ICRAES16 Proc Paper 9
ICRAES16 Proc Paper 9
Laboratory of Advanced Electronic Systems and Sustainable Energy, ENET’Com, University of Sfax, Tunisia
Tc : coil temperature (°C) On the other hand, Tunisia is known as great touristic
destination where leisure vehicles are widely needed,
15 (permanent)
Voltage (V) DC supply Nominal: 200 5. NO-LOAD STUDY
Variation range 90 → 220 This part focuses the machine cogging torque. First, we
torque (Nm) 145 (30 seconds) choose the remnant induction, then following studies are
130 (300 seconds) performed:
65 (permanent) - Variation of magnets thickness, where magnets
basic Speed (tr/min) 2250
height is fixed.
operating speed limit (tr/min) 8000 - Variation of magnets height, where magnets
thickness is fixed.
- Variation of magnets thickness where volume
4. MOTOR TOPOLOGY
magnets is fixed.
In this work, we propose a design procedure of a
A. Choice of magnets remnant induction
permanent magnet synchronous motor with flux
concentration arrangement. Considered topology is Fig.2 shows the output cogging torque when we used
illustrated in Fig.1. Besides, the initial dimensions of the respectively ferrite magnets (Br=0,37T) and Neodymium
machine are shown in Table II. Those dimensions will be Iron Boron magnets “NdFeB” (Br=1T). According to this
varied during the next studies in order to respond as best figure we can find that the use of ferrite magnets is not a
to required specifications. good solution despite of its low price. In fact, the output
torque in this case is almost zero, while the use of a TABLE III
NdFeB magnets give to machine the power to produce a Cogging torque (Haim=40mm)
cogging torque reaching 10 Nm.
Eaim=6mm Eaim=8mm Eaim=10mm
Moreover, Fig.2 and Fig.3 show that the angular period
of cogging torque is 10°. Max cogging 6.31Nm 10.27Nm 14.85Nm
torque
At this stage, we have tested the effect of the variation concentrated winding and distributed winding [10-11],
of the magnets dimensions on cogging torque. The Fig.8.
obtained resultants show that the analytical expression (1)
used in [9] to determine Eaimoptimal is false. Besides, we Fig.9 shows an output torque comparison of two
can find that a rise of magnet dimensions leads to an PMSM with flux concentration arrangement. The first
increase of the maximum cogging torque and also of encloses a concentrated winding and the second is a
magnets volume consequently a rise of magnet price, but classical one with a distributed winding. According to
to deciding about final dimensions of the magnets and find obtained curves, the concentrated winding should be the
a compromise between it and "J", load studies have to be best choice in our case. In fact, when J=10A/mm2, the
performed. output torque in case of concentrated windings can reach
108 Nm, while in case of distributed windings the output
torque cannot exceed 23Nm.
- Variation of magnets thickness, where magnets (respectively thickness) is high, more the
height is fixed (Haim=40mm), Fig. 10. current densities giving the required torques is
low.
- Variation of magnets height, where magnets
thickness is fixed (Eaim=8mm), Fig. 11. - The rise of magnets height (respectively
thickness) make the machine less susceptible
- Variation of magnets thickness where volume to saturation
magnets is fixed, Fig. 12.
Table VI and VII show the required currents density.
According to Fig.12 and TABLE VIII we can see that
the required current density increases with magnet
thickness when magnet volume is fixed, as a result of the
growth magnets short circuits, Fig.13.
TABLE VI
TABLE VIII
(2)
A. no-load study
According to Fig.14 we can show that the decrease of
lds caused the rise of torque ripple. Besides, Fig.15 and 16 Fig. 16. Field vectors no-load operation (lds=10mm).
show that the decrease of lds causes a high saturation at the B. load study
stator tooth. In fact, when lds= 20mm the flux density does
not exceed 1.81T, while when lds= 10mm it reached 2.19T. Fig.17 and table IX show that the decreases of lds
causes the rise of required current density and it make the
machine more susceptible to saturation.
Fig. 14. Cogging torque versus angular position, for lds from 10
to 20mm
Fig. 17. Electromagnetic torque versus current density for lds
Despite the bad effects of lds decrease presented in the from 10 to 20mm
no-load study, load studies have to be performed before
TABLE IX
Value (mm) 45 10 20 18 The iron losses in rotor are equal to zero because the
rotor’s magnetic material sees a fixed magnetic field.
However, the stator's magnetic material sees a magnetic
8. PERFORMANCE OF THE DESIGNED CONFIGURATION field turning to the electric rotation frequency, for this
reason iron losses in stator are the majority and it can be
At this level, we have achieved the studies of cogging calculated by the equation (6), (7), (8), (9).
and electromagnetic torques. According to the previous
obtained results the most convenient dimensions for
(7)
magnets and stator tooth are as shown in table X. Besides,
Fig.14 shows that the designed motor is able to produce
Where:
cogging torque higher than 20 (Nm). Thanks to this
property, the motor does not need a high current density to (8)
produce the electromagnetic torque required by
specifications, Fig.17 and Table IX shows the current (9)
density requested
(10)
In the next parts, we focus on the evolutions of iron
and copper losses which are the principal heat sources in Fig.19 shows the induction in a stator tooth versus
motor and can degrade her performances. angular position.
A. copper losses Fig.20 shows the iron losses versus speed, according to
Copper losses refer to the heat produced by electrical this figure we can find that the iron losses which dissipated
currents in the conductors of motor. as heat, increase with the speed rise.
Where:
(4)
(5)
Fig. 19. Induction in stator tooth versus angular position.
(6)