3712 IEEE TRANSACTIONS ON MAGNETICS, VOL. 34, NO.

5, SEPTEMBER 1998

Calculation of Synchronous Reactances of Small

Permanent-Magnet Alternating-Current Motors:
Comparison of Analytical Approach and
Finite Element Method with Measurements
Jacek F. Gieras, Ezio Santini, and Mitchell Wing

Abstract— The synchronous reactances of permanent magnet of the magnetic field. This distribution is very helpful to
(PM) motors have been determined using: 1) analytical method, correctly estimate the form factors of the rotor and stator
i.e., form factors of the stator field (armature reaction factors), magnetic flux densities.
2) finite element method (FEM), and 3) experimental tests on
a special machine set. The analytical method is widely used The finite element method (FEM) makes it possible to
in calculations of synchronous reactances of salient pole syn- find the - and -axis synchronous reactances and mutual
chronous machines with electromagnetic excitation. Rotors of (armature reaction) reactances by computing the corresponding
PM synchronous machines have more complicated structures, inductances, e.g., [3], [4], [6], [18]–[20]. It can be done by
hence it is more difficult to predict accurately the magnetic field using the flux linkage and magnetic vector potential concept or
distribution in their airgaps in order to find the form factors of
the stator field. Numerical methods of field analysis can easily energy stored in the winding. Recently, two modern FEM tech-
solve this problem. The FEM can predict both the synchronous niques in ac machines analysis have emerged: current/energy
and mutual (armature reaction) reactances in the d and q axes. perturbation method [7], [8], [15], [23] and time-stepping
The leakage reactance can then be evaluated as a difference analysis [1], [5]. These methods are especially suitable for
between synchronous and mutual reactances. As an example, a transient analysis of converter-fed PM synchronous machines.
small, three-phase, four-pole motor with SmCo surface mounted
PM’s (three parallel magnets per pole), and mild-steel pole shoes The measurement of the synchronous reactances for small
has been investigated. Such a complicated rotor structure has PM synchronous motors seems to be more difficult. There
been intentionally designed in order to be able to compare the are several methods for the measurement of synchronous
advantages and disadvantages of the analytical method and the reactances of medium and large synchronous machines but the
FEM. In the FEM, the reactances have been calculated using assumptions made do not allow one to apply these methods
both the flux linkage and current/energy perturbation method.
Synchronous reactances as functions of the stator current and to small PM synchronous motors. In this paper a special
load angle obtained analytically from the FEM modeling and laboratory setup is presented for the load angle estimation
from measurements have been compared. of small synchronous motors. The measured load angle, input
Index Terms—Analytical approach, FEM, measurements, per- voltage, armature current, armature winding resistance, and
manent magnet ac motors, synchronous reactance. power factor allow for finding the synchronous reactances
and on the basis of the phasor diagram.
I. INTRODUCTION
II. ANALYTICAL APPROACH

T HE accuracy of calculating the steady-state performance

of small permanent magnet (PM) synchronous motors
depends on the accuracy of calculating the synchronous reac-
As it is known the - and -axis synchronous reactances
are defined as
tances in the and axes. For typical medium power and large (1)
synchronous motors with electromagnetic excitation analytical
methods using form factors of the stator (armature) magnetic where is the -axis mutual (armature reaction) reactance,
flux density are good enough. Small PM synchronous motors is the -axis mutual reactance, and is the armature
sometimes have a complicated structure, and numerical or winding leakage reactance per phase. The reactance and
analog modeling is necessary to obtain an accurate distribution sometimes depend on the magnetic saturation due to
Manuscript received March 31, 1996; revised March 12, 1998. This work the main flux. The leakage reactance consists of the slot,
was supported by the Foundation for Research Development (FRD) of the differential, tooth–top, and end-connection leakage reactances.
Republic of South Africa, Italian Ministry of Scientific and Technological Only the slot and differential leakage reactances depend on the
Research, and the University of Rome, La Sapienza.
J. F. Gieras and M. Wing are with the Department of Electrical Engi- magnetic saturation due to leakage fields. It can be taken into
neering, University of Cape Town, Rondebosch 7700, South Africa (e-mail: account using Norman’s method [16].
[email protected]). The analytical approach to calculating the synchronous
E. Santini is with the Department of Electrical Engineering, University of
Rome, La Sapienza, 00184 Rome, Italy. reactances is based on the distribution of the stator winding
Publisher Item Identifier S 0018-9464(98)06332-8. normal component magnetic flux density. This distribution
0018–9464/98$10.00  1998 IEEE
GIERAS et al.: CALCULATION OF SYNCHRONOUS REACTANCES 3713

can be assumed to be a periodic function of the stator inner

perimeter or can be found using numerical or analog modeling.
The -axis mutual reactance and -axis mutual reactance are
expressed in terms of form factors of the stator field (armature
reaction factors) and i.e., [12]

(2)

The mutual reactance is the same as that for a cylindrical-

rotor synchronous machine

(3) (a)

where is the number of stator (armature) phases, is the

permeability of free space, is the input frequency, is the
number of stator turns per phase, is the stator winding
factor for the fundamental space harmonic, is the number of
pole pairs, is the pole pitch, is the effective length of the
armature core, is Carter’s coefficient for the airgap, and
is the airgap in the -axis. To obtain a saturated synchronous
reactance, the equivalent airgap should be multiplied by
the saturation factor of the magnetic circuit, i.e., to
obtain In most salient pole synchronous machines
with electromagnetic excitation, the magnetic saturation affects
only since the -axis airgap is comparatively very large. (b)
In some PM synchronous machines the magnetic saturation
affects both and
The form factors of the stator field are defined as the
ratios of the first harmonic amplitudes-to-the maximum values
of normal components of stator (armature) magnetic flux
densities in the -axis and -axis, respectively, i.e.,

(4)

The peak values of the fundamental harmonics and

of the stator magnetic flux density can be calculated
as coefficients of Fourier series for the fundamental harmonic,
i.e., (c)
Fig. 1. Distribution of the d-axis and q -axis magnetic flux density for
rare-earth PM rotors: (a) inset-type PM rotor, (b) surface PM’s, and (c) buried
PM rotor.

(5)

For the distributions of - and -axis, magnetic flux densities

according to Fig. 1, the first harmonics of the magnetic flux
densities are
i) in the case of inset-type PM’s [Fig. 1(a)] (7)

ii) in the case of surface PM’s [Fig. 1(b)]

(8)

(6) (9)
3714 IEEE TRANSACTIONS ON MAGNETICS, VOL. 34, NO. 5, SEPTEMBER 1998

TABLE I for salient pole synchronous motors with electromagnetic

REACTION FACTORS FOR PM SYNCHRONOUS MACHINES excitation [12].

III. FEM

A. Approach
The two following concepts are the most frequently used in
the FEM computations of the steady-state inductances:
1) the number of flux linkages of the coil, divided by the
current in the coil;
2) the energy stored in the coil, divided by one-half the
current squared.
Both concepts give identical results for linear inductances but
not the same for nonlinear inductances [13].
If the incorrect potential distribution does not differ very
greatly from the correct one, the error in energy is much
smaller than that in potential [21]. Therefore, the steady-state
inductances are often very accurately approximated even if the
potential solution contains substantial errors.
In order to predict accurately the dynamic behavior and
iii) in the case of buried PM’s [Fig. 1(c)] performance of a converter-fed PM brushless motor, one must
have accurate knowledge of values of the self- and mutual
winding dynamic inductance rather then the steady-
state value [15].
In the early 1980’s, the current/energy perturbation method
of calculating inductances was proposed [7], [8], [15], [23].
This method is based upon consideration of the total energy
(10) stored in the magnetic field of a given device comprising
windings.
In the papers [8], [15] the self and mutual inductance terms
of the various windings have been expressed as the partial
derivatives of the global stored energy with respect to
various winding current perturbations These derivatives
can, be expanded around a “quiescent” magnetic field solution
(11)
obtained for a given set of winding currents, in terms of
For inset type PM’s the coefficient expresses an increase various current perturbations and in the th and
in the -axis armature magnetic flux density due to a decrease th windings and the resulting change in the global energy.
in the airgap [Fig. 1(a)] from to where is the The details are given in [7], [8], [15], and [23]. For the
depth of the slot for the PM. Since the magnetic voltage two-dimensional (2-D) field distribution the current/energy
drop across is equal to the sum of the magnetic perturbation method does not take into account the end wind-
voltage drops across the airgap and ferromagnetic tooth ing leakage.
height , the following equality can be written: For the steady-state conditions a similar accuracy can be
Because obtained calculating first the synchronous reactance, then
, the coefficient of increase in the magnetic mutual reactance, and finally the slot and differential leakage
flux density due to a decrease in the airgap is reactance as a difference between synchronous and mutual
Of course, when , then , , reactances.
and It means that the machine behaves as In this paper the inductances will be found on the basis of
a cylindrical rotor machine. Similarly, for the surface magnet flux linkage, Stokes’ theorem, and magnetic vector potential,
rotor [Fig. 1(b)], since the relative magnetic i.e.,
permeability of rare-earth PM’s For the buried magnet
rotor the -axis armature flux density changes as and
the axis armature flux density changes as [Fig. 1(c)]. (12)
The parameter is the ratio of the PM or pole shoe
width to the pole pitch and can also be expressed as an where is the magnetic vector potential around the contour
angle. The synchronous reactance is when the total flux or
The coefficients and for different rotor configu- includes both mutual and leakage fluxes, i.e., the stator
rations are given in Table I. The last row shows and slot, tooth top, and differential leakage flux.
GIERAS et al.: CALCULATION OF SYNCHRONOUS REACTANCES 3715

The magnetic flux through the airgap does not include the
stator leakage, but simply the -axis and -axis linkage flux
or The first harmonics of these main fluxes give the
mutual (magnetizing) reactances [19], [20]. A combination of
the total fluxes , , and linkage fluxes and will
give the stator leakage reactance (excluding the end connection
leakage reactance).
If the armature current , then it follows that the
normal component of the rotor magnetic flux density deter-
mines the location of the -axis. A line integral through the
airgap gives the distribution of the magnetic vector potential.
The values of constant vector potential represent flux lines.
Numerical Fourier analysis of this vector potential yields an
analytical expression for the fundamental harmonic, i.e.,
(13)
where and The angle
relates to the -axis which is the -axis of the FEM model
since it shows the angle of zero crossing of the magnetic
vector potential through the airgap line contour. This angle is
usually found to be zero due to the symmetry in the machine.
Fig. 2. Phasor diagram of a synchronous motor.
The -axis is related to the -axis by a shift of thus

These permeabilities can then be used for the no-load linear

B. Synchronous Reactances calculation. This ensures that the saturation effect which occur
In a 2-D FEM model the -axis and -axis synchronous in the loaded model is not ignored in the no-load results.
reactances, respectively, excluding the end leakage flux are
C. Mutual Reactances
(14) The -axis and -axis fundamental components of magnetic
flux in the airgap can be derived by performing a Fourier
where and are the magnitudes of the total fluxes in the analysis on the vector potentials around the inner surface of
and -axis, respectively, and are the magnitudes of the armature core. In Fourier series the cosine term coefficient
the -axis and -axis stator currents, respectively, and is the expresses the quantity of half the -axis flux per pole and
input frequency. The -axis and -axis fluxes are obtained from the sine term coefficient expresses the quantity of half the
the combination of the phase belt linkages [3], [4], [6], [18]. -axis flux per pole [20]. Thus, the fundamental harmonic of
According to the phasor diagram for a synchronous motor, the the resultant flux per pole and the inner torque load angle
rotor excitation flux and -axis armature flux are in the same (Fig. 2) are
direction while the -axis armature flux is perpendicular.
The calculation of the synchronous reactances using (14) are
(16)
sensitive to the values of and , respectively. This is due
to the fact that in the full-load analysis of a PM synchronous
motor the values of and both approach zero at different This magnetic flux rotates at the synchronous speed
load angles. The rounding off errors that occur in the FEM, (in revolutions per second) and induces in each phase winding
amplify the error in the synchronous reactance as the armature the following internal EMF:
current components tend toward zero. The use of a constant
current disturbance, in both and , solves this problem.
The synchronous reactances are then (17)

(15) The phasor diagram shown in Fig. 2 gives the following

expressions for the -axis and -axis mutual reactances:
where and are the changes in flux linkages at
the load point in the -axis and -axis, respectively, and (18)
and are the values of the -axis and -axis stator current
disturbances, respectively. (19)
The magnetic saturation that occurs in the loaded finite ele-
ment solution should be transferred to the unloaded problem. Armature reaction affects saturation of magnetic circuit at dif-
This is done by storing the permeability of every element ob- ferent loads. To take into account this effect another equation
tained from the loaded nonlinear magnetic field computation. similar to (18) is used [20].
3716 IEEE TRANSACTIONS ON MAGNETICS, VOL. 34, NO. 5, SEPTEMBER 1998

Fig. 3. Laboratory setup for measuring the load angle  of PM synchronous

motors: TSM — tested synchronous motor, ASM — additional synchronous
motor with the same number of poles as TSM, PM — prime mover
(synchronous or dc motor), B — brake, and DBO — dual-beam oscilloscope.
Fig. 4. Cross section of the tested PM rotor.

D. Leakage Reactance
The armature leakage reactance can be obtained in two
ways:
1) from numerical evaluation of the energy stored in the
slots and in the end connections;
2) as the difference between the synchronous reactance and
armature reaction reactance, i.e.,
or (20)

In general, the calculation of the end connection leakage

reactance is rather difficult using a 2-D FEM program and (a) (b)
does not bring satisfactory results. For this purpose a three-
dimensional FEM software is recommended [23].

IV. EXPERIMENTAL METHOD

A reliable method of measurement of synchronous reac-
tances results from the phasor diagram (Fig. 2). Thus, the
following equations can be written:

(21)

(c) (d)
(22)
Fig. 5. Magnetic flux distribution (FEM) in the cross section of the tested
The -axis synchronous reactance can be found from (21) and motor: (a) at no-load, (b) with load at 10 Nm shaft torque, (c) d-axis stator
the -axis synchronous reactance can be found from (22), i.e., flux, and (d) q -axis stator flux.

(23) windings can be found. When the TSM is connected to a three-

phase power supply and the oscilloscope receives the signals
(24) and , the load angle can be measured, i.e.,
the phase angle between instantaneous values of and
It is easy to measure the input voltage , phase armature Equations (23) and (24) allows for the investigation
current , armature winding resistance per phase, and the of how the input voltage and load angle affect and
angle To measure the load angle It is necessary to say that the accuracy of measurement
it is recommended to use the equipment [17] as shown in of the angle much depends on the higher harmonic contents
Fig. 3. The EMF induced by the rotor flux is assumed to in and
be equal to the no-load EMF, i.e., at
The terminals of the two corresponding phases of the tested V. EXPERIMENTAL VERIFICATION OF
synchronous motor (TSM) and additional synchronous motor THE ANALYTICAL APPROACH AND FEM
(ASM) are connected to a dual-beam oscilloscope. The no-
load EMF’s of TSM and of ASM operating A. Construction of the Tested Motor
as generators should be in phase. Thus, the same positions The stator of a small 1.5-kW, 4-pole, 1410-rpm, 380-V,
of the TSM and ASM rotors with regard to the same phase three-phase, Y-connected, cage induction motor manufactured
GIERAS et al.: CALCULATION OF SYNCHRONOUS REACTANCES 3717

Fig. 6. Distributions of the d-axis and q -axis armature magnetic flux density in the airgap of the tested motor.

by GEC (South Africa) has been selected. The stator consists

of a laminated core made of cold-rolled electrotechnical sheet-
steel and a three-phase double-layer winding. There are 36
semi-open slots with the slot opening equal to 2.25 mm. The
number of turns per phase is and the coil span is
equal to the pole pitch. The stator inner diameter is 82.54 mm
and the effective length of the armature stack is mm.
The commercial cage rotor has been replaced with a surface
PM rotor with mild-steel pole shoes [10]. The yoke and shaft
have been made of one piece of a solid carbon steel. The
length of the rotor solid core is 100 mm. Three parallel 10-
mm wide and 5-mm thick surface SmCo PM’s per pole have
been designed as shown in Fig. 4. To reduce the airgap, make Fig. 7. Form factors kfd and kfq as functions of = bp = for the surface
PM rotor with thin mild-steel pole shoes.
the motor self-starting, and improve its stability, the PM’s have
been equipped with thin mild-steel pole shoes covering 83% of
the width of PM’s. The -axis airgap (mechanical clearance) mild-steel pole shoes are shown in Fig. 7. For the rotor of the
is mm, the -axis airgap is mm, and the tested motor Then the leakage
thickness of the mild-steel pole shoe is mm. The PM’s reactance and armature
have been magnetized radially. reaction reactanes and
Fig. 5 shows the magnetic flux distribution through the cross according to (2), (3), (25), and (26) have been calculated
section of the motor as obtained with the aid of the FEM. The for Hz. For A the leakage reactance
-axis and -axis magnetic flux density distributions in the This reduction is caused by the magnetic
airgap along the pole pitch are plotted in Fig. 6. saturation due to leakage fields [16]. In order to find the
leakage reactances, classical equations for slot (double-layer
winding), end-connection, differential, and tooth-top leakage
B. Analytical Approach permeances have been used [12].
The form factors of the armature reaction can be derived However, and have been obtained with the mag-
on the basis of Fig. 6, similarly to those in Section II. For the netic saturation due to the main flux being neglected. Fig. 8
surface PM rotor with mild-steel pole shoes the and -axis shows the mutual reactances and plotted against the
form factors are armature current for the analytical approach
The synchronous reactances according to (1) at Hz,
, and are and
(25)

(26) C. FEM
The computation on the basis of the FEM shows that
the mutual reactances are dependent on the armature current
The coefficient is evaluated in a similar (Fig. 8) which means that the magnetic saturation is included.
way as the coefficient in Section II, where is the Owing to the thin mild-steel pole shoe which is subject to
thickness of the mild-steel pole shoe and is the airgap in the magnetic saturation, the -axis mutual reactance is more
axis. For the tested machine The form factors sensitive to the armature current than the -axis reactance [see
and versus for the rotor with surface PM’s and also Fig. 5(c) and (d)].
3718 IEEE TRANSACTIONS ON MAGNETICS, VOL. 34, NO. 5, SEPTEMBER 1998

Fig. 8. Xad and Xaq as functions of the rms stator current Ia at f = 50 Hz. Computation results using analytical approach and FEM.

(a) (a)

(b) (b)
Fig. 9. Synchronous reactance Xsd at constant terminal voltage of 220 V Fig. 10. Synchronous reactances Xsq at constant terminal voltage of 220 V
and f = 50 Hz versus: (a) the stator current Ia and (b) the load angle : and f = 50 Hz versus: (a) the stator current Ia and (b) the load angle :

The calculation of the leakage reactances has been done E. Comparison

using (20). Figs. 9 and 10 show the -axis and -axis synchronous
reactances obtained from analytical, finite element, and exper-
imental methods. In both cases, the FEM results are closer
D. Experimental Tests to the test results than those obtained from the analytical
Experimental tests have been performed using the machine approach.
set according to Fig. 3. It was possible to test the motor only The load angle in the range of armature current
in the range of the stator current from 5.7–17.5 A. The no-load A can be both positive or negative. Consequently,
current is approximately 5.7 A. the stator current decreases and then increases as the angle
GIERAS et al.: CALCULATION OF SYNCHRONOUS REACTANCES 3719

recommended. In this way the difficult problem of laboratory

measurements of and for small synchronous motors
seems to be successfully solved.

REFERENCES
[1] A. Arkkio, “Time stepping finite element analysis of induction motors,”
presented at Int. Conf. Elect. Mach. ICEM’88, Pisa, Italy, 1988, vol. 1,
pp. 275–290.
[2] V. A. Balagurov, F. F. Galtieev, and A. N. Larionov, Permanent Magnet
Electrical Machines, (in Russian). Moscow, Russia: Energia, 1964.
[3] M. V. K. Chari and P. P. Silvester, “Analysis of turboalternator magnetic
fields by finite elements,” IEEE Trans. Power Appart. Syst., vol 92, pp.
454–464, 1973.
[4] M. V. K. Chari, Z. J. Csendes, S. H. Minnich, S. C. Tandon, and J.
Fig. 11. Stator leakage reactance X1 versus armature current Ia at f = 50 Berkery, “Load characteristics of synchronous generators by the finite-
Hz. element method,” IEEE Trans. Power Appart. Syst., vol. 100, no. 1, pp.
1–13, 1981.
[5] A. Demenko, “Time stepping FE analysis of electric motor drives with
semiconductor converters,” IEEE Trans. Magn., vol. 30, pp. 3264–3267,
between and increases from negative to positive values. Sept. 1994.
This is why two values of and for small values of the [6] N. A. Demerdash and H. B. Hamilton, “A simplified approach to deter-
armature current have been obtained [Figs. 9(a) and 10(a)]. mination of saturated synchronous reactances of large turbogenerators
under load,” IEEE Trans. Power Appart. Syst., vol. 95, no. 2, pp.
Classical approach neglects this effect. 560–569, 1976.
The computation of the leakage reactance on the basis of [7] N. A. Demerdash, F. A. Fouad, and T. W. Nehl, “Determination of
the FEM gives slightly different results when using winding inductances in ferrite type permanent magnet electric machinery
by finite elements,” IEEE Trans. Magn., vol. 18, pp. 1052–1954, Nov.
and Both results are very close to the analytical 1982.
results (Fig. 11). [8] N. A. Demerdash, T. M. Hijazi, and A. A. Arkadan, “Computation
of winding inductances of permanent magnet brushless dc motors
with damper windings by energy perturbation,” IEEE Trans. Energy
Conversion, vol. 3, no. 3, pp. 705–713, 1988.
[9] E. F. Fuchs and E. A. Erdélyi, “Determination of waterwheel alternator
VI. CONCLUSION steady-state reactances from flux plots,” IEEE Trans. Power Appart.
The analytical and finite element approach to calculating the Syst., vol. 91, pp. 2510–2527, 1972.
[10] J. F. Gieras and M. Wing, Permanent Magnet Motors Technology:
synchronous reactances have been compared with the exper- Design and Applications. New York: Marcel Deker, 1996.
imental tests on a new type of PM synchronous motor with [11] V. B. Honsinger, “Performance of polyphase permanent magnet ma-
chines,” IEEE Trans. Power Appart. Syst., vol. 99, no. 4, pp. 1510–1516,
surface PM’s and mild-steel pole shoes. A full comparative 1980.
analysis (FEM flux linkage, FEM current/energy perturbation, [12] M. Kostenko and L. Piotrovsky, Electrical Machines, vol. 2. Moscow,
classical approach, and measurements) of reactances for wide Russia: Mir Publishers, 1974.
[13] D. A. Lowther and P. P Silvester, Computer-Aided Design in Magnetics.
range of the load has been done. Berlin: Springer-Verlag, 1986.
Since the popularity of PM synchronous motors in modern [14] S. A. Nasar, I. Boldea, and L. E. Unnewehr, Permanent Magnet, Re-
electric drives has recently increased, there is a need for luctance, and Self-Synchronous Motors. Boca Raton, FL: CRC Press,
1993.
reliable methods of prediction of operating characteristics for [15] T. W. Nehl, F. A. Fouad, and N. A. Demerdash, “Determination
newly designed machines. The accuracy of calculating the of saturated values of rotating machinery incremental and apparent
inductances by an energy perturbation method,” IEEE Trans. Power
armature current, electromagnetic power, and electromagnetic Appart. Syst., vol. 101, no. 12, pp. 4441–4451, 1982.
torque of a synchronous motor depends first of all on the [16] H. M. Norman, “Induction motor locked saturation curves, Elec. Eng.,
correct evaluation of the synchronous reactances and pp. 536–541, 1934.
[17] I. L. Osin, V. P. Kolesnikov, and F. M. Yuferov, Permanent Magnet
Although simple analytical equations describing the form Synchronous Micromotors, (in Russian). Moscow, Russia: Energia,
factors of the armature field are adequate for evaluating the 1976.
synchronous reactances of typical salient-pole synchronous [18] D. Pavlik, V. K. Garg, J. R. Repp, and J. Weiss, “A finite element
technique for calculating the magnet sizes and inductances of permanent
machines with electromagnetic excitation, they cannot always magnet machines,” IEEE Trans. Energy Conversion, vol. 3 no. 1, pp.
bring good results in the case of small PM motors. PM ac 116–122, 1988.
motors often have intricate rotor structure and the airgap [19] M. A. Rahman and A. M. Osheiba, “Performance of large line-start per-
manent magnet synchronous motors,” IEEE Trans. Energy Conversion,
magnetic field is difficult to express by analytical equations. vol. 5, no. 1, pp. 211–217, 1990.
In the tested motor with surface PM’s and thin mild-steel [20] M. A. Rahman and P. Zhou, “Determination of saturated parameters of
PM motors using loading magnetic fields,” IEEE Trans. Magn., vol. 27,
pole shoes, the FEM gives better results than the analytical pp. 3947–3950, Sept. 1991.
approach (Figs. 9 and 10). Hence, the classical analytical [21] P. P. Silvester and R. L. Ferrari, Finite Elements for Electrical Engineers,
approach is not recommended in designing high-performance 2nd ed. Cambridge, U.K.: Cambridge Univ. Press, 1990.
[22] W. Szela̧g, “Numerical method for determining the parameters of perma-
PM synchronous machines for modern electric drive sys- nent magnet synchronous motors,” presented at 12th Symp. Electromagn.
tems. Phenomena Nonlinear Circuits, Poznan, Poland, 1991, pp. 331–335.
For experimental verification of synchronous reactances a [23] R. Wang and N. A. Demerdash, “Comparison of load performance and
other parameters of extra high speed modified Lundell alternators for
combination of the equation resulting from the phasor diagram 3-D-FE magnetic field solutions,” IEEE Trans. Energy Conversion, vol.
and measurement of the load angle on a multimachine set is 7, no. 2, pp. 342–352, 1992.
3720 IEEE TRANSACTIONS ON MAGNETICS, VOL. 34, NO. 5, SEPTEMBER 1998

Jacek F. Gieras graduated in 1971 from the Technical University of Lódź, Mitchell Wing received the B.Sc. degree in 1990, M.Sc. degree in 1992, and
Poland. He received the Ph.D. degree in 1975 and D.Sc. (Dr hab) degree in Ph.D. degree in 1996 in electrical engineering from the University of Cape
1980 from the University of Technology, Poznań, Poland. Town, South Africa.
From 1971 to 1987, he was with Poznań University of Technology, Poznań, Currently, he is a Postdoctoral Fellow in the Department of Electrical
and Academy of Technology and Agriculture, Bydgoszcz, Poland. In 1975 Engineering at the University of Cape Town. He co-authored about 30 papers
and 1976, he was a Visiting Researcher at the Czechoslovak Academy of on permanent magnet electrical machines and one book, Permanent Magnet
Sciences, Prague, Czechoslovakia. From 1983 to 1985, he was a Research Motor Technology: Design and Applications (New York: Marcel Deker, 1996).
Visiting Professor at Queen’s University, Kingston, Ontario, Canada. In 1987,
he was promoted in Poland to the rank of Full Professor (life title) in
Electrical Engineering. From 1987 to 1989, he was with the Department
of Electrical Engineering at Jordan University of Science and Technology,
Irbid, the Hashemite Kingdom of Jordan. From 1989 to 1998, he was with
the Department of Electrical Engineering at the University of Cape Town,
South Africa. In 1994, he was a Visiting Professor at the University of
Rome, La Sapienza, Italy. In 1996, he was a Japan Railway Central Company
Visiting Professor (Endowed Chair in Transportation System Engineering) at
the University of Tokyo, Japan. In 1996 and 1997, he was a Guest Professor at
Chungbuk National University, Cheongju, South Korea. He is now involved
in high technology research in the United States. He has authored and co-
authored 6 books and about 200 papers and patents. The best known books
are Linear Induction Drives (London, U.K.: Oxford Univ. Press, 1994) and
Permanent Magnet Motor Technology: Design and Applications (New York:
Marcel Deker, 1996) co-authored with Dr. M. Wing.