Paper accepted for presentation at the 2011 IEEE Trondheim PowerTech

Capacitances Calculation Using FEM for Transient

Overvoltage and Common Mode Currents
Prediction in Inverter-Driven Induction Motors
C. S. Chaves, J. R. Camacho, Senior Member IEEE, H. de Paula, M. L. R. Chaves and E. Saraiva

practical difficulties [5], which turn the procedure into

Abstract – The prediction of transient overvoltage and common- something generally impracticable. Whereas there are many
mode currents in inverter-driven induction motors is necessary to simplifications in analytical formulations and some specific
propose mitigating solutions, in order to prevent failures and
ensure the proper operation of the entire electrical system. The
motor constructive details must be known [3, 4, 5, 6].
motor model for these high-frequency studies is relatively simple, In this context, this paper presents a new alternative, based
however to find its parameters accurately is a very complex task. on the finite element methods (FEM), to compute the
In this context, this paper presents an alternative to estimate capacitances Csf, Csr and Crf through electrostatic analysis
leakage capacitances in induction motors, based on the finite provided by the software FEMM (Finite Element Methods
element techniques, using FEMM (Finite Element Methods Magnetics). In sequence, an application of such capacitance
Magnetics). Then, an example of the motor capacitances in an
appropriate model for common-mode quantities determination is
values in the motor equivalent circuit is shown for the analysis
shown, using the software ATP (Alternative Transients of the common-mode phenomena through simulations in ATP
Program). (Alternative Transients Program).

Index terms – ATP, common-mode currents, FEMM, high- II. MOTOR REPRESENTATION IN FEMM
frequency induction motor model, leakage capacitance
calculation. In this research, a 4 pole, 220 V and 3 HP induction motor
was used. The prediction of Csf, Csr and Crf in FEMM was
I. INTRODUCTION made based on a detailed representation of the motor in two
dimensions, taking into account all the relevant material
T HE power quality is a subject that does not only consider
the standard wave forms of the supplied voltage, but also
the good performance of electrical machines. These are always
properties (dielectric constants and the voltage applied on
conductors) for this calculation. In the complete motor
representation, it was considered the rotor, magnetic core and
being studied due to their great relevance and contribution to
windings of the stator (including the coil head), frame,
the electrical system. The study of side effects of the joint
bearings and materials responsible for the wire insulation,
operation between induction motors and inverters, even
such as the varnish and insulating film. However, when it
though it aims to reduce the electrical power consumption, is
comes to rotor, bearings and frame, only their surfaces were
necessary to prevent the decrease of motor lifetime, mistakes
considered in the representation, since the internal structure of
on protection equipment specification and EMC issues.
these components was not relevant to determine such
Regarding three-phase inverters (PWM), it is well-known
capacitances. It was made a sensitive research about the input
that the instantaneous combination of its output phase voltages
parameters in the motor representation, being shown in this
is a non-zero value. This resulting voltage is characterized by
section.
a step waveform, which variations correspond to one third of
the DC link voltage of the converter [1]. Each variation of this A. Dielectric Constant of Insulating Materials
common-mode voltage excites the capacitive couplings To make an electrostatic study in FEMM, the
between connecting cables, converter and motor to ground, representation must have conductors and dielectric materials,
providing the circulation of common-mode currents [1]. These which generally provides the insulation. Motors have
currents are responsible for the improper operation of fault insulating materials like varnish, impregnation resin and
relays to ground, electromagnetic interference (EMI), bearing insulated film. Each of them has a dielectric constant – K,
failures and other problems [2, 3, 4]. which is defined by the material atomic properties. This
Concerning the motor, it is included in its model for high- constant can be changed because of the material state and
frequency studies the capacitive couplings: capacitance motor operation conditions, like voltage frequency [7] and
between stator windings and frame (Csf), stator windings and temperature, being difficult to define an exact value for K in
rotor (Csr), rotor and frame (Crf) and bearings (Cb). Despite the this high-frequency study.
simplicity of the motor equivalent circuit to represent their In order to analyze the influence of K variation in the
capacitances, to obtain these parameters is very complex. leakage capacitance results, it was made a sensitive study on
Measurements require high-accuracy instruments and have

978-1-4244-8417-1/11/$26.00 ©2011
Authorized licensed use limited to: CENTRO FED DE EDUCACAO TECNOLOGICA DE MINAS GERAIS. Downloaded on May 22,2023 at 23:27:35 UTC from IEEE Xplore. Restrictions apply.
 2

TABLE I
VARIATION WITH FREQUENCY OF THE DIELECTRIC CONSTANT OF SOME
MATERIALS USED IN MOTOR INSULATION – REF. [7]
Material K (50/60 Hz) K (1 kHz) K (1 MHz)
Polyethylene (PE) 2.3 2.3 2.3
Air gap Rotor
Polyethylene (LD) 2.3 2.2 2.2
Polyethylene (HD) 2.35 - 2.34
Epoxy Resin 3.5 – 3.9 - 3.6
Polyester (PET) - 3.2 3.0

Insulating
film inside the slot
100
80 K without
variatons
Ceq (%)

60 Conductor
+ 22% K
40 (stator coil)

20 - 31% K
0
Frame

Fig. 1. Equivalent capacitances obtained (% of Ceq without K variations) in

FEMM for different values of K considering the insulating materials.
Fig. 2. 2D motor representation in FEMM – staror slot. Details showing the
the motor insulating materials. Table I shows a list of some leakage capacitances, insulating film, conductor (stator coil), frame, rotor and
materials used in motors insulation, the respective K and the the air gap between stator and rotor.
variation of this constant in some frequencies [7].
As can be seen, the dielectric constant has a slight variation
in high frequencies. However, it was made a comparison
between capacitances calculated in FEMM using different
values for K. Figure 1 shows the obtained results, founding
that even with considerable changes in the dielectric constant
values of insulating materials, the equivalent capacitance has
not changed significantly. Thus, the dielectric losses and other
effects which affect the value of K can be neglected. In
addition, it is satisfactory to use a typical value for K in the
absence of more specific information.
(a)
B. Wire Position
It was noticed during the research that more distant the Bundle of wires for the
conductors are from the slot boundaries, less capacitance is same phase
generate than if they were closer to those boundaries. For this
reason, a sensitive study was made to verify how much the
coil position inside the slot influence on capacitances final Insulating film
between phases
values in FEMM. (coil head)
It was made a representation that wires were distributed
randomly and close to the slot boundaries. The calculated
(b)
capacitance in FEMM of one slot was 141.346 pF, which is Fig. 3. (a) Side view of the motor and its interior and (b) detail of the
around 48.5% higher than if the wires were confined in the conductors and insulating film in the coil head.
center of the slot. This confirms the meaningful influence of
the coil position in leakage capacitances values. Nevertheless, In figure 2, this can be seen in the pointed out detail of the
the coils touch the insulating film and are also uniformly 2D motor representation created in FEMM. It can also be seen
distributed because of the winding process. From this in figure 2 the respective leakage capacitances, insulating film,
observation, it can be concluded that, in the motor frame, rotor and the air gap.
representation, wires must be disposed as close as possible of Figure 3 shows the representation created in FEMM of the
the stator limits (because circular structures cannot have any coil head and figure 4 shows the insulating film being inserted
interceptions to be possible the computation using finite in the coil head of a motor during its manufacturing process,
elements) and must have a uniform distribution, which being possible to see the disposition of conductors and
corresponds to a block of conductors centered inside the slot. insulating film in this part. To analyze the leakage
capacitances in the coil head region, it was used the side view

Authorized licensed use limited to: CENTRO FED DE EDUCACAO TECNOLOGICA DE MINAS GERAIS. Downloaded on May 22,2023 at 23:27:35 UTC from IEEE Xplore. Restrictions apply.
 3

of the motor, since that the 2D representation has some

limitations and, therefore, the coil head can not be considered 3500 3100
2952
in the front view. To adapt the circular geometry of the coil 3000
head, it was considered in the representation a linear 2500
disposition of the conductors along the depth (like a “linear Measurement

Csf (pF)
2000
motor”), with a length equal to the average length of the FEMM
1500
conductors in the coil head.
1000
500
0

Fig. 5. Obtained values through FEMM and measurements for the

capacitance Csf.

It is observed from figure 5 the close proximity between the

results. This shows that the methodology developed to
determine Csf in FEMM is effective and provides reliable
results. During the process, it was observed that the greatest
part of the electrostatic energy stored due to Csf is confined in
Insulating film
in the coil head the stator slots, since it is the region that the coils are closer to
the frame (grounded).
Fig. 4. Picture of the insulating film being inserted in a motor coil head B. Capacitance between Stator and Rotor – Csr
during its manufacturing process.
To obtain Csr from the 3 HP motor, it must be made a
It could be noticed that the number of conductors in stator subtraction between the two values found in the measurements
windings also has a great impact in leakage capacitances, taken in it with and without the rotor, because the result of this
because the final value is a sum of individual capacitances operation is precisely Csr [5]. It can be seen in figure 6 several
produced by each conductor. It is also the reason that it is not measurements taken for motors of different capacities, in order
possible to make an equivalent conductor in the to investigate how much the leakage capacitances increase
representation, since these individual capacitances are not with motor capacity and number of poles.
computed. It can be observed from figure 6 that the measurement
result for the 3 HP motor was equal to the previous one taken
III. LEAKAGE CAPACITANCES DETERMINATION in the motor without the rotor. After doing an average of the
measurements taken in identical motors, it was possible to
A. Capacitance between Stator and Frame – Csf
obtain Csr equal to 83 pF. Though, this method of
The capacitance Csf was calculated in FEMM after the measurement is not effective, since it involves a subtraction
creation of the motor representation. In this case, the rotor was between two values much larger than the expected result,
disregarded, since it is removed from the motor in the which can lead to dubious values. The limitations of the
measurement [5]. The result in FEMM was compared with the measuring instrument concerning its accuracy must be
value taken in the motor used in the research. considered, because in this situation they interfere
The capacitance measurements were performed with the considerably on the results. On the other hand, the calculation
instrument Minipa Digital MC - 152 (200 pF – 20000 μF). of Csr in FEMM will be more reliable, even though it involves
With the rotor removed and frame grounded, the equivalent the same subtraction. This can be justified because the
capacitance was measured by the stator terminals [5], which precision due to the finite element method is greater than that
were interconnected at a common point. This capacitance obtained by the measuring instrument.
represents Csf, since the other leakage capacitances were In order to have a reference value and be able to compare
excluded. Nevertheless, the result of this measurement was the FEMM results, the formulation proposed in [6] was used
almost the same value taken in the motor with the rotor. This to calculate analytically Csr. The value found was 67.4095 pF,
shows that Csf is much greater than Csr, because when the rotor while in FEMM the obtained Csr was 77.58 pF. The results for
is in the motor and it is stopped, the equivalent leakage this situation were quite close and they are also close to the
capacitance is a parallel association between Csf and Csr. value obtained by the measurement method. Therefore, it can
Figure 5 shows through the graph the difference between Csf
obtained through FEMM and the average of measurements
taken.

Authorized licensed use limited to: CENTRO FED DE EDUCACAO TECNOLOGICA DE MINAS GERAIS. Downloaded on May 22,2023 at 23:27:35 UTC from IEEE Xplore. Restrictions apply.
 4

IV. MOTOR EQUIVALENT CIRCUIT AND ELECTRICAL

17585
17970 SYSTEM CHARACTERISTICS TO STUDY THE COMMON MODE
CURRENTS
14585 This research uses the equivalent circuit proposed in [2] as
Ceq
it is known the required parameters to it. So, simulations in
11585 (pF) ATP were made to verify the common-mode current behavior.
The motor capacitive paths for the common-mode currents are
8585 illustrated in figure 7, while the corresponding equivalent
4910 circuit is shown in figure 8, where Cb is the bearing
5585 3960 capacitance and Zshaft is the half-impedance of the motor shaft.
3100
2585 The shaft impedance was measured in laboratory with the
2585 motor stopped. The instruments used were an analogical
1,5 HP 3 HP 5 HP 10 HP 100 HP Winston bridge and a digital indutimeter, taking the end-to-
Fig. 6. Measurements taken in motors with 4 poles, different capacities, end resistance and inductance of the motor shaft (considering
original coils and all its parts assembled (rotor inside the motor). the resistance and inductance of the linkage cables). Although
the measurement procedure and the instruments used were not
be concluded that the methodology developed in this work is the most indicated to take Zshaft, the importance of this
also valid for the determination of Csr, leading to more reliable impedance in ATP simulation is only to improve the results
results than the traditional methods of measurement. In this (remove numerical problems and others). Therefore, its exact
case, the region that stored the greatest amount of electrostatic value does not have relevance in this study. The source VCmot
energy due to Csr was the coil head, because the electrostatic represents the inverter, which supplies the motor with
interaction between coils and rotor is greater there. common-mode voltage, while the switches SW1 and SW2 are
used to represent the occurrence of dielectric breakdown in the
C. Capacitance between Rotor and Frame – Crf bearings. The opened switches indicate the sustenance of the
insulation, while the closed switches represent the grease
As shown so far, the FEMM simulations reproduce the
dielectric breakdown [2, 4, 5, 6]. There are nonlinear
measurement methods done in the motor. To measure Crf, it is
resistances in series with the switches SW1 and SW2 that
necessary to ground the frame and connect the rotor to a drive
could be ignored in the model as they do not offer any
motor, so the motor could move without applying voltage in
contribution to the investigation of the circulating current.
the stator windings [5]. This sometimes allows the formation
of a lubricant film between bearing balls and races, and that
forms a capacitor called Cb. For that reason, the capacitance
measured in this test is equal to an association between all
leakage capacitances of the motor, and thus Crf can be
obtained by the equation (1).
−1
§ 1 1 ·¸
C eq =¨ + + C rf + C b (1)
¨C ¸
© sf C sr ¹

The estimation of Crf from FEMM simulations aimed to

determine this capacitance directly, without the need of (1). It
is known that Cb varies with the motor rotation speed [4, 5, 6],
so its measure is uncertain. This was one of the reasons that
Fig. 7. Capacitive paths of the motor for the common-mode currents.
made impossible the measurement of Crf in the case of the
present work, and then the simulation result was compared
again with the value obtained by the analytical formulas
proposed in [6]. Using the equation, the value found for Crf
was 936.15 pF, while the result from the simulation was equal
to 860.93 pF. In the simulations, it could be observed that
most of the stored energy associated to the Crf also comes
from the stator magnetic region, since it is the area where the
rotor is closest to the ground. Consequently, in this case the
coil head could also be neglected to simplify the
representation and speed up the simulation. Fig. 8. Common-mode current equivalent circuit of the motor [2].

Authorized licensed use limited to: CENTRO FED DE EDUCACAO TECNOLOGICA DE MINAS GERAIS. Downloaded on May 22,2023 at 23:27:35 UTC from IEEE Xplore. Restrictions apply.
 5

When the rotor is not moving, the bearing spheres are any duration.
generally in contact with the races, having a resistive
characteristic. Bearings become capacitive during the rotor 15

operation, when it is created a grease film between balls and [V]

tracks. In order to reproduce this phenomenon, the switches 10

SC1 and SC2 were inserted in the equivalent circuit to control

5
the behavior of Cb. When these switches are closed, the balls
are in metallic contact with the tracks, creating a resistive path
0
for common-mode currents through the resistors r1 and r2.
When they are opened, the capacitive circuit takes place. -5
There is a difference between the bearings from the front
and back of the motor, leading to different dielectric -10
breakdown characteristics in each bearing [2]. Since the one in
the back is smaller, its lubricant film thickness might be -15
0,70 0,75 0,80 0,85 0,90 0,95 1,00 [ms] 1,05
thinner and, consequently, the capacitance is larger. Thus, the Fig. 9. Common-mode voltage at the bearings.
breakdown voltage is slightly lower in the back bearing and it
is where the dielectric rupture occurs first. In this study, it was 3

supposed that the capacitances and breakdown voltages of the [A]

bearings were the same, because the difference between them 2

is more relevant in studies of shaft end-to-end voltage, which

1
does not correspond to the present case. Moreover, it was
admitted to these parameters (Cb and disruption voltage of the 0
grease) typical values mentioned in other studies [2, 5], due to
the lack of specific information and measurement difficulties. -1
The source of common-mode voltage and the linkage
cables were designed from existing models [1]. The common- -2

mode voltage source has a transition time of 100 ns with steps

of 100 V. Each step has the same duration, resulting in a -3
0,70 0,75 0,80 0,85 0,90 0,95 1,00 [ms] 1,05
symmetrical waveform. The cable design was carried out by Fig 10. Discharge current in the bearings with the occurrence of dielectric
cells which have three phases, a ground conductor and 10 PI breakdown.
circuits, being considered in ATP a 20 meters long cable.
80
Furthermore, it is emphasized that the cable impedance is [mA]
distributed by the PI circuits, so it is possible to verify the 60

reflection phenomenon in the common-mode voltage 40

waveform at motor terminals. As a result, this has influence in
20
the occurrence of common-mode currents discharge through
the bearings, which will be verified in the next session. 0

-20

V. SIMULATIONS IN ATP AND RESULTS -40

The simulation in ATP was performed considering an -60
integration time step equal to 10-9 seconds. The motor
common-mode input voltage, shaft voltage and the common- -80
0,70 0,75 0,80 0,85 0,90 0,95 1,00 [ms] 1,05
mode currents (the total common-mode current and the portion Fig 11. Discharge current in the bearings when occurs the contact between
balls and bearing trails.
that circulates through the bearings) were analyzed. Figures
from 9 to 12 illustrate simulation results.
After doing some ATP simulations and taking into account
its results, it could be seen that the discharge currents did not
influence the total common-mode current, which is shown in
figure 12, as it does not depend on these parcels. It can be
verified in figure 9 that the bearings entered in contact with
the trails when the time was 0.8 ms, and after that the voltage
shaft felt down to zero because it was grounded by the bearing
balls. The moment of the grease film creation and when the
balls touch the trails, its duration and the number of times that
these events occur are random; it is observed from
measurements [4, 5] that they can happen anytime and have
Fig. 12. Total common-mode current to the ground.

Authorized licensed use limited to: CENTRO FED DE EDUCACAO TECNOLOGICA DE MINAS GERAIS. Downloaded on May 22,2023 at 23:27:35 UTC from IEEE Xplore. Restrictions apply.
 6

common-mode currents in the drive system is justified, and

In figure 10, it is seen that the discharge current due to the the present study aimed to determine the motor capacitances to
dielectric breakdown exceeded 2 A and it is much larger than make possible the determination of such currents. The
the one due to the bearing contact (showed in figure 11). software FEMM is very efficient on estimating those
Therefore, it is observed that the discharge current from the capacitances, providing more benefits than other methods.
dielectric breakdown might cause more damage to bearings This can be concluded by the fact that it is more accurate than
because they are greater than the one from the balls contact. the analytical formulations and sometimes simpler than the
Moreover, the use of inverters without preventive actions can measurement process.
lead to premature bearing failures in induction motors, Making a comparison between alternatives for the
reducing their lifetime. The distortions in the common-mode determination of the motor leakage capacitances, the following
voltage waveforms not only illustrate the effect of voltage findings can be highlighted:
reflection due the interconnecting cables, but also the
switching events of the source. 1) Motor constructive details. Either the computation
The results are in conformity with the expectations; then, (which uses the finite element method) or analytical
the entered values for leakage capacitances in the motor formulations require their previous knowledge.
equivalent circuit were reasonable to analyze the dielectric 2) Simplifications. The analytical formulations provide
breakdown and balls contact with trails event in the bearings, greater simplifications, which can compromise the results.
and the magnitude of common-mode currents. 3) Measurement effectiveness. By the traditional
measurement methods, it is possible to obtain the
VI. CONCLUSIONS equivalent leakage capacitance of the motor. However,
more complex/not feasible practical procedures are
In this research it was verified that the exact representation required to measure each capacitance separately.
of all the details and elements of the motor it is neither feasible
nor practical. The reason of this difficulty is the lack of Based on these observations, it can be concluded that the
information and imprecision in some constructive details. described measurement techniques are not always practical
Motors of the same manufacturing line and rated power may and the analytical formulations involve more simplifying
have some constructive differences and, consequently, slightly considerations than the computer simulations; in this way, it
different leakage capacitances. Though, it was possible to can be stated that from the proposed methodology arises an
obtain suitable results from the proposed approach. Moreover, attractive tool for the motor capacitances determination.
among the parameters considered, the one that most affected
the calculated capacitances was the distance between the VII. ACKNOWLEDGEMENT
excited conductor (the stator windings or rotor) and the
Our thanks to Hugo Gustavo Gomes and Eribert Augusto
grounded region (frame or rotor). The number of excited
Neves (WEG Motores) and MEC/CAPES, for the invaluable
conductors also affected a lot the capacitances final values,
collaboration and financial support.
being important in the representation to consider all of them.
The dielectric materials and their constant values did not have
a great impact in the capacitances final values, showing that
VIII. REFERENCES
the use of typical values or even disregarding some insulating
[1] H. de Paula et al., “Methodology for Cable Modeling and Simulation
materials are reasonable in the lack of more accurate or
for High-Frequency Phenomena Studies in PWM Motor Drives”, IEEE
specific information. Trans. on Power Electronics, pp. 744 – 752, March 2008.
It was done measurements in motors of different types and [2] U. T. Shami, H. Akagi. “Experimental Discussions on a Shaft End-to-
power to investigate better the leakage capacitance values. It End Voltage Appearing in an Inverter-Driven Motor”, IEEE
Transactions on Power Electronics, vol. 24, nº. 6, jun 2009.
could be noticed that motors that had their windings changed [3] A. V. Jouanne et al.., “An Evaluation of Mitigation Techniques for
(a very common procedure in some countries, when windings Bearing Currents, EMI and Overvoltages in ASD Applications”, IEEE
are in short-circuit, for example) have completely different Trans. on Industry Applications, vol. 34, n. 5, pp. 1113-1122, 1998.
[4] J. M. Erdman, R. J. Kerkman, D. W. Schelegel, G. L. Skibinski, “Effect
capacitances values, generally much higher than before. The of PWM Inverters on AC Motors Bearing Currents and Shaft
analysis about the large interference of the number and Voltages”, IEEE Trans. on Industry Applications, vol. 32, n. 2,
position of the excited conductors on FEMM results helps on March/April 1996.
[5] J. Erdman et al.., “System Electrical Parameters And Their Effects On
explaining this fact, because when the windings are changed Bearing Currents”, IEEE APEC Conference, San Jose, CA, 1996.
the number and position of conductors are generally changed [6] A. Muetze, A. Binder, "Calculation of Motor Capacitances for
too. Then, this shows that this procedure can lead to more Prediction of the Voltage Across the Bearings in Machines of Inverter-
Based Drive Systems," IEEE Trans. Ind. Appl., vol.43, nº.3 (2007)
losses and premature failures in motors, since it usually pp.665-672.
increases the motor leakage capacitances. [7] G. G. Raju. Dielectrics in Electric Fields – Power Engineering. New
The study performed in ATP demonstrated that PWM York: Marcel Dekker, 2003.
[8] R. R. Riehl, E. R. Filho, “High Frequency Capacitances Determination
inverters not only degrade the electrical power quality, but in Three-Phase PWM Inverter-Motor Drive System Using an
also bring undesired high-frequency phenomena to the driven Alternative Method Applied to EMI Effect Studies”, 9th COBEP,
motor. Therefore, the concern related to the effects of pp.137 – 141, 2007.

Authorized licensed use limited to: CENTRO FED DE EDUCACAO TECNOLOGICA DE MINAS GERAIS. Downloaded on May 22,2023 at 23:27:35 UTC from IEEE Xplore. Restrictions apply.
 7

[9] T. Humiston, “Parameter Measurements to Study Surge Propagation in

Induction Machines”. IEEE Transactions on Industry Applications,
vol.40, no.5 (2004), pp. 1341 – 1348.
[10] K. Maki, H. Funato, L. Shao, “Motor Modeling for EMC simulation by
3-D Electromagnetic Field Analysis”. IEEE International IEMDC '09,
Miami, FL, jun 2009.
[11] B. Mirafzal et al., “Universal Induction Motor Model with Low-to-
High Frequency Response Characteristics”. IEEE Transactions on
Industry Applications, vol. 43, pp. 1233 – 1246, September 2007.
[12] T. Halkosaari, H. Tuusa, “Reduction of Conducted Emissions and
Motor Bearing Currents in Current Source PWM Inverter Drives”. 30th
Annual IEEE Power Electronics Specialists Conference, 1999. PESC
99, vol. 2, pp. 959 – 964, Charleston, SC, aug 2002.
[I3] J. Ollila, T. Hammar, J. Iisakkala, H. Tuusa, “A New Reason for
Bearing Current Damages in Variable Speed AC Drives”. Proceedings
of the EPE'97, pp. 2.539-2.542, vol. 2, 1997.
[14] Clayton R. Paul. Introduction to Electromagnetic Compatibility. New
York: Wiley, 1992. pp. 449-490.
[15] E. Zhong, A. Lipo Thomas, “Improvements in EMC Performance of
Inverter-Fed Motor Drives”. IEEE Trans. on Industry Applications, pp.
1247-1256, vol. 31, no. 6, November/ December 1995.
[16] L. Ran et al., “Conducted Electromagnetic Emissions in Induction
Motor Drive Systems Part I: Time Domain Analysis and Identification
of Dominant Models”. IEEE Trans. on Power Electronics, pp. 751-767,
vol. 13, no. 4, July 1998.

Authorized licensed use limited to: CENTRO FED DE EDUCACAO TECNOLOGICA DE MINAS GERAIS. Downloaded on May 22,2023 at 23:27:35 UTC from IEEE Xplore. Restrictions apply.