produces a high torque. At higher speeds the 33. Damping Effect
low-resistance winding becomes effective. Another type of In the early days when prime movers consisted mostly
special winding is one that is insulated from the iron and of reciprocating engines the pulsating character of the
Chapter 6 Machine Characteristics 183

TABLE 3— CONSTANTS 0~ A SYNCHRONOUS CONDENSER As ratio is greater than 1.35 and less than 1.35 add 2 and 3
AFFECTED BY TYPE OF DAMPER WINDING (5000 KVA, percent, respectively, to the price of the machine. In con-
4000 VOLTS, 721 AMPERES)
sideration of the many complicated problems involved in
the selection of a damper winding it would appear, in view
of the low increase in cost of the connected damper, that
if any damper winding is thought necessary, the connected
type should be used.

X. SELF-EXCITATION OF SYNCHRONOUS
MACHINES

torque made parallel operation difficult. This was success- When a synchronous machine is used to charge an un-
fully solved by damper windings in that the damper wind- loaded transmission line whose charging kva is equal ap-
ing absorbed the energy of oscillation between machines proximately to the kva of the machine, the machine may
and prevented the oscillations from becoming cumulative. become self-excited and the voltage rise beyond control.
More recently in consideration of the stability problem The conditions that must be satisfied for this phenomenon
low-resistance damper windings have been advocated for to occur are made manifest by determining the machine
the same reason. While a low-resistance damper winding characteristics for a constant inductive reactive load.
will decrease the number of electro-mechanical oscillations
following a disturbance this effect in itself is not impor-
tant14 in increasing the amount of power that can be
transmitted over the system.
The general influence of damper windings, their negative-
sequence resistance and reactance, and also their purely
damping action, upon the stability problem, is discussed
in more detail in Chap. 13.

34. Other Considerations AfTecting Damper Wind-

ings
Synchronous generators feeding loads through trans-
mission lines having a high ratio of resistance to reactance
tend to set up spontaneous hunting.15 This tendency is
greater at light loads than at heavy loads, the criterion at
which it tends to disappear being when the angle between
the transient internal voltage and the load voltage equals
the impedance angle of the connecting impedance. There
need not be any periodic impulse, such as the pulsating
torque of a compressor, to initiate this phenomenon but,
it may very well aggravate the condition. Damper wind-
ings are very effective in suppressing such inherent, hunt-
ing conditions and also alleviate hunting produced by
periodic impulses, although the latter phenomenon is usu-
ally eliminated by altering the natural frequency of the
system by changing the fly wheel effect of the generator
or motor or both. Synchronous motors connected through
high resistance lines or cables also develop spontaneous
hunting but not so frequently as they are always provided
with a damper winding. Fig. 61—Construction of regulation curves for induction
loading.
Series capacitors in decreasing the effective series react-
ance increase the ratio of resistance to reactance and thus
tend to increase the likelihood of spontaneous hunting. In Fig. 61 the line OAG represents the no-load saturation
In general, where beneficial effects can accrue with the curve. Suppose the machine is loaded with a three-phase
use of damper windings, the benefits are greater for con- reactor equal to x, .ohms per phase. To determine the
nected or continuous dampers than for non-connected regulation curve for this impedance, that is, a curve of
dampers. Mechanically there is no choice as both types terminal voltage plotted against field current, proceed as
can be made equally reliable. The non-connected winding follows: Choose an armature current, such that x,i*, the
lends itself somewhat easier to the removal of a pole but terminal voltage, is approximately rated voltage. This
not to sufficient, extent to constitute a consideration in the voltage is given by the distance BD in Fig. 61. By adding
choice of type to install. A ratio of x~” to xd” as low as *In this discussion, the terminal voltage is regarded as the terminal-
about 1.35 can be obtained with non-connected and 1.1 to-neutral value. When terminal-to-terminal voltage is used the volt
with connected dampers. Damper windings for which this age drops considered will have to be multiplied by fi.
184 Machine Characteristics Chapter 6

to this distance the x,i drop, DA, the voltage behind Potier
reactance denoted by the point A is obtained. The mag-
netizing current to produce this voltage is given by the
distance OB. In addition to this, however, the field current
Ai is required to overcome the demagnetizing effect of the
armature current. For normal current, Ai is the distance
KN in the Potier triangle, OMN. In conclusion, to produce
the terminal voltage F, the field current OC is necessary.
The triangle BAC is a sort of Potier triangle, in which
the Potier reactance is replaced by a reactance equal to
(x,+x,). Thus by drawing any line HG parallel to CA and
GJ parallel to AF, the intersection with the vertical from
H determines the terminal voltage for the excitation H.
When the load consists of balanced capacitors having a
reactance xc in which xc is greater than x,, the impedance
as viewed from the voltage behind Potier reactance is ca-
pacitive and the armature current is magnetizing instead
of demagnetizing. This case can be treated in a manner
similar to that for an inductive-reactance load with some
modifications as is shown in Fig. 62. In this figure the
distance CF represents the terminal voltage produced by
the external drop x$. Since the current leads the terminal
voltage by ninety degrees the voltage behind Potier react-
ance for the assumed armature current is found by sub-

Fig. 63—Regulation curves for constant capacitive load of

such values as to give the loads at rated voltage indicated on
the curves. HG parallel to Oq. Point q represents excitation
and internal voltage, neglecting saturation, to produce rated
terminal voltage with 100-percent capacitive current.

tracting the drop x,i giving the distance CD or BA. To

produce this voltage the magnetizing current OB is re-
quired but since the armature current is magnetizing to
the extent of Ai, the actual field current necessary is only
OC. This determines F as a point in the regulation curve.
For other field currents such as the point H, draw HG
parallel to CA until it intercepts the no-load saturation
curve at G. Then draw GJ parallel to AF. The intersection
with the vertical from H determines the point J.
Fig. 63 depicts the regulation curves for different sizes of
capacitors. The number assigned to each curve represents
the percent kva delivered at rated voltage.
(XC-‘,b. At
The angle cx in Fig. 62 is equal to tan -l Ai

zero excitation it can be seen that if this angle is sufficiently

small, intersection with the no-load saturation curve is
possible, but as a increases a point is finally reached at
which intersection is impossible and the solution fails.
This signifies that when this point is reached self-excitation
does not occur. This critical condition occurs when the
(xc - x,)i
slope equals the slope of the no-load saturation
Ai
curve. In discussing the significance of xd use was made of
Fig. 10, where it was pointed out that DA is the current
Fig. 62—Construction of regulation curves for capacitive necessary to overcome the demagnetizing effect, Ai, of the
loading. armature current. The distance AB is the synchronous
Chapter 6 Machine Characteristics 185

reactance drop xdi and DC the Potier reactance drop.

Thus the slope of the no-load saturation curve is equal to
Xdi - xpi
~ The condition for self excitation is then that
Ai ’
(xc- x,) 2. < bd - x,>i

Ai Ai
or
%cxd (117)
Stated otherwise, the machine will become self-excited if
E2 .
the kva of the machine as defined by G IS less than the
E2
charging kva of the line ;. Since xd is, except for special

cases, of the order of 120 percent, danger may threaten

when the charging kva requirements of the line exceed
approximately 80 percent of the kva of the machine.

XI. CAPACITANCE OF MACHINE WINDINGS

A knowledge of the capacitance to ground of machine

windings is necessary for several reasons, among which are:
(a) Grounding of Generators. This is discussed in con-
siderable detail in the chapter on Grounding. The ca-
pacitance to ground of the windings must be known so
that the associated resistance can be selected.
(b) System Grounding. The capacitance must be
known so that the contribution of this element to the
ground current can be determined for single line-to-
ground faults. The contribution to the fault current for
this condition is equal to t/3 27rjCoE X lo+ where f is the
system frequency, E the line-to-line voltage and CO the
Fig. 64—Capacitance to ground of TURBINE-GENERATOR capacitance per phase in microfarads.
windings for 13 200-volt machines in microfarads per phase. (c) System Recovery Voltage. The capacitance of the
For other voltages multiply by factor K in insert. rotating machines may be an important element in the
determination of the system recovery voltage. It is cus-

Fig. 65-Capacitance to ground of SALIENT-POLE GENERATORS AND MOTORS in microfarads per phase.
 Machine Characteristics Chapter 6

capacitance per phase to ground in microfarads and E is

the applied voltage from winding to ground.
Figures 64 to 67 provide basic data calculated for West-
inghouse turbine generators and salient-pole generators
and motors. The generator data was obtained from refer-
ence 23 and the motor data from some unpublished ma-
terial of Dr. E. L. Harder. This information should be
typical of other machines to within about + 50 percent.
In general, it should be borne in mind that these character-
istics vary greatly between machines of different designs.
Fortunately, however, not very great accuracy is required
for the applications cited above.

XII. NATURAL FREQUENCY OF SYNCHRONOUS

MACHINE CONNECTED TO INFINITE BUS
A synchronous machine connected to an infinite bus
HORSE POWER possesses a natural period of oscillation which is given in
the ASA C50-1943 Rotating Electrical Machinery
Fig. 66—Capacitance to ground of 2300-volt SYNCHRONOUS
MOTORS in microfarads per phase to ground. For voltages
Standards as
between 2300 and 6600, the capacitance will not vary more 35 200 IP, Xf
than f 15 percent from the values for 2300 volt. fp- - cycles per minute (113)
(rpm)d WR2
where P, is the synchronizing power in kw per electrical
tomary to represent the machine capacitance in this
radian displacement,
work by placing one-half of the total capacitance to
f is the system frequency.
ground at the machine terminal. For details of this type
When given an angular displacement, the machine oscil-
of calculation refer to the chapter on Power-System
lates with this frequency and finally subsides unless sub-
Voltages and Currents During Abnormal Conditions.
jected to periodic impulse of proper magnitude. It is not
(d) Charging Kva. In testing the insulation of ma-
within the scope of this work to discuss this subject in its
chines, particularly in the field, it is sometimes necessary
entirety, but merely to derive the above expression.
to know the approximate charging kva of the windings
If an incremental displacement A0 be assumed, the cor-
so that a transformer of sufficiently high rating can be
responding synchronizing power is
provided beforehand to do the job. This is required
either for normal routine testing, for testing at time of AP = P,A8 in kw (114)
installation or for testing after rewinding. The charging and A0 is in degrees. From the Stability Chapter it can be
kva per phase is equal to 27rjCoE2X 10eg where Co is the seen that the acceleration of the rotor is
180 f
a = kvaHAP in deg/seG

Tf
= LAP in rad/sec2 (115)
(kva) H
where the kva refers to the rating of the machine and H
the inertia constant. Substituting H from Eq. (93)
n106 f
a=- AP in rad/se$ (116)
0.231 (WR2)(rpm)2

The sign of P, is actually negative as an increment in A0

produces a torque which tends to return the machine to the
Fig. 67—Capacitance to ground of 2300-volt INDUCTION operating angle. Thus, K is positive. Now
MOTORS in microfarads per phase. For voltages between
2300 and 6600, the capacitance will not vary more than d2(A0)
a=-= -KAB. (120)
f 15 percent from the values for 2300 volts. dt2
Chapter 6 Machine Characteristics 187

Further, let current equal to rated value is called the “rated-current

reactance.”
A0=A sin 27rjnt (121) A knowledge of these two values of xd’ is not sufficient
then substituting this relation into Eq. (120) for all applications for which xd’ is required. The rated-
current xd’, because of lower excitation, lends itself more
- (27rfJ2 A sin 2n-fnt= - KA sin 2rfnt
readily for determination from test. The rated-voltage xd’
from which is that required for short-circuit studies. Saturation within
the machine is a minimum for the former and a maximum
--
fG for the latter. The rated voltage value is sometimes called
n- 2n’
the “saturated value” and is the value usually given by the
Substituting K from Eq. (119) designer. Certain applications, such as stability studies,
demand a quantity determined under conditions for which
fn = sJ& cycles per set (122) the terminal voltage is near rated voltage and the armature
current is likewise near its rated current. Fig. 68 obtained
which converts to Eq. (113). from data presented by Kilgore 16 shows how the reactances
of typical machines of different classes vary if three-phase
short-circuits were applied from rated voltage no load, the
XIII. TYPICAL CONSTANTS AND COSTS
current being altered by introducing different external
Both the voltage and the current at which a machine reactances in the armature circuits. The rated-current
operates affect certain of the principal constants through figure is used as a base for all the curves. The particular
the variability of the permeability of the iron. In this reactance on the curves for rated current is the one that
sense, these so-called constants are not in reality constant. would have greatest utility for stability and regulation
Consider the transient reactance, xd’. If three-phase short- problems. No specific name has been assigned to this
circuits are applied to a machine from no load, the react- quantity.
antes so obtained vary with the excitation. Two of these Similar considerations apply to the subtransient react-
quantities have been given special designations. Thus the antes, with this difference, that the rated-current reactance
reactance obtained when the excitation is such as to pro- xd” is obtained from the same test as that for which the
duce rated voltage at no load before the short-circuit is rated-current reactance of xd’ was obtained. In this case
called the “rated-voltage reactance” and the reactance rated current refers to the transient component and not
obtained when the excitation is reduced so as to produce the subtransient component of current. Fig. 69 shows how
from no load a transient component of the short-circuit

Fig. 69—Saturation factors for subtransient reactance.

“Rated current” value used as base. All reactances from
Fig. 68—Saturation factors for transient reactance. Three- three-phase short circuits without external reactance. Satu-
phase short circuits from rated voltage no load. Current ration factors for salient-pole machine with damper winding
limited by series reactance. is equal to unity.
188 Machine Characteristics Chapter 6

xd” varies with the transient component of current, all

points being obtained from three-phase short-circuits with
no external reactance, the current being altered by the
excitation before the short-circuit.
In general, it is unnecessary to make this distinction for
the negative-sequence reactance. The AIEE code lo sug-
gests determination of x2 by means of the method discussed

Fig. 72—Variation of x, for turbine generators.

Fig. 70—Normal unsaturated transient reactance (au’) for

waterwheel generators.

under Negative-Sequence Reactance, the current during

the sustained terminal-to-terminal short-circuit being
limited to the rated current.
The normal value of x’du designed into waterwheel gen-
erators varies with the kva capacity and speed. These
values are plotted in curve form in Fig. 70. To obtain
lower values than those indicated usually involves an
increased cost.
Fig. 73—Open-circuit transient time constants of a-c gen-
The angular relations within the machine are determined
erators and motors.
to a large extent by x,. The variation, by test, of x, for
several salient-pole machines12*” is shown in Fig. 71.

Fig. 71—x, for salient-pole machines.

A = 7500 kva generator without damper winding.
B= 750 kva generator without damper winding.
C=331 kva motor with damper winding removed.
D= 100 kva generator with damper winding. Fig. 74—Open-circuit transient time constants of turbine
generators.

The zero-sequence reactance, as evidenced by Fig. 72 the wide variation of T dlo with the size of the unit the
taken from Wright’s paper,17 is not affected to any great curves of Figs. 73 and 74 taken from a paper by Hahn
extent in the region for which it has greatest use. and Wagner,18 are also included.
For practical purposes the effect of saturation upon the Table 4 gives both the range of typical constants that
open-circuit transient time constant Tdlo and the sub- are characteristic of normal designs and also an average
transient short-circuit time constant Td” can be neglected. that can be used for general purposes when the specific
In general, Td’ varies l7 in the same manner as xd’, so that value of a particular machine is not known. The negative-
sequence resistance is that obtained at a negative-sequence
the relation Tdf =x6Tdf,, is still maintained. Because of
current equal to rated current. It must be kept in mind
Chapter 6 Machine Characteristics

that the loss associated therewith varies as the second

power of iz for salient-pole machines either with or without
damper windings and as the 1.8 power of iz for turbine
generators. Column (9) in Table 4 refers to the a-c re-
sistance, rl, (which includes the effect of load losses) and
column (10) the d-c resistance, rs.
The inertia constant, H, which is discussed in Chap. 13
is likewise given in Table 4. The general variation of H of
turbogenerators and the corresponding figures for water-
wheel generators are given in Fig. 75. The effect upon H

Fig. 75—Inertia constants.

(a) Large turbine generators, turbine included.
(b) Large vertical type waterwheel generators, including allow-
ance of 15 percent for waterwheels.

of increasing the short-circuit ratio and changing the

power-factor is given in Fig. 76. The WR2 represented
by the curves of Figs. 75 and 76 are those obtained
from a normally designed machine in which no particular
effort has been made to obtain abnormally high H. When
magnitudes of WR2 in excess of these are desired a more
expensive machine results. The additional cost of the
additional WR2 is about proportional as shown in Fig. 77.
The cost per kva of water-wheel generators depends
upon its kva and speed. The extent of this variation is
shown in Fig. 78. Machines of higher short-circuit ratio
or power-factor are more expensive in the proportion shown
 Machine Characteristics Chapter 6

Fig. 79—Effect of short-circuit ratio upon cost (Normal 1.0

short-circuit ratio and 0.8 power-factor used as base).

Fig. 76—Effect of short-circuit ratio upon H.

Fig. 77—Effect of increasing H above the normal values given

by Fig. 75.

Fig. 8O—Cost of synchronous condensers including exciter

and autotransformer.

in Fig. 79. Naturally these figures will vary from year to

year with the cost of materials and labor.
The condenser cost per kva including the exciter, pilot
exciter, and auto-transformer is plotted in Fig. 80. The
exciter kw varies with the size of the unit, ranging from
1.2, 0.7, and 0.32 percent for a 1000, 5000, and 50 OOO-kva
unit, respectively.
The cost of normal exciters for water-wheel generators
varies from 7 to 13 percent of the cost of the generator
alone for slow speeds, and from 2.5 to 6 percent for high
speeds. The larger figures apply for units of about 3000
kva and the smaller figures for machines of about 50 000
kva. Direct-connected pilot exciters cost approximately
30 percent of that of the exciter.

XIV. INDUCTION MOTORS

Fig. 78—Cost of water wheel generators including direct-con- The equivalent circuitof the induction motor is shown
nected exciters only. l-s
in Fig. 31. The loss in the resistor -T, represents the
(0.8 power-factor and 1.0 short circuit ratio) S
(0.9 power-factor and 1.1 short circuit ratio) shaft power and since the circuit is on a per phase basis,
(1.0 power-factor and 1.25 short circuit ratio) the total shaft power is thus
Chapter 6 Machine Characteristics 191

example of this case is the phenomena that occurs during

Total shaft power = in watts (123) the interval between the transfer of power-house auxiliaries
from one source to another. In the second case, the tran-
=A 5(3r,ir2) in hp. (124) sient is determined by reactions involving both the stator
and rotor and the duration is quite short. Examples, of
The rotor copper loss is (3r,ir2). Therefore, neglecting this case, are the sudden energization of an induction motor
other losses, the efficiency is: or sudden short circuit across its terminals.
total shaft power 35 Contribution to System Short-Circuit Current
Efficiency =
total shaft power+rotor copper loss In the calculation of system short circuits only synchro-
l-s nous machines are usually considered but in special cases
S where induction machines constitute a large proportion of
=-=l-ss. (125) the load, their contribution to the short-circuit current
even if its duration is only a few cycles may be large enough
to influence the choice of the breaker from the standpoint
Thus, the efficiency decreases with increasing slip. For of its short-time rating, that is, the maximum rms current
10 percent slip the efficiency is 90 percent, for 90 percent the breaker can carry for any time, however small.
slip the efficiency is 10 percent. Similarly, the rotor copper As a first approximation the short-circuit current sup-
loss is directly proportional to slip; being 10 percent for plied by an induction motor can be resolved into an alter-
10 percent slip and 90 percent for 90 percent slip. nating and a unidirectional component much like that for
The total shaft power can also be expressed in terms of a synchronous machine. The initial rms magnitude of the
torques. Thus, alternating component is equal to the terminal voltage
to neutral divided by the blocked rotor impedance per
Total shaft horse power = ~ I phase. The time constants are namely,
in lb ft> @pm> 8Yn. c1 -s>-
332;oo(f for the alternating component,
(126)
(blocked rotor reactance per phase in ohms) in cycles
Equating (124) and (126), the torque is .
27r (rotor resistance per phase in ohms)
1 (3rrir2) in watts
for the unidirectional component,
T=7.04(rpm),yn.
lb ft. (127)
Sin per unit
(blocked rotor reactance per phase in ohms) in cycles
The equivalent circuit of Fig. 31 can be simplified con- .
siderably by shifting the magnetizing branch to directly 27r (stator resistance per phase in ohms)
across the terminals. The resultant approximate circuit Fig. 82 shows the short-circuit current of a 25-horse-power,
is shown in Fig. 81. This approximation permits of 550-volt squirrel-cage motor. The dotted line in the upper

Fig. 81—Approximate equivalent circuit of induction motor.

relatively simple determination of i,, so that Eq. (127)

becomes

lb ft. (128)

Most transients involving induction motors fall within

one of two categories; first, those in which the machine is
disconnected from the source of power and, second, those
in which the machine remains connected to the source of
power. In the first case the transient is determined largely Fig. 82—Short-circuit currents in armature of squirrel-cage
by changes in magnetization and may be quite long. An induction motor.
192 Machine Characteristics Chapter 6

curve indicates the computed value of the envelope of the

alternating component of short-circuit current. The am-
plitude shows a substantial check but the computed time
constant was low. This can probably be attributed to using
the a-c resistance of the rotor rather than the d-c resistance.
The dotted line in the lower curve is the computed value of
the unidirectional component which checks quite well.
Wound-rotor motors, operated with a substantial amount
of external resistance, will have such small time constants
that their contribution to the short-circuit can be neglected.

36. Electra-Mechanical Starting Transient

Fig. 31 shows the conventional diagram of an induction
motor. In the present discussion the per unit system of
units will continue to be used, in which unit current is the
current necessary to develop the rated power at the rated
voltage. The unit of both power and reactive volt-amperes
will be the rated kva of the motor and not the rated power
either in kilowatts or horse power. This convention is
consistent with the choice of units for the impedances. At
rated slip the volt-amperes input into the stator must be
l-s
equal to unity but the power absorbed in the resistor -T
s *
will be less than unity and will be equal numerically to Fig. 83—Illustrating calculation of speed-time curve of an
the ratio of the rated power of the motor to the rated kva. induction motor upon application of full voltage.
The unit of shaft torque requires special comment. The
shaft power can be expressed as lr
-
l-s S
Shaft Power in kw = kvarated I$,- (129) T in per unit = ei (133)
s ’
(2,+z,)2+ r.+? *
In terms of torque the shaft power is equal to ( S )
Shaft Power in kw The crosses close to the torque curve in Fig. 83 were
computed by this expression.
EO 746T in ib ft 2a(rpm),,ncdl-~s)m (13o)
. In Fig. 83 is also shown the torque requirements of a
33 000 particular load such as a blower. Upon applying voltage
Equating, there results that to the motor the difference between the torque developed
by the motor and that required by the load is the torque
33 000 available for acceleration of the rotor. To convert to accel-
kvarated I$. (131)
in lb ft = 2r(O.746) (rpm) synch. eration it is convenient to introduce a constant, H, which
is equal to the stored energy in kw-sec. per kva of rating
If unit torque be defined as that torque required to pro-
at synchronous speed. H may be computed by means of
duce a shaft power equal to rated kva at synchronous
Eq. (93). WR2 must, of course, include the WR2 of the con-
speed, then from (130), the unit of torque is
nected load.
33 000 Suppose that one per unit torque is applied to the motor
kvarated which means that at synchronous speed the power input
2~(0.746) (w-d synch.
into acceleration of the rotor is equal to rated kva, and
and equation (131) in per unit becomes suppose further that the rotor is brought to synchronous
speed in one second. During this interval the acceleration
T in p.u. = I::. (132) is constant (1 per unit) and the power input increases
linearly with time so that at the end of one second the
For the purpose of determining the nature of electro- stored energy of rotation is (s kva) in kw-sec. Thus 1 per
mechanical transients upon starting a motor from rest, the unit of torque produces 1 per unit of acceleration if the
first step involves the calculation of the shaft torque as a inertia is such that s kva of stored energy is produced in
function of the speed. Either the conventional method of one second. From this it can be seen that if the inertia is
the circle diagram or expression (132) can be used. In such that at synchronous speed the stored energy is H,
using the latter method it is only necessary to solve the then to develop this energy in one second, the same acceler-
network of Fig. (31) and substitute the solution of I, there- ation but a torque 2H times as great is required. Therefore,
from into Eq. (132). A solution of a typical motor is shown there results that
in Fig. 83. For most motors the magnetizing branch can be -- T- TL
(134)
neglected, for which case the torque expression becomes a- 2H
Chapter 6 Machine Characteristics 193

do
Acceleration can be expressed as x and its reciprocal

dt
as do. Thus from (134)

dt
-=- 2H
(135)
do T-T=
This function is likewise plotted in Fig. 83. The utility of
this form of the expression may be seen at once from the
.-I4
fact that E is known as a function of o and the time to
Fig. 84—Decay of residual voltage% of a group of power house
reach any value of o can be determined by a simple auxiliary motors.

t=
integration. Thus

(136)

By summing up areas (such as indicated by the shaded

portion) in a vertical direction, the time to reach any
speed is obtained. The curve of time so obtained is
plotted in Fig. 83.
The following formula can be used to form an approxi-
mate idea of the time required to accelerate a motor,
whose load varies as the square or cube of the speed, to
half speed

Time to half speed= H(zs+zr)2in seconds (137)

w2t
All of the above units must be expressed in per unit. Re-
member also that xB should include any external react-
ance in the stator back to the point where the voltage
may be regarded as constant and et should be that con-
stant voltage.
Fig. 85—Typical time constants for 2300-volt squirrel cage
37. Residual Voltage
induction motors.
If an induction motor is disconnected from its supply, it
rotates for some time, the rate of deceleration being de- appears at the open terminals of the machine. If the volt-
termined by the inertia of its own rotor and the inertia of age source is reapplied when the source voltage and residual
the load and also by the nature of the load. Because of voltage of the motor are out of phase, currents exceeding
the inductance of the rotor, flux is entrapped and voltage starting values may be obtained.

Fig. 86—Approximate variation of price with voltage and horsepower of squirrel-cage

induction motors. These values apply approximately for 8 poles or less for 60-cycle
motors. Most economical used as base price.
194 Machine Characteristics Chapter 6

Figure 84 shows the decay of a group of power-house 9. Two-Reaction Theory of Synchronous Machines, II, by R. H.
auxiliary motors2s. The group had a total rating of 2500 Park, A.I.E.E. Transactions, Vol. 52, June 1933, p. 352.
kw of which the largest was 1250 hp. This curve includes 10. A.I.E.E. Test Code for Synchronous Machines. A.I.E.E. Publi-
cation No. 503, June 1945.
not only the effect of magnetic decay but the reduction in
11. Discussion, by C. F. Wagner, A.I.E.E. Transactions, July 1937,
voltage due to decrease in speed. The open-circuit time
p. 904.
constant for individual 2300-volt machines is given in Fig. 12. Unsymmetrical Short-Circuits in Water-Wheel Generators
85. There is a great variance in this constant between Under Capacitive Loading, by C. F. Wagner, A.I.E.E. Transac-
different designs but these curves give an idea of the mag- tions, November 1937, pp. 1385-1395.
nitude for squirrel-cage induction machines. 13. Overvoltages on Water-Wheel Generators, by C. F. Wagner,
The Electric Journal, August 1938, p. 321 and September 1938,
38. Cost of Induction Motors p. 351.
The price of induction motors of a given rating varies 14. Damper Windings for Water-Wheel Generators, by C. F. Wag-
with the voltage. As the rating increases the most eco- ner, A.I.E.E. Transactions, Vol. 50, March 1931, pp. 140-151.
15. Effect of Armature Resistance Upon Hunting of Synchronous
nomical voltage also increases. To form a basis of judg-
Machines, by C. F. Wagner, A.I.E.E. Transactions, Vol. 49,
ment of the effect of voltage upon size the curve in Fig. 86
July 1930, pp. 1011-1024.
was prepared. 16. Effects of Saturation on Machine Reactances, by L. A. Kilgore,
A.I.E.E. Transactions, Vol. 54, 1935, pp. 545-550.
17. Determination of Synchronous Machine Constants by Test,
REFERENCES
by S. H. Wright, A.I.E.E. Transactions, Vol. 50, 1931, pp.
1. Power System Transients, by V. Bush and R. D. Booth, 1331-1350.
A.I.E.E. Transactions, Vol. 44, February 1925, pp. 80-97. 18. Standard Decrement Curves, by W. C. Hahn and C. F. Wagner,
2. Further Studies of Transmission Stability, by R. D. Evans and A.I.E.E. Transactions, 1932, pp. 353-361.
C. F. Wagner, A.I.E.E. Transactions, Vol. 45, 1926, pp. 51-80. 19. Approximating Potier Reactance, by Sterling Beckwith,
3. Synchronous Machines-I and II-An extension of Blondel’s A.I.E.E. Transactions, July 1937, p. 813.
Two Reaction-Theory-Steady-State Power Angle Character- 20. Auxiliary Power at Richmond Station, by J. W. Anderson
istics, by R. E. Doherty and C. A. Nickle, A.I.E.E. Transac- and A. C. Monteith, A.I.E.E. Transactions, 1927, p. 827.
tions, Vol. 45, 1926, pp. 912-942. 21. Preferred Standards for Large 360%RPM 3-Phase 60-Cycle
4. Synchronous Machines-III. Torque Angle Characteristics Condensing Steam Turbine-Generators, AIEE Standards Nos.
Under Transient Conditions, by R. E. Doherty and C. A. 601 and 602, May 1949.
Nickle, A.I.E.E. Transactions, Vol. 46, 1927, pp. 1-14. 22. Regulation of A-C Generators With Suddenly Applied Loads,
5. Synchronous Machines, IV, by R. E. Doherty and C. A. Nickle, by E. I,. Harder and R. C. Cheek, A.I.E.E. Transactions, Vol.
A.I.E.E. Transactions, Vol. 47, No. 2, April 1928, p. 457. 63, 1944, pp. 310-318.
6. Synchronous Machines, V. Three-Phase Short Circuit Syn- 23. Regulation of A-C Generators with Suddenly Applied Loads-
chronous Machines, by R. E. Doherty and C. A. Nickle, II, by E. L. Harder and R. C. Cheek, A.I.E.E. Transactions, 1950.
A.I.E.E. Transactions, Vol. 49, April 1930, p. 700. 24. Practical Calculation of Circuit Transient Recovery Voltages,
7. Definition of an Ideal Synchronous Machine and Formula for by J. A. Adams, W. F. Skeats, R. C. Van Sickle and T. G. A.
the Armature Flux Linkages, by R. H. Park, General Electric Sillers, A.I.E.E. Transactions, Vol. 61, 1942, pp. 771-778.
Review, June 1928, pp. 332-334. 25. Bus Transfer Tests on 2300-Volt Station Auxiliary System, by
8. Two-Reaction Theory of Synchronous Machines-I, by R. H. A. A. Johnson and H. A. Thompson, presented before AIEE
Park, A.I.E.E. Transactions, Vol. 48, No. 2, July 1929, p. 716. Winter Meeting, Jan. 1950.