INTERMAG -MMM 2019

Efficiency Comparison Between a

Permanent Magnet Generator and a
Pseudo Direct Drive

Prof. Dr. Carlos Guilherme C. Neves

Universidade Federal de Pelotas – UFPEL/Ceng
Prof. Dr. Ály F. F. Filho
Dep. Eng. Elétrica/DELET
Universidade Federal do Rio Grande do Sul - UFRGS
 INTRODUCTION

 In this work, a comparison of efficiency and power

density between a Permanent Magnet Generator (PMG)
and a Pseudo Direct Drive (PDD) is carried out.
 The finite element technique is applied to calculate the
copper losses, iron losses in order to determine the
efficiency of both machines.
 Generic Electrical Machines and specific Pseudo Direct
Drive techniques are applied to sizing both machines.
CONVECTIONAL PERMANENT MAGNET
GENERATORS DRIVE-TRAIN

 PMG are employed in conjunction with mechanical

gearboxes to match the speed of the generator to the
primary machine.
 The mechanical gears require maintenance, they are
subject to frequent failures and have poor durability.
 PSEUDO DIRECT DRIVE

Outer Stator (OS)

lamination pack,
MR
copper windings
and magnets;
Inner Rotor (IR)
magnets rotating
Magnetic Gear at high speed;
Components Modulator Ring
(MR)
ferromagnetic
pole-pieces
IR connected to the
driven turbine.
OS

 Unique device for matching the low-speed wind

turbine and the high-speed low-torque generator to
Convectional further reduce the overall size and weight.
PM Generator
 PDD WORKING PRINCIPLE (DIDACTICAL APPROACH)

Without modulators is The modulators will

not possible to produce alternately “short-circuit”
torque. N-S pole-pairs of outer
rotor to produce torque.
 PDD WORKING PRINCIPLE (HARMONIC APPROACH)

The magneto motive force (MMF) of the The product of Fr(θ) and Pr(θ) yields the
inner rotor and the permeance of the outer air-gap flux distribution :
modulators are mathematically
expressed by : 
AP0 Fundamental

A
r     (2 h - 1)
sinC1  inner t   (FC)
Fr    
h 1
(2 h - 1)
sin 2h  1 pinner   innert  h 1


Pr    P0  
B
(2 - 1)
 
sin 2  1n p    p t  
cos C  C 

  
C 21  C1inner  
t  Harm.
1    2 1 
1  C  C   H1 e
 C0
2 h 1   

1  cosC  C   2 1
C 
2
 C
1
 

1 inner   H2
Where A is the permanent magnet 2 1  t
  C  C 
MMF amplitude, B , is the modulators  2 1

permeance amplitude and P0 is the AB

C0  C1  2h  1 pinner C 2  2  1n p
permeance of vacuum. 2h  12  1
The angular velocities of H1 and H2 are different from that of Fr(θ).
 When the inner rotor moves, FC follows it, but H1 and H2 do not.
 Thus the number of external rotor poles must be equal to H1 ou H2 to synchronize.
 If equal to H2  pouter=np+pinner
 If equal to H1 (higher outer air-gap flux density)  pouter=np-pinner or pinner + pouter= np
 GEAR RATIO
If each term of pinner + pouter= np is multiplied by its respective speed, it
gets the following expression :
pinnerinner  pouterouter  n p p

when the outer magnet array is held stationary , and the modulators
are driven by a turbine, the gear ratio becomes
inner np
Gr  
p pinner
If it is considered there are no losses, the power in inner rotor is
equal to power in the modulators, then
Tp inner np
   Gr
Tinner p pinner
 FINITE ELEMENTS COMPUTED FLUX AT OUTER AIR-GAP
(pinner = 2 and np =21)

Radial field density at outer air-gap due Radial field density harmonic content at
to the inner magnets only, with outer air-gap due to the inner magnets
modulating rotor. only, with modulating rotor.

As espected the dominant components are 2nd, 19th and 23th, which
corresponds to FC, H1 and H2, since pinner=2 and np=21, when h    1
GENERATOR’S SIZING
Output Coefficient

Pout 2
p  2
0.5 kw1AmBmgcos
Di Lins

Pout
Di2Li 
pns

Pout
Di  3
pns

Stack
0.56
p length ratio
for a
certain
number of
pole-pairs
PDD COMPONENTS SPECIFIC SIZE (1):
Number of pole-pairs and modulators

inner 1000
  8.3
p 120

np 25
  8.3
pinner 3

pouter np pinner 22

pinner  3 1000 rpm 50 Hz

 PDD COMPONENTS SPECIFIC SIZE (2):
Based on the dimensionless Vernier Pseudo Repeat Pattern
parametersα, λ e β defined as:

 2*2 l  2.4  *( Dinner g1  g2 1 )

2 
2* pouter
  a l 0.2 ond
e *( Dinner g1 )
1 
 la l 0.5 2* mod
 PDD FINITE ELEMENTS MODEL (1)

Load PDD’s Electrical

Circuit
 PDD FINITE ELEMENT MODEL (2)
In order to simulate the movement of both rotors in relation to the
winding and modulators and obtain the flux densities values to
perform the losses calculation, two movement bands are applied.
 LOSSES CALCULATION – WINDING COPPER
LOSSES
The instantaneous DC copper loss is calculated directly from the finite
element model by

The total copper loss

is obtained by
summing the
contributions from
all elements.
 LOSSES CALCULATION – CORE LOSSES (1)
The core loss for electrical steel is based on

Kh, Kc and Ke were found from loss curves for E170 steel provided by an
electrical manufacturer
 LOSSES CALCULATION – CORE LOSSES (2)

B–P curves measured over a

range of frequencies from 50
Hz to 1000 Hz and in flux
density increments of 0.1 T.

Epstein Tests Curves

LOSSES CALCULATION – MECHANICAL LOSSES
The mechanical loss of the gear coupled to PMG is calculated
by
Wgear  0.02qPmR

The maximum practical gear ratio per stage is approximately 6:1,

so two or three stages of gears are typically required.
 Then this design which requires 8.3:1 gear ratio should use two
stages.
 Considering a PmR of 10.25 kW and two stages, then Wgear is 410
W.
 RESULTS (1)
This figure compares the copper and iron losses obtained
from 2D FE transient analyses of a PGM and a PDD for the
same electrical output.
 RESULTS (2)
Table IV presents the values of input speeds and torques,
losses, electrical output that result in the efficiencies for both
machines.

 One can notice that the amount of copper and iron losses are
greater in the PDD than in the PGM, but due the mechanical loss of
the gear coupled to PMG, the efficiency of whole set decreases,
hence the PDD efficiency is greater than the PGM.
 RESULTS (3)
Table V shows the dimensions and power densities.

 The power density of the PGM is 20.4 % greater than the PDD
regardless of the mechanical gear volume.
 CONCLUSION
 In this paper the comparison of losses and efficiency of a
Permanent Magnet Generator and Pseudo Direct Drive
was performed.
 It was concluded that the amount of copper and iron
losses are greater in PDD than PGM, but due the
mechanical loss of the gear coupled to PMG the PDD is
more efficient.
 The power density of PGM is greater than PDD excluded
the mechanical gear.
 REFERENCES
[1] L. Jian, G. Xu, Y. Gong, J. Song, J. Liang, and M. Chang, “Electromagnetic
design and analysis of a novel magnetic-gear-integrated wind power
generator using time-stepping finite element method,” Prog. Electromagn.
Res., vol. 113, December 2010, pp. 351–367, 2011.
[2] K. Atallah, J. Rens, S. Mezani and D. Howe, “A Novel ‘pseudo’ direct-drive
brushless permanent magnet machine”. IEEE Trans. on Magn.vol. 44, no. 11,
pp. 2195– 2198, 2008.
[3] G. L. Johnson,”Wind Energy Systems”, Englewood Cliffs, NJ: Prentice-Hall,
1994.
[4] S. A. Nasar, I. Boldea, L. E. Unnewehr, “Permanent Magnet, Reluctance, and
Self-Synchronous Motors”, Boca Raton: CRC Press, 1993.
[5] D. Matt, J. Jac, N. Ziegler, and S. Multiplier, “Design of a Mean Power Wind
Conversion Chain with a Magnetic Speed Multiplier Advances in Wind Power –
Chapter X Design of a Mean Power Wind Conversion Chain with a Magnetic,”
Adv. Wind Power, 2012.
[6] D.Matt, J.F. Llibre, “Performances comparées des machines à aimants et à
réluctance variable”, J. Phys. III France 5, October 1995, 1621-1641.
[7] I. Tarımer,, S. Arslan, M. E Güven, “Investigation for Losses of M19 and
Amorphous Core Materials Asynchronous Motor by Finite Elements Methods”.
Research Journal Elektrotecnika, vol. 18, No. 9, pp. 1392-1215, 2012.