energies
Article
Control of a Dual-Stator Flux-Modulated Motor for
Electric Vehicles
Xinhua Guo 1, *, Shaozhe Wu 1,2 , Weinong Fu 2 , Yulong Liu 2 , Yunchong Wang 2 and
Peihuang Zeng 1
1
2
*
College of Information Science and Engineering, Huaqiao University, Xiamen 361021, China;
hxwx@foxmail.com (S.W.); zph1990@foxmail.com (P.Z.)
Department of Electrical Engineering, The Hong Kong Polytechnic University, Kowloon 999077, Hong Kong,
China; eewnfu@polyu.edu.hk (W.F.); iyulong2008@gmail.com (Y.L.); wangycee@gmail.com (Y.W.)
Correspondence: guoxinhua@hqu.edu.cn; Tel.: +86-159-8583-3956
Academic Editor: K. T. Chau
Received: 29 April 2016; Accepted: 28 June 2016; Published: 2 July 2016
Abstract: This paper presents the control strategies for a novel dual-stator flux-modulated (DSFM)
motor for application in electric vehicles (EVs). The DSFM motor can be applied to EVs because
of its simple winding structure, high reliability, and its use of two stators and rotating modulation
steels in the air gap. Moreover, it outperforms conventional brushless doubly-fed machines in terms
of control performance. Two stator-current-oriented vector controls with different excitation in the
primary winding, direct and alternating current excitation, are designed, simulated, and evaluated
on a custom-made DSFM prototype allowing the decoupled control of torque. The stable speed
response and available current characteristics strongly validate the feasibility of the two control
methods. Furthermore, the proposed control methods can be employed in other applications of
flux-modulated motors.
Keywords: dual-stator; flux-modulated; vector control (VC); electric vehicles (EVs); decouple control
1. Introduction
With the emergence of an energy crisis, ways of reducing air-pollution have become the great
challenge [1]. Electric vehicles (EVs) contribute significantly to energy saving and environmental
protection, and on account of these benefits, they constitute today’s direction for the automotive
industry [2]. A compact and highly-reliable electric traction motor, which has high efficiency over a
wide speed range with load-adaptive controllable speed-torque contours, is crucial for any electric
propulsion system [3,4]. Traction motors of EVs should have high torque density in terms of weight
and volume, a wide range of torque-speed characteristics, high operating efficiency, and high reliability.
The novel structure of a brushless, doubly-fed generator (BDFG) proposed in [5], which has the
aforementioned advantages, can be efficiently applied to EVs. To distinguish the different applications
of such machines, the concept of dual-stator flux-modulated (DSFM) motors for EVs was presented.
Figure 1 illustrates a novel machine, which has a three-phase, 36-slot, 12-pole winding on the outer
stator and three-phase, 24-slot, 10-pole winding on the inner stator. Compared with most permanent
magnet synchronous motors (PMSMs), the DSFM motor with a wide range of speeds and torques and
no flux-weakening control yields higher performance in EVs [5].
Energies 2016, 9, 517; doi:10.3390/en9070517
www.mdpi.com/journal/energies
Energies 2016, 9, 517
Energies 2016, 9, 517
2 of 20
2 of 19
Energies 2016, 9, 517
Outer winding
2 of 20
Outer stator
Outer winding
Outer stator
Iron
segments
Iron
Inner
segments
Inner
Inner
winding
winding
Inner
stator
stator
(a)
(b)
(b)
(a)
FigureFigure
1. Structure
of the
motor
with
ananiron
rotor:(a)
(a)3D
3Dmodel;
model;
and
model.
1. Structure
of DSFM
the DSFM
motor
with
ironsegment
segment rotor:
and
(b)(b)
2D 2D
model.
Figure 1. Structure of the DSFM motor with an iron segment rotor: (a) 3D model; and (b) 2D model.
Conventional
brushless,
doubly-fed
suchasasBDFG,
BDFG,have
have
been
designed
Conventional
brushless,
doubly-fedmachines
machines (BDFMs),
(BDFMs), such
been
designed
Conventional
brushless,
doubly-fed
machines
(BDFMs),
such as BDFG,
have
beenfor
designed
and
and investigated
only
wind
energy
systems[6–10].
[6–10].
In particular,
control
strategies
BDFMs,
and investigated
only
for for
wind
energy
systems
In
particular,
control
strategies
for
BDFMs,
such
asonly
direct
control[11],
[11],voltage
voltage
vector-oriented
andsuch
investigated
fortorque
wind
energy
systems
[6–10].
In particular,
controlcontrol
strategies
for[12,13],
BDFMs,
such
as direct
torque
andand
fluxflux
control
vector-oriented
control(VC)
(VC)
[12,13],
and
field-oriented
(FOC)
[14,15],
havebeen
been investigated.
investigated.
The
schemes
werewere
as direct
torquecontrol
andcontrol
flux
control
[11],
voltage
vector-oriented
control
(VC)
[12,13],
and
field-oriented
field-oriented
(FOC)
[14,15],
have
The FOC
FOCand
andVCVC
schemes
compared
in
[15]
and
the
advantages
and
disadvantages
of
the
two
control
methods
were
presented.
control (FOC)
[14,15],
have
been investigated.
The FOC and
VCtwo
schemes
were
compared
[15] and
compared
in [15]
and the
advantages
and disadvantages
of the
control
methods
werein
presented.
The structural diagram of the brushless, doubly-fed reluctance machine (BDFRM (G)) drive setup is
the advantages
and disadvantages
of the two
control methods
were
presented.
The structural
The
structural diagram
of the brushless,
doubly-fed
reluctance
machine
(BDFRM
(G)) drive diagram
setup is
depicted in Figure 2 [13].
of the brushless,
doubly-fed
depicted
in Figure
2 [13]. reluctance machine (BDFRM (G)) drive setup is depicted in Figure 2 [13].
(2Va-Vb-Vc)/3=Vα
ωp
(2Va-V(V
b-V
c)/√3=V
αβ
b-V
c)/3=V
(Vb-Vc)/√3=Vβ
ωp
Shaft
Sensor
Frame Angle
iα=ia
sa bi)/sb√3
iβ=(iai+2i
θrm
VαCalcsV：βA ro B θrm
Pr
Pr
a
θr iαis=i
α
iab
SVPWM
isβ SV-
iβ=(ia+2ib3/2
)/√3 PWM
e-jθ
θ
θpr ×θsisα
~
DC Link+
Brake chopper
ωs
isa isb
Shaft
Sensor
Vα Vβ
~
DC Link+
Brake chopper
ωs
1/U
1/U
SV-PWM Vector
Controller
iab
SV-PWM Vector
Controller
Vabc
Vabc
isβ
e-jθ *isd isq V
i sd × PI sd
Q=⅔(iα*uβ-iβ*uα) × PI
i*sdsq isq e-jθ 2/3
i
P=⅔(iα*uα-iβ*uβ) × PI
iβ=(iV
a+2ib)/√3
* × PI Vsd
αβ
iαβ
P**β-iβ*uα) × PI i* sd × PI V -jθ
Q=⅔(iα*u
iα=ia
2/3
i sq × PI sqe
P=⅔(iαQ
*u*α-iβ*uβ) × × PIPI
iβ=(ia+2ib)/√3
ω
ω
*
m
m
iαβ
P
Vsq
Q*
×
PI
*
ωm
ωm
Frame Angle Vαβ
Calcs：A roiαB=ia
θp ×θs
3/2
Figure 2. A structural diagram of the BDFRM (G) drive setup.
2. A structural
diagram of themotor
BDFRM
(G) drive
setup.
However, theFigure
application
of the flux-modulated
to EVs
has not
been fully investigated,
Figure 2. A structural diagram of the BDFRM (G) drive setup.
especially the control strategies. In this paper, the air gap structures of the DSFM motor and BDFRM
were compared
and dynamic
were investigated,
andtothe
operating
of this
motor
However,
the
of models
the flux-modulated
motor
EVs
has not
notprinciple
been fully
fully
investigated,
However,
the application
application of
the
flux-modulated motor
to EVs
has
been
investigated,
was
verified
through
the
open-loop
variable
voltage
and
variable
frequency
(VVVF)
control.
especially the
the control
control strategies.
strategies. In
In this
this paper,
paper, the
the air
air gap
gap structures
structures of
the
DSFM motor
motor and
and BDFRM
BDFRM
especially
ofvector
the DSFM
Moreover,
two dynamic
closed loop
control
methods,
stator-current-oriented
control (SCOVC)
with was
were
compared
and
models
were
investigated,
and
the
operating
principle
of
this
motor
were compared
and
dynamic
models
were investigated,
the operating
principle
of thisthat
motor
direct current
(DC)
excitation
and alternating
current (AC)and
excitation,
were proposed
to ensure
verified
through
the
open-loop
variable
voltage
and
variable
frequency
(VVVF)
control.
Moreover,
two
was verified
open-loop
variable
voltage
and variable
frequency
(VVVF)results
control.
the motorthrough
operatesthe
at different
speeds
and load
conditions.
Simulation
and experiment
closed looptwo
control methods,
stator-current-oriented
vector control (SCOVC)
with
direct current with
(DC)
Moreover,
loop control
methods,
stator-current-oriented
vector
control
obtained forclosed
a customised
prototype
completely
validated the feasibility of
the two
control(SCOVC)
strategies.
excitation
and (DC)
alternating
current
excitation,
were(AC)
proposed
to ensure
the motor
operates
at
direct
current
excitation
and(AC)
alternating
current
excitation,
werethat
proposed
to ensure
that
different
speeds
and
load
conditions.
Simulation
experiment
results obtained
for a customised
2. Dynamic
Modelling
of Dual-Stator
Flux-Modulated
(DSFM) Simulation
Motors
the
motor
operates
at
different
speeds
and loadand
conditions.
and experiment
results
prototype
completely
validated
the
feasibility
of
the
two
control
strategies.
obtained for
a customised
validated
the feasibility
the two
strategies.
Figure
3 presentsprototype
the air gapcompletely
structures in
the BDFRM
and DSFMof
motor
withcontrol
an ideal
stator
surface around the rotor. The air gaps in the two motors were assumed to be uniform, implying that
2. Dynamic Modelling of Dual-Stator Flux-Modulated (DSFM) Motors
the air Modelling
gap length between
the statorFlux-Modulated
and the rotor teeth
is ge and
that between the stator and the
2. Dynamic
of Dual-Stator
(DSFM)
Motors
rotor
yoke
is
g
e
+
h
m
in
Figure
3a,
that
the
air
gap
length
between
the
two
stator
and the
iron
Figure 3 presents the air gap structures in the BDFRM and DSFM
motor
with
ansegments
ideal stator
Figure 3 presents the air gap structures in the BDFRM and DSFM motor with an ideal stator
surface around the rotor. The air gaps in the two motors were assumed to be uniform, implying that
surface around the rotor. The air gaps in the two motors were assumed to be uniform, implying that
the air gap length between the stator and the rotor teeth is ge and that between the stator and the rotor
the air gap length between the stator and the rotor teeth is ge and that between the stator and the
yoke is ge + hm in Figure 3a, that the air gap length between the two stator and the iron segments on
rotor yoke is ge + hm in Figure 3a, that the air gap length between the two stator and the iron segments
Energies 2016, 9, 517
3 of 19
Energies 2016, 9, 517
Energies 2016, 9, 517
3 of 20
3 of 20
theonrotor
is go +
thatthat
between
thethe
outer
stator
and
Figure
the rotor
is ggio, +and
gi, and
between
outer
stator
andthe
theinner
innerstator
statorisisggoo +
+ ggii ++hhmmininFigure
3b.3b.
on the rotor is go + gi, and that between the outer stator and the inner stator is go + gi + hm in Figure 3b.
Based
on
the
preceding
analyses,
the
air
gap
functions
of
the
two
motors
are
illustrated
in
Figure
Based on the preceding analyses, the air gap functions of the two motors are illustrated in Figure 4.4.
Based on the preceding analyses, the air gap functions of the two motors are illustrated in Figure 4.
Outer stator
Outer stator
core
core
Stator core
Stator core
ge
ge
hm
hm
αpm
αpm
π
π
0
0
Rotor core
Rotor core
αpm
αpm
go
go
φ
φ
hm
hm
αpm
αpm
0
0
gi
gi
αpm
αpm
π
π
φ
φ
Inner stator
Inner stator
core
core
Steel
Steel
Steel
Steel
Rotor
Rotor
steels
steels
(a)
(a)
(b)
(b)
Figure3.3.Air
Air gapstructures
structures ofthe
the (a) BDFRM
BDFRM motor
Figure
motor and
and (b)
(b)DSFM
DSFMmotor.
motor.
Figure 3. Airgap
gap structuresof
of the (a)
(a) BDFRM motor
and
(b)
DSFM
motor.
g
g
g
g
hm
hm
hm
hm
go+gi
go+gi
ge
ge
θrm
θrm
θrm
θrm
(a)
(a)
(b)
(b)
Figure 4. Air gap function of the (a) BDFRM motor and (b) DSFM motor.
Figure 4. Air gap function of the (a) BDFRM motor and (b) DSFM motor.
Figure 4. Air gap function of the (a) BDFRM motor and (b) DSFM motor.
According to Figure 4a,b, the air gap function of the DSFM motor and that of the BDFRM are
According to Figure 4a,b, the air gap function of the DSFM motor and that of the BDFRM are
both
periodictorectangular
fundamental
the two
air gap
functions
be
According
Figure 4a,b,waves.
the air The
gap function
of thewaves
DSFMof
motor
and that
of the
BDFRM can
are both
both
periodic rectangular
waves.
The
fundamental
waves
of
the two
air gap
functions
can
be
expressed
as periodic
sinusoidal
waves usingwaves
Fourier
transform
[16].
Therefore,
in
thebesimplified
periodic
rectangular
waves.
The
fundamental
of
the
two
air
gap
functions
can
expressed
expressed as periodic sinusoidal waves using Fourier transform [16]. Therefore, in the simplified
on dynamic
model
of the
DSFM
motor, the
inverse
machine
air simplified
gap function
can be
as discussion
periodic sinusoidal
waves
using
Fourier
transform
[16].
Therefore,
in the
discussion
discussion
on dynamic
model
of the
DSFM
motor, the
inverse
machine
air gap function
can be
modelled
as
follows
[17]:
onmodelled
dynamicasmodel
of[17]:
the DSFM motor, the inverse machine air gap function can be modelled as
follows
1
follows [17]:
1 (,  )  m  n cos pr (  rm )
g
(1)
g
n cos pθ
p (   )
(1)(1)
´
1 (, rm
g pθ, θrmrm
q )“mm`ncosp
r r ´ θrm qrm
where pr is the number of rotor poles, θ is the mechanical angle around the periphery of the
where
pr the
is the
number
of rotor
poles,
θ ismechanical
the mechanical
the periphery
of the
where
pr is
number
of rotor
poles,
θ is the
angle angle
aroundaround
the periphery
of the machine
machine
with respect to the primary A phase axis, and θrm is the rotor mechanical angle with
machine
with
respect
to
the
primary
A
phase
axis,
and
θ
rm is the rotor mechanical angle with
with
respecttotothe
theprimary
primaryAA
phase
axis,
and
rotor mechanical
angle
with reference
to the
rm is
reference
phase
axis.
The
mθ
and
n the
in Equation
(1) are two
constants
in the Fourier
reference
to
the
primary
A
phase
axis.
The
m
and
n
in
Equation
(1)
are
two
constants
in
the
Fourier
primary
phase
m and
n indifferent
Equationair
(1)gap
arestructures,
two constants
in means
the Fourier
series,
they
series, A
and
theyaxis.
differThe
from
the two
which
that the
two and
motors
series,
and
they
differ
fromair
the
two
different which
air gapmeans
structures,
which
means
that
thesimilar
two motors
differ
from
the
two
different
gap
structures,
that
the
two
motors
have
air
gap
have similar air gap functions regardless of the constants.
have similar air gap functions regardless of the constants.
functions
regardless
of
the
constants.
According to the first principle-based rigorous development of dynamic equations for the
According to the first principle-based rigorous development of dynamic equations for the
According
to the firstfor
principle-based
rigorous
development
dynamic
equations
fora the
BDFRM,
BDFRM,
the equations
the primary and
secondary
windingsofof
the DSFM
motor in
stationary
BDFRM, the equations for the primary and secondary windings of the DSFM motor in a stationary
thereference
equations
for
the
primary
and
secondary
windings
of
the
DSFM
motor
in
a
stationary
reference
frame (denoted with the sub-subscript s) are [17]:
reference frame (denoted with the sub-subscript s) are [17]:
frame (denoted with the sub-subscript s) are [17]:
d
d
ˇ ps

R
 j p 
dλ ps ˇ ps
ps 
p ii ps 


R

 ps
dt
υ ps p“
R p ipps p`
`jωj
s
s
p λpps ps
dt dtˇθ const
const
p
p
ˇ  const
p
dλss ˇ
υss “ Rs iss ` ddt
` jω λ
ˇs
d  sssθs const  js ss


R
i

s
s
˚
s
jθrjs  ss
s
s
ss λR
i ss Lp i pdt
` L ps iss e
s ss
pss “
s
dt
s const
(2)
(2)
(2)
s const
˚ jθr
λs s “ L s i s s ` L
* psjips e

i*s e jrr
 ppss 
L
Lpp ii ppss 
L
Lps
ps i sss e
where the primary winding is the winding of outer *stator,
jr and the secondary winding is the winding
*

 ss 
L
Ls ii ss 
L
L ps ii ps ee jr
of inner stator in this paper.
ss
s ss
ps ps
The torque
expression
in the
reference
frame
is [17]:
where
the primary
winding
is stationary
the winding
of outer
stator,
and the secondary winding is the
where the primary winding is the winding of outer stator, and the secondary winding is the
winding of inner stator in this paper.
´
¯
winding of inner stator in this paper. 3
´ jθr
Te “
j pr L ps reference
i˚ps i˚ss e jθr frame
´ i ps isiss e[17]:
(3)
The torque expression in the
stationary
The torque expression in the stationary
reference frame is [17]:
4
Energies 2016, 9, 517
4 of 19
Two dq reference frames are needed to convert the equations in the stationary frame to the
corresponding rotating frame equations because the frequencies of the two windings in the DSFM
motor are different. Assuming that the angle of the primary reference frame is θ, the angle of the
secondary reference frame is θr ´θ, the conversion formulas can be written as:
λ pr “ λ ps e´ jθ
υ pr “ υ ps e´ jθ
i pr “ i ps e´ jθ
ri “ i˚ e jpθr ´θq
sr
ss
(4)
λsr “ λss e´ jpθr ´θq
υsr “ υss e´ jpθr ´θq
isr “ iss e´ jpθr ´θq
ri “ i˚ e jθ
ps
pr
Substituting the above equations into Equation (2), the dynamic model of the DSFM motor in the
rotating dp qp and ds qs reference frames through strict deductions can be represented as:
υp “ Rpi p `
dλ p
dt
` jωλ p
dλs
dt
υs “ R s i s `
` j pωr ´ ωq λs
´
¯
`
˘
λ p “ L p i pd ` ji pq ` L ps looooomooooon
isd ´ jisq
looooomooooon
ip
(5)
i˚
s
´
¯
`
˘
λs “ Ls looooomooooon
isd ` jisq ` L ps i pd ´ ji pq
looooomooooon
is
i˚
p
The torque expression is:
Te “
¯
3 L ps ´
pr
λ pd isq ` λ pq isd
2 Lp
(6)
3. Experiments Conducted Using the Variable Voltage and Variable Frequency (VVVF) Control
The prototype of the DSFM motor was developed experimentally; the test machine and its
controller are shown in Figure 5. The DSFM motor specifications are summarized in Table 1. The
back-to-back controller is comprised of two converters for simultaneously controlling the outer and
inner
Energieswindings.
2016, 9, 517
5 of 20
(a)
(b)
Figure
back-to-back controller.
controller.
Figure 5.
5.Prototype
Prototypeof
ofthe
theDSFM
DSFMmotor
motor and
and its
its controller:
controller: (a)
(a) test
test machine;
machine; and
and (b)
(b) back-to-back
Table 1. DSFM motor prototype parameters and ratings.
Parameter
Inner winding pole pairs
Outer winding pole pairs
Rotor pole pairs
Value
5
6
11
(a)
(b)
Figure
of the DSFM motor and its controller: (a) test machine; and (b) back-to-back controller.5 of 19
Energies
2016, 5.
9, Prototype
517
Table 1. DSFM motor prototype parameters and ratings.
Table 1. DSFM motor prototype parameters and ratings.
Parameter
Parameter
Inner winding
pole pairs
Outer
winding
polepairs
pairs
Inner winding pole
Outer
winding
Rotor
pole pole
pairspairs
Rotor
pole pairs(Ω)
Inner
resistance
Inner resistance (Ω)
Outer
resistance (Ω)
Outer resistance (Ω)
Outer
(mH)
Outerinductance
inductance (mH)
Inner
(mH)
Innerinductance
inductance (mH)
Mutualinductance
inductance (mH)
Mutual
(mH)
Rated
speed (rpm)
Rated speed
(rpm)
Maximum speed (rpm)
Maximum speed (rpm)
Value
Value
5
65
6
11
11
0.08
0.08
0.1
0.1
44
3
0.8
924
924
1300
1300
The VVVF control was implemented to verify the operating principles of the DSFM motor, and
the result is
is shown
shown in
in Figure
Figure 6.
6. The rotor speed
speed was 212 rpm
rpm when
when the frequencies of the outer
outer and
and
inner winding current were
f
=
20
Hz
and
f
=
20
Hz,
respectively.
Several
frequency
supplies
were
ss
were fpp
then tested
tested using
using the
the VVVF
VVVF control.
control. The
The results
results are
are summarized
summarized in
in Table
Table2.2.
The line voltage of A and B phase
The outer winding current
The inner winding current
Amplitude (V/A)
100
50
0
-50
-100
0
0.02
0.04
0.06
0.08
0.1
Time (s)
Figure 6.
6. Experimental
Experimental results
results of
of the
the VVVF
VVVF control.
control.
Figure
Table 2. Test results of the VVVF control.
Frequency (Hz)
Theoretical Speed (rpm)
Real Speed (rpm)
fp = fs = 9
fp = fs = 20
fp = fs = 54
fp = fs = 72
fp = fs = 85
fp = fs = 89
98
218
589
785
927
970
96
212
560
748
883
924
From the test results shown in Table 2, one can get the relation among the frequencies of the two
windings and the rotor speed:
`
˘
nr “ 60 f p ` f s {pr
(7)
The differences between the theoretical and real speeds listed in Table 2 were due to the inaccurate
speed control of the open loop VVVF control.
From the test results shown in Table 2, one can get the relation among the frequencies of the two
windings and the rotor speed:
nr  60  f p  f s  pr
(7)
Energies
2016,
9, 517
The
differences
19
between the theoretical and real speeds listed in Table 2 were due to6 ofthe
inaccurate speed control of the open loop VVVF control.
4. Stator-Current-Oriented Vector Control (SCOVC)
4. Stator-Current-Oriented Vector Control (SCOVC)
Two additional formulas can be deduced from Equation (7).
Two additional formulas can be deduced from Equation (7).
`
˘
ω
rm “ ω
p `ωs {ppr
rm

p
s

(8)
(8)
r
where ω
is the
the angular
angular frequency
frequency of
of the
the primary
primary winding
winding current,
current, ω
is the
the angular
angular frequency
frequency of
of the
the
where
ωpp is
ωss is
secondary winding
winding current,
current, and
and ω
isthe
themechanical
mechanicalangular
angularvelocity
velocityof
ofthe
therotor.
rotor.
rmis
secondary
ωrm
θ
θss
r r“θ
p`
p 
(9)
where θθpp,, θθs ,s,and
andθrθare
r are
electric
angles
ofprimary
the primary
winding
current,
secondary
thethe
electric
angles
of the
winding
current,
secondary
windingwinding
current,
current,
and
rotor, respectively.
and rotor,
respectively.
According to the
the analysis
analysis of
ofthe
thevector
vectorrelationship
relationshipbetween
betweenthe
thetwo
twostator
stator
windings
and
rotor
windings
and
rotor
in
in
DSFM
motor,
vectordiagram
diagramofofthe
theDSFM
DSFMmotor
motorcan
canbe
berepresented
represented as
as shown in Figure 77 [17].
thethe
DSFM
motor,
thethe
vector
[17].
qs

qp
p
ip
s
is

p
ds
r  
s

dp
r  

Figure
7. Reference
frames
current
vectors
used
in DSFM
motor
equations.
Figure
7. Reference
frames
andand
current
vectors
used
in DSFM
motor
equations.
The decoupled control
control of
ofthe
theDSFM
DSFMmotor
motororients
orientsthe
theprimary
primary
winding
current
vector
ip to
winding
current
vector
ip to
thethe
dp
daxis,
p axis,
which
implies:
which
implies:
θ “ θp
(10)
  p
(10)
ω “ ωp
(11)
ωr ´ ω “ 
ωr´ω
p p “ ωs
(11)
(12)
θ s “ θr ´ θ p
(13)
where ωr = pr ˆ ωrm , and:
Thus, the secondary winding current vector ip can be automatically oriented to the ds axis.
The decoupled voltage equations of the DSFM motor become:
υp “ Rpi p `
υs “ R s i s `
dλ p
dt
dλs
dt
` jω p λ p
` jωs λs
(14)
The primary equations are obviously in the ωp frame, and the secondary ones are in the ωs frame.
Therefore, this selection moves each of the equations into their “natural” reference frames where
the respective vector components appear as DC in the steady state. This decoupled control method
proposed in this paper is called SCOVC.
Based on the preceding basic analysis of the dynamic model and vector relationship of the DSFM
motor, two SCOVC methods with different excitation were investigated and implemented as follows.
Energies 2016, 9, 517
7 of 19
4.1. SCOVC with DC Excitation
4.1.1. Theoretical Analysis
If ωp in Equation (11) is equal to 0, which implies that a DC current is inputted in the primary
winding, the angle of the primary reference frame θ is equal to 0, then according to Equation (4), the
conversion formulas of the primary frame become:
υ pr “ υ p s
(15)
i pr “ i p s
Meanwhile, ipq * = 0 control was implemented to realize θp = θ = 0. Substituting this value into
Equation (13) yields:
θ s “ θr
(16)
and the conversion formulas of the secondary frame become:
υsr “ υss e´ jθr
(17)
isr “ iss e´ jθr
That is, a stationary dq reference frame for the primary equation and a rotor reference frame for
the secondary equation. The SCOVC method with DC excitation is simply called DC excitation control
Energies
2016, 9, 517 text. The DC excitation control method for the DSFM motor is illustrated in Figure
8 of 20
in the following
8.
ipα
ipβ
ipd
i*pd
ipq
*
i pq=0
*
ωr
ωr
PI
i*s
upd
upq
isd
PI
usα
usq InvPark usβ SVPWM
isq
isα
isβ Clarke
Park
K
IGBT
SVPWM
Current Vector
Oriented Control
usd
PI
MTPA
PI
PI
Clarke
iA
iB
iC
θr
IGBT
ia
ib
ic
outer
stator
rotor
inner
stator
outer
stator
Figure 8. DC excitation control method.
Figure 8. DC excitation control method.
With the
the direct
direct given
given current
current vector
vector in
in the
the primary
primary winding,
winding, the
the flux
flux linkage
linkage produced
produced by
by the
the
With
primary
winding
current
was
stationary
and
constant,
and
its
value
could
be
adjusted
by
varying
the
primary
current was stationary and constant, and its value could be adjusted by varying
*
*. The
ipd .ipdThe
primary
winding
current
vector
was was
set stationary
and the
winding
currentcurrent
vector
the
primary
winding
current
vector
set stationary
andsecondary
the secondary
winding
was
oriented
by
θ
.
Moreover,
the
maximum
torque
per
ampere
(MTPA)
strategy
was
used
to
control
r by θr. Moreover, the maximum torque per ampere (MTPA) strategy was used to
vector was oriented
* . For* the DC excitation control, Equation (6) becomes:
the secondary
winding
currentcurrent
is d * andisdi*sqand
control
the secondary
winding
isq . For the DC excitation control, Equation (6) becomes:
33
TTe e“ prpLr ps
i pdi pd
isqisq
Lps
22
(18)
(18)
The decoupled
decoupled control
control of
of T
was realized
according to
to Equation
Equation (18);
(18); moreover,
moreover, the
the
The
Tee was
realized through
through iisq
sq according
torque capacity
capacity could
could be
be increased
increased by
by controlling
controlling the
the DC
DC excitation
torque
excitation in
in the
the primary
primary winding.
winding.
4.1.2. Simulation
Simulation Studies
Studies
4.1.2.
The simulation
proposed
DCDC
excitation
control
method
were carried
out in Matlab
The
simulation studies
studiesofofthe
the
proposed
excitation
control
method
were carried
out in
SimulinkSimulink
using theusing
DSFMthe
motor
data motor
from Table
obtained
by Maxwell
model.
The simulation
results
Matlab
DSFM
data1 from
Table
1 obtained
by Maxwell
model.
The
simulation results in Figure 9 have been successfully experimentally verified by the corresponding
experiments. Both the simulation results and the experimental results will be jointly discussed in the
following section.
120
6
Te =
3
pr Lps i pd isq
2
(18)
The decoupled control of Te was realized through isq according to Equation (18); moreover, the
torque capacity could be increased by controlling the DC excitation in the primary winding.
Energies 2016, 9, 4.1.2.
517 Simulation Studies
8 of 19
The simulation studies of the proposed DC excitation control method were carried out in
Matlab Simulink using the DSFM motor data from Table 1 obtained by Maxwell model. The
simulation
in Figureexperimentally
9 have been successfully
experimentally
verified by the corresponding
in Figure 9 have
been results
successfully
verified
by the corresponding
experiments. Both
experiments. Both the simulation results and the experimental results will be jointly discussed in the
simulation results and the experimental results will be jointly discussed in the following section.
following section.
6
5
Torque (Nm)
4
3
2
1
0
-1
0.3
Energies 2016, 9, 517
0.4
0.5
(a)
120
0.6
Time (s)
(b)
0.7
0.8
9 of 20
130
125
110
Speed (rpm)
Speed (rpm)
120
100
90
80
115
110
105
100
95
70
90
60
85
50
0.3
0.4
0.5
0.6
0.7
80
0.3
0.8
0.4
Time (s)
(c)
0.6
0.7
0.8
0.7
0.8
0.7
0.8
Time (s)
(d)
8
20
6
15
Secondary Current (A)
Secondary Current (A)
0.5
10
5
0
-5
-10
-15
4
2
0
-2
-4
-6
-20
0.3
0.4
0.5
0.6
0.7
-8
0.3
0.8
Time (s)
0.4
(e)
(f)
5
4.5
4.5
4
Primary Current (A)
Primary Current (A)
0.6
Time (s)
5
3.5
3
2.5
2
1.5
1
0.5
0
0.3
0.5
4
3.5
3
2.5
2
1.5
1
0.5
0.4
0.5
0.6
0.7
0.8
0
0.3
0.4
0.5
0.6
Time (s)
Time (s)
(g)
(h)
Figure 9. Simulation results of the DC excitation control for the DSFM motor: (a) curve of the given
Figure 9. Simulation
results
of the
DC
excitation
control
for the
DSFM
motor:
(a)
curve
speed variation;
(b) curve
of the
load
torque variation;
(c) speed
curve
for speed
varying
from
55 to of the given
110 rpm;
speed of
curve
torque
varyingvariation;
from 0 to 5 Nm;
(e) secondary
curve for
speed from 55 to
speed variation;
(b)(d)
curve
theforload
torque
(c) speed
curvecurrent
for speed
varying
55 to 110 rpm; (f) secondary current curve for torque varying from 0 to 5 Nm; (g)
110 rpm; (d)varying
speedfrom
curve
for torque varying from 0 to 5 Nm; (e) secondary current curve for speed
primary current curve for speed varying from 55 to 110 rpm; and (h) primary current curve for
varying fromtorque
55 tovarying
110 rpm;
current curve for torque varying from 0 to 5 Nm; (g) primary
from (f)
0 tosecondary
5 Nm.
current curve for speed varying from 55 to 110 rpm; and (h) primary current curve for torque varying
Experimental Results
from 0 to4.1.3.
5 Nm.
Figure 10a,b shows the actual DSFM motor test rig and schematic of its constituent parts,
respectively.
the
Energies 2016, 9, 517
9 of 19
4.1.3. Experimental Results
Figure 10a,b shows the actual DSFM motor test rig and schematic of its constituent
parts,
respectively.
Energies
2016, 9, 517
10 of 20
Energies 2016, 9, 517
10 of 20
Dual-Fed Flux-Modulated
motor
Dual-Fed Flux-Modulated
motor
Servo motor
Servo motor
Servo motor
（load）
Servo
motor
（load）
Torque
sensor
Torque
sensor
Signal Resolver
U1V1W1 U2V2W2
transducer
Signal
U1V1W1 U2V2W2
transducer
Back to Back
converter
+12V Back
U+ to Back
DC
converter
U+12V
U+
Power
DC
UPower
U V W
Controller
Controller
Torque sensor
Torque sensor DC power supply
Oscilloscope Multimeter
DC power supply
Oscilloscope Multimeter
U Vmotor
W
Servo
U1+
controller
Servo
motor
U1U1+
controller
U1-
DC
U1+
Power
DC
U1U1+
(torque)
Power
U1(torque)
DSFM motor
DSFM motor
Resolver
DC
Link
DC
Link
Can Bus
PC
Can Bus
PC
(a)
(a)
(b)
(b)
Figure10.
10.Experimental
Experimental environment
environment of
of the
the DSFM
DSFM motor:
motor: (a)
Figure
(a) laboratory
laboratory test
test facility;
facility;and
and(b)
(b)block
block
Figure
10.ofExperimental
environment
of the
DSFM motor: (a) laboratory test facility; and (b) block
diagram
the
DSFM
motor
experimental
setup.
diagram of the DSFM motor experimental setup.
diagram of the DSFM motor experimental setup.
25
25
25
20
25
Harmonic
Amplitude
(A)(A)
Harmonic
Amplitude
Secondary
Current
(A)(A)
Secondary
Current
The DSFM motor can operate stably with DC excitation in the primary winding. The current
The
motor
can
stably
in the
the primary
primary
winding.The
Thecurrent
current
TheofDSFM
DSFM
motor
can operate
operate
stably
excitation
in
winding.
curves
the DSFM
motor
at
200 rpm
and with
3 NmDC
areexcitation
shown in Figure
11.
curves
Figure 11.
11.
curvesof
ofthe
theDSFM
DSFM motor
motor at
at 200
200 rpm
rpm and 3 Nm are shown in Figure
20
15
15
10
10
5
5
0
0
-5
-5
-10
-10
-15
-15
-20
-20
-25
0
-25
0
0.02
0.04
0.02
0.04
0.06
0.08
0.1
Time (s)
0.06
Time
(s)
(a)
(a)
0.08
0.1
20
20
15
15
10
10
5
5
0
0
0
0
1
2
3
4
5
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15
Harmonic
6 7 8 9 Order
10 11 12
Harmonic
Order
(b)
(b)
13 14 15
Figure 11. Winding currents of the DSFM motor at 200 rpm: (a) secondary winding current; and (b)
Figure
11.Winding
Winding
currents
ofthe
theDSFM
DSFMmotor
motoratat
200
rpm:
secondary
winding
current;
and
the harmonic
content
of the of
current.
Figure
11.
currents
200
rpm:
(a)(a)
secondary
winding
current;
and
(b)(b)
the
the
harmonic
content
of
the
current.
harmonic content of the current.
As shown in Figure 11a, the secondary current was asymmetric sine wave and was stable at 36.6
As shown
in Figure
11a,
current was
wave
andwas
was200
stable
at 36.6
Hz when
the DC
current
in the
the secondary
primary winding
was asymmetric
8 A and thesine
given
speed
rpm.
The
As
shown
in
Figure
11a,
the
secondary
current
was
asymmetric
sine
wave
and
was
stable
at
Hz
when
the
DC
current
in
the
primary
winding
was
8
A
and
the
given
speed
was
200
rpm.
The
harmonic content of the current was low as shown in Figure 11b.
36.6
Hz
when
the
DC
current
in
the
primary
winding
was
8
A
and
the
given
speed
was
200
rpm.
The
harmonic
content the
of the
current wasoflow
shown incontrol
Figure method
11b.
To evaluate
performance
theasproposed
of the DSFM motor, the speed
harmonic
content the
of the
current wasoflow
as
shown incontrol
Figuremethod
11b.
To
evaluate
performance
the
proposed
of thethe
DSFM
speed
response was analysed by changing the given speed or increasing
load motor,
torque,the
and
the
To
evaluate
the
performance
of
the
proposed
control
method
of
the
DSFM
motor,
the
speed
response
was
analysed
by
changing
the
given
speed
or
increasing
the
load
torque,
and
the
corresponding phase currents in both the primary and secondary windings were measured. The
response
was analysed
by changing
the primary
given speed
or increasing
the load
andThe
the
corresponding
phaseare
currents
the
and secondary
windings
were torque,
measured.
experimental results
shownin
inboth
Figure
12.
corresponding
phase
currents
in
both
the
primary
and
secondary
windings
were
measured.
The
experimental results are shown in Figure 12.
experimental results are shown in Figure 12.
Energies 2016, 9, 517
10 of 19
11 of 20
130
7
120
6
110
5
100
Torque (Nm)
Given Speed (rpm)
Energies 2016, 9, 517
90
80
70
60
4
3
2
1
0
50
-1
40
30
0
2
4
6
8
10
12
-2
5
14
10
15
Time (s)
(b)
Time (s)
(a)
120
130
125
110
Speed (rpm)
Speed (rpm)
120
100
90
80
115
110
105
100
95
70
90
60
85
50
0
2
4
6
8
10
12
80
5
14
10
Time (s)
(c)
15
Time (s)
(d)
10
10
Secondary Current (A)
Secondary Current (A)
8
6
4
2
0
-2
-4
-6
-8
-10
0
5
0
-5
-10
2
4
6
8
10
12
5
14
10
50
40
40
Primary Current (A)
Primary Current (A)
50
30
20
10
0
-10
30
20
10
0
-10
-20
0
15
Time (s)
(f)
Time (s)
(e)
-20
2
4
6
8
Time (s)
(g)
10
12
14
5
10
15
Time (s)
(h)
Figure 12. Experimental results of the DC excitation control for the DSFM motor: (a) curve of the given
speed variation; (b) curve of the load torque variation; (c) speed curve for speed varying from 55 to
110 rpm; (d) speed curve for torque varying from 0 to 5 Nm; (e) secondary current curve for speed
varying from 55 to 110 rpm; (f) secondary current curve for torque varying from 0 to 5 Nm; (g) primary
current curve for speed varying from 55 to 110 rpm; and (h) primary current curve for torque varying
from 0 to 5 Nm.
Energies 2016, 9, 517
11 of 19
According to Figure 12a, the given speed was increased from 55 to 110 rpm at 1 s; the final speed
was stable at 110 rpm. In Figure 12c, the rotor speed was changed by a slow progressive change of
the PI control. The secondary winding current was changed during the adjustment period, and the
stability was restored by increasing the frequency only twice from 10 to 20 Hz, as shown in Figure 12e.
The primary winding current was maintained at a constant value, as shown in Figure 12g, because
a DC current was supplied to the primary winding all the time. The steady state errors of the speed
were less than 5 rpm.
By maintaining the given speed at 109 rpm and increasing the load torque from 0 to 5 Nm
at 5 s shown in Figure 12b, the speed of the DSFM motor changed to 109 rpm accurately through
a period of adjustment. The speed curve is presented in Figure 12d. The plot in Figure 12f illustrates
that the amplitude of the secondary winding current was increased with this load torque, and its
frequency was maintained at 20 Hz. The increase in the secondary winding current resulted from
the increase in isq , thus enhancing the electromagnetic torque of the DSFM motor Te as predicted by
Equation (18). The current in the primary winding was maintained at a constant value, as shown in
Figure 12h. The measured and simulated results in Figure 9 are mostly identical to the experimental.
The simulated speed responses were better than the experimental because the PI parameters for speed
control experiment were not ideal.
4.1.4. Discussion
According to the preceding simulation and experiment results, the proposed DC excitation control
can realize the decoupled control for the current of the two stator windings and the electromagnetic
torque Te . For this control method, the DSFM motor can always operate under motoring conditions
similar to PMSMs, which allows the application of the DSFM motor to EVs.
4.2. SCOVC with AC Excitation
4.2.1. Theoretical Analysis
If ωp in Equation (11) is not equal to 0 then, according to Equation (4), the conversion formulas of
the primary frame become:
υ pr “ υ ps e´ jθ p
(19)
i pr “ i ps e´ jθ p
and the conversion formulas of the secondary become:
υsr “ υss e´ jθs
isr “ iss e´ jθs
(20)
This relationship must be satisfied at any time so that the secondary frame can be oriented
according to the rotor angle θr and angle of the primary winding current θp . The SCOVC with AC
excitation is referred to as AC excitation control in the following text. The AC excitation control method
for the DSFM motor is illustrated in Figure 13.
The current closed loop control with a given fp was implemented to produce a rotating magnetic
field with angular velocity ωp , and the primary winding current phase angle θp was established after
the integral link. The secondary winding current phase angle θs was then obtained according to
Equation (13). Thus, the secondary winding current vector is was oriented to the ds axis by θs . The
two winding current vectors were oriented throughout the AC excitation control method. Moreover,
the frequency of the secondary winding changed to fs with a known θr , which implied that the speed
control of the DSFM motor was indirectly conducted using the frequency of the secondary winding
current. For example, if ωr * = 218 rpm and fp = 20 Hz, fs = 20 Hz can be exactly acquired using
the algorithm of the proposed control method, and the rotor can be operated accurately at 218 rpm.
− jθ p
i pr = i psiepβ Clarke
Park
iB
iC
and the conversion
formulas
ipd of the
updsecondary become:
i*pd
up α − j θ
PI
s
υsr =uυpsβ
e SVPWM
ipq
upq InvPark
*
s
i pq=0
PI
i sr = i ss e − jθs
∫ θ
Energies 2016, 9, 517
fp 2π
IGBT
(20)
outer
12 of 19
Current Vector
ωp
p
stator
Oriented Control
rotor
This relationship
must be satisfied
at any time so that the secondary frame can
be oriented
usd
usα
i* the
=0rotor angle
PI
inner
according tosd
θ
r and angle of the primary winding current θp. The SCOVC
with AC
isd was used
usβdecoupled
IGBT of T , which was expressed
SVPWM
usq to InvPark
Moreover,excitation
the MTPA
control
realize
control
stator
e
is* referred
to as AC
excitation
controlthe
in the following text.
The AC excitation
control
PI
ωr
method(18).
for the
motor
using Equation
ia
outer
isq is illustrated in Figure i13.
ωDSFM
r
sα
stator
ib
Park i
sβ Clarke
ic
θr
K
Figure 13. AC excitation control method.
The current closed loop control with a given fp was implemented to produce a rotating magnetic
field with angular velocity ωp, and the primary winding current phase angle θp was established after
the integral link. The secondary winding current phase angle θs was then obtained according to
Equation (13). Thus, the secondary winding current vector is was oriented to the ds axis by θs. The
two winding current vectors were oriented throughout the AC excitation control method. Moreover,
the frequency of the secondary winding changed to fs with a known θr, which implied that the speed
control of the DSFM motor was indirectly conducted using the frequency of the secondary winding
current. For example, if ωr* = 218 rpm and fp = 20 Hz, fs = 20 Hz can be exactly acquired using the
algorithm of the proposed control method, and the rotor can be operated accurately at 218 rpm.
Moreover, the MTPA control was used to realize the decoupled control of Te, which was expressed
Figure
ACexcitation
excitation control
method.
Figure
13.13.
AC
control
method.
using Equation (18).
The current closed loop control with a given fp was implemented to produce a rotating magnetic field with
4.2.2.
Simulation
Results
4.2.2.
Simulation
Results
angular velocity
ωp, and the primary winding current phase angle θp was established after the integral link. The
secondary winding current phase angle θs was then obtained according to Equation (13). Thus, the secondary
To To
fully
test
the
performance
ofofthe
proposed
control strategy,
strategy, start-up
start-upofofthe
themotor
motorunder
under
load
fully
test
the
performance
the
proposed
load
winding
current
vector
is was oriented
to the
ds axis by θs. The two
winding current vectors
were oriented
was
simulated,
and
the
results
are
shown
in
Figure
14.
The
given
speed
was
500
rpm
and
the
load
was simulated,
theexcitation
resultscontrol
are shown
Figurethe14.
The given
speed was
500 changed
rpm and
throughoutand
the AC
method. in
Moreover,
frequency
of the secondary
winding
to fsthe load
with
known
torque
was
1.4a1.4
Nm.
torque
was
Nm.θr, which implied that the speed control of the DSFM motor was indirectly conducted using the
Speed (rpm)
500
400
300
frequency of the secondary winding current. For example, if ωr* = 218 rpm and fp = 20 Hz, fs = 20 Hz can be
600
exactly acquired using the algorithm of the proposed control method,
and the rotor can be operated accurately
at 218 rpm. Moreover, the MTPA control was used to realize the decoupled control of Te, which was expressed
500
using Equation (18).
4.2.2. Simulation Results
Speed (rpm)
600
400
300
To fully test the performance of the proposed control strategy,
start-up of the motor under load was
simulated, and the results are shown in Figure 14. The given speed was 500 rpm and the load torque was 1.4 Nm.
200
200
100
100
0
0
-100
0
0.2
0.4
0.6
0.8
1
-100
0
1.2
0.1
0.2
Energies 2016, 9, 517
0.3
0.4
0.5
Time (s)
(b)
Time (s)
(a)
0.6
14 of 20
7
2.5
6
Torque (Nm)
Torque (Nm)
2
1.5
1
5
4
3
2
0.5
1
0
0
0.2
0.4
0.6
0.8
1
0
0
1.2
25
20
20
Figure 14. Cont.
10
5
0
-5
ndary Current (A)
ndary Current (A)
25
15
0.1
0.2
0.3
Time (s)
(d)
Time (s)
(c)
15
10
5
0
-5
0.4
0.5
0.6
T
T
2
0.5
1
0
0.4
0.6
0.8
1
0
0
1.2
0.2
25
20
20
15
10
5
0
-5
-10
-15
0.4
0.5
0.6
13 of
19
15
10
5
0
-5
-10
-15
-20
-20
0.2
0.4
0.6
0.8
1
-25
0
1.2
0.1
0.2
0.3
0.4
0.5
0.4
0.5
0.6
Time (s)
(f)
Time (s)
(e)
6
20
15
Primary Current (A)
4
Primary Current (A)
0.3
Time (s)
(d)
25
-25
0
0.1
Time (s)
(c)
Secondary Current (A)
Secondary Current (A)
Energies02016, 9,0.2
517
2
0
-2
10
5
0
-5
-10
-4
-15
-6
0
0.2
0.4
0.6
Time (s)
(g)
0.8
1
1.2
-20
0
0.1
0.2
0.3
0.6
Time (s)
(h)
*pd = 5
Figure
ofof
thethe
motor
under
load:
(a) speed
curve
for ifor
Figure 14.
14. Simulation
Simulationresults
resultsofofthe
thestart-up
start-up
motor
under
load:
(a) speed
curve
i* pdA;=(b)
5
*pd = 15 A;
*pd = 5 A; (d)
*pd = 15 A; (e)*secondary
*
*
speed
curve
for
i
(c)
torque
curve
for
i
torque
curve
for
i
A; (b) speed curve for i pd = 15 A; (c) torque curve for i pd = 5 A; (d) torque curve for i pd = 15 A;
current
curve for
i*pd = curve
5 A; (f)forsecondary
current
curve forcurrent
i*pd = 15curve
A; (g)for
primary
current
for
(e) secondary
current
i* pd = 5 A;
(f) secondary
i* pd = 15
A; (g)curve
primary
*
*
*
*
icurrent
pd = 5 A;
and(h)
current
curve
for i current
pd = 15 A.
curve
forprimary
i pd = 5 A;
and(h)
primary
curve for i pd = 15 A.
Note from Figure 14 that when the primary current i**pd increased, the starting torque increased
Note from Figure 14 that when the primary current i pd increased, the starting torque increased
and the starting time decreased. The start-up simulation illustrates that Te can be controlled not
and the starting time decreased. The start-up simulation illustrates that Te can be controlled not only
only by the secondary current
i*sq, but also by the primary current i*pd, one can choose the better
by the secondary current i* sq , but also by the primary current i* pd , one can choose the better control
control method according to the actual situation.
method according to the actual situation.
More simulation studies of the proposed AC excitation control method were carried out and
More simulation studies of the proposed AC excitation control method were carried out and
experimentally verified by the corresponding experiments. The simulation results are shown in
experimentally verified by the corresponding experiments. The simulation results are shown in
Figure 15, and agree well with the experimental results. The corresponding plots will be jointly
Figure 15, and agree well with the experimental results. The corresponding plots will be jointly
discussed in the next section.
discussed in the next section.
Energies 2016, 9, 517
14 of 19
Energies 2016, 9, 517
15 of 20
6
220
Torque (Nm)
Given Speed (rpm)
5
200
180
160
140
3
2
1
120
100
0.2
4
0
0.3
0.4
0.5
0.6
0.7
-1
0.3
0.8
Time (s)
(a)
0.4
0.5
0.6
0.7
0.8
Time (s)
(b)
240
240
235
220
Speed (rpm)
Speed (rpm)
230
200
180
160
225
220
215
210
205
140
200
120
195
100
0.2
0.3
0.4
0.5
0.6
0.7
190
0.3
0.8
0.4
Time (s)
(c)
0.6
0.7
0.8
0.7
0.8
Time (s)
(d)
15
20
15
Secondary Current (A)
Secondary Current (A)
0.5
10
5
0
-5
-10
-15
10
5
0
-5
-10
-20
0.2
0.3
0.4
0.5
0.6
0.7
-15
0.3
0.8
0.4
0.5
8
4
6
3
4
2
0
-2
-4
-6
-8
0.2
0.6
Time (s)
(f)
Primary Current (A)
Primary Current (A)
Time (s)
(e)
2
1
0
-1
-2
-3
0.3
0.4
0.5
Time (s)
(g)
0.6
0.7
0.8
-4
0.3
0.4
0.5
0.6
0.7
0.8
Time (s)
(h)
Figure 15. Simulation results of the AC excitation control for the DSFM motor: (a) curve of the given
Figure 15. Simulation results of the AC excitation control for the DSFM motor: (a) curve of the given
speed variation. (b) curve of the load torque variation; (c) speed curve for speed varying from 110 to
speed variation. (b) curve of the load torque variation; (c) speed curve for speed varying from 110 to
220 rpm; (d) speed curve for torque varying from 0 to 5 Nm; (e) secondary current curve for speed
varying from 110 to 220 rpm; (f) secondary current curve for torque varying from 0 to 5 Nm; (g) primary
current curve for speed varying from 110 to 220 rpm; and (h) primary current curve for torque varying
from 0 to 5 Nm.
Energies 2016, 9, 517
16 of 20
220 rpm; (d) speed curve for torque varying from 0 to 5 Nm; (e) secondary current curve for speed
varying from 110 to 220 rpm; (f) secondary current curve for torque varying from 0 to 5 Nm; (g)
primary
curve for speed varying from 110 to 220 rpm; and (h) primary current curve for
Energies
2016, 9, current
517
15 of 19
torque varying from 0 to 5 Nm.
Experimental Results
4.2.3. Experimental
different speeds with different load torques. The current
current curves
The DSFM motor can operate at different
of the DSFM motor at 924 rpm and 3 Nm are shown in Figure 16.
20
15
Secondary Current (A)
Primary Current (A)
15
10
5
0
-5
10
5
0
-5
-10
-15
-10
-20
-15
-0.1
-0.05
0
0.05
-25
-0.1
0.1
-0.05
0
Harmonic Amplitude (A)
Harmonic Amplitude (A)
0.1
20
14
12
10
8
6
4
2
0
0
0.05
Time (s)
(b)
Time (s)
(a)
18
16
14
12
10
8
6
4
2
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15
Harmonic Order
(c)
0
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15
Harmonic Order
(d)
Figure
16. Winding
Windingcurrents
currents
of the
DSFM
motor
924(a)rpm:
(a) primary
current;
(b)
Figure 16.
of the
DSFM
motor
at 924at
rpm:
primary
winding winding
current; (b)
secondary
secondary
winding
current;
(c)
the
harmonic
content
of
the
primary
current;
and
(d)
the
harmonic
winding current; (c) the harmonic content of the primary current; and (d) the harmonic content of the
content
of the
secondary current.
secondary
current.
As shown in Figure 16a,b, the two winding currents are asymmetric sine waves and are stable
As shown in Figure 16a,b, the two winding currents are asymmetric sine waves and are stable
at 85 Hz when the given speed was 924 rpm. The harmonic content of the currents were both low as
at 85 Hz when the given speed was 924 rpm. The harmonic content of the currents were both low as
shown in Figure 16c,d.
shown in Figure 16c,d.
Then the speed responses and phase currents are illustrated in Figure 17.
Then the speed responses and phase currents are illustrated in Figure 17.
Energies 2016, 9, 517
16 of 19
17 of 20
240
6
220
5
200
4
Torque (Nm)
Given Speed (rpm)
Energies 2016, 9, 517
180
160
140
3
2
1
120
0
100
80
0
2
4
6
8
10
12
-1
5
14
10
Time (s)
15
Time (s)
(a)
(b)
240
240
235
220
230
Speed (rpm)
Speed (rpm)
200
180
160
225
220
215
210
205
140
200
120
195
100
0
2
4
6
8
10
12
190
5
14
10
(c)
10
20
15
Secondary Current (A)
Secondary Current (A)
8
6
4
2
0
-2
-4
-6
10
5
0
-5
-10
-15
-8
-10
15
Time (s)
(d)
Time (s)
0
2
4
6
8
10
12
-20
14
5
6
7
8
9
10
11
12
13
14
15
11
12
13
14
15
Time (s)
Time (s)
(e)
(f)
10
10
Primary Current (A)
Primary Current (A)
8
5
0
-5
6
4
2
0
-2
-4
-10
-6
-15
-10
-8
0
2
4
6
8
Time (s)
(g)
10
12
14
5
6
7
8
9
10
Time (s)
(h)
Figure 17. Experimental results of the AC excitation control for the DSFM motor: (a) curve of the
given speed variation. (b) curve of the load torque variation; (c) speed curve for speed varying from
110 to 220 rpm; (d) speed curve for torque varying from 0 to 5 Nm; (e) secondary current curve for
speed varying from 110 to 220 rpm; (f) secondary current curve for torque varying from 0 to 5 Nm;
(g) primary current curve for speed varying from 110 to 220 rpm; and (h) primary current curve for
torque varying from 0 to 5 Nm.
Energies 2016, 9, 517
17 of 19
There are two processes in the AC excitation control experiments as shown in Figure 17a. First,
the given speed was increased from 110 to 220 rpm at 3 s. Second, the frequency of primary winding
current was increased from 10 to 20 Hz at 8 s. According to Figure 17c, the final speed was stable
at 220 rpm throughout a slow progressive change of the PI control. The secondary winding current
changed during the two periods of adjustment. First, from 3–8 s, its frequency increased from 10 to
30 Hz. Second, from 8–10 s, its frequency decreased from 30 to 20 Hz, and was finally stable after 14 s.
The secondary current curve is shown in Figure 17e. The amplitude of the primary winding current
restored stability and the frequency was changed to 20 Hz through an adjustment period of 2 s from
8–10 s as shown in Figure 17g. The steady state errors of the speed were less than 5 rpm.
By maintaining the given speed at 218 rpm and increasing the load torque from 0 to 5 Nm at 5 s
shown in Figure 17b, the speed of the DSFM motor changed to 218 rpm accurately through a period
of adjustment. The speed curve is presented in Figure 17d. The plot in Figure 17f illustrates that the
amplitude of the secondary winding current was increased using this load torque, and its frequency
was maintained at 20 Hz. The current in the primary winding was controlled using the current closed
loop control presented in Figure 11 and maintained constant, as shown in Figure 17h. The same speed,
frequency and load torque change was implemented as shown in Figure 15, and the results are also
the same. The speed can be strictly controlled by the speed closed loop control, and the frequency of
the two winding currents can be automatically adjusted through the AC excitation control method.
4.2.4. Discussion
The experimental results showed high speed response with the change in speed and load torque,
which illustrated that the DSFM motor can operate stably and reliably under different conditions. The
wide speed range of the DSFM motor can be achieved without flux-weakening control such that the
copper loss can be reduced and the efficiency can be increased at a high speed compared with PMSMs.
Thus, the proposed control can meet the requirement of high-speed operation and efficient torque
control, and would have great application prospects in EVs.
5. Conclusions
This study explored and investigated the control strategies for the application of the DSFM motor
to EVs. Two available vector control methods were proposed and control system effects were validated
using a customised DSFM prototype.
1.
2.
3.
4.
The differences in air gap structures of the DSFM and BDFM motors were compared and
analysed, and the mathematical model of the DSFM motor was used as the basis of the decoupled
control method.
The VVVF control was implemented to explore and verify the operating principles of the DSFM
motor. According to the experimental results of the VVVF control, the speed and torque of the
DSFM motor can be controlled using the frequency of the current in the two stator windings.
The SCOVC with DC excitation control method was realized theoretically and experimentally, and
high speed and torque performance were obtained through the variable speed and torque tests.
The SCOVC with AC excitation control method was realized theoretically and experimentally, and
high speed and torque performance were obtained through the variable speed and torque tests.
The DSFM motor and its control strategies proposed in this paper can serve as a basis for further
research on application of the flux-modulated motor to EVs, and can be crucial for more applications
of such motors.
The DC excitation control had a limited range of frequency, while the AC one did not. Thus, the
two closed loop controls can be flexibly applied to different situations. For example, in generator mode,
the DSFM motor can be used as a DC generator with the DC one or an AC generator with the AC one,
and in motor mode, it can be used as a slow speed motor with the DC one or a high speed motor with
Energies 2016, 9, 517
18 of 19
the AC one. The proposed control methods will be implemented on an electric bus being developed in
our laboratory.
Acknowledgments: This study was supported in part by the National Science Foundation of China (51477058),
Science and Technology Plan Project of Xiamen (3502Z20153029), and Promotion Program for Young and
Middle-aged Teacher in Science and Technology Research of Huaqiao University (ZNQ-YX304).
Author Contributions: Xinhua Guo and Weinong Fu conceived and designed the experiments; Shaozhe Wu,
Yunchong Wang, and Peihuang Zeng performed the experiments; Shaozhe Wu performed the simulations;
Shaozhe Wu and Yulong Liu analysed the data; and Xinhua Guo wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.
Abbreviations
The following abbreviations are used in this paper:
DSFM
BDFMs
BDFG
BDFRM
EV
SCOVC
DC
AC
VC
FOC
PMSMs
VVVF
MTPA
Dual-stator flux-modulated
Brushless doubly-fed machines
Brushless doubly-fed generator
Brushless doubly-fed reluctance machine
Electric vehicle
Stator-current-oriented vector control
Direct current
Alternating current
Voltage vector-oriented control
Field-oriented control
Permanent magnet synchronous motors
Variable voltage and variable frequency
Maximum torque per ampere
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Long, B. Energy-regenerative braking control of electric vehicles using three-phase brushless direct-current
motors. Energies 2014, 7, 99–114. [CrossRef]
Cheng, M.; Sun, L.; Buja, G.; Song, L.H. Advanced electrical machines and machine-based systems for
electric and hybrid vehicles. Energies 2015, 8, 9541–9564. [CrossRef]
Chan, C.C. The state of the art of electric, hybrid, and fuel cell vehicles. Proc. IEEE 2007, 95, 704–718.
[CrossRef]
Chan, C.C.; Wong, Y.S. The state of the art of electric vehicles technology. In Proceedings of the 4th
International Power Electronics and Motion Control Conference, Xi’an, China, 14–16 August 2004; pp. 46–57.
Wang, Y.C.; Ho, S.L.; Fu, W.N.; Shen, J.X. A novel brushless doubly fed generator for wind power generation.
IEEE Trans. Magn. 2012, 48, 4172–4175. [CrossRef]
Knight, A.; Betz, R.; Dorrell, D.G. Design and analysis of brushless doubly fed reluctance machines.
IEEE Trans. Ind. Appl. 2013, 49, 50–58. [CrossRef]
Hsieh, M.F.; Lin, I.H.; Dorrell, D.G. Magnetic circuit modeling of brushless doubly-fed machines with
induction and reluctance rotors. IEEE Trans. Magn. 2013, 49, 2359–2362. [CrossRef]
Roberts, P.; Long, T.; McMahon, R.; Shao, S.; Abdi, E.; Maciejowski, J. Dynamic modelling of the brushless
doubly fed machine. IET Elect. Power Appl. 2013, 7, 544–556. [CrossRef]
Knight, A.M.; Betz, R.E.; Song, W.K.; Dorrell, D.G. Brushless doubly-fed reluctance machine rotor design.
In Proceedings of the IEEE Energy Conversion Congress and Exposition (ECCE), Raleigh, NC, USA,
15–20 September 2012; pp. 2308–2315.
Logan, T.; McMahon, R.; Seffen, K. Noise and vibration in brushless doubly fed machine and brushless
doubly fed reluctance machine. IET Elect. Power Appl. 2013, 8, 50–59. [CrossRef]
Jovanovic, M. Sensored and sensorless speed control methods for brushless doubly fed reluctance motors.
IET Elect. Power Appl. 2009, 3, 503–513. [CrossRef]
Ademi, S.; Jovanovic, M. Vector control strategies for brushless doubly-fed reluctance wind generators.
In Proceedings of the 2nd International Symposium on Environment-Friendly Energies and Applications
(EFEA), Newcastle upon Tyne, UK, 25–27 June 2012; pp. 44–49.
Energies 2016, 9, 517
13.
14.
15.
16.
17.
19 of 19
Ademi, S.; Jovanovic, M. Theoretical and experimental evaluation of vector control for doubly-fed reluctance
generators. In Proceedings of the International Conference on Electrical Machines (ICEM), Berlin, Germany,
2–5 September 2014; pp. 936–942.
Ademi, S.; Jovanovic, M. Maximum torque per inverter ampere control of brushless doubly-fed reluctance
generators for wind turbines. In Proceedings of the International Symposium on Power Electronics, Electrical
Drives, Automation and Motion (SPEEDAM), Ischia, Italy, 18–20 June 2014; pp. 883–888.
Ademi, S.; Jovanovic, M.; Hasan, M. Control of brushless doubly-fed reluctance generators for wind energy
conversion systems. IEEE Trans. Energy Convers. 2015, 30, 596–604. [CrossRef]
Neti, P.; Nandi, S. Determination of effective air-gap length of synchronous reluctance motors (SynchRel)
from experimental data. IEEE Trans. Ind. Appl. 2006, 4, 454–464. [CrossRef]
Betz, R.E.; Jovanovic, M.G. Introduction to the space vector modeling of the brushless doubly fed reluctance
machine. Elect. Power Compnt. Syst. 2003, 31, 729–755. [CrossRef]
© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).