Paper Final
Paper Final
net/publication/261468821
CITATIONS READS
4 572
3 authors, including:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Nejila Parspour on 20 June 2016.
Abstract — A variety of efficient control strategies for behavior, it is still an effort to express flux and voltages, which
synchronous motors have been proposed in recent years. Many of are time and position dependant functions, as vectors. This
them use the rotary or stationary two dimensional axes in order leads to a lack of information which is necessary for analytical
to make the control algorithm easier. However, trying to describe calculation and consideration of motors. For example the
the motor behavior using just two axes is either impossible or at effects of teeth and displacement of torque-zero axis due to the
least not complete. Neglecting the real flux density distribution of cogging torque cannot be observed in transformed models.
motors, leads even to wrong or at least naive interpretation about There is also a misunderstanding of how the field weakening
the functionality of the machine. In this paper a new approach influences the torque development in motors. In this paper an
for understanding the torque development of a permanent
analytical model is presented to determine the distribution of
magnet synchronous motor in the field weakening region is
developed and verified with experimental results. The description
induced voltage and torque. This leads to a better
of induced voltage and torque using flux density distribution in understanding of the motor operation using field oriented
the air gap offers a deep insight into the functionality and control (FOC) and causes in a wider insight into the motor
behavior of a electric machine. behavior. A brief explanation of the field oriented control in the
first section is followed by the classical derivation of the motor
Keywords — field-oriented control; fieldweakening; MTPA; equations. After that, the induced voltage is explained and the
PMSM; FEM. torque development in the motor, based on the flux density in
the air gap, is discussed. The experimental results are described
I. INTRODUCTION at the end.
Permanent magnet synchronous motors (PMSM) are used
more and more in the industry as servo motors. They are II. FIELD ORIENTED CONTROL
established as drives of electric vehicles, household remedies The derivation of the voltage equations of PMSM using the
and in a big variety of other devices and applications. They Park transformation can be looked up in [4] and [5]. The three
receive a great standing due to their acceptable price, low phase equations are transformed in the flux axis (d) and the
construction complexity and high efficiency without any torque axis (q). The electromagnetic torque is
requirement of complex control strategies. These simple
control algorithms are achieved using proper transformations
which make the controllability of a PMSM very similar to that
of a DC-Motor. There are the stationary transformation known
as Clarke transformation and the rotary which is recognized as where and are flux linkages in d and q axis. It
Park transformation. These are well-known three to two phase follows
transformations which are applied since many years. The field
oriented control is based on the Park transformation and the
direct torque control uses the stationary transformation. By
using the rotary transformation it is possible to consider the Now the torque equation can be presented as
flux along the d-axis and torque along the q-axis. With a proper
current control it will be possible to control torque and flux and
also allows operating the motor in the field weakening region
where the speed is higher than the nominal values [1].
Regardless of control principles such as PI, state space, sliding If there is a symmetrical rotor, where , the torque
mode or genetic algorithm, the transformation need to be used can be expressed simplified
to apply the mentioned control methods. However, the
transformation bases on the presumption of sinusoidal flux
density and induced voltage distribution of the motor which is
not always the case. Recently, transformations which consider
and the voltage equations in steady state are
the harmonics and also the effects of saturation have been
presented [2], [3]. Although this results in a better control
U
Based on this model the FOC is implemented so that and V
W
are controlled using the rotor position . At each position the
three phase currents are measured and transformed to dq-axes
using . Fig. 1 shows the boarders for and depending on
the maximum current (green) and the induced voltage (blue).
The curve of maximum permitted current is a circle:
1
0.8
0.7
Flussdichte im Luftspalt
0.6
0.5
0.3
0.2
0.1
0 10 20 30 40 50 60 70 80 90
Fig. 1. Curves of maximum current and voltage of motor
B B B
Fig. 3. Stator flux density distribution as a convolution result between air gap flux density and stator winding
So the induced voltage can be differentiated from the stator With this insight, it is possible to explain the development
flux: of the torque in a motor based on the interaction between these
two flux functions. Considering the vectors as functions over
∅ ∅
the angle, the cross product is proportional to the difference
between the red and green areas of these two functions, which
is shown in Fig. 4.
and are the number of windings and the motor speed
respectively. As it is shown in (12) the induced voltage When two vectors are aligned, the difference between
depends directly on the motor speed and air gap flux density. these two areas which are characterized by different colors is
zero. It can be observed how the variation of the amplitude
IV. TORQUE and the angle between two vectors change the difference
The electromagnetic force can be derived from between the discussed areas. The torque can be considered as
Hamiltonian and conservation of momentum [7], [8]. This the cross multiplication of two vectors or as a difference of the
leads to Lorentz force or Maxwell Stress Tensor [9]. Based on integral of the fluxes over a pole pitch as it is defined in Fig. 4.
Lorentz force, a current carrying conductor which lays in a Therefore, an expression based on the air gap flux density is
magnetic field is exposed to a force in the direction which is developed here. Fig. 5-a shows a conductor which lays in an
defined by the right-hand rule. As the current in the conductor external magnetic field and carries no current. As it is shown
produces also a magnetic field based on Ampere’s circuital in Fig. 5-b, when a current flows in direction out of the paper,
law, the Lorentz force can be expressed as the reaction there is a distortion in the total flux lines. Hence the Lorentz
between these two fluxes. This is useful for the explanation of force can be explained as a function of distortion in a magnetic
the torque in a PMSM. The force, or rather torque, can be field. The resultant force tries to move the conductor in that
described as the cross multiplication of the flux of rotor and direction, in this case left, in which the distortion would be
stator. This is also the expression which is used to explain the vanished. This is used to explain the torque production of a
operation of FOC. However, as it is mentioned before this PMSM. Considering the flux density distribution of a pole
vector demonstration of the torque restricts the complete which is shown in Fig. 2, the flux is totally symmetrical to the
understanding of the torque behavior of a motor. A better middle of the pole. That can be compared with Fig. 5-a. Only
interpretation can be achieved by observing the flux with an asymmetry in this flux density distribution, what can
distribution in the air gap. Therefore, the relation between two be caused by a stator current, a torque can be developed in the
vectors and their sinusoidal distribution over the angel is motor. A simplified illustration of this effect can be observed
indicated in Fig. 4. in Fig. 5-b. With these descriptions of induced voltage and
electromagnetic torque it is possible to explain what happens
in the field weakening region and how torque and induced
voltage are influenced by the currents.
V. FIELD WEAKENING the time. While the current is constant at 10 A, the current
is decreased from 0 A to -10 A over the time. Due to the
In this section the effects of d- and q-currents are
explanation in section IV and V, there should be no
discussed. Figure 6 shows finite element results of the
variation of the torque measured over the same time.
behavior of flux density as a function of and .
Decreasing reduces the air gap flux density. Attention Fig. 8 shows the measured behavior of torque over time.
should be paid to the fact that the reduction happens According to the expectations it is nearly constant. This
especially in the middle and rarely at the edges of the verifies the proposed explanation of torque and induced
permanent magnets. The symmetry of the air gap flux voltage. The small variation in the distribution of the torque
density has not been distorted; hence there is no torque can be explained by a small difference between the d and q-
production. Now it should be remembered from section III axes inductance ( ). According to (3), any difference
that the induced voltage is the convolution between the air between the d and q-axis inductances results in a reluctance
gap flux and the function of the windings . Considered torque as a function of and .
the windings as a unit function, the integral of flux density
and therefore the result of convolution are reduced.
Consequently the induced voltage is decreased too.
Therefore the negative current in d-axis reduces the induced
voltage but does not produce any torque. The effect of is
even more interesting. Fig. 6 shows that the current in the q-
axis changes the flux density distribution at the edges of the
permanent magnets and rarely at the middle. Since the
symmetry of the flux density has been distorted, there is a
torque development in the motor. This torque production is
a response to overcome the asymmetry in the air gap flux
density. On the other hand, the area under the flux curve is
not changed, which means that the integral of the flux
density and therefore the induced voltage is unchanged.
VI. EXPERIMENTAL RESULTS
The proposed idea is verified using an experimental test
bench. The FOC is implemented on a dSpace 1104-card.
Currents are measured using LEM transducers. A DC-Motor
is used as mechanical load. The PMSM and DC-Motor are
coupled using a torque gauge bar of the type 0260-DM by Fig. 7. Current profile of iq and id using FOC
the manufacturer Staiger-Mohilo. By implementing the
FOC with enough reserve of current it is possible to
decrease step by step, while is maintained as constant
value.
In Fig. 7 and Fig. 8 the experimental results are
determined. Fig. 7 shows the currents in d and q-axes over
Conference on Power Electronics and Motion Control, 2012, pp.
2353-2356.
[7] K. Delaere, P. Sas, R. Belmans. “ Finite Element analysis of
Electrical Machine Vibrations caused by Lorentz, Maxwell and
Magnetostriction Forces,” IEEE Seminar on Electromagnetical
design and analysis, 2000, pp. 1-4.
[8] W. zhu, S. Pekarek, B. Fahimi. “Investigation of Force Generation in
a Permanent Magnet Synchronous Machine.” IEEE Transactions on
Energy Conversion, vol. 22, pp. 557-565 , 2007.
[9] Z. Ren, A. Razek. “Force Calculation by Maxwell Stress tensor in 3D
Hybrid finite Element-boundry integral formulation.” IEEE
Transactions on Magnetics, vol. 26, pp. 2774- 2776, 1990.
BIOGRAPHIES
Amir Ebrahimi received the B.S degree from Shiraz University, Shiraz,
Iran in 2006 and the M.S. degree from Iran university of Science and
technology, Tehran, Iran in 2008 both in Electrical Engineering.
He is currently perusing the Ph.D. degree in electrical engineering at
University of Stuttgart, Germany. His research interests include calculation
and control of electrical machinery and power electronics.