The Reactances of Synchronous Machines

BY R. H. PARK1

B. L. ROBERTSON1

and

Associate, A. I. E. E.

Associate, A. I. E. E.

Synopsis.-Until somewhat recently, synchronous machine

theory has been satisfied with a relatively few characteristic constants,
or reactances, in terms of which the behavior of machines has been
calculated. Present theory, however, reqtires many more coefficients. There are now generally recognized two values each of
leakage, synchronous, and transient reactance wt,hich correspond
to the two symmetrical axes of magntetization of the armature current
and which refer to balanced operation. Negative and zero phasesequence reactances are also employed to determine operation under
unbalanced conditions, and it is possible and desirable to distinguish
other reactances. In view of the increasing complexity of the subject
it is felt that a critical survey of it is in order and the object of the
paper has been to provide that survey.
The paper has been divided into two parts. Part I is descriptive,
and treats the subject with regard to thosefactors which are important
to application or operating engineers, and to designers. In particular, the major types of reactances which include the synchronous,
transient, and phase-sequence reactances, are discussed. These
quantities are defined and their methods of test outlined. It appears
necessary to consider a second type of transient reactance, namely
sub-transient reactance. Both reactances may be determined from
short-circuit oscillograms as illustrated in the paper. A table is
included which gives the numerical range of reactanzces for the
various types of synchronous machines.
Part II discusses the theoretical considerations, with a view to
broadening and classifying existing conceptions of reactance. It
includes the effect of external reactance on negative phase-sequence

reactance, and the variation in this latter quantity depending upon

whether current or voltage is impressed on the machine. An important aspect of the division of synchronous reactance into armature
reaction and leakage reactance is discussed. Transient reactance is
shown to be the difference between synchronous reactance and the
ratio of the mutual reactance between armature and field and the total
field reactance. Calcutlations are included to show that the shortcircuit and open-circuit time-constants are related to each other in a
simple manner.
The appendixes cover the followting subjects:
a. Application of the Principle of Superposition to Synchronous
Machine Analysis.
b. Replacing the Effect of Induced Field Currents by Employing
Transient instead of Synchronous Reactances.
c. Significant Rotor Circuits in Addition to the Main Field
lVinding (which affect transient reactances).
d. The Negative Phase-Sequence Reactance of a Synchronous
Machine with Negative Phase-Sequence Voltage Impressed.
e. Construction of Equivalent Circuits: Concept of Field Leakage

THE REACTANCES OF SYNCHRONOUS MACHINES

IN the course of the development of the theory of
synchronous machines there has come into use. along
with an increased knowledge of the behavior of such
machines under varying conditions of operation, a correspondingly increased number of characteristic constants
or coefficients referring generally to reactive voltages,
in terms of which this behavior is calculated.
Thus, in place of the concept which was formerly held
of a single leakage reactance and armature reaction,
there are at present generally recognized two values
each of leakage, synchronous, and transient reactances,
corresponding to the two symmetrical axes of magnetization of armature current. In addition to these
reactances, which refer to balanced operation, there are
also values of negative and zero phase-sequence
reactance, which are employed to determine operation
under unbalanced conditions. It is also possible and
desirable to distinguish other reactances, particularly
those which obtain with the rotor stationary, and those
which are encountered by harmonic currents. At the
same time, improvements have recently been made in
methods of calculating leakage reactance and armature
reaction. Therefore, on account of the increasing

complexity of the subject it is felt that a critical survey

of it is in order. The object of the paper has been to
provide a survey of this character which would be of
value both from the standpoint of system analysis and
from that of fundamental theory.
Since the requirements of these two points of view are
to a considerable extent distinct, the paper has been
divided into two parts. Part I treats the subject with
regard to those factors which are important to designing
and application engineers, and to operating engineers.
Part II discusses in detail the theoretical considerations
involved, with a view to broadening and classifying
existing conceptions of reactance.
PART I. REACTANCES PARTICULARLY AFFECTING
OPERATION
In the analysis of system stability, and in the calculation of the effect of short-circuits, the factors of
interest to operating engineers are those which relate
to the behavior of the machine as viewed from the
armature terminals. The most significant of these
factors are the armature reactances of machines to
normal frequency current having any distribution
between phases, any power-factor, and whether transient or sustained; also, in connection with transient

Reactanlce.
f.
g.
Field
h.
i.
j.

Calculation of Total Field Reactance.

Relation of the Mutual Reactance Between Armature and
to the No-Load Excitation Current.
Relation Between Three-Phase and Single-Phase Reactances.
Discutssion of the System of Notation Used in the Paper.
Per-Unit Representation of Quantitities.
* * * * *

1. Both ofthe Geni. Engg. Dept.,General ElectricCo.,Schen- components of current, their rates of decay, or
d
e c r e
m
e
n t s.
eatady, N. Y.
Types of Reactances. The major types of armature
Presented at the Wrinter Convention of the A. I. E. E., New York,
Nr. V., Feb. 1S-17, 1928.

reactances then are to be distinguished according to:

514
28-521

Feb. 1928

PARK AND ROBERTSON: SYNCHRONOUS MACHJNEt;

A. Distribution; that is, the relative distribution of

current between phases.
Any distribution of armature current may be
expressed as the superposed sum of three symmetrical
components:'
a. Balanced three-phase currents of normal phase
rotation, or positive phase-sequence,
b. Balanced three-phase currents of reverse phase
rotation, or negative phase-sequence, and
c. Balanced three-phase currents of equal time
phase, or zero phase-sequence.
B. Method of application in time of positive
phase-sequence currents.
Here it is desirable to distinguish between:
a. Steadily applied, or sustained currents,
b. Suddenly applied, or transient currents.
In the case of armature reactances these differences
are due to the transient currents induced in the rotor
circuits when armature current is suddenly applied.
if, as is generally so, there are closed circuits on the
rotor in addition to the field winding, as for example an
amortisseur, it is found that some of the currents in the
rotor circuits may die away very rapidly. In order to
distinguish between the effect of these currents and the
effect of those which die away slowly, it is desirable to
establish two (or more) transient reactances. The
higher reactance, i. e., the reactance which depends on
currents that die away slowly, is then referred to as the
transient reactance of the machine. The lower reactance
may be referred to as the sub-transient reactance on
account of its lower value.
C. Position of the rotor with respect to axis of
magnetization of positive phase-sequence currents.
When the rotor is moving synchronously, the positive
phase-sequence current can be resolved into two
components, one of which magnetizes in the axis of the
poles, and the other in the inter-polar space. Accordingly, these components are referred to as direct and
quadrature,4 and the corresponding reactances are:
a. Direct, or
b. Quadrature.
Thus, to summarize, the types of armature reactance
which have been considered so far are to be distinguished according to whether they are:
A. Positive, negative, or zero phase-sequence,
B. Sustained, transient, or sub-transient,
C. Direct or quadrature.
The determination of a suitable notation for these
reactances should depend more upon present and future
requirements than merely upon previous practise.
After careful consideration of the subject, it seemed
that a consistent notation could be obtained by denoting the various types of reactance according to the
method outlined below:
a

3. Reference (11).
4. A considerably broader conception of direct and quadrature
quantities is developed in Part II.

515

A. Distribution.
Armature reactances
Positive phase-sequence xi, or no subscript
Negative phase-sequence x2
Zero phase-sequence x0
B. Application in time.
Sustained No special indication.
Transient One prime, i. e., x'
Sub-Triansient-Two primes, i. e, x"
If it is desired to distinguish other degrees of transiency, additional primes may be added.
C. Position of rotor.
Direct Xd
Quadrature xq.
The various armature reactances of the types discussed are then represented by the notation shown in
Table I.
TABLE I

ARMATURE REACTANCES

Direct synchronous, positive phase-sequence.

Xd' Direct transient, positive phase-sequence.
xqI Quadrature transient, positive phase-sequence.
Xd' Direct sub-transient, positive phase-sequence.
x
Quadrature sub-transient, positive phase-sequence.
phase-sequence.
X2 Negative
Zero phase-sequence.
x0
Definitions and Methods of Test. Synchronous reactance: The problem of determining the relations between
fundamental components of armature voltage and
current during steady operation has been thoroughly
analyzed by Blondel,5 Arnold,6 and Doherty and
Nickle.7 The accepted theory may be briefly summarized as follows:
First, the balanced three-phase system of armature
currents is resolved into two component three-phase
systems: one in which the current in each individual

Xd
x4

Quadrature synchronous, positive phase-sequence.

phase reaches a maximum at the instant that the axis

of the field pole coincides with the axis of magnetization
of the phase under consideration; and another in which
the current in each individual phase reaches maximum
at the instant the axis of magnetization is in line with
the axis midway between poles, that is, one-quarter
cycle later. The former is called the direct component
because it produces direct component of armature
reaction. The latter is the quadrature component.
Then coefficients are defined expressing the ratio of
reactive component of voltage to armature current for
each type of current. These coefficients are the
direct and quadratu7re synchronous reactances of the
machine.
These reactances may be expressed either in ohms,
or as a ratio of their reactance in ohms to normal
ohms, where normal ohms is the ratio of normal voltage
5. Reference (13).
6. Reference (12).
7. Reference (8).

Transactions A. T. E. E.

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

516

and current. When so expressed, they are referred to In the figures:

e () = normal armature voltage.9
as per-unit reactances.8 Thus:
i
(o) = normal armature current.
reactance in ohms
at no-load.
to produce
Per-unitreactance
= normal ohms I(o) == field
Per-unit reactance
flux
to forcee(0)normal
required
currentrequired
fieldcurrent
across the air-gap.
= field current required to produce normal
armature current at short-circuit.
Consequently, per-unit direct synchronous reactance is Synchronous reactance, in accordance with established
defined as the per-unit fundamental component of practise, is understood to be calculated on the basis of
reactive armature linkages, due to unit sustained direct no saturation, and all quantities are referred to the airgap line. Hence, the direct component of synchronous
reactance is measured as the ratio of I, and I,, i. e.,

Normal ohms =

normal line-to-neutral voltage

.

Ic

e(,)

Ig'

/,

:1(.)

|
/!

Iz

The quadrature synchronous reactance, x0, of a threephase machine may be determined by the following
method."0 The machine, coupled in electrical quadrature to a relatively very large synchronous motor and
connected to the motor lines, is operated unexcited.
The ratio of the armature voltage, e, to the corresponding armature current, i, is the quadrature synchronous

FIG. 1-SATURATION CU.RVE

component of armature current. Quadrature synchronous reactance is defined similarly. The resultant
terminal voltage is then found by subtracting the
reactance drop in each axis from the no-load terminal
voltage corresponding to the existing field current,
_
assuming no saturation. Hence, if the machine is dead
FIG. 3 SLP METROD"OF OBTAINING Zd AND .Cq
short-circuited at normal voltage, the reciprocal of the
per-unit sustained armature current is equal to the reactance in ohms. Expressed as a per-unit quantity,
it is.
per-unit direct synchronous reactance.
e
*
x,= t .ee(0)

~ (o)-

~~~

The disadvantage of this scheme is the limitation it

imposes on the sizes of the machines which may be
tested, due to the fact that the driving motor must be
/
ilo)
of sufficient size so that its angular displacement will
not be appreciable.
A method, referred to as the "slip method," affords,
however, a practical means of obtaining xq, and may
/ IIZ
also be used to obtain Xd. The machine to be tested is
left unexcited and a three-phase voltage is applied to
its armature. The rotor is coupled to a driving motor,
of sufficient capacity to overcome the reluctance torque
at the reduced voltage, and is run at a low value of slip.
In this manner the poles are slipped past the m. m. f.
Synchronous reactances can be measured accurately wave. The magnetizing current is then a function of
by test. Thus, direct synchronous reactance, Xd, is the rotor position with respect to the m. m. f.
The oscillogram of Fig. 3 is the result of such a test.
conveniently determined from the saturation and
synchronous impedance curves of the machine, which
normal
refers to used
parentheses,
9. The
represent.
2, respectively,
respectively, represent.
and 2,
Fsynh1
(0),beenclosed inwith
Figs. I and
to
the
andsubscript,
should not

8. The desirability of using the term per-unit to characterize

reactance, when expressed in terms of unity as a base, is discussed
in Appendix J.

values,

confused

subscript, 0,

indicate zero phase-sequence quantities.

10. Two alternative methods of obtaining x, are given in References (3) and (7) respectively.

Feb. 1928

PARK AND ROBERTSON: SYNCHRONOUS -MACHINES

It will be noticed that both the current and voltage are

modulated harmonically as the pole position changes,
the variation in voltage being due to reactance drop in
the source of power employed.
When the poles line up with the axes of the phases,
position (a) on the above curve, the current is a minimum. At this position the ratio of voltage to current
is the direct synchronous reactance. Similarly, when
the axes of the phases are midway between the poles,
position (b), the ratio of voltage to current is the

quadrature synchronous reactance.

This last means of testing for the synchronous
reactances is simple and has several points of advantage.
Thus, the power supply to the tested machine can be
small, just enough so that the meter readings are
easily obtained. Corresponding readings, or oscillograms, of voltage and current can be taken, and the
calculated reactances plotted as a function of time or
pole position. Also, the driving motor need only be
of sufficient capacity to supply the reluctance torque.
Moreover, if the motor speed is well regulated, so as to
produce a low value of slip, there will be no difficulty
in taking readings directly from indicating instruments,
since the swings of the meter needles can be easily
followed. In some cases, it may be desirable to take
oscillograms of both voltage and current.
Transient reactance. When a machine is subjected to
a three-phase short-circuit from an initial condition of
no-load, the flux-linkages in every rotor circuit must
initially stay constant. But, since the armature current
tends to demagnetize these circuits, it is necessary for
the currents in them to increase, in order that the condition of constant flux-linkages in each circuit may be

fulfilled.

517

to the field, this initial value may not satisfactorily

represent the performance of the machine on account
of the fact that the current induced in some or all of
these additional circuits may die away very rapidly.
It is, therefore, desirable to establish the conception of
transient reactance proper, as the value of apparent
transient reactance which applies to the current variation after the rapidly decaying components of current
have died away. It can be obtained, as shown in Fig. 4,
by projecting the envelope of the current wave to the
instant of short circuit, neglecting the first one or two
peaks. Point (b) of Fig. 4 corresponds to the transient
reactance proper. The reactance determined by the
projected value of the envelope of the first few peaks of
the current wave, point (a), is then referred to as the
sub-transient reactance. In practise, the low value of
this reactance may be due in part to saturation in the
leakage paths.
Practically, the value of short-circuit current can be
expressed as the sustained value plus one or more components, each of which die away exponentially. Tran-

FIG.

4-DETERMINATION OF Xd' AND Xd

PROJECTION

BY

METHOD OF

This increase in rotor m. m. f. is responsible for the sient reactance proper is then the reciprocal of the cur-

familiar fact that the initial short-circuit current of a rent that would exist at t = 0 i- the high-speed terms
machine is greater than that obtained under sustained were absent. These relations may be made clear by
conditions, after the induced d-c. currents in the field reference to Fig. 5, which shows the observed values of
and additional rotor circuits have died away. The symmetrical component of armature current from a
resultant armature current is calculated in terms of the three-phase short-circuit test on a large water wheel
voltage before the short-circuit, as the ratio of that generator. The test was taken at reduced voltage so
voltage to a value of reactance referred to as the as to exclude saturation, and the per-unit values of
transient reactance of the machine.
current have been multiplied by a factor so as to refer
More specifically, this reactance is, of course, the to a test at normal voltage without saturation.
direct component of transient reactance, since it inThe per-unit current is expressible in this case as,
i = 1.54 + 1.83 e-653t + 1.06 e-'" ,
volves only reactive or direct axis component of current. Thus, direct transient reactance, Xd', iS
the term with high decrement factor being due
to the presence of a copper short-circuiting collar at
e
Xd
the top of the pole. The values of direct transient
reactance are then,
where e is the voltage preceding short-circuit and i is
1.0
the symmetrical component of armature current just Xd' = direct sub-transient = 1.54 + 1.83 + 1.06
after the short-circuit.
= 0.226
The exact interpretation of this definition would
imply that the current i is to be measured as the value and,
of the envelope of the wave of symmetrical component
1.0
of current, as projected to the instant of short circuit. Xd' = transient = 154 + 183 = 0.297
But, in machines with closed rotor circuits in addition

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

518

Transactions A. I. E. E.

So tar, transient reactance has been interpreted only

in terms of the no-load short-circuit current of the
machine. But, clearly, it may also be regarded as the
reciprocal of the ratio of suddenly applied balanced
current, to the reactive voltage due to this current.
It is, therefore, similar to synchronous reactance,
except referring to suddenly applied instead of sustained currents. From this standpoint it will of course
have both a direct and quadrature value, each of which
in turn can be transient or sub-transient. If, however,
as is often the case, there is no really effective winding
in the quadrature axis of the machine, the transient
reactance in this axis should be taken equal to synchronous. Thus, the term sub-transient reactance will
always refer to the "high-speed" phenomena in either

defined as the per-unit value of positive phase-sequence

voltage, due to the sudden application of a positive
phase-sequence current of normal amount; while
transient reactance is the corresponding apparent value
that obtains after the "high-speed" transients have died
out. The four coefficients will, therefore, be
Xd' direct transient reactance
xq' quadrature transient reactance
Xd" direct sub-transient reactance
x q"' quadrature sub-transient reactance
It has been pointed out that the direct sub-transient
reactance can be determined from short-circuit tests.
It is also possible to measure it by suddenly opening up a
sustained short circuit, that is, applying the equivalent
of a negative current equal in magnitude to the sustained current. Also, either direct or quadrature
- I - t - cL1bransient reactance can be found by impressing a direct
<
I;
;FI
-- -_g_
_ l _ ,or quadrature voltage, respectively, on the terminals,
I-directly or through a reactance. In all of these cases
--__ _
-jLL- - the ratios of voltage to current, as shown on oscillo4------- - ---{
grams, vary progressively from an initial value equal to
14-t-t tt+-lt
+t- .t -tl1

sub-transient reactance, through transient, to a final

equal to synchronous reactance.

X < A sTvalue
-1 - !-..1;
Another method which is sometimes useful in deter-

)z-3
T -> <

X r- p
Iiit1< -;
+4t-l
_ __i-4 7it _ _
- 2 0
-L t =

1
X X
tt-t

-X
| X---

mining Xd' is to determine oscillographically the timeconstant of the decay of armature voltage and current,
respectively, with the armature open-circuited and
If these time-constants are respecshort-circuited.
11X
X
tively To and T0', Xd' may be found from the relation"1

-1r 1-;S : tt r [ + provided thatX dis already known.

t I I I I \ I tA convenient method, which may be used in testing

for both Xd"'2 and xq"13 is to apply a single-phase line-toXL1L -XIl
-l l I_
line voltage to the machine, the rotor being held still.
L
0o
I_ 1.0 - 1.5. J X
0
'0-5
The field is shorted through an ammeter and the rotor
SECONDS
FIG. 5-DECREMENT OF SYMMIETRICAL COMPONENT OF moved until the meter reading is a maximum, at which
CURRENT THREE-PHASE SHORT CIRCUIT ON A 20,000-KV-A. position the rotor is directly under the center of the
WATERW-HEREL GENERATOR
phase. The direct sub-transient reactance, in ohms,
the ratio oft the armature. voltage to the
is
axis. Moreover, if, as in the case of the machine to
which the data of Fig. 5 refer, there happens to be no
The pole is moved slowly until the field ammeter
rotor winding in the quadrature axis, the sub-transient
in reads zero, showing that the pole is in quadrature with
quadrature reactance will also equal the synchronous
axis of magnetization. Then,
the quadrature subtthe
transient
reactance, in ohms, is again one-half14 the
ratio of the armature voltage to the armature current.
xq" = Xq1 = Xq.
On the other hand, for machines equipped with amorFor delta-connected machines the line-to-line
tisseurs, the value of the sub-transient quadrature reactances are similarly divided by two and the equivareactance will be different from the corresponding syn- lent Y-circuit used.
11. See Part II.
chronous value, although, as before,

t--tl

I-

one-half'4

quadrature reactance,in whil calsoequal

'= Xq

12. If there are no auxilia,ry windings on the rotor this reac-

since there is no really effective winding in this axis. tance is transient, i. e., xci'. W=ithout such windings the method
In practise, an amortisseur always has a decrement

may also be used to obtain synchronous reactance by leaving the

suffcietlyighsotat ts ractnce s sb-trnsint.

13. If amortisseur windings are not present this reactance
Thus, the per-unit sub-transient reactance of a is the quadrature synchronous reactance.
14. See Appendix H.
machine, in either the direct or quadrature axis, is

Negative phase-sequence reactance. In the case of a

machine rotating synchronously the application of
fundamental negative phase-sequence current gives
rise to negative phase-sequence fundamental and
positive phase-sequence third harmonic voltages from
line-to-line and from line-to-neutral. The per-unit negative phase-sequence reactance of a machine is thus equal
to the per-unit fundamental phase voltage, or the perunit line-to-line voltage, due to normal negative phasesequence current supplied at the machine terminals.
On the other hand, if the rotor is not moving exactly
synchronously the line-to-line voltages will not be
purely fundamental, but, as is shown in Part II, theywi1l
contain slip frequency components. On this account,
the direct determination of negative phase-sequence
reactance by test is often inconvenient.
However, it can be determined very easily by static
test since the negative phase-sequence reactance of a
machine is very nearly equal to the average"5 of Xd' and
Xq/ i. e.,
Xd' + XQ
X2

where a is the sustained component of current, and b, c,

etc., are the transient components of current with timeconstants Tb, Tc, etc., respectively. In practise all but
one of these time-constants will be small, i. e., less than
about one-twentieth sec., while for dead short-circuits
the remaining time-constants will vary in large machines
from about 0.5 to 2.5 sec.
Thus, a few cycles after the occurrence of a short,
from an initial condition of no-load and voltage e, the
armature current will be
1
1
r 1
-+
i = e
Xd
Xd
LXd
where To' is the short-circuit time-constant of the
machine.
If the machine is operating at no-load and its field
winding is short-circuited, the variation of armature
voltage, shortly after the beginning of the transient, will
follow a decrement having a time-constant To, which is
larger than the short-circuit time constant To'. These
time-constants are related by the simple expression

Zero phase-sequence reactance. The application of

fundamental zero phase-sequence current to a synchronous machine gives rise to pulsating third harmonic
m. m. f. in the air-gap and end-windings, and to a slot
flux the magnitude of which varies widely with the winding pitch. Thus, with other than full pitch coils, the slot
flux is diminished by the presence, in the slots, of coil
sides carrying current in opposite directions. Zero
phase-sequence slot reactance is thus very sensitive to
pitch, and because of the fact just previously mentioned
is a minimum at 2//3 pitch. The zero phase-sequence
air-gap leakage also varies accordingly and disappears
at 2/3 pitch, since in this case there is no air-gap m. m. f.
There is no armature reaction m. m. f. due to zero
phase-sequence currents. Hence zero phase-sequence
reactance is very small, say, from 15 to 60 per cent of the
direct sub-transient reactance. The effect of the
motion of the rotor is very small, and, as a consequence,
there are no appreciable harmonic voltages. Per-unit
zero phase-sequence reactance is defined as the per-unit
phase voltage with normal zero phase-sequence current
applied. While its definition implies a condition under
which the rotor is moving at normal speed, the difference
in test result, if the rotor is stationary, is inappreciable.
DECREMENTS
.
Symmetrical component of current. As pointed out in
the discussion of transient reactance, the decrement

curve of the transient armature current of a machine on

short-circuit will, in general, be composed of several
simple decrement terms, or exponentials. That is, the
current will be expressible as a series
_ __

15.

519

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

Feb. 1928

= a+b e Tb +cC
-See Appendix D.

T2c +

To' =-

Xd_

To.

As it happens that the open-circuit time-constant of

most large machines is about five sec., the above relation provides an easy means of determining Tot when the
other two quantities are known. The time constant
T for a short circuit through an external reactance x is
then given as
XdI + x
To.
T=Xd + X
This also applies to the case of single-phase short circuits. Thus, on single-phase short circuits, the positive
phase-sequence component of current is to be calculated
as the current that would exist when an equivalent16
three-phase short circuit is applied.
In the case of short circuits under load the timeconstants of the rotor circuits in the quadrature axis
are also involved. Except for solid rotor turbo alternators, these are all so fast, however, as not to merit
attention from an operating standpoint.
The case of turbo alternators is also complicated
to some extent by the large amount of saturation which
exists in the rotor leakage paths. The effect of saturation in such paths is not confined to turbo alternators
to some extent in short circuits
but is encountered
only,
of
salient
pole machines. Its effect does not, however,
greatly modify the general conclusions stated.
16. The equivalent reactance of a line-to-line short is the negof short,
thethepoint
reactance
ative for
phase-sequence
and
negative
sum of
iS the from
shorts itviewed
line-to-neutral
while
phasezero
zero phase-sequence reactances. The negative and
sequence currents can, of course, be found by applying the con-

~~~~~~~ditions which hold at any instant,-that for line-to-line faults, the

positive and negative components of current are equal in the

and that for line-to-neutral faults, all three components

~~~~~~~~~~~~~~~fault;
are equal in the fault.

Transactions A. I. E. E.

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

520

D-c. component. On dead short circuits, the timeconstant of the d-c. component is
1

T=

vl

_X2

_l_._

r seconds

while if the machine is short-circuited through an

external reactance, x, the time-constant is
1

X2 + xO

1w7rfx2f +
r

22
c

_-

seconds

where r is the armature resistance as a per-unit

quantity.
The time-constant on single-phase short-circuit is
calculated as for the equivalent three-phase shortcircuit.
Range of magnitude of reactances and time-constants.
It is noticeable that in any certain type of machine the
average value of each of the particular reactances
appears to be a rather constant quantity, the deviation
from which is not very great. Thus it is possible to
obtain tables showing the general limits and the average
values of the reactances, such as Table II.
The first figures in any group of three quantities
indicate the lower limit, the second figures are the
average, and the third figures, the upper limit. It is
understood that they are not absolute values, but are
representative of most machines. Where only two
quantities are mentioned, the lower and upper limits are
meant. If only one term is given it is the average

--I

cl

_
n

cc

to

CC

C1

O C

LO

o
v

E-,;|Z|

reactance.

PART II
THEORETICAL CONSIDERATIONS
The general problem of calculating the performance
of a synchronous machine does not admit of any but
the most difficult forms of analysis, unless the effects of
saturation and hysteresis are neglected. Therefore,
they are not considered in the analysis which follows.
Practically, these effects are small in a large number of
cases. In those cases where their consideration be|
comes necessary, the final result, or the coefficients, may
be "shaded."
Since saturation and hysteresis are neglected, linear
relations exist between current and magnetic flux. It
then follows, as is shown in detail in Appendix A, that
for a given motion or position of the rotor, the currents
and voltages in any part of the machine, due to any
currents or voltages impressed at its terminals, may be
found by superposing the effects that would result from
the action of each separately. Thus, there will exist.
factors of proportionality between the voltage or current
impressed at any terminal, and the voltages and currents at all other points of the machine.
For instance, if the armature of the machine iS opencircuited, a steady voltage impressed on the field will
give rise to a steady direct current in the field, and to
fundamental and harmonic voltages at the armature iaW a
terminals. On the other hand, if, say, one phase of the

1~ C 1:
N

00

oo

|
z

I*
N j
--

..

|
X

..~
Q ,
,
S .t,t
;=@
m

.
*lc
;at .
*

Feb. 1928

521

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

armature is short-circuited, a steady voltage applied to

the field will give rise to steady and harmonic currents
in the field, and harmonic currents and voltages,
respectively, in the shorted and open armature phases.
The ratios of these various currents or various voltages
to the impressed field voltage may be referred to as the
characteristic coefficients of the machine with reference
to the particular condition of operation and the particular terminal quantity impressed, which in the illustrative case under consideration happens to be field

voltage.
In general, characteristic coefficients can be established to cover any combination of impressed and resultant quantities, of any frequency, in any of the four
distinct terminal circuits17 of the machine comprising
the three armature phases and the field winding.
These coefficients may be further distinguished from
each other according to the rate and time-phase of
application of the impressed quantities, the character
of the terminal circuits on which these quantities may
be impressed, the motion and position of the rotor, and
the time-phase of the resultant quantities relative to the
impressed quantities.
The general problem of classifying and establishing a
nomenclature for all of the various types of coefficients
would be difficult, especially as there is no generally
accepted name for the voltage, due to a voltage, or the
current due to a current.18 Moreover, the necessity
for a careful study of the broad subject is not pressing
at present, although it is anticipated that as progress
is made in the analysis of machine performance, the
recognition and employment of an increasing number of
characteristic coefficients will take place.
However, the specific problem of classifying and
studying that particular class of coefficients, which
refers to the reactive voltages and currents in synchronous machines, is relatively straightforward: at
the same time it is important, in view of the practical
significance of this type of coefficient in the study of the
phenomena occasioned by short circuits. It is this
comparatively restricted aspect of the general subject
which is considered in the present paper.
Following conventional practise, these particular
characteristic coefficients which express the factors of
proportionality between reactive voltage and current,
are referred to, generally, as reactances.

fundamental and a positive phase-sequence third

harmonic. Neglecting the effect of the resistance of
the closed rotor circuits as regards double frequency
current, the per-unit value of the fundamental component of current with normal negative phase-sequence
voltage applied is approximately
1 \
1
1
, + D,J
(1)
2
Xq
dX
corresponding to a reactance:
2 Xc" X"
Xd

xq

(2)

X" + x(
If this voltage were applied through an external
reactance, x, the total reactance of the circuit would be:
2 (Xd" + x) (Xq' + X)
x +x+x"+x
q
Xd
Thus, the effective negative phase-sequence reactance
of the machine would be:
2 (Xcd" + x) (Xq," + x)
x2 =
Xd" + Xq" + 2 x -x
2 Xd" Xq,/' + X (Xd" + XqX0)
Xd" + Xql + 2.

(3)

If x is sufficiently great this reactance becomes:

1
(4)
x2 = 2 (Xcd" + X8") *

Physically, the effect of the external reactance is

to force the current in the machine to be sinusoidal, and
to permit the existence of a third harmonic voltage
across each phase. This results in a change in the
effective negative phase-sequence reactance of the
machine. This change may be considerable. For
instance, the ratio of the two extreme types of reactance
given by equations (2) and (4) might be as high as five
to one.19
In the case of sustained single-phase line-to-line
short circuit, on the other hand, or in the equivalent
case of fundamental voltage impressed from line-to-line,
the per-unit fundamental component of current is shown
by Doherty and Nickle to be :20

19. In the case of the small machine employed in the tests illus-

trated in Doherty and Nickle's paper Synchronous Machines

the negative phase-sequence coefficient is 0.40 when negative
Dependence of Reactance on Terminal Circuit. Unlike IV
phase-sequence voltage is impressed directly, and is 1.99 when the
re
are not voltage is impressed through a large reactance, or a ratio of five,
these

aimdeplende
terminal
the character of thetereatanc
independent of thechractaeer,

circuit.

based on the smaller quantity. This, however, is a very extreme

___________
17. Three-phase machines are referred to. The armatures of
single-phase machines may be regarded as three-phase armatures
with one leg left idle in operation.
18. The term short-circuit ratio, expressing the per-unit armature current on sustained short circuit due to the applica.tion
ef no-load field current, is an exception to this rule.

Synchronous Motors and Condensers .. 1 .04

Turbine Generators.1 .00
The exact difference in these two types of reactances is discussed in detail in Appendix D.
20. This form follows from equation (27b) of the paper mentioned by substituting XD = 2 Xdi, XD' = 2 xd"', and XQ' =2 Xq".

Thus, when negative phase-sequence voltage is applied case.

In machines of normal design, the ratio would be apdirectly at the terminals of a three-phase machine, the
phase current contains a negative phase-sequence proximately,W~~~~~~~~~~~~~~~~~biater-wheel generators .........1 .30

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

522

+ V/ Xd' Xq
where e is the per-unit voltage before the short. Thus e
is also the per-unit value of the equivalent line-to-line
voltage impressed. But, in the analysis provided by
the method of symmetrical coordinates, this current is:
N73 e
(5)
Xd + x2
Thus, for the particular type of short-circuit under
consideration, the effective negative phase-sequence
reactance is
(6)
x2 = '/ Xd' Xq
that is, the geometric average of XI' and Xq". If, however, the short circuit is effected through an external
reactance, x, per phase, the reactance of the complete
circuit is:
Xd

X
(Xd"

X) (Xq"

Xd"ft+X"
+
xx

+ 22
1

d"+Xq"

Xi

reactances. Thus, when fundamental current is "impressed" on a synchronous machine, the rotor of which
is moving normally, harmonic as well as fundamental
terminal voltages are produced. If, however, the
machine is supplied from an infinite bus these harmonic
voltages are short-circuited. But an nth harmonic
current of proper phase-sequence, taken in combination
with a winding which does not exclude nth harmonic
m. f., results
m.
rotation
thisa m.
m. f.,pole
in
synchronism
withinthethepoles.
Also,ofwith
salient
machine, an nth harmonic m. m. f. over the pole can
give rise to a flux of fundamnental pitch, and thus
generate fundamental voltage. It follows that if the
harmonic voltages which exist with current impressed
are short-circuited by impressing fundamental voltage
at the terminals of the machine, the effect of the resulting harmonic currents will be to generate some fundamental voltage. This slightly modifies the proportion

X)

___

and the effective reactance of the machine is:

(Xd + X) (Xq + x) - x
X2 =
Expanding (7) by the binomial theorem gives:

X2

q2

Transactions A. I. E. E.

Xd"
x2
.

(7)

l
etc

]-x

When x is sufficiently large this becomes

x2 =

(Xd' + Xq )

(8)

asin (4).
In practise short circuits are usually effected through
some external reactance. For this reason, and in order
to retain the principle that the reactance of a machine
and any external reactance in series21 is the sum of the
machine and external reactances, it is desirable to
normally regard the reactance of a machine, when
referred to without qualification, as that obtaining with
a large external reactance in each phase, or the reac-

FIG. 6

FI0,000KV-A. WATERWRREL GENERATOR.

CUITS: FIELD COLLAR

ROTRIR Cut-

between positive phase-sequence current and voltage

from the value which would obtain with current impressed at the machine terminals.

tance that obtains with fundamental current rather than DIVISION OF SYNCHRONOUS REACTANCE INTO ARMATURE
REACTION AND LEAKAGE REACTANCE
voltage impressed. Thus, negative phase-sequence
reactance is to be interpreted normally as the effective
Armature reaction. The synchronous reactance of a
phase-sequence reactance with current rather than machine is ordinarily calculated as the sum of the

voltage impressed and, numerically, may be taken

as armature reactance due to armature reaction flux, and

the average of Xd" and x,".
the armature reactance due to leakage flux, but the
So far, it has been pointed out that the effective exact line of demarcation between reactive and leakage
negative phase-sequence reactance of a synchronous flux has not always been very carefully drawn.
machine varies with the nature of the circuit external
Originally, armature reaction was apparently conto the machine. This effect, however, is not limited to sidered equivalent merely to the action of a lumped coil
the negative phase-sequence reactance, but applies, to moving along the armature surface and subtracting its
a sligfht extent, to positive and zero phase-sequence amnpere-turns directly from the ampere-turns on the
21. Wihen reactances are to be taken in parallel the principlepoe Lar,iwsralzdttitaseiabeo
is reversed. However, it is felt that usually the series principle modify the effective amount of armature reaction to
is the more useful.
take account of the facts ;-that the m. m. f. of armature

Feb 1928.

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

reaction existed at the armature surface and, hence, a

given amount of it did not have the same effect as if it
were due to the field winding; and that the principal
component of m. m. f., and the only one that is stationary with respect to the poles, is distributed sinusoidally
along the armature surface.

Anortisseur

C_ Short Circuiting Field Collar

523

slip frequency in these circuits. Therefore, it appears

logical that, in general, armature reaction should be
taken to include all of the m. m. f. that is stationary with
respect to the rotor, since, and especially in induction
motors or turbo alternators, the end-winding m. m. f.
can produce a flux which links the rotor circuits.
Hence, this point of view, i. e., that armature reaction
exists in the end windings as well as along the stacked
length of the machine, is favored by the authors.
However, from a practical point of view, the net
effect of the armature reaction in the end-windings of a
salient pole machine is usually negligible as far as
regards the induction of flux in any of the rotor circuits.
Consequently in the present state of the art, at least,
it is legitimate to regard all the end-winding flux of
salient pole machines as leakage flux.
In the case of turbo alternators, this assumption is
less correct, although the action of the iron end-rings
employed on these machines tends to reduce the mutual

FIG. 7-150-KV-A. HIGH-SPEED SYNCHRONOUS MOTOR. ROTOR

CIRCUITS: AMORTISSEUR, FIELD COLLAR

Thus, in recent theory, armature reaction is understood to refer to the space fundamental or stationary
component of m. m. f. due to the armature. However,
some question arises as to whether the term should be
interpreted to apply to the total synchronous m. m. f.,
due to the armature, thus including the synchronous
m. m. f. in the end-windings, or whether it should be
limited so as to refer only to the m. m. f. along the
stacked length of the machine.
Obviously, the answer to this question should depend
on the relative value of each alternative as regards

AmortiBseur

Amortisur

Fic. 9-25,000-KV-A. SYNCHRONOUS CONDENSER.

CIRCUITS: AMORTISSEUR, FIELD COLLAR

ROTOR

flux between armature and field, and, thus, lessen the

degree of approximation involved. However, these
machines have a relatively small leakage reactance,
so that the error involved, practically, in treating endflux as all leakage is small here also.
On the other hand, in the case of induction motors,
there can exist a comparatively large amount of endflux. Since a considerable part of this flux may be
mutual with the rotor winding, the approximation
that end-flux is all leakage cannot be satisfactorily

FIG. 8--240-KV-A. LOW-SPEED SYNCHRONOUS MOTOR. ROTOR

CIRCUITS: AMORTISSE UR

calculating the performance of the machine. But the

chief purpose of separating out the voltage due to
armature reaction from the total synchronous reactance
drop is to consider separately those fluxes which can
normally penetrate the rotor windings, and which can,
under abnormal conditions, induce direct current or

employed.

Thus, on account of physical appropriateness, and to

obtain consistency, it is desirable, normally, to consider
all armature m. m. f. which is stationary with respect
to the rotor as armature reaction. In practical
computations, however, it is desirable, at present, to
limit the term to the m. m. f. in the stacked length of
the machine, except in the case of induction motors and
special synchronous machines.

524

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

In addition to establishing the conception of armature

reaction as denoting a particular type of m. m. f., it has
appeared desirable to broaden the scope and application of the term so that it may apply not only to m. m. f.
but also to fluxes and their linkages. Thus, it is proposed to characterize by the qualifying term armature
ratoall
reactionall ths phnmn whic deen upo th
action of the synchronous component of armature
M. m. f. In this way there are:
A. M. m. f. of armature reaction, or armature reaction
proper; all the synchronous armature m. m. f.
B. Flux or flux density of armature reaction; that
component of flux, or of flux density at any point of the
air-gap or in the rotor, which is due to the armature
reaction m. m. f.; under normal conditions, all that flux
or flux density which is stationary with respect to the

Transactions A. 1. E. E.

far as possible so that this effect is small in order to

eliminate losses in these windings.
Methods of calculating leakage reactance are given
by Mr. P. L. Alger in a companion paper on this subject.
THE IDEAL SYNCHRONOUS MACHINE

The ideal synchronous machine has a

thosehenomenahichdepenupontheopen-circuit
voltage wave, and leakage

sinusoidal
reactance

unaffected by rotor position, i. e., only armature reaction

flux is affected by pole position. Practically, the
divergence of machines from this ideal type is often
important; for example, as a cause of circulating currents, or, in the case of motors, of locking at part speed.
In a large number of cases, however, it is legitimate to
neglect these differences. The caloulation of the
electrical behavior of an ideal machine of this type can
be accomplished in terms of a relatively small number of
rotor.
characteristic constants. Thus, it may be shown24 that
C. Linkages of armature reaction; the linkages with the per-unit linkages in phase a of an ideal synchronous
any circuit on the armature, or rotor, of the fluxes of machine are expressed as
armature reaction.
D. Reactances of armature reaction; the reactive 4{a = Idos 0-I,q sin 0- 3 (ia + i6 + ic)
coefficients expressing the voltages of armature reaction
(i. e., linkages) in any circuit on the armature, or
Xd Xq+r
]c
La 2rotor, due to normal armature current.
The reactance of armature reaction, or in the case of
induction motors, the magnetizing reactance, will,
Xd - xq r
L ia Cos (2 0) + ib cos (2 0 - 120)
however, when referred to without qualification, mean
3
the per-unit armature fundamental linkages due to
normal positive phase-sequence armature current and
+ ic cos (2 0 + 120)
(10)
will have direct and quadrature components, thus:
where,
Xad = direct reactance of armature reaction
iJa, tb, and tc are the instantaneous currents in phases
Xaq = quadrature reactance of armature reaction. a, b, andc.
c.
Methods of calculating armature reaction have been a, b, and
is
the
nominal voltage, or per-unit rotor current
Id
of
Methods
calculatin
given by R. W. Wieseman.armat2
in the direct axis.
I q is the nominal voltage, or per-unit rotor current in
Leakage Reactance. Leakage reactance is defined as
that part of synchronous reactance which is not in- the quadrature axis, ahead of the poles.
cluded in the reactance of armature reaction. Hence,
0 = electrical degrees in direction of normal rotation
there are two components
between axis of poles and the magnetic axis of phase a.
Xld = Xd - Xad
Axis of phase b is 120 deg. ahead of the axis of phase a.
=
of phase c is 120 deg. behind the axis of phase a.
Axis
Xld Xq Xaq.
The linkages in phase b are similar except that
They are the reactances due to the flux in the slots and
to the flux produced by the moving harmonic of (0 - 120) is substituted for 0 in the first two terms,
m. m. f., which latter is that flux in the air-gap or end- ib for ia in the fourth term, and (120 - 0) for 0 in the
winding which is not normally stationary with respect fifth term. The linkages in phase c are obtained in a
to the poles. Actually, the flux due to the nth harmonic corresponding way. From this, the voltage of any
phase can be obtained from the relation
1
times
)
ddtm. m. f. pulsates over the rotor at (14a
(11
na
n
=
normal frequency and, therefore, is subject to the
demagnetizing influence of the field winding and all
other closed rotor circuits. Consequently, in determining the permeance23 of the air-gap to nth harmonic
m. m. f. the effect of these circuits should be considered.
In practise, however, the rotor circuits are adjusted as
22. Reference (5).
23. Reference (8).

~~~~~~~~ea

In order to determine the reactance between the

armature and rotor, it is of great benefit to resolve the
armature quantities under consideration with respect
to the two axes of symmetry of the poles, a procedure
introduced by Blondel and further developed by Arnold,
and Doherty and Nickle.
.24. Reference (14).

Feb. 1928

525

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

I = X I -Xm id
in which4, X, Xm, and I refer to thiscircuit.
In order to retain the valuable principle that the
per-unit nominal voltage is equal to the total rotor
current, it is desirable to define unit current in any
circuit as that current which acting by itself produces
normal space fundamental flux at the armature surface.
In this way the nominal voltage at any instant is always
to be identified as the instantaneous rotor current.
Thus, in the direct axis, as
(14)
+ 12d +
= I
I d+dIld
(12)
+Ind
Transient Reactance. In analogy with the procedure
2
[ia sin 0 + ib sin (6 - 120) + i, sin (6 + 120)] employed in determining the effect of the armature
iq
circuit on the rotor, it is desirable to establish the conThe currents, id and iq then determine the reaction ception of instantaneous direct and quadrature comof the armature on the rotor at any instant. In general, ponents of armature linkages. If these components
there will be more than one rotor circuit. Consider, are defined as
however, any particular circuit, say circuit h of the
2
Ad = 4'a cos 0 + lb COS (6- 120) + f,c cos (0 + 120)]

To this end, and in analogy with their recognized

significance under steady conditions of operation, let
the direct and quadrature components of current at
any instant, and under any condition of operation, be
taken equal to the instantaneous amplitudes of the
direct and quadrature components of armature m. m. f.
It may be shown25 that, at any instant, the magnitudes
of these m. m. fs. are equal to the quantities
2
id = [ia cos 0 + ib cos (6 - 120) + ic cos (6 + 120)]

(15)
V'q I

2-k[ta sin 0+4b sin (0-120)+I, sin (0+120)]

it may be verified that the simple relations, which

obtain between armature linkages and armature and
rotor currents under steady conditions, now hold at
any instant. That is, under all conditions,
i/d

FIG.

10-10,000-KV-A. TUtRBO ALTERNATOR. ROTOR CIRCUITS: ROTOR IRON, METAL WEDGES

direct axis; that is, circuit h d. There will be factors

of proportionality between the linkages, Thd, in this
circuit, the currents in every other rotor circuit in the
direct axis, and the instantaneous direct component of
armature current. Thus, there is
+
Xhid =lid
Xdfhd = Xh,f d I 4- Xhl d Ild + ... XhfdI -I- .Xd
(13)
Ind - Xhmd Id
+ Xhhd Ihd + .
where
I - per-unit field current
per-unit current in circuit h of the direct axis
lhd
=
Xhfd per-unit mutual reactance of field with circuit
.

X
Xhmd

of circuit* d t
reactance
Xh~d
per-unit mutual
d
ecac.o lclthdw
pecrcunit nuua

Id- Xd id

Aid
A A1d A Id - Xd Atid
But,
where A Id is the corresponding increase in direct axis
rotor current. Therefore,
AI

______~~~~~~~~~~~~i

25. Reference (14).

(18)

and similarly in the quadrature axis. The value of

A Id may be found by solving the n equations
(19)
4' = 'ifld 4=T2d = . . . "nd = 0
found
A may be
subject to the conditions id = A

id. Iq

the
similarly. We shall only consider in detail here (19)
case of a single rotor circuit. The equations

per-unit mutual reactance of armature with become in this case merely

= XAoo I- XmTAhd = 0
'c
circuit h d.
Similar relations exist in the quadrature axis except that from which it follows that, in this case,
field current is not involved.X
Xd' X=x- Xm
If there is onlyone rotor circuit, equation (13) becomes

merelyX

(16)

litq I- Xq i'q
Transient reactance will then be the constant of
proportionality between V/ and i when i is suddenly
impressed. Thus, if A 4Ad iS the sudden change in APd
when a current A id is suddenly impressed, there is
(17)
X ' = -

(20)

where Xm is the mutual reactance of the main field wind-

526

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

Transactions A. I. E. E.

ing with respect to the armature, and X is the total field in order to obtain symmetry with the equations for
reactance.
armature voltage a different procedure is here employed.
Calculation of Rotor Reactances. Before calculating In the following, normal voltage of any rotor circuit is
these quantities in per-unit terms it is necessary to that voltage which would exist in this circuit if normal
decide on an appropriate value of normal or unit link- flux changed through it at normal frequency. Normal
ages. It has been found convenient to specify normal resistance is then the ratio of normal I R drop to this
linkages in each rotor circuit as the product of normal voltage. In this way the per-unit voltage of any field
flux by the number of series turns in the circuit. On circuit, say h d, is
this basis it is shown in Appendix G that, for any rotor
d hd
circuit h,
(25)
Ehd = dt + Rhd Ihd
A(0d
)
X htn =
(21) while for the field circuit proper there is
Fh
(0)

where,

E = dt + RI

(26)

A (o) = 2.12 X (effective full-load armature ampereDecrements. Only the case of a machine with the field
turns per pole, per phase).
=
as
the only rotor circuit will be considered. In this
h
no-load
of
circuit
normal
ampere-turns
Fh (0)
case,
if the armature is open-circuited, there is
with no saturation.
' = XI
Thus, the determination of the mutual reactances of
Xmh can be reduced to the determination of no-load and
dI
excitation ampere-turns in each rotor circuit.
dt
The self and mutual rotor reactances at present
require flux plots for their determination, except in the
that the open-circuit time-constant of the field is
case ofthe
totalfield
reactance.
case of the total
field reactance.so
As shown in Appendix F, the latter is given by the
(27)
To = X
R
expression
F (o)
If the machine is short-circuited through a reactance
X
and
the effect of the resistance of the armature circuit
x,
(=)
is negligible, as is normally the case, there is
where K0 and L are given by formulas and curves in
Xm
Doherty and Shirley's 1918 paper on reactances, 1 is
[r
Xd'X XI
I
Xd + x
Xd + x
the stacked length of the machine, and 4 (o) is normal
no-load armature flux. The transient reactance of a
dlI
Xd'+ x
machine with no additional rotor circuits is then given as
X t +RI
E=
dt
Xd+X
A(0)
1
X
Xd' 2d- F(o) K
(23) so that the time-constant of the field, in this case, is
F (o)
4

(o)

Xd' + X
T
To
~~~~~~~(28)
Xd + x
The time-constant on dead short-circuit is
X
(29)
To
To' =
Xd
The effect of armature resistance, or of machines
operating in parallel, is to modify these results in a
complicated way which will not be considered here.
If, as is ordinarily the case, the resistance of the rotor
circuit is negligible, the time-constant of the direct
current in the armature depends upon the inductance
of the armature to direct cuYrrent in the same way as
the time-constant of direct current in the field depends
on the inductance of the field to direct current. The
~~~~inductance of the armature to direct current is, however,
the average of Xd"/ and xqtt and is thus equal to the
negative phase-sequence reactance. Thus the timeconstant for the d-c. component in the armature is

Equivalent Circuits and Field Leakage Reactance.

This general subject is treated mathematically in
Appendix E. It is shown that any number of equivalent circuits can be constructed to represent the relation between armature and rotor currents and linkages.
Of the various circuits which may be so constructed
there is one in which the armature branch is Xld, and the
magnetizing branch is Xad. The field branch of this
particular circuit has been referred to as the field leakage
reactance, and is denoted by Xi.26 It is shown in
Appendix E thatX1 may be expressed in terms of the
quantities already described by the relation,
( Xad X
\
X Xad
1
(24)
/
Field Voltage. It sometimes is desirable to define
normal field voltage as the voltage which will just force
normal current through the resistance of the field, but
26. ReferenGe (6).
simply

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

Feb. 1928

527

may be found by superposing the voltages due to any

components into which this current may be resolved.
To illustrate, the armature voltage due to a field
current I, and an armature current i, may be found as
as stated in Part I.
the sum of the armature voltages due to the currents I
ACKNOWLEDGMENTS
The authors wish to thank Messrs. R. E. Doherty, and i acting separately. That is, the voltage under load
from the no-load voltage
subtracting
C. A. Nickle, P. L. Alger, and Professor W. V. Lyon for may be found by
to the
existing field
current, the voltage
d t t
their valuable suggestions, and Mr. R. G. Lorraine for corresponding
the
to
caused
by the armature
drop
mpedance
his assistance in the preparation of the figures. *
~~current. Or, the voltage due to a current
Appendix A
i = i, sin (t + a)
APPLICATION OF THE PRINCIPLE OF SUPERPOSITION TO can be found by superposing the voltages due to
currents
THE ANALYSIS OF SYNCHRONOUS MACHINES
x - ia cos a sin t
In a synchronous machine, as in a static device, the
flux density at any point will be in linear relation to and,
i-aisln a cOs t
the total number of currents in every circuit of the
machine, providing saturation and hysteresis are acting alone.
So far it has merely been pointed out under what
neglected. However, in a synchronous machine the
factors of proportionality between flux density and conditions the voltages due to a given current districurrent depend upon the angular position of the rotor. bution can be found by superposing the voltages due to
Thus, if the circuits in the machine are numbered 1, 2, any components of the distribution. Suppose, howetc., the component of flux density in a given direction ever, that a given distribution of voltage is impressed
at any point of the machine will be related to the at the machine terminals. Let this distribution be
currents in a manner shown by the following equation: characterized by the symbol e (t), and let ex (t) and ey (t)
into which e (t) may be
be component distributions
0 = fI (0) iI + f2 (0) i2 + . . . etc.
so that e (t) = ex (t) + ey (t). Let the currents
for agvepoiresolved
Consequently,
t)
thedutoheeisrbinsei(),i.t)adi
of the rotor,
Consequently, for a given position
flux at any point in the machine may be found by Then theremustrbe
superposing the fluxes due to the current in every circuit
i (t) = 1x (t) + t (t)
acting separately. Also, the flux due to the current in
all the conditions imposed by the
fulfills
result
this
since
fluxes
due
the
superposing
be
found
by
may
any circuit
terminal conditions of each
and
equations
differential
be
current
this
may
into
which
to any components
that the position of the
provided
machine,
the
of
circuit
resolved, taken separately.
0
is the same with e (t),
of
time,
(t),
function
rotor
as
a
Thus,
linkages.
The same rule will of course apply to
(t)
impressed.
ey
ex (t), and
the linkages in circuit I will be represented as
.
.
Thus, it follows that the current distribution due to a
.
etc.
=
i2
X12
+
X11
+
il
(0)
411
(0)
given distribution of armature voltage may be found
any
due to may
distributions
current
superposing
by
of circuit
sinductance
isthe inductance
the voltage
distribution
intothe
which
components
1 with re
of circuit
the mutual
be resolved, provided that the motion of the rotor, as a
function of time, is the same for the components as for
d 41
elr1Tl + d t
the resultant.
For example, if a machine is supplying current to a
reactor and is then short-circuited, the current at any
/
di, d X12 (0) d 0
dXi, (0) d 6 \ +
dt + d 6 77712 time after the short circuit may be found by super=r+ d 6 d t )+ Xi
posing on the initial current distribution the fictitious
currents which would be due to the application of a
d t,9
etc.
negative voltage at the point of short circuit. This
+ N12 () dt +
negative voltage is equal and opposite to that which
would have existed at the point of short circuit had the
But, since if i = ix + iy
short circuit not occurred, and it is assumed to be
dj
d i2
dit
applied with the machine unexcited.
d t = d t+ d t
On the other hand, if the machine is supplying
it follows that, for a given position and speed of the rotor, a kilowatt load, it would be only possible to calculate
the voltage in any circuit of the machine may be found the currents at the instant of short circuit in such
by superposing the voltages due to the current in every manner. This is because of the fact that generally,
other circuit acting separately. Also, under the same in this case, the motion of the rotor would be different
conditions, voltage due to the current in any circuit before and after the short circuit on account of the change
T

x2

(radians)

cdue

whee

and

is

528

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

Transactions A. I. E. E.

in load effected by the short circuit. In the first instant, the currents in the "additional rotor circuits" have died
however, the inertia of the rotor would prevent a change away.
in its motion.
b. Alternating so slowly that the demagnetizing
When current is impressed on the open circuits of a action of the "additional rotor circuits" is slight.
machine, in which closed circuits also exist, the same
But if, with closed circuits on the rotor, the armature
limitations apply as in the case for voltage. In both currents are sustained and of such a nature that the
events the currents and voltages due to the impressed armature m. m. f. is stationary and steady over the
quantities depend on the complete variation in time of poles, Xd and xq are used as before.
those quantities. Superposition is still legitimate,
Finally, if there are closed circuits on the rotor, but
subject to the condition that the motion is the same for neither of the above conditions is fulfilled, it is generally
the components as for the resultant.
possible to divide the currents into two parts; first, those
The general statement of the possibi'ity of super- which meet sub-transient or transient reactance, and
position is, then, that for a given motion or position of the second, those which meet synchronous reactance.
rotor, the currents and voltages in any part of the The appropriate factor is then used with each type of
machine, due to any currents or voltages impressed at current. For example, a single-phase line-to-line
its terminals, may be found by superposing the effects current
that would result from the action of each taken
- = io cos
i
separately.
is resolvable into positive and negative phase-sequence

Appendix B

components as

REPLACING THE EFFECT OF INDUCED FIELD CURRENTS

BY EMPLOYING TRANSIENT INSTEAD OF
SYNCHRONOUS REACTANCES
Equation (10), of Part II, gives the complete expression for the phase linkages of a synchronous machine
under all conditions of operation. However, in the case
of rapidly changing or suddenly impressed currents,
there is a definite relation between the currents in the
field and armature circuits.97 As a consequence, the
field current terms may be replaced by terms involving
armature currents.
Thus, if the field circuit is closed, and the effects of a
given armature current are to be calculated
a. only in the first instant after their sudden

application,

b. under sustained conditions of such a nature that

all m. m. fs. over the poles are pulsating, or moving at
sufficient speed that the demagnetizing action of the
rotor circuits is practically complete,
it is merely necessary to substitute Xd' and xq"' or Xd'
and xq' for Xd and xq in order to eliminate the terms

IdandI..

The linkages in phase a thus become

Xd"f + xq" r.
o
b +
{a=
3 (ia + ib+c)
[a

- Xd

ia

ib

A/3
1

io [cos (t - 30) + cos (t + 30)] = io cos t

to

cos

(t + 150)]

= -

io cos

(2b)
V3
For normal motion of the rotor, the positive phase-

i =c

io [cos (t -1- 90) + cos (t - 90)] = 0

sequence component of current meets synchronous

reactance, and the negative phase-sequence meets

sub-transient reactance. Thus,

Xd +
2

+
)/a =-

+_ i]

Xqcos (t - 30)

-q cos (t + 20 + 30)
2

Xd"+ x"l

Xd'
+

Xq" [ ia cos (2 0) + ibcos (2 0- 120)

+ i. (cos (2 6 + 120) 1

[cos (t - 150) +

2 t cos (t + 30)

Xd -xq"

cos (3 t + 2 0 + 30)
(3b)

(ib)

Xd' and Xq' replace Xd" and Xq" when the m.m. f.over
the poles is
a. Suddenly impressed, but steady thereafter, and
it iS desired to determine the effects of currents after
27. Tha,tis, Id = (Xd -Xd"t) id and 'q = (xq - Xq") iq.

SIGNIFICANT ROTOR CIRCUITS IN ADDITION TO THE

FIELD
J~~~~~~~~~~~~~

The principal circuits in the rotor of a machine,

besides the field winding proper, which are capable of
significantly affecting the behavior of the machine
under practical operating conditions, are listed in
Table III.

529

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

Feb. 1928

Xd"t + Xq"

TABLE III
Significant rotor circuits
in addition to field
Type of Machine

23

+Xd
Amortisseur, if used.
Field collars, if used.
Spider, unless laminated.
Metal Wedges
Rotor Iron, unless laminated.

'Pb =

Figs. 6 to 10 show these circuits pictorially for

representative machines of each type.

Water-wheel Generators
Synchronous Motors and
Condensers.
Turbine Generators

Xq

Cos

(3 t + a)

Cos

(2 0

it

Xd"t + Xq"
2

3t

(4d)

a)

cos (3 t-120 + a)

(5d)

Appendix D
THE NEGATIVE PHASE-SEQUENCE REACTANCE OF A
SYNCHRONOUS MACHINE WITH NEGATIVE PHASESEQUENCE VOLTAGE IMPRESSED
With normal negative phase-sequence current impressed on the terminals of a machine the phase currents are,
ta = cos t
COS(t 120)
ib=
ic = cos (t - 120)
It follows from equation (10), of Part II, and from
the fact that all the resultant fluxes are varying rapidly
over the rotor,28 that the flux linkages in each phase
are, under these circumstances,
Xd"

+ xq

Xd

Xq f

etc.

If fundamental negative phase-sequence voltage is

maintained at the armature terminals, and if the
resistance of the armature is negligible, the armature
linkages will also be negative phase-sequence. This
will require that both negative phase-sequence fundamental
to each other.
a definite ratio third-harmonic
exist,positive
having phase-sequence
currents and
Thus, let i2(i) be the magnitude of the negative phasesequence fundmental current, and i1(3) be the magnitude of the positive phase-sequence third-harmonic
current. Also, suppose that the two are acting together, and let 0 = 0 + t. Then, combining equations (Id) and (4d),
t
Xd + Xq

xd" +Xq Mi
'Pa

Xd"

os(t+12

+2

+is
(2 0 + t-120)(2d

+ i2(1)

Cos(20-Ht--

+x2

+ i'(2

(2d)

120)

cos (3 t + a)

2
X d"

Xd

X01'

Xit

cos (3t+2
00 )

Xdl+Xd"
+

20

2 aeCos (2 0 + t + 120)

-d

Xq"

(3d)

Jf

a=20S

1(3) d"
"((7d)
. __X
__d 2 ___
The variation in linkages will give rise to a voltage and, and 22(1)
of
a
and
harmonic
fundamental
of
frequency
(2 d O/d t + 1) times fundamental frequency. For 'Pa will contain only a fundamental component, i. e.,
= i2(1) L Xd" + Xq
(Xd" Xq/)2
normal motion of the rotor these voltages will be funda2 (Xd" + Xq") ] cost
2
mental and third harmonic of which the fundamental 'Pa
one
may be found by analyzing the voltage wave over
2 Xd" Xq"
cycle. However, if the speed of the rotor is not
(8d)
COS t
= X2Xdt +
exactly normal, the third harmonic in the terminal
voltage will pulsate at the slip frequency.
'Pa = cos t
If normal third harmonic positive phase-sequence If
11{ 1
.
current is impressed, i. e., if
then q2(1) - 2~ Xd" + xq"J
ta= COS (3 t ae)
x
1
1
1
ib = COS(3 t-120+ac)
and
xc= cos (3 t + 120 + a)
---)= ( -+-g) Xd +Xq
X"+q
2 \ d X '
the phase linkages become,
specified,
But under the conditions
28. See Appendix B.
X

530

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

lb

cos

(t + 120), i

(t - 120),

cos

so that the current distribution,

1
1
1
C t
i= y (
)cos
+
Xq
Xd _ z
Xdlt + X/cos (3 t + 2 0)]
Xd!' + Xq

1
1
2

)[cos (t-120)
Xd(xq

2
+

1
it

Xq
xv

where

0- 120)j

-Xd"Xd" -+Xq"
X q cos (3 t + 2
(

are determined by an external circuit such as an

impedance, for example, a reactance and resistance in
series. Thus, it is possible to calculate the power
factor of a current of any frequency induced in the field,

or to determine the effect of external field reactance, etc.

The field voltage equation is
(9d)
d__

ib =

Xd"- X"
-x Xd"^! +
+ xql/ cos (3 t + 2 0O + 120)

(lId)

is just adequate to give rise to balanced normal negative

phase-sequence terminal voltage. It follows that the
application of balanced normal negative phase-sequence

(Xtfn-i)

Ifi

1.0
Fic.

dt

R
E
=
E(f) = X and R(X)

and resistancein of
the field
(lOd) of Reactance
divided are,
by
and circuit
course, to be
calculated per-unit

,, )[cos (t + 120)

(x -1)

Transactions A. I. E. E.

'(f)

Xm. N'ormal or per-unit resistance is the product of

normal field current and resistance in ohms, divided
by the voltage due to unit rate of change of normal
flux linkages.
While the circuit just discussed is perfectly practical,
it is not at all the only circuit that can be constructed,
nor is it in every way the best, since it may involve
negative reactances. These latter, while not inconvenient mathematically, are not entirely satisfactory
from a physical standpoint.
A particular circuit previously employed29 is that
in which the armature branch of the equivalent circuit
is made equal to the leakage reactance of the armature.
In this case write:
=d= A

11-EQUIVALENT CIRCUIT OF SYNCHRONOUS MACHINE

DIRECT Axis, MAIN FIELD ONLY

xd id

(6e)

r I 1
voltage will give rise to a fundamental negative phaseB'J=ABXL A J-BXmid
(7e)
sequence current of amount
1/1
1\
where A and B are arbitrary factors to be determined in
+
2 \ Xd"
I
xqt J
The voltage may be applied either from line-to- such a way thatidand A can be employed in the
line or from line-to-neutral.
equivalent circuits. Then, it is necessary that
Appendix E
(8e)
Xd - A -x
CONSTRUCTION OF EQUIVALENT CIRCUITS: CONCEPT OF
A= Xd Xl = Xad
or,
FIELD LEAKAGE REACTANCE
also there must be, on account of the nature of the
Consider
machie
Consider a machine without any "additional
rotor wBut,
circuits." The equations relating per-unit armature equivalent circuit:
AB m
e
and field linkages are, in this case:
2

Xd"

I-Xdid
XI- Xmid
The second equation may be put in the form,
'i'(f) =X(f) I -i
V/d =

(le)

(2e)
(3e)

Xad

Xm
X*

The equations may then be written:

(4e
where, t(t) = X ~~~~~~~'
m
X() X_

where,

f-d

trf

Xad I[f]- Xd

XLffrIr[

id

X-7 1-7

~~~~~~~~~~~I

I[]= -

But, equations (le) and (3e) are satisfied by the equiva'F[fJ = Xm 'I
_____
lent circuit of Fig. 11.
This circuit also holds good when the linkages to(f
29. Reference (6).

(10e)
(lie)

(12e)

(13e)
(14e)

XVf

(Xad)2
X

(15e)

The equivalent circuit in this case is shown in Fig. 12.

In this case the field voltage equation is
d
E(f) - d t T[fI + R[f IIf

where

531

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

Feb. 1928

Xad

Xad

E[f l = Xm E and Ri =

The term

X[f

- Xad has been referred to as field

leakage reactance-thus,
Xl = X[f]- Xad = Xad

rxad

1xd

- ij=j [Xad Xm]

surface, in the slots or inter-polar space, and at the ends

of the poles. This corresponds to the division of
armature leakage into air-gap, slot, and end leakage.
The reactance X is then expressible as the sum of three
factors relating to the armature, inter-polar space, and
pole ends respectively. Thus:
X = Armature factor + inter-polar, or slot-factor
+ end factor.30
Armature Factor. That component of flux, due to
current in any field circuit, which enters the armature
surface will be assumed to link all of that circuit, a
condition which, it happens, is very generally fulfilled.
Then the armature factor in field reactance, expressed
as a per-unit quantity, is the ratio of the total flux to
entering the
the space fundamental component of flux
armature along one pole pitch. This factor was first
employed by Doherty and Shirley in their paper on
Reactance,30 and has received the designation K,.
As pointed out by them, it is capable of being expressed
as the sum of the per-unit harmonics of flux, due to

(16e)
The formula for transient reactance may be obtained
by shorting the field side of the equivalent circuit.
Referring to Fig. 12 this gives:
Xm normal current in the circuit under consideration,
1
(X(f)-1) _ divided by the order of the harmonic. Thus:
=
Xd
X
Xd' = Xd l
-X
X(f)
X1(f)

or in terms of field leakage reactance

Xd' =

XoXac

+XXl+Xad

(17e)
(18e)

k3

1+ 3 +

ks

+.

kn

+ -

where the corresponding per-unit flux density wave is:

sin 3

k5 sin 5

...etc.,

k3
y+
y+
from the quadraangle
measured
zy being the+electrical
1

IXm \
ture axis of the machine.
The form ( Xd- X ) is probably the easiest for
It is interesting to realize, then, that harmonics in
the no-load flux wave of a machine are of the nature of
leakage flux. This leakage flux can, however, be
calculation since;
positive or negative according to whether the wave of
Xd is calculated by standard methods.
Xm is calculated from a knowledge of no-load normal no-load flux is flat-topped or peaked. This would be
indicated by negative constants in the above equation.
field current.
Charts for determining K. for ordinary field windings
XI
available in the above mentioned paper on Reare
IXad
XXf
+
*7
actance.
If [f]
SXaS
iZd
1d
Interpolar, or Slot, and End Factors. These factors
-- lcomprise the field linkages due respectively to all the
Xd
FIG. 12-EQUIVALENT CIRCUIT OF SYNCHRONOUS MA-ACHINE flux in the stacked length of the machine that does not
penetrate the armature surface and to all the endDIRECT AxIS, MAIN FIELD ONLY
flux.31 A formula for the field-flux linkages of salientX may be calculated without great difficulty. It is pole machines, due to flux which does not penetrate
merely the ratio of the total field linkages to the the armature surface, has been given by Doherty and
linkages the field would enclose if normal space Shirley.32 Using the quantity L given in formula (33a)
fundamental armature flux linked all of the field of that paper, the inter-polar and end factor becomes
equal to:
turns.
N I(o) L 1
The construction of equivalent circuits when additional rotor circuits are to be considered is more difficult
t
than in the simple case discussed and will not be treated
Where l(o) = no-load field current in amperes
here.
(o~) = no-load space fundamental flux per pole
Appendix F
CALCULATION OF TOTAL FIELD REACTANCE1

grsstcelnthocr.

1242.
30. Referetncct(10)al the
The calculation of total field reactance, i. e., the
quantity or quantities X, is conveniently accomplished mental at the armature winding.
32. Reference (10), p. 1279.
by considering separately the fluxes at the armature

hihi ntspc f

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

532

3
2

Thus, there is:

X

Ko +

N I(o) L 1

o(oA

Appendix G

or,

Xhm

Transactions A. 1. E. E.

nfi(o)

Nh(o) Ih(o)

AF(o)

(6g)

RELATION BETWEEN Xm AND NO-LOAD EXCITATION

where,
CURRENT
= 2.12 (effective armature times per pole per phase)
A(0)
Ordinarily, the leakage reactance of a machine does
X (normal r. m. s. armature current).
not have a transient value perceptibly different from
air-gap ampere-turns of any circuit .h
no-load
its sustained one because the currents induced in the Fh(o) =
In particular,
various rotor circuits by the leakage fluxes are sensibly
reactive in character, i. e., are nearly out of phase with
A(0)
the voltage producing them. Consequently, it is Xm F() = per-unit mutual field reactance (7g)
possible to neglect the changes in leakage reactance with
method of application of current and to regard all the where,
change that is necessary to account for the difference in F(0) = no-load air-gap ampere-turns of field.
Using this relation, it is possible to reduce the calcusynchronous and transient reactance as being due to
the difference in the reactances of armature reaction lation of transient reactance to the following items:
A. Determination of synchronous reactance,
in the two cases. Therefore, in calculating the values
of the constants Xm, Xmid, etc., it is only necessary to
B. Determination of the "no-load field currents"
consider the flux linkages of armature reaction, that is, of each rotor circuit, i. e., determination of current
the flux linkages due to the space fundamental comnormal fundamental component of
to
ponent of armature m. m. f.ponent of armature m. m. f.necessary
n
amtr lx give
But the effect of normal space fundamental m. m. f.
C. Determination of self and mutual coefficients
acting in one axis of the rotoris just two-thirds33 of the
the rotor circuits between themselves.
current.
of
effect of the maximum value of normal phase
Appendix H
That is, if i(o) is the current flowing in a single-phase
in
winding when the rotor is stationary and its axis
question is directly in line with the axis of magnetizaBE
RE ES
PHASE REACTANCES
tion of the excited armature phase, the effect of normal
In accordance with the procedure discussed in Part
2
II, the impedances of a machine will be regarded as
i(o.
space fundamental m. m. f. is
3
the ratios of the sinusoidal components of voltage and
Now consider any rotor circuit, for example, circuit h. current due to sinusoidal current in the windings.
The linkages in this circuit, due to a current of normal peak Since only sinusoidal currents of definite frequency, and
value i(o) in one armature phase in line with circuit h, are only sinusoidal components of terminal voltage of
corresponding frequency, are to be considered, it follows
2
that
it is legitimate to employ vector notation.
(1g)
Xhm
N,h
q(0)
3
Thus, any current i = i0 sin (t + a) will be repreThus the mutual inductance of circuit h and the arma- sentable as a vector i = io /a and the phase currents
will be representable as
ture phase is
a=il 1.2 + to
2/3 Nh O(O) Xhm
(lh)
= \ 120+2/ 120 +0
(2g)
tb = i

/120 +t2\12O +i0

The linkages in the armature phase due to a normal The phase voltages are similar.
It may be readily verified that with a line-to-line
value of current Ih(o) in circuit h are, from the definition
connection between phases b and c,
of normal current in any rotor circuit, simply34
_
nT tm(a)
(3g) l
/ 90
1 =-2
=(2h)
Thus the mutual inductance of the armature phase and
+
circuit h is
where i. } = line-to-line current of phases b and c,
nl /(o)
~~~~~(4g)while with a line-to-neutral connection from phase a
to neutral,
Ih^(o)
But since (2g) must be equal to (4g) it follows that
33. The effect offundamental m.m. f.only is tobe considered.
34. Since only the sinusoidally distributed component of the

armature winding is considered, harmonic linkages are neglected.

t1= t 2 =

(3h3ta)

where 4Oa = current from phase a to neutral.

Feb. 1928

533

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

Under normal conditions of operation, however, it

follows from the symmetry of every phase, that the
fundamental components of the voltages due to any
phase-sequence current distribution must be of the same
phase-sequence as the current. Thus, subject to the
condition of normal motion of the rotor, there is,
ei = zi

xo

3 X0d - Xsd =3 xaq- Xsq

Xsd
Xd" = 2
x

(13h)
(14h)

(15h)

i,

as mentioned in Part I.
(4h )
e2 = Z2 X 2
The values of per-unit single-phase reactances which
obtain with no closed rotor circuits may also be identieO = z0 i,
but the line-to-line voltage of phases b and c is
fied as the per-unit single-phase armature inductances.
Thus, in treating the subject of short-circuit currents,
eb., eb -e = 3 (ei - e2) \90
(5h) Messrs. & Doherty and Nickle35 employ the relations,
(Zl + Z2) is bc Zs Is bc
or line-to-line impedance = z, = Z, + Z2
+ XSQ XSD - XSQ
2
cos (2 0')
+
2
Similarly it follows that, subject to the conditions line-to-line inductance=
stated, line-to-neutral impedance is
zoline-to-neutral inductance
Z1 +AZ2+zn
zs +3
3(6h)
3
20
X:D + XZQ X+ZD X
2
2
However, these relations do not obtain without
6' = 0 + 300
special interpretation unless the rotor is moving where,
The idea of transient inductance has also been
normally. In case of an "ideal machine" with the rotor
stationary, thevaluesof single-phase reactance are given, introduced. For example, the transient line-to-line
by the value of Pa 4-'b with ia = i-b = 1, ic = 0, and inductance is,
Vla with ia = 1, ib =i = 0, respectively, for a line-toXSD' - XSQ/
XQD + SQOS (2 0_)
line or line-to-neutral connection.
2
2
Thus, with all rotor circuits closed, the line-toIn case machines with auxiliary rotor circuits were
line static reactance is
to be considered, the conception of the subtransient
Xd"- xq" cos(2 0) - 2 cos (2 0 -120) inductance
and the use of the doutle primed notation
Xs Xd"I + Xq" +
would also be in order. Thus, for the transient
reactances,
+ cos (2 0 + 120)]
XSQ = 2
XSD
2 Xd'd
(7h)
= (Xd" + X,z) + (Xd" - Xq") cos (2 0 + 60)
, = x-+2 Xq
, X0 + 2XI
The value of x, when the axis of magnetization of the
X2D
XIQ3
3
armature coincides with the direct axis of the rotor is
(8h) and for the sub-transient reactances,
Xs = Xsd = 2 Xd"
XSD" = 2 Xd", etc.
When the axis of the armature winding coincides with
the quadrature axis the value of x, is
Appendix I
(9h)
Xs = Xs q = 2 x q /
OfeIsN asMPLOYE
T
o
The corresponding line-to-neutral reactances are,
of
The
devising a simple scheme, of notation
problem
with all rotor circuits closed,
as much as possible with existing
be
in
accord
will
which
closed,
circuits
with
alldrotor
been
not
has
easy. The difficulty is due
+ 3
(10h) nomenclatureterms
cos 2 6
X.,
3
3
3
which pertain to synchronous mato the many
3
When the axis of the armature winding is in line with chines.
The notation of the paper follows the somewhat
the
poles
-

xq

the poles

X,dgenerally established principle that all armature quan-

+ Xd 2 Xd' (llh) tities are denoted by small letters or symbols, and all
X X, = 0 + 2 Xd"t _0O
x
- xd Ih3 3 field terms by large letters or symbols. The subscripts
and when the armature axis is in quadrature with the d and q refer to the direct and quadrature axes of magnetiThe corresponding single-phase line-to-line
poles
~~~~~~~~~~zation.
poles
line-to-neutral quantities are recognized by the
(12h) subscripts s and a, respectively. Positive phaseX5X,q = X+ 2 xq" _Xo +Xsq
x<,-x,, sequence terms may be indicated by the subscript1, but
3
3
These relations are useful in determining the values are usually given no special notation. Negative and
35. Reference (2).
of xO, Xd", and Xqt". Thus
O

~~~~~~~~~~~and

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

534

zero phase-sequence terms are specified by the subscripts 2 and 0, respectively, following the notation
introduced by Mr. C. L. Fortescue.36 All transient
quantities are signified by primes.
These compose, in the main, the group of indexes
required, but various modifications and special designations were necessary. The notation is not peculiar to
this paper but is largely the same as the system used by
Doherty and Nickle in their recent treatments of synchronous machines. Although it is not urged that
this notation be standardized, since the primary
purpose has been to fill the requirements of the present
article, the system may be helpful to others.

Transactions A. I. E. E.

c. Use of zero as subscript.

Except in connection with time-constants, 0 will
indicate zero phase-sequence quantities, and (0) will
imply normal, or per-unit quantities.
To = field time-constant with all armature phases open.
Tro = armature time-constant with all rotor circuits
open.
C. PRIME QUANTITIES
.

a. Wth reference to armature reactances:

sustained no special indication.
transient-one prime, i. e., x'.
sub-transient-two primes, i. e., x".
BE
With reference to time-constants:
A.ticle,thesystem
MAJO QAIe TOlpful otheNTIb.
field time-constant with all armature phases
Circuits
shorted-Tof
armature time-constants with all rotor circuits
Armature Rotor
.
Voltage
Linkages...........
Current.
M. m. f.A
Reactance....
Resistance ...........
...

Time-constant ........
Flux density .........

e
V

.
x

I
...i

Time
t
t
(Electrical radians)
Electrical space angle along armature surface
with respect to phase "a"
ly
Electrical degrees in direction of normal rotation
between axis of poles and magnetic axis of
0
phase "a"

B. USE OF SMALL SUBSCRIPTS IN NOTATION

a. Direct and quadrature quantities.

d
direct
quadrature q

These apply to:

armature quantities
positive phase-sequence
quantities
rotor quantities
static quantities
single-phase
*
b.b DAstrebutuon.
1. Armature quantities.positive phase-sequence 1, or no subscript
negative phase-sequence 2
0
zero phase-sequence
s
single-phase line-to-line

rostu

Distribution.niie

single-phase line-to-neutral oa, b, c

separate phases
2. Rotor quantities.
d
direct
quadrature q
Where there are several circuits in any axis they are
to be distinguished from each other by numeral subscipts, as X2d, except that the symbol, f, may be
retained for the field winding proper. If no subscript
is employed reference to the main field winding is

implied._
36.

Reference (11).

closed-r0'

D. ALL SUBSCRIPTS CAPITALIZED

In the present scheme of notation, this form is

reserved for the particular single-phase static reactances
used by Doherty and Nickle in their paper Synchronous
Machines IV, and discussed here in Appendix B.
These are:

XD
total static reactance
static reactance due to space fundamental XAD
XLD
static leakage reactance
XDI
static transient reactance
XAD'
static transient fundamental reactance
As these reactances are single-phase they are also
characterized by S and ; to indicate whether they are
line-to-line or line-to-neutral, thus
XSD, or X2D.
Corresponding quadrature values are designated
similarly by the subscript Q instead of D.
E. MISCELLANEOUS TERMS
E
ICLAEU EM
It is necessary to include a number of special terms
which have not been discussed above. The notation
used for these quantities iss more
u
rbtrr but
es arbitrary,
or orr less

an attempt has been made to employ accepted symbols

as much as possible.

h = a subscript denoting any rotor circuit. In

conjunction with d or q it signifies that that
circuit is in the direct or quadrature axis.
K 37 = a per-unit quantity expressing the ratio of the
total flux to the space fundamental component of flux entering the armature along
one pole pitch.
I = used as asubscript to denote leakage reactance,
or used alone to denote length of armature
stacking.
L3 = effective leakage lines per unit length of stacking, linking the entire field winding when the
field m. m. f. is one ampere-turn.
37. First employed by Doherty and Shirley in their paper
Reactance, reference 10.

Feb. 1928

PARK AND ROBERTSON: SYNCHRONOUS MACHINES

535

Originally, the method of representation used was

a subscript denoting mutual reactance between
referred to as "percentage representation of quantities,
armature and any field winding.
n
a subscript, similar to h, denoting any field as a decimal," but this name was to some extent a source
circuit, or used alone to denote effective of confusion in view of the accepted meaning attached
to the word "percentage." Recently the term "perarmature turns per pole per phase.
unit representation of quantities" has been suggested
N -turns per pole of the main field winding.33
A = a coefficient employed to denote increment as a designation for the decimal percentage system.
As this name appears to solve the difficulties previously
change in any quantity.
it has been used throughout the paper.
encountered,
devise
to
necessary
found
been
it
has
In Appendix E
they
that
special notations for certain quantities in order
References
fit the construction of the equivalent circuits. These
Thences
Calculation of the Reactance of
1. 1928 Machines.
Alger, P. L.,SeeThep. 493.
notations employ parentheses and brackets. Their Synchronous
use may be made clear by reference to that appendix.
2. 1928, Doherty, R. E. and Nickle, C. A., Synchronous
Machines IV. See p. 457.
PRINCIPAL DERIVED UNIT QUANTITIES
Schaefer, R. A., A Study of Transverse Armature
The
niqanii3.
F. are
Reaction1927,
F. The principal derived unit quantities
in Synchronous Machines, JOURNAL A. I. E. E., August,
1927, pp. 811-814.
,(o) = normal armature flux-linkages
4. 1927, Wagner, C. F. and Dovjikov, A., "Impedance of a
= e(o)
Rotating Synchronous Machine to a Negative Sequence Voltage"
10
o
Electric Jl., March 1927, pp. 117-21.
2 7r f
of
Determination
R. W., Graphical
5. 1927,
normal
Synchronous
to Salient-Pole
Practical Application
Fields.Wieseman,
Magnetic
normal armature)flux
armature flux
(0)=
Machine Design, JL. A. I. E. E., Vol. 46, 1927, pp. 430-31;
O(o)
m

prncpa

erve

Discussion, pp. 614-20.

6. 1927, Doherty, R. E. and Nickle, C. A., Synchronous

Tto) normal flux-linkages of the main field winding Machines III, Torque-angle Characteristics Under Transient
Conditions. TRANS. A. I. E. E., Vol. XLVI, 1927, p. 1.
= N 0(o)
7. 1927, Douglas, J. F. H., Transverse Reaction in Synchrovariation
by
the
produced
voltage
field
-per-unit
E(o)
at normal frequency of normal flux-linkages nous Machines, JL. A. I. E. E., February 1927, pp. 109-15.
8. 1926, Doherty, R. E. and Nickle, C. A., Synchronou8
in the main field
Machines I and II, An Extension of Blondel's Two Reaction
Theory, A. I. E. E. TRANS., Vol. 45, pp. 912-47.
= wf''(0) 1O-8
9. 1926, Biermanns, J., Uberstrome in HochspannungAppendix J
sanlagen, 1926, p. 154.
PER-UNIT REPRESENTATION OF QUANTITIES
PER-UNIT REPRESENTATION OF QUANTITIES
10. 1918, Doherty R. E. and Shirley, 0. E., Reactance of
Messrs. Doherty and Nickle, in their recent series of Synchronous Machines and Its Application, A. I. E. E. TRANS.,
papers on Synchronous Machines, have employed the 1918, Vol. 37, part 2, pp. 1209-97; discussion pp. 1298-1340.
11. 1918, Fortescue, C. L., Method of Symmetrical Coordinates
practise of representing quantities as a decimal fraction
to the Solution of Polyphase Networks, A. I. E. E. TRANS.,
shares Applied
method shares
value. This
This methodl
or base
base value.
some normal
normal or
ofof some
1918 vol. 37,p 1037.
1 B.,
with the percentage method of representation, the 19 12. . 13
., Te Wechselstromtechn2k, 1913, Vol. IV,
d,p
1913, Arnold,
marked advantage that it eliminates the necessity of p. 214.
13. 1913, Blondel, A., Synchronous Motors and Converters,
employing unwieldy conversion factors in purely
analytical work. At the same time, it is definitely a translation by C. 0. Mailloux, 1913, Part III.
14. General Equations of an "Ideal Synchronous Machine."
superior to the percentage system in that it is unnecessary to continually consider the factor "one hundred." An article by one of the authors to appear shortly in the G. E.
In the percentage system this factor must be specially Review.
inserted or divided out whenever the operations of
Discussion
division or multiplication are performed. Therefore,
A. E. Kennelly: At the present time, according to the
on account of the convenience of the decimal fraction
information which has been presented to us, there are possibly
system it will be used hereafter in the paper.
thirty-one varieties of reactance in electrical machinery.
normal
a
have
will
each
in
quantity
Thus, the paper,
to in
It seems important that these useful concepts referred
be
time
will
at
of
the
quantity any
value, and the value
reactance," "differential
flthese papers as "reactanee,"on,"leakage
and
set
forth
be
should
and
so
clea.rly
value,
leakage
of
its
normal
reacetance,"
expressed as a decimal fraction
uses
when
a
speaker
in
the
future,
because
evidently,
as
defined
are
linkages
armature
tabula,ted,
normal
For example,
=

the linkages which generate normal voltage when varied

the word "reactance," his audience will want to say, "Which?"

are 45 per cent of this normal value,

value seems to me of great importance. Percentage, of course,

at normal frequency. Thus, ,6 = 0.45 means that It is increasingly important that this matter be thoroughly
the armature linkages at the instant of consideration classified.
The concept which has been presented to us of the per-unit

38. Effective turns in the case of turbo alternators and in- is an expression that we are constantly employing and is very
.useful, but in computations, percentage is not so convenient,
duction motors.