energies

Article
Optimal Pole Number and Winding Designs for Low
Speed–High Torque Synchronous
Reluctance Machines
Gurutz Artetxe 1,2, * ID
, Jesus Paredes 1,2 , Borja Prieto 1,2 , Miguel Martinez-Iturralde 1,2 and
Ibon Elosegui 1,2
1 Ceit, Manuel Lardizabal 15, 20018 Donostia/San Sebastián, Spain; jppuente@ceit.es (J.P.);
bprieto@ceit.es (B.P.); mmiturralde@ceit.es (M.M.-I.); ielosegui@ceit.es (I.E.)
2 Tecnun, Universidad de Navarra, Manuel Lardizabal 15, 20018 Donostia/San Sebastián, Spain
* Correspondence: gartetxe@ceit.es; Tel.: +34-943-212-800

Received: 16 November 2017; Accepted: 31 December 2017; Published: 5 January 2018

Abstract: This paper studies the feasibility of using synchronous reluctance machines (SynRM) for
low speed–high torque applications. The challenge lies in obtaining low torque ripple values, high
power factor, and, especially, high torque density values, comparable to those of permanent magnet
synchronous machines (PMSMs), but without resorting to use permanent magnets. A design and
calculation procedure based on multistatic finite element analysis is developed and experimentally
validated via a 200 Nm, 160 rpm prototype SynRM. After that, machine designs with different
rotor pole and stator slot number combinations are studied, together with different winding
types: integral-slot distributed-windings (ISDW), fractional-slot distributed-windings (FSDW) and
fractional-slot concentrated-windings (FSCW). Some design criteria for low-speed SynRM are drawn
from the results of the study. Finally, a performance comparison between a PMSM and a SynRM is
performed for the same application and the conclusions of the study are summarized.

Keywords: AC drives; magnetless machine; synchronous reluctance; synchronous reluctance

machines (SynRM); motor design; fractional slot concentrated winding; fractional-slot
concentrated-windings (FSCW)

1. Introduction
Permanent magnet synchronous machines (PMSM) are the most widely used machines for low
speed–high torque applications. They have the advantage of having a high torque density capability,
at the expense of needing large amounts of rare-earth permanent magnets, whose price represents a
high percentage of the total cost of the machine.
However, the sudden rise of the price of rare-earth magnets that took place at the beginning of
the decade greatly increased the cost of PMSMs. Thus, the interest in other technologies has increased.
Among them, the synchronous reluctance machine (SynRM) arises as a good alternative.
Many studies have shown that SynRMs offer several advantages compared to induction
machines (IM), such as a higher efficiency and torque density, lower rotor temperature and, potentially,
lower price [1–4]. Some advantages compared to PMSMs are also mentioned: no rotor temperature
restriction, better fault-tolerant behavior and lower price [2,5–7]. In some applications, SynRMs have
been deemed a viable alternative for IMs and even for PMSMs, e.g., in traction or industrial
applications [2–4,8].
Replacing PMSMs by SynRMs in low speed–high torque applications is especially interesting
and challenging. Moreover, using high pole number or FSCWs are common design strategies on
PMSMs for these applications, but they are not for SynRMs. Only a few examples for SynRMs can

Energies 2018, 11, 128; doi:10.3390/en11010128 www.mdpi.com/journal/energies

Energies 2018, 11, 128 2 of 21

Energies 2018, 11, 128 2 of 21

be found in the bibliography [9–15], but they have not been investigated together, or have evaluated
PMSMs for these applications, but they are not for SynRMs. Only a few examples for SynRMs can be
different winding types for a specific application.
found in the bibliography [9–15], but they have not been investigated together, or have evaluated
Consequently, the objective of this paper is to study the feasibility of SynRMs for low speed
different winding types for a specific application.
applications. For that purpose, firstly a procedure for designing SynRMs is developed and validated
Consequently, the objective of this paper is to study the feasibility of SynRMs for low speed
by theapplications. For and
manufacture testing of
that purpose, a 200
firstly Nm 160 rpm
a procedure prototype.
for designing Following
SynRMs this procedure,
is developed several
and validated
machines
by theare designed and
manufacture and testing
analyzed of ato200find
Nmthe 160optimal pole number
rpm prototype. andthis
Following stator slot combinations,
procedure, several
together with are
machines evaluations
designed andof various
analyzedwinding
to find the types: fractional-slot
optimal pole number concentrated-winding (FSCW),
and stator slot combinations,
together with evaluations of various winding types: fractional-slot
fractional-slot distributed-winding (FSDW) and integral-slot distributed-winding (ISDW).concentrated-winding (FSCW),
fractional-slot
This paper isdistributed-winding
organized as follows: (FSDW) andthe
first integral-slot
developed distributed-winding (ISDW). procedure is
design and calculation
This paper is organized as follows: first the developed design and
presented in Section 2 and validated in Section 3; in Section 4, the machine specifications calculation procedure
and theislimits
presented in Section 2 and validated in Section 3; in Section 4, the machine specifications and the
of the study are explained; in Section 5, the results of the studied machines are deeply analyzed and
limits of the study are explained; in Section 5, the results of the studied machines are deeply analyzed
some design criteria are obtained; in Section 6, a SynRM is designed based on the previously obtained
and some design criteria are obtained; in Section 6, a SynRM is designed based on the previously
design criteriadesign
obtained and itscriteria
performance is comparedis to
and its performance that of atoPMSM
compared that of candidate for the same
a PMSM candidate application.
for the same
Finally, the mainFinally,
application. conclusions of the
the main study are
conclusions discussed
of the indiscussed
study are Section 7.in Section 7.

2. Design and and

2. Design Calculation Procedure
Calculation Procedure
In this section,
In this a procedure
section, a proceduretotoobtain
obtain the
the full machinedefinition
full machine definition from
from thethe specifications
specifications is is
developed.
developed. Algebraic,
Algebraic, finite
finite element
element method
method (FEM)and
(FEM) andautomatic
automaticoptimization
optimization methods
methodsareareused
used for
for this.are
this. These These are developed
developed in a tool in generated
a tool generated ® that
in MATLAB
in MATLAB ® that works together with the open
works together with the open source
source FEM
FEM software FEMM . software
® FEMM ® .
Firstly, the winding configuration is defined. The winding type (ISDW, FSDW or FSCW), the
Firstly, the winding configuration is defined. The winding type (ISDW, FSDW or FSCW), the pole
pole pair number (p), the stator slot number (Qs), the slot layer number (nl,w,s) and the short pitching
pair number (p), the stator slot number (Qs ), the slot layer number (nl,w,s ) and the short pitching (τ a ) are
(τa) are chosen by the designer. Some winding configuration examples are shown in Figure 1 for a
chosen by the designer. Some winding configuration examples are shown in Figure 1 for a two-pole
two-pole pair machine, with different winding types, and thus a different number of slots and
pair machine, with different winding types, and thus a different number of slots and winding layers.
winding layers.

(a) (b)

(c) (d)

(e) (f)
Figure 1. Different examples of four pole machine winding configurations (a) integral-slot
Figure 1. Different examples of four pole machine winding configurations (a) integral-slot
distributed-winding (ISDW) (p = 2, Qs = 12, nl,w,s = 1); (b) ISDW (p = 2, Qs = 24, nl,w,s = 1); (c) ISDW (p =
distributed-winding (ISDW) (p = 2, Qs = 12, nl,w,s = 1); (b) ISDW (p = 2, Qs = 24, nl,w,s = 1); (c) ISDW
(p = 2, Qs = 24, nl,w,s = 2, τa = 1); (d) fractional-slot distributed-winding FSDW (p = 2, Qs = 18, nl,w,s = 2);
(e) fractional-slot concentrated-winding FSCW (p = 2, Qs = 6, nl,w,s = 1); (f) FSCW (p = 2, Qs = 6, nl,w,s = 2).
Energies 2018, 11, 128 3 of 21
Energies 2018, 11, 128 3 of 21
2, Qs = 24, nl,w,s = 2, τa = 1); (d) fractional-slot distributed-winding FSDW (p = 2, Qs = 18, nl,w,s = 2); (e)
fractional-slot concentrated-winding FSCW (p = 2, Qs = 6, nl,w,s = 1); (f) FSCW (p = 2, Qs = 6, nl,w,s = 2).
Then, the main machine dimensions are calculated based on the sizing Equations (1) and (2) [16].
Air-gapThen, the main
volume machine
(Vagap dimensions
) and stator current are calculated
(I) are calculated basedfor on the sizing
a given torque Equations
(Mem,ref ),(1) and (ω)
speed (2) [16].
and
Air-gapspecifications
voltage volume (Vagap)(U and
n,refstator
), and current (I) are calculated
other parameters that arefor a given
imposed torque
based on (Mtheem,ref ), speed (ω)
designer’s and
criteria
voltage
(e.g., specifications
electric (Un,ref), and
loading, current other
density, parameters that are imposed based on the designer’s criteria
etc.):
(e.g., electric loading, current density, etc.):
Mem,re f
Vagap = √ 𝑀𝑒𝑚,𝑟𝑒𝑓 (1)
𝑉𝑎𝑔𝑎𝑝 = 2·k w,s1 · Are f · Bagap,re f · PFre f ·ηe,re f (1)
√2 · 𝑘𝑤,𝑠1 · 𝐴𝑟𝑒𝑓 · 𝐵𝑎𝑔𝑎𝑝,𝑟𝑒𝑓 · 𝑃𝐹𝑟𝑒𝑓 · 𝜂𝑒,𝑟𝑒𝑓
Pout,re f
I= √ 𝑃𝑜𝑢𝑡,𝑟𝑒𝑓 (2)
𝐼 = 3·Un,re f · PFre f ·ηe,re f (2)
√3 · 𝑈 A · 𝑃𝐹 𝑟𝑒𝑓 · 𝜂𝑒,𝑟𝑒𝑓
where kω,s1 is the first harmonic winding factor,𝑛,𝑟𝑒𝑓 ref is the reference electric loading, B agap,ref is the
reference
where kω,s1 air-gap
is theflux
firstdensity,
harmonic PFrefwinding
the expected
factor,power
Aref isfactor and η e,ref the
the reference expected
electric efficiency.
loading, Bagap,ref is the
Innerair-gap
reference stator flux
diameter
density,andPFstack
ref the length
expected arepower
obtained factorfromandtheηe,refair-gap volume.
the expected The needed
efficiency.
slot section
Inner statoris calculated
diameter and basedstackonlength
the stator current,
are obtained fromthethe slot fill factor
air-gap volume.and The the current
needed slotdensity.
section
The outer stator
is calculated diameter
based on the and statorthecurrent,
slot dimensions
the slot fill arefactor
obtainedand from the slotdensity.
the current section,The so the
outerstator
statoris
completely defined.
diameter and the slot dimensions are obtained from the slot section, so the stator is completely defined.
The
The considered
considered rotor rotor geometry
geometry is is the
the one
one with
with elliptic-shaped
elliptic-shaped barriers,barriers, allall of
of them
them withwith the
the same
same
width
width (see Figure 2), which has been proven to be a good solution [5,17,18]. αdd represents the ratio
(see Figure 2), which has been proven to be a good solution [5,17,18]. α represents the ratio
between
between the the iron
ironarc
arcininthe
theddaxesaxes(the(thearc between
arc between thethe first barrier
first barrierof two
of two poles
polesat the air-gap)
at the air-gap)andandthe
polar arc atarc
the polar theatair-gap. αq represents
the air-gap. αq represents the ratio
the between
ratio between the iron thearciron
in the
arcqinaxesthe (the
q axesarc(the
between the two
arc between
extremes on the last barrier of a pole at the air-gap) and the polar arc at the
the two extremes on the last barrier of a pole at the air-gap) and the polar arc at theintair-gap. βint refersair-gap. β refers to the ratio
between
to the ratiothebetween
inner ironthelength
inner in thelength
iron centerin ofthe
thecenter
pole (the of thequadrature
pole (theaxis), and theaxis),
quadrature totalandrotorthe height.
total
β ext refers
rotor to the
height. ratio
βext between
refers to the theratio
outerbetween
iron length theinouter
the center ironoflength
the pole in (the
the quadrature
center of the axis),pole
and(the
the
total rotor height.
quadrature axis), and γ indicates the portion of the barrier (air) arc for each barrier
α the total rotor height. γα indicates the portion of the barrier (air) arc for each plus island (iron) arc
at the air-gap.
barrier plus islandγβ indicates
(iron) arc theatportion of the γ
the air-gap. barrier length
β indicates forportion
the each barrier
of theplus island
barrier length
length forineach
the
center of the pole.
barrier plus island length in the center of the pole.

Figure 2.
Figure Elliptic-shaped barrier
2. Elliptic-shaped barrier rotor
rotor geometry.
geometry.

rotor geometry
The rotor geometry is is defined
defined so
so the
the air/iron
air/iron proportion
proportion in radial and circumferential directions
constant regardless
is constant regardlessofofthe
thenumber
numberofofpoles.
poles.Table
Table 1 shows
1 shows thethe rotor
rotor geometric
geometric parameters’
parameters’ values
values for
for the example pre-defined
the example pre-defined rotor. rotor.
Energies 2018, 11, 128 4 of 21
Energies 2018, 11, 128 4 of 21

Table 1. Pre-defined
Table 1. Pre-defined rotor
rotor geometric
geometric parameters.
parameters.

Parameter Value
Parameter Value
αd 0.2
αdαq 0.2
0.2
αq 0.2
βint 0.2
βint 0.2
β
βextext 0.2
0.2
γαγα 0.5
0.5
γβγβ 0.5
0.5
Number of barriers
Number of barriers 44

With the
With thewinding
windingconfiguration,
configuration, stator
stator geometry
geometry andand a pre-defined
a pre-defined rotor rotor geometry,
geometry, the firstthe first
design
design
of of the machine
the machine is fully is fully defined.
defined.
completely defined machine,
For the completely machine, full performance characteristics are calculated with
multistatic FEM simulations, where the current angle is calculated by applying the maximum torque
(MTPA) control
per ampere (MTPA) control strategy.
strategy. The flux linkage and flux densities are obtained from these
simulations, and the rest of the machine
machine performance characteristics are calculated based on the
SynRM vector diagram, FigureFigure 3,
3, and
and algebraic
algebraic formulation,
formulation, Equations
Equations (3)–(6)
(3)–(6) [19]:
[19]:

Figure 3. Synchronous reluctance machine (SynRM) vector diagram [19].

Figure 3. Synchronous reluctance machine (SynRM) vector diagram [19].

33
M𝑀 == p 𝑝(𝐿
L 𝑑−−L𝐿q 𝑞 I)𝐼sin
sin(2𝜃) (3)


em𝑒𝑚 (2θ ) (3)
22 d
em = −ωLq iq + jωLd id (4)
𝑒𝑚 = −𝜔𝐿𝑞 𝑖𝑞 + 𝑗𝜔𝐿𝑑 𝑖𝑑 (4)
s
sin(2θ )
PF = (ξ − 1) (5)
2(tan θsin(2𝜃)
+ ξ 2 cot θ )
𝑃𝐹 = (𝜉 − 1)√ (5)
 2(tan 𝜃  +−𝜉12 cot 𝜃)
Ploss
ηe = 1 + (6)
ω · Mem −1
𝑃𝑙𝑜𝑠𝑠
where Ld and Lq the direct and quadrature𝜂𝑒inductances,
= (1 +
𝜔 · 𝑀θ𝑒𝑚is)the current angle, em is the electromotive
(6)
force, id and iq the direct and quadrature currents, ξ is the saliency ratio and Ploss is the value of the
where Ld and Lq the direct and quadrature inductances, θ is the current angle, em is the electromotive
total machine losses.
force, id and iq the direct and quadrature currents, ξ is the saliency ratio and Ploss is the value of the
The stator geometry is then modified to meet some reference flux density values. To do so,
total machine losses.
an iterative optimization process is used to obtain the slot dimensions and the stator outer diameter.
The stator geometry is then modified to meet some reference flux density values. To do so, an
The tooth width, the slot height and the yoke height are automatically modified and the machine
iterative optimization process is used to obtain the slot dimensions and the stator outer diameter. The
tooth width, the slot height and the yoke height are automatically modified and the machine
Energies 2018, 11, 128 5 of 21

Energies 2018, 11, 128 5 of 21

performance characteristics are calculated in every iteration with the proposed calculation procedure.
performance
The final characteristics
stator geometry is theareone
calculated in every iteration
that minimizes with the proposed
the difference between calculation
target andprocedure.
calculated flux
densities. Additionally, the stator phase current is calculated every iteration for the new flux
The final stator geometry is the one that minimizes the difference between target and calculated slot area,
densities. Additionally, the stator phase current is calculated every iteration for the new slot area, the
the current density and the number of conductors. The number of iterations for a given specification
current density and the number of conductors. The number of iterations for a given specification
should be between 20 and 25.
should be between 20 and 25.
As anAs example
an example of ofthe
theresults
results from
from this procedure,the
this procedure, theresults
results obtained
obtained for three
for three different
different
configurations of machine and the same specifications are shown in Figure 4
configurations of machine and the same specifications are shown in Figure 4 and Table 2. and Table 2.

(a) (b) (c)

Figure 4. Sample machine geometries (a) ISDW (p = 4, Qs = 24); (b) FSDW (p = 4, Qs = 36); (c) FSCW
Figure 4. Sample machine geometries (a) ISDW (p = 4, Qs = 24); (b) FSDW (p = 4, Qs = 36); (c) FSCW
(p = 4, Qs = 12, nl,w,s = 2).
(p = 4, Qs = 12, nl,w,s = 2).
Table 2. Sample machine characteristics.
Table 2. Sample machine characteristics.
Parameter Type Parameter Machine 1 Machine 2 Machine 3
Parameter Type Winding
Parameter type ISDW
Machine 1 FSDW
Machine 2 FSCW 3
Machine
Machine topology Pole number 4 4 4
Winding type ISDW FSDW FSCW
Slot number 24 36 12
Machine topology Pole number 4 4 4
Slot depth 23.8 mm 24.2 mm 25.4 mm
Slot number 24 36 12
Tooth width, parallel tooth 19.4 mm 13.4 mm 33.9 mm
Slotopening
Slot depth 23.8
3 mmmm 24.2
3 mmmm 25.4 mm
22.5 mm
Tooth width, parallel tooth 19.4 mm 13.4 mm 33.9 mm
Geometry Copper slot fill factor 0.45 0.45 0.59
Slot opening 3 mm 3 mm 22.5 mm
Geometry Rotor barrier number 4 4 4
Copper slot fill factor 0.45 0.45 0.59
Barrier geometry
Rotor barrier number α d = αq = βint = βext = 0.2; γα = γβ = 0.5
4 4 4
Rotor barrier
Barrier bridges
geometry αd0.8
= αmm 0.8 mm 0.8 mm
q = βint = βext = 0.2; γα = γβ = 0.5
Stator
Rotorteeth fluxbridges
barrier density 0.81.52
mm T 1.55
0.8 mmT 1.6mm
0.8 T
Stator yoke flux density 1.56 T 1.45 T 1.64 T
Stator teeth flux density 1.52 T 1.55 T 1.6 T
Maximum
Statorrotor
yokebarrier flux density
flux density 2.31TT
1.56 2.12 T
1.45 T 2.19
1.64 TT
Machine performance MaximumCurrent density
rotor barrier flux density 82.31
A/mm T
2 8 2.12
A/mm T
2 8 A/mm
2.19 T
2

Machine performance L
Current density
d 53.6
8 A/mm mH 2 44.9
8 A/mm mH 2 98.5
8 A/mm2mH
LLdq 16.8mH
53.6 mH 13.2 mH
44.9 39.7
98.5 mHmH
SaliencyLqratio (ξ) 16.83.19
mH 13.2 mH
3.39 39.7
2.48mH
Saliency ratio (ξ) 3.19 3.39 2.48
3. Validation
3. Validation
A SynRM is designed, manufactured and tested in order to validate the procedure presented in
Section 2. The specifications of the designed machine are based on a known PMSM for a hoisting
A SynRM is designed, manufactured and tested in order to validate the procedure presented in
application, shown in Table 3.
Section 2. The specifications of the designed machine are based on a known PMSM for a hoisting
application, shown in Table 3. Table 3. Design specifications.

TableParameter Value
3. Design specifications.
Rated torque (Nm) 200
Maximum voltage
Parameter (V) 400
Value
Maximum current (A) 20
Rated torque (Nm) 200
Rated speed (rpm) 160
Maximum voltage (V) 400
Maximum current (A) 20
Rated speed (rpm) 160
Energies 2018, 11, 128 6 of 21

The main characteristics of the designed machine are listed in Table 4.

Energies 2018, 11, 128 6 of 21
Table 4. Prototype SynRM characteristics.
The 2018,
Energies main11,characteristics
128 of the designed machine are listed in Table 4. 6 of 21

The main characteristicsTable Parameter

of the4.designed Value
Prototypemachine are listed
SynRM characteristics. in Table 4.
Electrical characteristic
Table 4.Parameter Value
Prototype SynRM characteristics.
Rated torque
Electrical (Nm)
characteristic 200
RatedParameter
voltage (V) 335
Value
Rated torque (Nm) 200
Rated current
Electrical (A)
characteristic 14.6
Rated
Rated voltage(rpm)
speed (V) 335160
Ratedcurrent
Rated torque(A) (Nm) 14.6200
RatedGeometric
Rated
speed voltage characteristics
(rpm) (V) 160335
Ratedofcurrent
Number
Geometric pole (A)
pairs
characteristics 14.6
7
Rated
Number of
Number speed
of pole (rpm)
statorpairs
slots 7 160
42
Stack
Number length
Geometric
of stator (mm)
characteristics
slots 42150
Stator outer
Number diameter
of
Stack length (mm) pole (mm)
pairs 150300
7
Stator inner diameter
Number of stator (mm)
slots 215
42
Stator outer diameter (mm) 300
AirStack
gap length
length(mm) (mm) 0.8
Stator inner diameter (mm) 215150
Stator
Air gapouter diameter
length (mm) (mm) 0.8300
Stator inner
For the manufactured prototype, the geometry of the diameter (mm)rotor215 is optimized based on a multi objective
For the manufactured prototype, Airthe
gapgeometry
length (mm) of the rotor 0.8is optimized based on a multi
differential evolution (MODE) approach [20–24] in order to maximize the mean torque and reduce
objective differential evolution (MODE) approach [20–24] in order to maximize the mean torque and
the torque ripple, the resulting in non-equal barriers, as cantheberotor
observed in Figure 5,onwhere the rotor
reduce For manufactured
the torque prototype,
ripple, resulting the geometry
in non-equal barriers,of is optimized
as can be observed based
in Figure a the
5, where multi
lamination
rotor and differential
objective the whole
lamination rotor
wholeof
and theevolution theofprototype
(MODE)
rotor are[20–24]
theapproach
prototype shown.
are in order to maximize the mean torque and
shown.
reduce the torque ripple, resulting in non-equal barriers, as can be observed in Figure 5, where the
rotor lamination and the whole rotor of the prototype are shown.

(a) (b)
Figure 5. (a) Prototype rotor lamination; (b) Prototype rotor.
(a) 5. (a) Prototype rotor lamination; (b) Prototype
Figure (b)rotor.
Figureof
An experimental testing 5. (a)
thePrototype
prototyperotor lamination;
machine (b) Prototype
is carried rotor. the setup shown in
out, with
Figure
An 6. The maintesting
experimental performance
of thecharacteristics obtainedis
prototype machine bycarried
the developed algebraic–magnetostatic
out, with the setup shown in Figure 6.
An experimental
FEM methods testing
areof
and test characteristics
results the prototype
compared machine
in the Table 5. isdeveloped
carried out, with the setup shown in
The main performance obtained by the algebraic–magnetostatic FEM
Figure 6. The main performance characteristics obtained by the developed algebraic–magnetostatic
methods andmethods
FEM test results areresults
and test compared in the Table
are compared 5. 5.
in the Table

Figure 6. Experimental setup for a prototype SynRM.

Energies 2018, 11, 128 7 of 21

Table 5. Prototype results comparison.

Parameter Matlab-FEMM Experimental Test Difference (%)

Electromagnetic torque (Nm) 200 200 0.0
RMS phase current (A) 14.1 13.6 3.7
RMS line voltage (V) 347 344 0.9
Apparent power (kVA) 8.48 8.12 4.4
Input power (kW) 4.82 4.46 8.1
Output power (kW) 3.35 3.35 0.0
Power factor 0.55 0.55 0.0
Efficiency 0.70 0.75 6.7

The difference between both calculation methods is lower than 10% in all of the cases, which proves
that the results given by the developed calculation procedure are accurate.

4. Machine Specifications and Limits of the Study

The objective of the performed analysis is to find some pattern that relates the main performance
characteristics with the pole number, stator slot number and the winding type. For this purpose,
machine designs with different combinations are studied for the same specification.
The specifications and restrictions for which the machines are designed are based on the PMSM
for a hoisting application mentioned in Section 3. These are shown in Table 6.

Table 6. Design specifications and restrictions.

Parameter Value
Electrical Specifications
Rated speed (rpm) 160
Maximum voltage (V) 400
Maximum current (A) 20
Geometrical Restrictions
Stack length (mm) 125
Stator outer diameter (mm) 300
Stator inner diameter (mm) 215
Air-gap (mm) 1
Rotor inner diameter (mm) 85
Winding Restrictions
Fill factor, distributed winding 0.45
Copper slot fill factor, single layer concentrated winding 0.65
Copper slot fill factor, double layer concentrated winding 0.6
Current Density and Reference Flux-Densities
Current density (A/mm2 ) 8
Stator yoke peak flux density (T) 1.6
Stator teeth peak flux density (T) 1.6
Air-gap peak flux density (T) 0.85

The limits for the studied parameters are the following: pole pair numbers from 4 to 10 have
been considered, because high pole number machines are typically used for low speed applications;
the maximum stator slot number value has been set to 72, in order to avoid too short slot spans;
different short pitching and winding layer values have been considered. Only winding layouts with
winding factor values above 0.85 have been considered.
Energies 2018, 11, 128 8 of 21
Energies 2018, 11, 128 8 of 21

For the sake of focusing on machine parameters related to the winding design, no rotor
For the sake of focusing on machine parameters related to the winding design, no rotor
optimization as for the prototype shown in Figure 6 is carried out for the machines designed and
optimization as for the prototype shown in Figure 6 is carried out for the machines designed and
analyzed in this work, the rotor geometry being as defined by the values in Table 1. Rotor geometry
analyzed in this work, the rotor geometry being as defined by the values in Table 1. Rotor geometry
optimization via a MODE or similar approach in order to maximize the mean torque and reduce the
optimization via a MODE or similar approach in order to maximize the mean torque and reduce the
torque ripple or improve other machine performance metrics should be a necessary subsequent step
torque ripple or improve other machine performance metrics should be a necessary subsequent step
after the analysis presented in this work.
after the analysis presented in this work.
Following the design procedure of the Section 2, 70 machines with all the possible combinations
Following the design procedure of the Section 2, 70 machines with all the possible combinations
of the studied parameters have been designed.
of the studied parameters have been designed.

5. 5. Results
Results Analysis
Analysis and
and Design
Design Criteria
Criteria
InIn this
this section,
section, thethe results
results ofof the
the 7070 machine
machine designs
designs with
with the
the pre-defined
pre-defined rotor
rotor geometry
geometry areare
shown
shown and
and analyzed.
analyzed. The The main
main machine
machine performance
performance characteristics
characteristics consist
consist ofof mean
mean torque
torque (M(M em),
em ),
torque ripple (M em,ripple), power factor (PF) and electromagnetic efficiency (ηe). Then, the causes of the
torque ripple (Mem,ripple ), power factor (PF) and electromagnetic efficiency (η e ). Then, the causes of the
obtained
obtained results
results areare explained
explained through
through studies
studies onon other
other parameters
parameters such
such asas
thethe saliency
saliency ratio
ratio (ξ)(ξ)
andand
the air-gap flux-density waveform (B agap) among others.
the air-gap flux-density waveform (B ) among others. agap

5.1.
5.1. Mean
Mean Torque
Torque
Figure
Figure 7 shows
7 shows thethe dependency
dependency of of mean
mean torque
torque (M(M em) on the pole pair number and the winding
em ) on the pole pair number and the winding
type.Each
type. Eachdot
dotininthe
thegraph
graphrepresents
represents the
the result
result ofof a single
a single machine.Different
machine. Differentwinding
winding types
types areare
plotted with different colors. Different dots with the same pole number and winding
plotted with different colors. Different dots with the same pole number and winding type correspond type correspond
toto the
the results
results ofof machines
machines with
with different
different stator
stator slot
slot number,
number, winding
winding layer
layer or or short
short pitching
pitching values.
values.

Figure Mean
7. 7.
Figure electromagnetic
Mean torque
electromagnetic (M(M
torque em )emversus pole
) versus pair
pole number
pair (p)(p)
number and winding
and type.
winding type.

AsAs cancanbebe seen

seen inin Figure
Figure 7, 7,
thethe first
first conclusion
conclusion is is that,
that, with
with the
the same
same volume,
volume, none
none ofof
thethe SynRM
SynRM
designs meet the mean torque
designs meet the mean torque specification. specification.
There
There is is
anan influence
influence ofof pole
pole number
number ononthethe mean
mean torque:
torque: thethe lower
lower the
the polenumber,
pole number, the
the greater
greater
the mean torque. Upon analyzing the influence of the winding
the mean torque. Upon analyzing the influence of the winding type on the mean torque, it is foundtype on the mean torque, it is found
that
that thethe machines
machines with
with thethe highest
highest mean
mean torque
torque values
values forfor each
each pole
pole number
number are
are obtained
obtained byby FSCW
FSCW
machines.
machines. ForForthetherestrest
of the ofFSCW
the FSCWmachines,machines,
the mean thetorque
mean values
torqueare values are comparable
comparable to those obtainedto those
obtained
with FSDWwith and ISDWFSDWmachines.
and ISDW machines.
If If the
the machines
machines with
with FSCW
FSCW areare further
further analyzed,
analyzed, it it
cancan
bebe observed
observed that
that thethe combinations
combinations that
that
lead to the highest mean torque values correspond to double layer slot
lead to the highest mean torque values correspond to double layer slot windings (nl,w,s = 2) with 3 slotswindings (n l,w,s = 2) with 3 slots
per
per pole
pole pair
pair (qs(q=s =0.5),
0.5),Figure
Figure8a.8a.TheThedots
dotsrepresent
represent the the mean
mean torque
torque results
results ofof
allall
the theFSCWFSCW machines.
machines.
The
The obtained
obtained results
results forfor different
different slotslot layer
layer numbers
numbers areare plotted
plotted with
with different
different colors.
colors.
Among the machines with distributed winding (i.e.,
Among the machines with distributed winding (i.e., ISDW and FSDW), the highest ISDW and FSDW), the highest mean
mean torque
torque
values
values forfor
thethe same
same pole
pole number
number areare obtained
obtained forfor
one one slot
slot perper pole
pole and and phase
phase (qs(q=s 1),
= 1), Figure
Figure 8b.8b.
Energies 2018, 11, 128 9 of 21
Energies 2018, 11, 128 9 of 21

Energies 2018, 11, 128 9 of 21

Energies 2018, 11, 128 9 of 21

(a) (b)
Figure 8. Mean electromagnetic torque (Mem) versus number of stator slots per pole and phase (qs) (a)
(a)
Figure 8. Mean electromagnetic torque (Mem ) versus number of stator slots(b)per pole and phase (qs )
and slot layer number (nl,w,s) for FSCW machines; (b) and pole pair number (p) for ISDW and FSDW
(a) andFigure
slot layer number(a) (nl,w,s ) for FSCW machines; (b) and pole pair (b) number (p) for ISDW and
machines.8. Mean electromagnetic torque (Mem) versus number of stator slots per pole and phase (qs) (a)
FSDWand machines.
slot layer number (nl,w,s) for FSCW machines; (b) and pole pair number (p) for ISDW and FSDW
Figure 8. Mean electromagnetic torque (Mem) versus number of stator slots per pole and phase (qs) (a)
Although
machines.
and thenumber
slot layer torque(ncapability is related to the difference between direct and quadrature
l,w,s) for FSCW machines; (b) and pole pair number (p) for ISDW and FSDW
inductances
Although
machines. (Ld −torque
the Lq) [1,2,8], the correlation
capability is related betweento thethe difference
studied variable (Ld − Ldirect
between q) and the andobjective
quadrature
Although
variable (Mem )Lforthethetorque
studiedcapability
machines is isrelated
weak to the difference
(Figure between is
9). The correlation direct and quadrature
analyzed using the
inductances (L d − q ) [1,2,8], the correlation between the studied variable (L d − L q ) and the objective
inductances
coefficient of
Although (Ldthe
− Ltorque
q) [1,2,8],or
determination, theR 2 correlation
capabilityvalue,is of the
relatedbetween
leastto thedifference
squares
the studied variable
regression. Linear
between (Ldirect
− Lquadratic
dor q) and the
and objective
functions
quadrature
variable (Mem ) for the studied machines is weak (Figure 9). The correlation is analyzed using the
variable
have been(M
inductances em) for the
considered studied
depending
(Ld − Lq) [1,2,8], the machines
on the data
correlation
2 value,
isbetween
weak shape.
series (Figure R29). Thevariable
ranges
the studied correlation
from 0 to
(L1, is
Lq)analyzed
d −being and0 no using the
correlation
the objective
coefficient of determination,
coefficient of determination, or Ror R ofof
2 value, the theleast
least squares
squares regression.
regression. Linear
Linear or or quadratic
quadratic functions
functions
and 1 perfect
variable (Memcorrelation.
) for the studied machines is weak (Figure 9). 2 The correlation is analyzed using the
have been
haveconsidered
beenof considereddepending
depending on the data
on the data series shape. R 2 ranges from0 0toto1, 1, being 0 correlation
no correlation
coefficient determination, or R2 value, of theseries
least shape.
squaresRregression.
ranges from Linear orbeing 0 no
quadratic functions
and 1have
perfect
and 1beencorrelation.
perfect correlation.
considered depending on the data series shape. R2 ranges from 0 to 1, being 0 no correlation
and 1 perfect correlation.

Figure 9. Mean electromagnetic torque (Mem) versus the difference between direct and quadrature
inductances (Ld − Lq).
9. Mean
FigureFigure electromagnetic
9. Mean electromagnetictorque (M(M
torque ) )versus
emem versus the
the difference betweendirect
difference between direct and
and quadrature
quadrature
Another
inductances
inductances (L −
parameter
(L
Ldq−
).Lq). related to the torque capability is the air-gap flux density (Bagap). Figure 10a,b
Figure 9. dMean electromagnetic torque (Mem) versus the difference between direct and quadrature
shows the relation
inductances (Ld between
− Lq). the first harmonic of the air-gap flux density (Bagap1) and the mean torque
Another
for the parameter
machines related to the
with distributed and torque capabilitywindings,
concentrated is the air-gap flux density
respectively. (Bagap
For the). Figure 10a,b
distributed
Another
shows
parameter
the relation
related the to the torque capability is the flux
air-gap flux(Bdensity (B agap ). Figure 10a,b
winding machines
Another abetween
parameter strongrelated first
to theharmonic
correlation torque of the first
exists:capability
higher air-gap
harmonic
is the density
air-gap air-gap flux) and
agap1
flux density the
density
(Bagap mean
valuestorque
). Figure lead
10a,b
showsforthethe
relation between the first harmonic of the air-gap flux density (Bagap1 )Forandthe thedistributed
mean torque
to
shows themachines
higher mean torque
relation with
between distributed
values.theThe and concentrated
firstcorrelation
harmonic is
ofnot
the clearwindings,
forflux
air-gap respectively.
concentrated
density (Bwinding
agap1) and machines.
the mean torque
for thewinding
machines with distributed
machineswith a strong and concentrated
correlation windings, respectively. For the distributed
valueswinding
for the machines distributed andexists: higher first
concentrated harmonic
windings, air-gap fluxFor
respectively. density
the distributed lead
machines
to a strong
higher mean correlation
torque exists:
values. The higher first
correlation isharmonic
not clear air-gap
for flux
concentrateddensity
winding
winding machines a strong correlation exists: higher first harmonic air-gap flux density values lead values lead
machines. to higher
meanto
torque
highervalues. The correlation
mean torque values. The is not clearisfor
correlation notconcentrated windingwinding
clear for concentrated machines.
machines.

(a) (b)
Figure 10. Mean electromagnetic torque (Mem) versus the first harmonic of the air-gap flux density
(a) (b)
(Bagap1) for (a) ISDW and FSDW machines; (b) FSCW machines.
(a)
Figure 10. Mean electromagnetic torque (Mem) versus the first harmonic of(b)
the air-gap flux density
(B agap1) for (a) ISDW and FSDW machines; (b) FSCW machines.
Figure 10. Mean electromagnetic torque (Mem) versus the first harmonic of the air-gap flux density
Figure 10. Mean electromagnetic torque (Mem ) versus the first harmonic of the air-gap flux density
(Bagap1) for (a) ISDW and FSDW machines; (b) FSCW machines.
(Bagap1 ) for (a) ISDW and FSDW machines; (b) FSCW machines.
Energies 11, 128
2018, 11,
Energies 2018, 128 10
10 of
of 21
21
Energies 2018, 11, 128 10 of 21

Going deeper
Going deeper into
into the
the torque
torque capability
capability study,
study, another
another parameter
parameter related
related toto the
the air-gap
air-gap flux
flux
Going deeper into the torque capability study, another parameter related to the air-gap flux density
density is
density is studied:
studied: the
the magnetomotive
magnetomotive force force (N·
(N· I), N
I), N being
being the
the number
number of of series
series turns
turns per
per phase.
phase.
is studied: the magnetomotive force (N·I), N being the number of series turns per phase. Figure 11a,b
Figure
Figure 11a,b
11a,b shows
shows the
the relation
relation between
between the
the magnetomotive
magnetomotive force
force and
and the
the mean
mean torque
torque for
for the
the
shows the relation between the magnetomotive force and the mean torque for the machines with
machines with with distributed
distributed and
and concentrated
concentrated windings,
windings, respectively.
respectively. The R R2 values
2 values for
for each
each pole
pole pair
pair
machines
distributed and concentrated windings, respectively. The R2 valuesThe for each pole pair data group
data
data group
group are are shown,
shown, from p = 4 to p = 10. Note that data series with one or two machines will always
are shown, from p = 4from
to p =p =10.
4 toNote
p = 10. Note
that that
data datawith
series series with
one or one
twoor two machines
machines will always
will always have
have
2
have RR22 == 11 value.
value.
R = 1 value.

(a) (b)
(b)
(a)
Figure 11.
Figure 11. Mean electromagnetic
Mean electromagnetic torque (Mem)) versus
electromagnetic torque the
versusthe magnetomotive
themagnetomotive force (N
magnetomotive force
force (N· I) for
·I)I) for (a)
(a) ISDW and
ISDW and
and
Figure 11. Mean (Mem
em) versus (N· for (a) ISDW
FSDW machines;
FSDW machines; (b)
(b) FSCW
FSCW machines.
machines.
FSDW machines; (b) FSCW machines.

The relation between

The between the mean mean torque and and the magnetomotive
magnetomotive force is is more direct
direct if the first
first
The relation
relation between the the mean torque
torque and the the magnetomotive forceforce is more
more direct ifif the
the first
harmonic
harmonic winding factor (k ω,s1) is taken into account. Figure 12a,b shows the relation between the
harmonic winding
winding factor
factor (kω,s1 ) )isistaken
ω,s1 takeninto
intoaccount.
account.Figure
Figure12a,b
12a,bshows
shows the
the relation
relation between
between thethe
mean
mean torque and N· I· kω,s1 for the machines with distributed and concentrated windings respectively.
mean torque
torque and N· N·I·I·kkω,s1
ω,s1for
forthe
the machines
machines with
withdistributed
distributedand
and concentrated
concentrated windings
windings respectively.
respectively.
The adjustment values
The values show that that there is is a strong correlation
correlation for the
the distributed winding
winding machines,
The adjustment
adjustment values show show that there there is aa strong
strong correlation for
for the distributed
distributed winding machines,
machines,
but
but not in the case of concentrated winding machines.
but not
not in
in the
the case
case of concentrated winding machines.

(a)
(a) (b)
(b)
Figure 12. Mean electromagnetic torque (Mem) versus the magnetomotive force multiplied by the first
Figure
Figure 12.
12. Mean
Mean electromagnetic
electromagnetic torque
torque (M(Mem ) versus the
themagnetomotive
magnetomotiveforce
force multiplied
multiplied by
by the
the first
harmonic winding factor (N· I·
kω,s1) for (a) ISDWem ) versus
and FSDW machines; (b) FSCW machines.
first
harmonic winding factor (N·
I·
harmonic winding factor (N·I·kk ω,s1) for (a) ISDW and FSDW machines; (b) FSCW machines.
) for (a) ISDW and FSDW machines; (b) FSCW machines.
ω,s1

The relation
The relation between
between the the mean
mean torque
torque and and the the magnetomotive
magnetomotive force force is is defined
defined as as the
the torque
torque
The relation
constant, k T = Mem/(N·between
I). the mean
Figures 11a torque
and 12a and thethat
shows magnetomotive
the torque force is
constant defined
value is as theconstant
almost torque
constant, kT = Mem/(N· I). Figures 11a and 12a shows that the torque constant value is almost constant
constant,
for the kT = Mem
the machines
machines with ·I). Figures 11a
/(Ndistributed and 12a
windings andshows that the
the same
same poletorque
number. constant
Moreover, valuethe is almost
higherconstant
the pole
pole
for with distributed windings and the pole number. Moreover, the higher the
for the
number, machines
the lower withthe distributed
torque windings
constant is. Theand the
Figure same
13a pole
plots number.
the Moreover,
dependency of the
the higher
torque the pole
constant
number, the lower the torque constant is. The Figure 13a plots the dependency of the torque constant
number,
(kττ)) on
on thethestator
the lower slot
the torque
number constant
per pole
poleis. and
The Figure
phase, 13a andplots the
poledependency
the pole pair number
number of the torque
for the constant
the distributed
distributed
(k stator slot number per and phase, and the pair for
(k )
winding
τ on the stator slot
machines. Figure number
Figure 13b per pole
13b plots and
plots the phase,
the dependency and
dependency of the pole
of the pair
the torque number
torque constantfor the
constant divided distributed
divided by winding
by the
the first
first
winding machines.
machines.
harmonic Figure 13b
winding factor plots
(kτ/kthe
ω,s1) dependency
on the same of the torque
parameters. On constant
the one divided
hand, it by clearly
can the firstbe harmonic
seen that
harmonic winding factor (kτ/kω,s1) on the same parameters. On the one hand, it can clearly be seen that
winding
the dependencyfactor
dependency on (k /k
on the
τ the pole
ω,s1 ) on the
pole number same
number is parameters.
is high.
high. OnOn the On
the other the
other hand, one
hand, the hand, it
the statorcan
stator slot clearly
slot number be
number per seen that
per pole
pole andthe
and
the
dependency
phase has has its on the
its influence: pole
influence: the number
the higher
higher theis high.
the stator On
stator slot the other
slot number,
number, the hand,
the lower the
lower thestator slot
the torque number
torque constant. per
constant. Attendingpole
Attending toand
to
phase
phase
Figure has
13a, itsthis
influence:
effect thebe
can higher the stator
explained by slotinfluence
the number, of thethe lower the torque
winding factor. constant. Attending
Nevertheless, to
Figure
Figure 13a, this effect can be explained by the influence of the winding factor. Nevertheless, Figure
Figure
13b shows 13a, that
shows this effect canthe be only
explained
factorby the influence of the winding factor. Nevertheless, Figure 13b
13b that itit is
is not
not the only factor explaining
explaining this trend.
this trend. In any
In any case,
case, aa low
low stator slot
stator slot number
number is
is
shows
recommendedthat it is not
beyond the only
the factor
effect of explaining
the winding this trend.
factor. In any case, a low stator slot number is
recommended beyond the effect of the winding factor.
recommended beyond the effect of the winding factor.
Energies 2018, 11, 128 11 of 21
Energies 2018, 11, 128 11 of 21
Energies 2018, 11, 128 11 of 21

(a) (b)
(a) (b)
Figure 13. 13. (a) Torque
Torque constant
constant (k
(kT);
); (b) Torque
Torque constant
constant divided
divided by
by the
the first
first harmonic
harmonic winding
winding factor
factor
Figure 13.(a)(a)
Figure Torque constant (kT);(b)
T (b) Torque constant divided by the first harmonic winding factor
(k τ/kω,s1) versus the number of stator slots per pole and phase (qs) and pole pair number (p) for ISDW
(kτ(k/k ) versus the number of stator slots per pole and phase (qs )s)and
τ/kω,s1) versus the number of stator slots per pole and phase (q
ω,s1 andpole
polepair
pairnumber
number(p)
(p)for
forISDW
ISDW
and FSDW
and FSDW machines.
machines.
and FSDW machines.

The main conclusions of this subsection are summarized below:

The
Themain
mainconclusions
conclusionsofofthis
thissubsection
subsectionarearesummarized
summarized below:
below:
 Low pole numbers are recommended due to their high torque constant;
• Lowpole
Low polenumbers
numbersareare recommended
recommended due
due toto their
their high
high torque
torque constant;
constant;
 Distributed winding machines have higher torque constant and winding factor than
• Distributed
Distributed winding
winding machines
machines have higher
have higher torque and
torque constant constant
winding and winding
factor factor than
than concentrated
concentrated winding machines. However, concentrated winding machines allow a higher slot
concentrated
winding winding
machines. machines.
However, However, concentrated
concentrated winding machines winding machines
allow allow
a higher slota fill
higher slot
factor,
fill factor, so overall torque capability is similar to distributed winding machines;
fill factor, so overall torque capability is similar to distributed
so overall torque capability is similar to distributed winding machines; winding machines;
 Concentrated winding machines with qs = 0.5 and nl,w,s = 2 have the highest torque capability.
• Concentrated windingmachines
Concentrated
Mean torque iswinding
machineswith
considerably higher
withqsq= s = 0.5and
than the0.5
and l,w,s
nl,w,s==22have
rest of nthe
have the highest torque capability.
machines; the highest torque capability.
Mean
Mean torque
torque is considerably
is considerably higher than the rest of the machines;
 Among distributed winding higher thanlow
machines, the stator
rest ofslot
the number
machines; are recommended for their high
• Amongdistributed
Among distributedwinding
windingmachines,
machines,low lowstator
statorslot
torque constant. The cases with qs = 1 have the highest torque
slotnumber
numberare arerecommended
recommendedfor
capability among
fortheir
theirhigh
high
distributed
torqueconstant.
constant.The The cases with qs = 1 have the highest torque capability among distributed
torque
winding machines. cases with qs = 1 have the highest torque capability among distributed
winding machines.
winding machines.
5.2. Torque Ripple
5.2.Torque
5.2. Torque Ripple
Ripple
Regarding the torque ripple, Figure 14a shows the dependency of torque ripple (Mem,ripple) on the
Regardingthe
Regarding thetorque
torqueripple,
ripple,Figure
Figure14a
14ashows
shows the
the dependency
dependency of torque rippleripple (M(Mem,ripple ) on
em,ripple the
) on
pole pair number and the winding type, and Figure 14b on the stator slot number. It can be seen that
pole
the pair
pole number
pair number and thethe
and winding
windingtype, and
type, andFigure 14b14b
Figure ononthethe
stator slotslot
stator number.
number.It can be be
It can seen that
seen
the pole number has no influence on the torque ripple, while the winding type and the stator slot
thethe
that pole number
pole number has nonoinfluence
has influenceon
onthe
thetorque
torqueripple,
ripple, while
while the winding
winding type
typeand
andthethestator
statorslotslot
number have a great influence. The lower ripple values are obtained with FSDWs, closely followed
number
number have
have a great
a great influence.
influence. TheThe lower
lower ripple
ripple values
values are are obtained
obtained withwith
FSDWs,FSDWs, closely
closely followed
followed by
by ISDWs, while FSCWs obtain high torque ripple values.
by ISDWs,
ISDWs, whilewhile
FSCWsFSCWs
obtainobtain
high high
torquetorque
rippleripple values.
values.

(a) (b)
(a) (b)
Figure 14. Torque ripple (Mem,ripple) versus (a) pole pair number (p) and winding type; (b) stator slot
Figure 14. Torque
Torque ripple
ripple (M(M em,ripple) versus (a) pole pair number (p) and winding type; (b) stator slot
Figure
number14.
per pole and phase (qs). ) versus (a) pole pair number (p) and winding type; (b) stator slot
em,ripple
number
number per
per pole
pole and
and phase
phase (qs(q).s).
ISDW machines deserve further study of the torque ripple. Figure 15a shows the torque ripple
ISDW machines deserve further study of the torque ripple. Figure 15a shows the torque ripple
results for machines
ISDW deservewith
ISDW machines further study of
different the torque
stator ripple. Figure
slot numbers 15a and
per pole shows the torque
phase ripple
and short results
pitching
results for ISDW machines with different stator slot numbers per pole and phase and short pitching
for ISDW
values. machines
It can be seenwith
thatdifferent statorripple
lower torque slot numbers
values areperobtained
pole andwhen
phasestator
and short pitchingper
slot number values.
pole
values. It can be seen that lower torque ripple values are obtained when stator slot number per pole
It
andcanper
be phase
seen that lower torque
is greater ripple
than unity (qs values
> 1). In are obtained
these when
cases, the stator
torque slot number
ripple values arepercomparable
pole and perto
and per phase is greater than unity (qs > 1). In these cases, the torque ripple values are comparable to
phase is greater than unity
FSDW machines’ values. Moreover, (qs > 1). In these cases, the torque ripple values are comparable to
the short pitching has almost no influence on the torque ripple. FSDW
FSDW machines’ values. Moreover, the short pitching has almost no influence on the torque ripple.
machines’ values. Moreover, the short pitching has almost no influence on the torque ripple.
Energies 2018, 11, 128 12 of 21
Energies 2018, 11, 128 12 of 21

Energies 2018, 11, 128 12 of 21

Energies 2018, 11, 128 12 of 21

(a) (b)
Figure 15. Torque ripple (Mem,ripple) versus stator slot number per pole and phase (qs) (a) and short
Figure 15. Torque ripple (Mem,ripple ) versus stator slot number per pole and phase (qs ) (a) and short
pitching (τa) for ISDW (a)machines; (b) and slot layer number (nl,w,s) for FSCW (b) machines.
pitching (τ a ) for ISDW machines; (b) and slot layer number (nl,w,s ) for FSCW machines.
(a) (b)
Figure 15. Torque ripple (M em,ripple) versus stator slot number per pole and phase (qs) (a) and short
Returning to the FSCW machines, Figure 15b shows the torque ripple for all the concentrated
Figure
pitching 15. aTorque ripple (Mem,ripple(b)
) versusslot stator slot number per pole andmachines.
phase (qs) (a) and short
winding
Returning to(τthe
) for ISDW
machines depending
FSCW machines;on theand
machines, stator slot
Figure layernumber
15b number
showsper (nl,w,s
thepole
pitching (τa) for ISDW machines; (b) and slot layer number (nl,w,s) for FSCW machines.
) for
torque FSCW
and ripple
phase andfor slot
all thelayer number.
concentrated
Previously,
winding machines Figure 8 has shown that the ones with double
depending on the stator slot number per pole and phase and slot layer layer windings and q s = 0.5 arenumber.
the
Returning to the FSCW machines, Figure 15b shows the torque ripple for all the concentrated
combinations
Previously, Figure with
8 has highest
shown mean
that torque values. In Figure 15b, it can be seen that they
qconcentrated are the
Returning
winding machines to the FSCW
depending on thethe
machines, stator ones
Figure with
slot15b double
shows
number thepole
per layer
torque windings
andripple
phasefor andalland
thelayer
slot s =number.
0.5 are
simultaneously
winding machines the ones
depending with the highest
on torque
the torque ripple values.
combinations
Previously, with
Figure highest
8 has mean shown thatstator slot number
thevalues.
ones per pole
withIndouble
Figure layer andwindings
15b, phase
it canand beslot
and qs layer
seen = 0.5 number.
thatarethey
the are
If the causes
Previously, Figure of8 the
has torque
shown ripple
that are
the studied,
ones the double
with origin oflayer the torque
windings ripple
andin SynRMs
q = 0.5 is inthe
are the
simultaneously
combinations thewith
ones highest
with themean highest torque torque ripple
values. In values.
Figure 15b, it can be seen that they are s
content of harmonics of the air-gap flux-density waveform (B agap), which are mainly caused by the
Ifcombinations
simultaneously
the causes of
relative position
with
the
the highest
ones withmean
torque
between
the highest
ripple
stator
torque
and rotor
values.
torque
are studied, In values.
ripple
the
slotsripple
Figure
origin of
and barriers,
15b, theit torque
can be ripple
seen that they areis in
in SynRMs
and the stator magnetomotive force
simultaneously
If the the ones
causes of theoftorquewith the highest
ripple are torque values.
studied, thewaveform
origin of the
the content of harmonics
waveform. Air-gap the air-gap
flux-density flux-density
waveforms andthetheir harmonic (Btorque
spectrum
ripple in
agap ), which are
for in
SynRMs
threemainly iscaused
machines
in the by
are
If the
content of causes
harmonics of the of torque
the ripple
air-gap are studied,
flux-density origin
waveform of
(B the torque ripple SynRMs
agap), which are mainly caused by the
is in the
the relative
shown position
below: between
the machine stator
in and rotor
Figure 16 is slots
a and barriers,
four-pole, 24-stator and
slot the
ISDW stator magnetomotive
machine, with a torqueforce
content
relative of harmonics
position betweenof the air-gap
stator andflux-density
rotor slots and waveform
barriers,(Bagapand ), which are mainly
the stator caused byforce
magnetomotive the
waveform.
ripple
relative
Air-gap
of 11.6%;
position
flux-density
the one
between in waveforms
Figure
stator and 17 is a and
four-pole,
rotor slots
their harmonic
36-stator slot FSDW spectrum
machine, for three
with a machines
torque ripple are
waveform. Air-gap flux-density waveforms andandtheirbarriers,
harmonic andspectrum
the statorfor magnetomotive
three machinesforce are
shown of 2.9%;
below: the
waveform. the other one,
machine
Air-gap in Figure
in Figure
flux-density 18, is a four-pole,
16 is a four-pole,
waveforms 12-stator
and their 24-statorslot,
harmonic double layer FSCW
slot ISDWformachine,
spectrum machine,
three with with
with aare
machines a
torque
shown below: the machine in Figure 16 is a four-pole, 24-stator slot ISDW machine, a torque
torque
rippleshown
of 11.6%;ripple
below:the of
the47.2%.
one in
machineFigure in 17
Figureis a four-pole,
16 is a 36-stator
four-pole, slot
24-stator FSDW
slot ISDWmachine,
machine,with witha torque
a torque ripple
ripple of 11.6%; the one in Figure 17 is a four-pole, 36-stator slot FSDW machine, with a torque ripple
of 2.9%; the
ripple other
of 11.6%; one,
the in
oneFigure
in 18,
Figure is
17 a
is four-pole,
a four-pole, 12-stator
36-stator slot,
slot double
FSDW
of 2.9%; the other one, in Figure 18, is a four-pole, 12-stator slot, double layer FSCW machine, with a layer
machine, FSCW
with a machine,
torque with
ripple a
of 2.9%;
torquetorque the
rippleripple other
of 47.2%. one,
of 47.2%. in Figure 18, is a four-pole, 12-stator slot, double layer FSCW machine, with a
torque ripple of 47.2%.

(a) (b)
Figure 16. Air-gap flux density (Bagap) (a) waveform; (b) harmonic spectrum; for the p = 4, Qs = 24
machine. (a) (b)
(a) (b)
Figure 16. Air-gap flux density (Bagap) (a) waveform; (b) harmonic spectrum; for the p = 4, Qs = 24
FigureFigure Air-gap
16. 16. flux density (Bagap ) (a) waveform; (b) harmonic spectrum; for the p = 4,
machine. Air-gap flux density (Bagap) (a) waveform; (b) harmonic spectrum; for the p = 4, Qs = 24
Qs = 24 machine.
machine.

(a) (b)

Figure 17. Air-gap flux density (Bagap) (a) waveform; (b) harmonic spectrum; for the p = 4, Qs = 36
machine. (a) (b)
(a) (b)
Figure 17. Air-gap flux density (Bagap) (a) waveform; (b) harmonic spectrum; for the p = 4, Qs = 36
Figure 17. Air-gap flux density (Bagap) (a) waveform; (b) harmonic spectrum; for the p = 4, Qs = 36
Figuremachine.
17. Air-gap flux density (Bagap ) (a) waveform; (b) harmonic spectrum; for the p = 4,
machine.
Qs = 36 machine.
Energies 2018, 11, 128 13 of 21
Energies 2018, 11, 128 13 of 21

(a) (b)
Figure 18. Air-gap flux density (Bagap) (a) waveform; (b) harmonic spectrum; for the p = 4, Qs = 12 and
Figure 18. Air-gap flux density (Bagap ) (a) waveform; (b) harmonic spectrum; for the p = 4, Qs = 12 and
nl,w,s = 2 machine.
nl,w,s = 2 machine.
As can be observed in Figures 16–18, the ISDW and FSDW machines have more sinusoidal
shaped
As can bewaveforms
observedcompared
in Figures to16–18,
the FSCW
the machine’s
ISDW andwaveform.
FSDW machinesMoreover, the more
have harmonic content shaped
sinusoidal for
the FSCW
waveforms machine to
compared air-gap flux-density
the FSCW waveform
machine’s is proven to
waveform. be greater the
Moreover, thanharmonic
for the other machines.
content for the
Considering that all the machines with the same pole number have
FSCW machine air-gap flux-density waveform is proven to be greater than for the other machines. exactly the same rotor
geometry, it can be stated that the differences between the machines is caused by the waveform of
Considering that all the machines with the same pole number have exactly the same rotor
magnetomotive force created in the stator.
geometry, it can be stated that the differences between the machines is caused by the waveform of
The magnetomotive force created by FSDW machines is very sinusoidal, and have high order and
magnetomotive
low amplitudeforce created Thus,
harmonics. in thethe stator.
overall harmonic content is very low, and so is the torque ripple.
The magnetomotive force created
ISDW machines with one stator by slot
FSDW permachines
pole pair is very
and sinusoidal,
phase and
have fifth have
and high order
seventh order and
low amplitude harmonics. Thus, the overall harmonic content is very low, and
harmonics with a considerable amplitude, and lower amplitude 11–13 and 17–19 order harmonics. so is the torque ripple.
ISDW machines
When the number with one
of stator slotstator slot
per pole per
pair andpole
phasepair and phase
is higher, the 5–7have fifth are
harmonics andconsiderably
seventh order
reduced.
harmonics Thea harmonic
with considerablecontent for ISDW machines
amplitude, and lower with stator slot11–13
amplitude number higher
and 17–19than the unity
order is
harmonics.
low enough to ensure low torque ripple values.
When the number of stator slot per pole pair and phase is higher, the 5–7 harmonics are considerably
In the
reduced. The case of FSCW
harmonic content machines,
for ISDW low machines
order and high
withamplitude
stator slotmagnetomotive
number higher force harmonics
than the unity is
are produced. Moreover, the open slots of these machines lead to a negative effect on the torque
low enough to ensure low torque ripple values.
ripple. In the case of machines with three stator slots per pole (qs = 0.5), 2–4 order harmonics are
In the case of FSCW machines, low order and high amplitude magnetomotive force harmonics
produced. The great amplitude of the second harmonic produces a very detrimental effect on the
are produced. Moreover, the open slots of these machines lead to a negative effect on the torque ripple.
torque ripple.
In the caseThe
of machines with three
main conclusions of thisstator slots are
subsection persummarized
pole (qs = 0.5), 2–4 order harmonics are produced.
below:
The great amplitude of the second harmonic produces a very detrimental effect on the torque ripple.
 The pole number has no effect on the torque ripple;
The main conclusions of this subsection are summarized below:
 The winding type and the stator slot number per pole and phase have a great influence on the
torque ripple, due to their effect on the magnetomotive force waveform and its harmonic content
• The pole number has no effect on the torque ripple;
(Figures 16–18).
• The winding type and the stator slot number per pole and phase have a great influence on the
o ripple,
torque FSDW due
machines obtain
to their very
effect onlow
thetorque ripple results,
magnetomotive duewaveform
force to their lowand
magnetomotive
its harmonicforce
content
harmonic content;
(Figures 16–18).
o ISDW machines obtain low torque ripple results, due to their relatively low magnetomotive
force harmonic content. When the number of stator slots per pole and phase is greater than
# FSDW machines obtain very low torque ripple results, due to their low magnetomotive force
unity (qs > 1), the torque ripple is especially low, because the 5–7 order harmonics are
harmonic
removed;content;
# oISDWFSCWmachines
machinesobtain low
obtain hightorque
torqueripple
rippleresults, due to their
results, especially relatively
when lowslots
three stator magnetomotive
per pole
force
areharmonic content.
used (qs = 0.5), due to When themagnetomotive
their high number of statorforce slots per content
harmonic pole and andphase is greater
slot opening
effect.
than unity (qs > 1), the torque ripple is especially low, because the 5–7 order harmonics
are removed;
5.3. Power Factor
# FSCW machines obtain high torque ripple results, especially when three stator slots per
Concerning
pole are usedthe(qpower factor
s = 0.5), due(PF), Figurehigh
to their 19 shows its dependency
magnetomotive on the
force pole paircontent
harmonic number and
and slot
the winding
openingtype. It can be seen that both pole number and winding type have an evident influence.
effect.
On the one hand, the influence of the pole number shows that better power factor results are obtained
5.3. Power Factor
Concerning the power factor (PF), Figure 19 shows its dependency on the pole pair number and
the winding type. It can be seen that both pole number and winding type have an evident influence.
On the one hand, the influence of the pole number shows that better power factor results are obtained
Energies 2018, 11, 128 14 of 21
Energies 2018, 11, 128 14 of 21

Energies 2018, 11, 128 14 of 21

with lower
with lower pole
Energies pole128numbers.
2018, numbers.
11, On the
On the other
other hand,
hand, ISDWand
ISDW andFSDW
FSDW machines
machines obtain
obtainbetter power
better powerfactor
14 offactor
21
resultsresults
with lower
than than
pole
FSCW FSCW machines.
numbers.
machines. On the other hand, ISDW and FSDW machines obtain better power factor
with lower
results pole numbers.
than FSCW machines.On the other hand, ISDW and FSDW machines obtain better power factor
results than FSCW machines.

Figure 19. Power

Figure factor
19. Power (PF)
factor versus
(PF) pole
versus polepair
pairnumber
number (p) and
andwinding
windingtype.
type.
Figure 19. Power factor (PF) versus pole pair number (p) and winding type.
Figure 20 isFigure
presented for afactor
19. Power deeper
(PF)study
versusof thepair
pole FSCW machines.
number The power
(p) and winding type.factor for FSCW
Figure 20 is presented for a deeper study of the FSCW machines. The power factor for FSCW
machines for different stator slot number per pole and phase,
Figure 20 is presented for a deeper study of the FSCW machines. The power factorand pole pair number is plotted.
for FSCWAgain,
machinesit for
can bedifferent
Figureseen20 is
that stator
presented
the slotfor
number
combinationsa deeperperstudy
that poletoofand
lead the
the phase,
FSCW
highest and pole
machines.
power pair
factorThe number
power
values isthe
factor
(for plotted.
for
same FSCW Again,
pole
machines for different stator slot number per pole and phase, and pole pair number is plotted. Again,
it can be seen that the combinations that lead to the highest power factor values (for the same
= 0.5).pole
it machines
number)
can be seen forthat
different
correspond tostator
doubleslotlayer
the combinations number per
slotlead
that pole
windings
to theand
(n phase,
highest
l,w,s and pole
= 2)power
with three
factorpair number
stator slots
values isthe
(forperplotted.
pole
same (qsAgain,
pole
number) it
number)correspond
Incan becorrespond
these seen that
cases, to the
the double
power layer
combinations
factor
to double slot
values
layer windings
that
slot lead
are to (n
slightly
windings thelower
(nl,w,s
l,w,s ==2)2)
highest with
power
than
with thosethree
three factor
of stator
ISDW
stator slots
values
and
slots (for
per per
FSDW thepole
pole same (qpole
machines.
(q s = 0.5).
s = 0.5).
In theseThecases,
number)
rest the
of thepower
correspond factor
to double
concentrated values
layer
winding are
slot slightly
windings
machines lower
(n
obtain = than
2) with
appreciably
l,w,s those
three
In these cases, the power factor values are slightly lower than those of ISDW and FSDW machines.lowerof ISDW
stator
power and
slots per
factor FSDW
pole (q
values. smachines.
= 0.5).
The rest
TheIn of
these
rest ofcases,
the the powerwinding
concentrated
the concentrated factor
winding values are slightly
machines
machines obtainlower
obtain than those
appreciably
appreciably of power
lower
lower ISDW
powerandfactor
factorFSDW machines.
values.
values.
The rest of the concentrated winding machines obtain appreciably lower power factor values.

Figure 20. Power factor (PF) versus stator slot number per pole and phase (qs) for FSCW machines.
Figure 20. Power
Figure 20. Power factor (PF)
factor (PF)versus
versusstator
statorslot
slot number
number per perpole
poleandand phase
phase (qs)(q s ) FSCW
for for FSCW machines.
machines.
IfFigure
the causes
20. Powerbehindfactorthe
(PF)power
versusfactor
stator results are going
slot number to and
per pole be studied,
phase (qs)itfor is FSCW
necessary to analyze
machines.
theIfsaliency
the causes ratio (ξ). The
behind relation
the power of the
factor power
results are factor
going with
to be the saliency
studied, ratio has been
it is necessary clearly
to analyze
If demonstrated
the causes behind
If the causes the
behind power factor results arearegoing to bethat
studied, it itisisnecessary to analyze
analyze the
the saliency ratio in (ξ).
the The the
literature power offactor
[1,8,15,25,26].
relation results
the power Figure 21going
factor withtothe
shows be studied, rationecessary
the correlation
saliency has thetomachines
forbeen clearly
saliencythe ratio (ξ). in
saliency The
ratiorelation ofstrong,
the power factor with thehigher
saliency ratio has been clearly demonstrated
considered
demonstrated in the (ξ).
this work Theis
literature relation of the
and
[1,8,15,25,26]. power
confirms
Figure factor
that
21 shows with the
thesaliency
saliency
that ratios
correlation ratioforhas
provide been
higher
the clearly
power
machines
in theconsidered
literature
demonstrated
factor results [1,8,15,25,26].
in the
independently
in this work Figure
literature
is strong, 21
theand shows
of [1,8,15,25,26].
winding
confirms that
Figure
type that the
and21 correlation
shows
pole
higher that the
number.
saliency for the machines
correlation
ratios providefor considered
the machines
higher power in
this work
factor is strong,
considered
results in thisand work
independently confirms
is of thethat
strong, andhigher
winding typesaliency
confirms and that
pole ratios
higher
number. provideratios
saliency higher power
provide factor
higher results
power
factor results
independently of theindependently
winding type of the
andwinding type and pole number.
pole number.

Figure 21. Power factor (PF) versus saliency ratio (ξ).

Figure 21. Power factor (PF) versus saliency ratio (ξ).
Figure Power
21. 21.
Figure factor
Power (PF)
factor (PF)versus
versussaliency ratio(ξ).
saliency ratio (ξ).
Energies 2018, 11, 128 15 of 21
Energies 2018, 11, 128 15 of 21
Energies 2018, 11, 128 15 of 21
In addition,
In addition,thethe relationship
relationship ofpole
of the thenumber
pole number
with the with
saliencytheratio
saliency
has alsoratio
beenhas also been
demonstrated
by In addition,
demonstrated
many authors the
by in
many relationship of
authors in[3,25].
the literature the pole
the literature
Figure 22anumber
[3,25].
shows with
Figure the saliency
22a shows of
the dependency ratio
thethe has
dependencyalso ofbeen
saliency ratio the
on
demonstrated
saliency ratio by
on many
the poleauthors
pair in
numberthe literature
and the [3,25].
winding Figure
type. It 22a
can shows
be seen the
that
the pole pair number and the winding type. It can be seen that the lower the pole number, the higher dependency
the lower of
the the
pole
saliency
number,
the ratio
saliency on the
the ratio,
higher pole
the
and pair number
saliency
consequently andconsequently
ratio,higher
and the winding
power type.
factorhigher
valuesItpower
canobtained,
are befactor
seen that
valuestheare
as shown lower the pole
obtained,
in Figure as
21.
number,
Moreover, the
shown in Figurehigher the
the influence saliency
21. Moreover, ratio, and
the influence
of the winding consequently
type onofthe higher
thesaliency
winding type
ratio power
can factor
onclearly values
the saliency are obtained,
ratio can
be observed as
clearly22b:
in Figure be
shown
observed
the in Figure
highest 21. 22b:
insaliency
Figure Moreover,
the
ratios the influence
arehighest
obtained with of
saliency the winding
ratios
FSDW type
are obtained
and ISDW on theFSDW
with
machines; saliencyandratio
noticeably can machines;
ISDW
worse clearly be
saliency
observed
noticeably
ratios in Figure 22b:
worse saliency
are obtained the
with FSCW highest saliency
ratiosmachines. ratios are obtained
are obtained with FSCW machines. with FSDW and ISDW machines;
noticeably worse saliency ratios are obtained with FSCW machines.

(a) (b)
(a) (b)
Figure 22.
Figure 22. Saliency
Saliency ratio
ratio (ξ)
(ξ) versus
versus (a)
(a) pole
pole pair
pair number
number (p)
(p) and
and winding
winding type;
type; (b)
(b) stator
stator slot
slot number
number
Figure
per 22.and
pole Saliency
phase ratio
(q (ξ) pole
s) and versus (a)number
pair pole pair number (p) and winding type; (b) stator slot number
(p).
per pole and phase (qs ) and pole pair number (p).
per pole and phase (qs) and pole pair number (p).
The main conclusions of this subsection are summarized below:
The main
The main conclusions
conclusions of of this
this subsection
subsection are are summarized
summarized below: below:
 The pole number has a great influence on the power factor. Low pole numbers are recommended
• The
due pole number
to their high has a great
saliencygreat influence
influence on the
ratios; the power
power factor.
factor. Low
Low pole
pole numbers
numbers are are recommended
recommended
 due to their winding
Distributed high saliency ratios;have higher saliency ratios than concentrated winding machines.
machines
• Distributed
Among concentrated winding have
Distributed winding
winding machines
machines have higher
higherthe
machines, saliency
ones ratios
saliency ratiosqthan
with than concentrated
and nl,w,s = winding
concentrated
s = 0.5 2winding
have the machines.
machines.
highest
Among
Among concentrated
saliencyconcentrated
ratios. winding machines, the ones with q = 0.5 and n
winding machines, the ones with qs = 0.5 and nl,w,s = 2 have the highest
s l,w,s = 2 have the highest
saliency
saliency ratios.
ratios.
5.4. Efficiency
5.4. Efficiency
Finally, the electromagnetic efficiency (ηe) is studied. Due to the calculation procedure’s
Finally,
simplifications, theonlyelectromagnetic
electromagnetic
electromagnetic efficiency
efficiency
losses are(η e) isis studied.
(ηeconsidered
studied. Due
Due
in this to the
to
work the
(i.e.,calculation
copper losses procedure’s
(PCu) and
simplifications,
simplifications,
iron losses (PFe)). only
only electromagnetic
electromagnetic losses
losses areare considered
considered in in
thisthis
work work (i.e.,
(i.e., coppercopper losses
losses (PCu)(P Cu )
and
and iron
iron The losses
losses (PFe)).(PFe )). of the efficiency on the pole pair number and the winding type is shown in
dependency
The dependency
dependency
Figure 23a. The influence of the
of of efficiency
the pole numberon the shows
pole pair thatnumber and the winding
better efficiency results are type is shown
obtained in
with
Figure 23a.
23a. The
The influence
influence of
of the
the pole
pole number
number shows
shows that better
that efficiency
better efficiencyresults
low pole numbers. The influence of the winding type is not that clear, but it can be seen that in mostresults are obtained
are with
obtained low
with
pole
low numbers.
casespole FSCWThe
the numbers. influence
The influence
machines of thelower
obtain ofwinding type type
the winding
efficiency isresults.
notisthat
not clear, but itbut
that clear, canitbecan seen
be that
seeninthat
most cases
in most
the FSCW
cases machines
the FSCW obtainobtain
machines lowerlower
efficiency results.
efficiency results.

(a) (b)
(a) (b)
Figure 23. Electromagnetic efficiency (ηe) versus (a) pole pair number (p) and winding type; (b) stator
Figure 23.
23. Electromagnetic
slot number efficiency
per pole and phase
Electromagnetic (ηe))pole
(qs) and
efficiency (η versus (a)
pair pole
pole pair
number (p).number
number (p)
(p) and
and winding
winding type;
type; (b)
(b) stator
Figure e versus (a) pair stator
slot number per pole and phase (qs) and pole pair number (p).
slot number per pole and phase (q ) and pole pair number (p).
s
Figure 23b shows the dependency of the efficiency on the stator slot number per pole and phase,
and the pole23b
Figure shows
pair the dependency
number. of the
It can be seen thatefficiency
the higheronefficiency
the statorvalues
slot number perpole
for each polenumber
and phase,
are
and the pole pair number. It can be seen that the higher efficiency values for each pole number are
Energies 2018, 11, 128 16 of 21

Figure
Energies 23b128
2018, 11, shows the dependency of the efficiency on the stator slot number per pole and phase, 16 of 21
Energies 2018, 11, 128 16 of 21
and the pole pair number. It can be seen that the higher efficiency values for each pole number are
obtained with ISDW machines with one stator slot per pole and per phase
obtained with ISDW machines with one stator slot per pole and per phase (qs = 1). general (qs = 1). In general terms,
obtained
distributed with ISDW machines
winding machines obtain
with one stator
higher slot per results
efficiency pole and
thanperconcentrated
phase (qs = 1). In general
winding terms,
machines.
distributed winding machines obtain higher efficiency results than concentrated winding machines.
distributed winding
The efficiency machines
efficiency of the machines obtain higher efficiency results than concentrated winding machines.
The machines depends
dependson onthe
thelosses
lossesand
andthetheoutput
output power.
power. Since thethe
Since speed
speedfor
all The
the efficiency
machines is of the machines
constant, the depends
output on the
power losses and
exclusively the output
depends on power.
the mean Since the speed
torque. Thus, for
the
for all the machines is constant, the output power exclusively depends on the mean torque. Thus,
all the machines
efficiency is constant,
is studied with themean
output power exclusively depends on losses
the mean torque. Thus, The
the
the efficiency is studied withthethe meanelectromagnetic
electromagnetictorque,
torque,copper
copper losses and iron iron losses.
losses. The
efficiency
Figure 24 is studied
24 shows
shows the with
the copper the mean
copper losses
losses (a)electromagnetic
(a) and
and the
the iron torque,
iron losses
losses (b) copper
(b) for losses
for different and
different pole iron
pole pair losses.
pair number The
number andand
Figure
Figure 24
windingtypes.shows
types. the copper losses (a) and the iron losses (b) for different pole pair number and
winding
winding types.

(a) (b)
(a) (b)
Figure 24.
Figure 24. (a)
(a)Copper
Copperlosses (PCu
losses (P); (b)
); and iron iron
(b) and losseslosses
(PFe) versus pole pair
(PFe ) versus polenumber (p) and winding
pair number (p) and
Figure
type. 24. (a) Copper losses (PCu);Cu(b) and iron losses (PFe) versus pole pair number (p) and winding
winding type.
type.
The iron losses depend on the pole pair number, as Figure 24b shows, due to is dependency on
The iron losses depend on the pole pair number, as Figure Figurethan24b shows,
24b copper due
due to
shows, losses, is dependency on
to so
frequency. In any case, the iron losses are considerably lower the attention in
frequency.
frequency. In any
In any case,
case,on the
the iron
iron losses
losses are considerably
are considerably lower than copper losses, so the attention in
this study must focus the copper losses.
this study must focusfocus on on the
the copper
copper losses.
Since the current density is set,losses.
the copper losses mainly depend on the total slot area of the
Since the current density is set, the copper losses mainly depend on the total slot area of the
machines and the stator slot fill factor. This explains why concentrated winding machines have higher
machines
machines and
and the stator slot fill factor. This explains why concentrated winding machines have higher
copper losses, while distributed winding machines have similar copper losses, as it can be seen in the
copper
copper losses,
losses, while
while distributed
distributed winding
winding machines
machines havehave similar
similar copper
copper losses,
losses, as
as itit can
can bebe seen
seen in the
Figure 24a.
Figure
Figure 24a.
24a.
The Figure 25a shows the relationship between the magnetomotive force (which depends on the
The Figure 25a 25a shows
showsthe the relationshipbetween
between the magnetomotive force (which depends on
slot area and the fill factor, inrelationship
the same way as the the magnetomotive
copper losses), and force
the(which
copperdepends on the
losses for the
the
slot slot
area area
and and
the the
fill fill factor,
factor, in in the
the same same
way way
as as the
the coppercopper losses),
losses), and and
the the
copper copper losses
losses for for
the
distributed winding machines, depending on the pole pair number. It can be seen that the copper
the distributed
distributed winding
winding machines,
machines, depending
depending onforceonpole
the the pair
polenumber.
pair number.It can It can bethat
seen
thethat the
losses are proportional to the magnetomotive for the machines with thebe seenpole
same copper
number. The
copper
losses losses
are are proportional
proportional to thepole to the magnetomotive
magnetomotive force for the machines with the same pole
difference between different numbers liesforce forlosses
in the the machines
producedwith theend
in the same pole
windings: number. The
the higher
number.
difference The difference
betweenthe between
different poledifferent
numbers pole numbers
lieslength. lies in the losses produced in the
in the losses produced in the end windings: the higher end windings:
the pole number, lower the end winding
the higher the polethe
pole number, number,
lower thethe end
lower the endlength.
winding winding length.

(a) (b)
(a) (b)
Figure 25. (a) Copper losses (PCu) versus magnetomotive force (N· I) and pole pair number (p); (b) and
Figure
Figure 25. (a)
25. (a) Copper
electromagnetic Copper losses(η(P
losses
efficiency (P Cu) versus magnetomotive force (N·
e)Cu ) versus
versus magnetomotive
magnetomotive force
force (N·I)
I) and pole
andby
multiplied thepair
pole firstnumber
pair number (p);winding
(p);
harmonic (b) and
(b) and
electromagnetic
electromagnetic
factor (N· I· kω,s1) andefficiency
efficiency (η
pole pair e) versus magnetomotive force multiplied by the first harmonic winding
(η e )number
versus magnetomotive force
(p); for ISDW and multiplied
FSDW by the first harmonic winding
machines.
factor (N
factor (N··II·
·kkω,s1
ω,s1
) and
) and pole
polepair
pair number
number (p);
(p);for
forISDW
ISDWand
andFSDW
FSDW machines.
machines.
It has been proved in Section 5.1 that the mean torque is also proportional to the magnetomotive
forceIt(especially
has been proved in Section
if the effect of the5.1 that thefactor
winding meanis torque
takenisinto
alsoaccount,
proportional
as in to the magnetomotive
Figure 12). Thus, high
force
magnetomotive force values obtain both high mean torque and copper losses. Figure 25bThus,
(especially if the effect of the winding factor is taken into account, as in Figure 12). showshigh
the
magnetomotive
relation between force
the values obtain bothforce,
magnetomotive high mean torquebyand
multiplied thecopper losses.
winding Figure
factor (N· 25b
I·kω,s1shows
) and the
the
relation between the magnetomotive force, multiplied by the winding factor (N· I·kω,s1) and the
Energies 2018, 11, 128 17 of 21

It has been proved in Section 5.1 that the mean torque is also proportional to the magnetomotive
force (especially if the effect of the winding factor is taken into account, as in Figure 12). Thus,
high magnetomotive force values obtain both high mean torque and copper losses. Figure 25b shows
the relation between the magnetomotive force, multiplied by the winding factor (N·I·kω,s1 ) and the
electromagnetic efficiency. It is proved that, for the same pole number, machines with higher N·I·kω,s1
values obtain higher efficiency results.
The main conclusions of this subsection are summarized below:

• A low pole number is recommended to obtain a higher efficiency, because they can produce
higher output powers with similar losses;
• The winding type has a considerable effect on the efficiency.

# Distributed winding machines have higher efficiencies than concentrated winding machines;
# Concentrated winding machines have both, lower torque production and higher losses,
in general;
# Among concentrated winding machines, the ones with qs = 0.5 and nl,w,s = 2 have efficiencies
comparable to distributed winding machines.

5.5. Design Criteria

Based on the conclusions of the analysis, the main key steps for a good SynRM design are
summarized in this subsection. Then, the feasibility of using SynRMs for low speed applications
is studied.
The conclusions of the main parameters studied in this work (i.e., pole number, stator slot number
per pole and phase, and the winding type) show that:

• On the one hand, a low pole number is recommended for high mean torque, power factor and
efficiency. On the other hand, the pole number has negligible effect on the torque ripple;
• ISDW machines have good torque capability, efficiency and power factor. The torque capability
and efficiency are especially high when one stator slot per pole and phase is used (qs = 1), but the
torque ripple is considerable. Very low torque ripple values are obtained when the number of
stator slots per pole and phase is greater than the unity (qs > 1), although it must be noted that
further design measures to reduce the torque ripple have not been considered;
• FSDW machines exhibit satisfying torque capability, power factor and efficiency and excellent
torque ripples;
• FSCW machines, in most cases lead to moderate torque capability and poor torque ripple, power
factor and efficiency results.

# Among fractional slot concentrated winding machines, the ones with qs = 0.5 and nl,w,s =
2 show excellent torque capability and efficiency. The power factor for these machines is
slightly lower than for distributed winding machines and the torque ripple is very high.

In addition to the design criteria related to the main studied parameters, some other aspects must
be taken into account:

• The rotor geometry has not been studied in this work, although the literature has demonstrated
that it has a deep effect on the torque ripple and the saliency ratio;
• The mean torque of the machines has proven to be linearly related to the air-gap flux density first
harmonic. Design strategies that lead to higher air-gap flux densities are recommended, such as
the reduction of the air-gap thickness.
 Energies 2018, 11, 128 18 of 21

6. PMSM vs. SynRM Comparison

In this section, a SynRM is designed based on the previously shown design criteria and the results
are compared to those of the original PMSM.
A two-pole pair ISDW machine has been chosen, with 48 slots and two slots short pitching.
The machine has been designed and calculated applying the procedure explained in Section 2, with the
design specifications and restrictions shown in Table 6. Due to the absence of permanent magnets,
Energies
placing 2018,
the11,rotor
128 inside the stator is easier for SynRMs, so the air-gap of the designed machine
18 ofhas
21
been reduced to 0.8 mm. Figure 26 shows one half of the designed machine geometry.

Figure 26. p = 2, Qs = 48, τ a = 2 SynRM geometry.

Figure 26. pof= both

The performance characteristics 2, Qs =machines
48, τa = 2 SynRM geometry.
have been calculated under the conditions in
which the steady-state temperatures do not exceed the maximum allowable temperatures. The thermal
The performance
requirements characteristics
for the PMSM of both machines
allow temperatures have◦ Cbeen
up to 155 in thecalculated
windingsunder theinsulation)
(F class conditionsand
in
which
120 ◦ Cthe steady-state
in the magnets (N40H temperatures do not exceed
grade magnet). Since thetheabsence
maximum allowable magnets
of permanent temperatures. The
in SynRMs
thermal requirements for the PMSM allow temperatures up to 155 °C in
allows less stringent thermal restrictions, the maximum allowable temperature for the SynRM isthe windings (F class
insulation)
imposed toand 180120
◦ C °C
(Hin the insulation).
class magnets (N40H grade magnet). Since the absence of permanent magnets
in SynRMs
Models allows less stringent
by MotorCAD thermal
software restrictions,
for both machinesthe are maximum
used for theallowable temperature
thermal calculation, for the
employing
SynRM is imposed to 180 °C (H class insulation).
natural convection as cooling method. The results for both machines with the same volume are
Models
shown in theby MotorCAD
Table software
7. The results show for
that,both
evenmachines
for a higher aretemperature
used for the thermal
restriction in calculation,
the winding,
employing natural convection as cooling method. The results for both machines
the performance characteristics of SynRMs are unable to reach the specifications of the PMSM with the same
with
volume are shown in the Table
the same geometric restrictions. 7. The results show that, even for a higher temperature restriction in
the winding, the performance characteristics of SynRMs are unable to reach the specifications of the
PMSMTable
with7.theSynRM
same geometric restrictions.
and permanent magnet synchronous machines (PMSM) characteristics for the
same volume.
Table 7. SynRM and permanent magnet synchronous machines (PMSM) characteristics for the same
volume. Parameter SynRM PMSM
Electromagnetic
Parameter torque (Nm) 103
SynRM 162
PMSM
Torque ripple (Nm) 4.9 0.6
Electromagnetic torque (Nm)
Line voltage (V)
103
319
162
330
Torque ripple (Nm)
Phase current (A) 4.9
5.7 0.6
6.5
Line voltage (V)
Apparent power (kVA)
319
3.15
330
3.71
Phase current (A)
Input power (kW) 5.7
2.73 6.5
3.35
Apparent power (kVA)
Output power (kW) 3.15
1.73 3.71
2.71
InputPower
power (kW)
factor 2.73
0.85 3.35
0.90
OutputEfficiency
power (kW) 1.73
0.63 2.71
0.81
Power factor 0.85 0.90
Efficiency 0.63
The increase needed in the volume for meeting torque specifications 0.81 is calculated. The stack
length has been increased from 125 mm to 200 mm, which means that the total volume of the machine
Thelarger.
is 60% increase
Theneeded
resultsin the
are volume for
compared meeting
to the torque
original specifications
PMSM performanceis calculated. The in
characteristics stack
the length
Table 8.
has been increased from 125 mm to 200 mm, which means that the total volume of the machine is 60%
larger. The results are compared to the original PMSM performance characteristics in the Table 8.

Table 8. SynRM and PMSM characteristics for the same torque requirement.

Parameter SynRM PMSM

Electromagnetic torque (Nm) 162 162
Torque ripple (Nm) 7.5 0.6
Energies 2018, 11, 128 19 of 21

Table 8. SynRM and PMSM characteristics for the same torque requirement.

Parameter SynRM PMSM

Electromagnetic torque (Nm) 162 162
Torque ripple (Nm) 7.5 0.6
Line voltage (V) 332 330
Phase current (A) 8.05 6.5
Apparent power (kVA) 4.63 3.71
Input power (kW) 3.92 3.35
Output power (kW) 2.71 2.71
Power factor 0.82 0.90
Efficiency 0.69 0.81

It can be concluded that, for a comparison in which fair thermal restrictions are applied to both
machines, the SynRM is able to provide around 35% lower torque than PMSMs with the same volume.
For the same torque specifications, about 60% longer stacks are needed for SynRMs. In both cases,
the power factor is slightly lower and efficiency is around 20% worse for the SynRM.
Nevertheless, it is expected that a proper design with an optimized rotor geometry could greatly
improve the presented results. This should be studied in future works.
Only the main performance characteristics of the machines have been compared in this work,
but SynRMs have some advantages over PMSMs that have not been quantified. These advantages
are derived from the absence of permanent magnets: there is no demagnetization risk, they exhibit
good fault tolerance, and, most importantly, manufacturing cost is reduced. It is expected that,
for applications in which the machine size is not restrictive, a properly designed SynRM with a longer
stack could substitute a PMSM, with a reduction of costs.

7. Conclusions
In this paper, the influence of the rotor pole number, the stator slot number and the winding type in
SynRMs is studied, in order to obtain the optimal combination for low speed–high torque application.
This study has proven that high pole numbers are not suitable for SynRMs, and that FSCW
machines should not be used as well in general terms. However, FSCW machines with qs = 0.5 and
double layer winding have proven to be good candidates in terms of mean torque capability for
applications in which having a small torque ripple is not an essential requirement.
A performance comparison between SynRM and PMSM has been conducted. It is found that,
for the evaluated machine, the torque capability for the same geometric restrictions and limited
maximum temperature is around 35% lower for the SynRM. For the same torque requirement and
limited maximum temperature, the needed volume is around 60% larger for the SynRMs. Moreover,
it is found that the power factor and the efficiency are lower for the SynRM in both cases.
The proposed calculation procedure has been validated by FEM simulations and experimental
measurements, meaning that the results and conclusions of the study are validated as well.
The literature has proved that some performance characteristics of the machines are heavily
dependent on the rotor geometry, and the machines in this work have been designed with the same
pre-defined rotor geometry. In order to know the potential of these machines, the optimization of the
rotor geometry is crucial. New machines with optimized rotor geometries should be considered in
future works, with the aim of verifying the conclusions of this work, and studying the feasibility of
using SynRMs for low-speed applications.

Author Contributions: All the authors contributed substantially to the work. Gurutz Artetxe performed most of
the work and wrote the paper. Jesus Paredes and Borja Prieto contributed in the development of the designing
tool and also in the experimental testing. Miguel Martinez-Iturralde and Ibon Elosegui supervised all the stages
of the work and revised the manuscript.
Conflicts of Interest: The authors declare no conflict of interest.
Energies 2018, 11, 128 20 of 21

References
1. Lipo, T.A. Synchronous reluctance machines—A viable alternative for AC drives? Electr. Mach. Power Syst.
1991, 19, 659–671. [CrossRef]
2. Miller, T.J.E.; Hutton, A.; Cossar, C.; Staton, D.A. Design of a synchronous reluctance motor drive. IEEE Trans.
Ind. Appl. 1991, 27, 741–749. [CrossRef]
3. Vagati, A.; Fratta, A.; Franceschini, G.; Rosso, P. AC motors for high-performance drives: A design-based
comparison. IEEE Trans. Ind. Appl. 1996, 32, 1211–1219. [CrossRef]
4. Boglietti, A.; Cavagnino, A.; Pastorelli, M.; Vagati, A. Experimental comparison of induction and synchronous
reluctance motors performance. In Proceedings of the Fourtieth IAS Annual Meeting. Conference Record
of the 2005 Industry Applications Conference, Kowloon, Hong Kong, China, 2–6 October 2005; Volume 1,
pp. 474–479.
5. Pellegrino, G.; Jahns, T.M.; Bianchi, N.; Soong, W.; Cupertino, F. The Rediscovery of Synchronous Reluctance and
Ferrite Permanent Magnet Motors: Tutorial Course Notes; Springer: Berlin, Germany, 2016.
6. Roshanfekr, P.; Lundmark, S.; Thiringer, T.; Alatalo, M. A synchronous reluctance generator for a wind
application-compared with an interior mounted permanent magnet synchronous generator. In Proceedings of
the 7th IET International Conference on Power Electronics, Machines and Drives (PEMD 2014), Manchester, UK,
8–10 April 2014; pp. 1–5.
7. Schmidt, E. Comparison of different designs of synchronous reluctance machines with high-anisotropy rotors.
In Proceedings of the 2012 XXth International Conference on Electrical Machines (ICEM), Marseille, France,
2–5 Sepetember 2012.
8. Staton, D.A.; Miller, T.J.E.; Wood, S.E. Maximising the saliency ratio of the synchronous reluctance motor.
IEE Proc. B-Electr. Power Appl. 1993, 140, 249–259. [CrossRef]
9. Gamba, M.; Pellegrino, G.; Vagati, A. A new PM-assisted Synchronous Reluctance machine with a
nonconventional fractional slot per pole combination. In Proceedings of the 2014 International Conference on
Optimization of Electrical and Electronic Equipment (OPTIM), Bran, Romania, 22–24 May 2014; pp. 268–275.
10. Park, J.M.; Kim, S.I.; Hong, J.P.; Lee, J.H. Rotor design on torque ripple reduction for a synchronous
reluctance motor with concentrated winding using response surface methodology. IEEE Trans. Magn. 2006,
42, 3479–3481. [CrossRef]
11. Spargo, C.M.; Mecrow, B.C.; Widmer, J.D.; Morton, C. Application of fractional-slot concentrated windings
to synchronous reluctance motors. IEEE Trans. Ind. Appl. 2015, 51, 1446–1455. [CrossRef]
12. Donaghy-Spargo, C.; Mecrow, B.C.; Widmer, J.D. Electromagnetic Analysis of a Synchronous Reluctance
Motor with Single Tooth Windings. IEEE Trans. Magn. 2017, 53, 8206207. [CrossRef]
13. Reddy, P.B.; Grace, K.; El-Refaie, A. Conceptual design of sleeve rotor synchronous reluctance motor for
traction applications. IET Electr. Power Appl. 2016, 10, 368–374. [CrossRef]
14. Vagati, A.; Boazzo, B.; Guglielmi, P.; Pellegrino, G. Design of ferrite-assisted synchronous reluctance machines
robust toward demagnetization. IEEE Trans. Ind. Appl. 2014, 50, 1768–1779. [CrossRef]
15. Boazzo, B.; Vagati, A.; Pellegrino, G.; Armando, E.; Guglielmi, P. Multipolar ferrite-assisted synchronous
reluctance machines: A general design approach. IEEE Trans. Ind. Electron. 2015, 62, 832–845. [CrossRef]
16. Pyrhonen, J.; Jokinen, T.; Hrabovcova, V. Design of Rotating Electrical Machines; John Wiley & Sons: Hoboken,
NJ, USA, 2009.
17. Bomela, X.B.; Kamper, M.J. Effect of stator chording and rotor skewing on performance of reluctance
synchronous machine. IEEE Trans. Ind. Appl. 2002, 38, 91–100. [CrossRef]
18. Pellegrino, G.; Cupertino, F.; Gerada, C. Barriers shapes and minimum set of rotor parameters in the
automated design of Synchronous Reluctance machines. In Proceedings of the 2013 IEEE International
Electric Machines & Drives Conference (IEMDC), Chicago, IL, USA, 12–15 May 2013; pp. 1204–1210.
19. Moghaddam, R.R. Synchronous Reluctance Machine (SynRM) Design. Mater’s Thesis, Royal Institute of
Technology, Stockholm, Sweden, 2007.
20. Mahadevan, B.; Nagarajan, G. Design optimization of ferrite assisted synchronous reluctance motor using
multi-objective differential evolution algorithm. COMPEL Int. J. Comput. Math. Electr. Electron. Eng. 2017, 36,
219–239.
21. Cupertino, F.; Pellegrino, G.; Gerada, C. Design of synchronous reluctance motors with multiobjective
optimization algorithms. IEEE Trans. Ind. Appl. 2014, 50, 3617–3627. [CrossRef]
Energies 2018, 11, 128 21 of 21

22. Gamba, M.; Pellegrino, G.; Cupertino, F. Optimal number of rotor parameters for the automatic design
of synchronous reluctance machines. In Proceedings of the 2014 International Conference on Electrical
Machines (ICEM), Berlin, Germany, 2–5 September 2014; pp. 1334–1340.
23. Wang, Y.; Ionel, D.M.; Jiang, M.; Stretz, S.J. Establishing the relative merits of synchronous reluctance and
PM-assisted technology through systematic design optimization. IEEE Trans. Ind. Appl. 2016, 52, 2971–2978.
[CrossRef]
24. Wang, Y.; Ionel, D.; Dorrell, D.G.; Stretz, S. Establishing the power factor limitations for synchronous
reluctance machines. IEEE Trans. Magn. 2015, 51, 1–4.
25. Matsuo, T.; Lipo, T.A. Rotor design optimization of synchronous reluctance machine. IEEE Trans. Energy Convers.
1994, 9, 359–365. [CrossRef]
26. Wang, K.; Zhu, Z.Q.; Ombach, G.; Koch, M.; Zhang, S.; Xu, J. Optimal slot/pole and flux-barrier layer
number combinations for synchronous reluctance machines. In Proceedings of the 2013 8th International
Conference and Exhibition on Ecological Vehicles and Renewable Energies (EVER), Monte Carlo, Monaco,
27–30 March 2013; pp. 1–8.

© 2018 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/).