energies

Article
Performance Analysis and Optimization of a Novel Outer Rotor
Field-Excited Flux-Switching Machine with Combined
Semi-Closed and Open Slots Stator
Siddique Akbar 1, * , Faisal Khan 1 , Wasiq Ullah 1 , Basharat Ullah 1 , Ahmad H. Milyani 2,3
and Abdullah Ahmed Azhari 4

1 Department of Electrical and Computer Engineering, COMSATS University Islamabad,

Abbottabad 22060, Pakistan
2 Department of Electrical and Computer Engineering, King Abdulaziz University, Jeddah 21589, Saudi Arabia
3 Center of Research Excellence in Renewable Energy and Power Systems, King Abdulaziz University,
Jeddah 21589, Saudi Arabia
4 The Applied College, King Abdulaziz University, Jeddah 21589, Saudi Arabia
* Correspondence: engrsiddique97@gmail.com

Citation: Akbar, S.; Khan, F.; Ullah,
W.; Ullah, B.; Milyani, A.H.; Azhari,
A.A. Performance Analysis and
Optimization of a Novel Outer Rotor
Field-Excited Flux-Switching
Machine with Combined
Semi-Closed and Open Slots Stator.
Energies 2022, 15, 7531. https://
doi.org/10.3390/en15207531
Academic Editors: Vladimir Prakht,
Mohamed N. Ibrahim and Vadim
Kazakbaev
Abstract: Slotting effect in electric machines reduces flux per pole that effect magnetic flux density
distribution in the air gap which induces harmonics in magnetic flux density causing flux pulsation,
that in turn generates dominant torque pulsation in the form of cogging torque and torque ripples.
To overcome the abovesaid demerits, a novel outer rotor field-excited flux-switching machine (OR-
FSFSM) with a combined semi-closed and open slots stator is proposed in this study. The developed
OR-FEFSM offers a high-power factor, due to the utilization of the semi-closed slot for armature coils.
The open slot stator structure was chosen for the field excitation coil, which effectively suppresses
leakage reluctance that causes flux pulsation. Thus, the influence of torque ripples is reduced, and
the average torque is improved. In order to investigate the effectiveness of the proposed OR-FEFSM,
a detailed study of stator slot and rotor pole combinations are performed. Based on simplified
mathematical formulation, 12S/7P (stator slot/rotor poles), 12S/11P, 12S/13P, and 12S/17P are the
most feasible combinations. Finite Element Analysis (FEA) based on comprehensive electromagnetic
performance is performed on each combination, and found that 12S/13P offers the highest average
torque of 4.62 Nm, whereas 3.72 Nm, 2.72Nm, and 1.68 Nm average torque is offered by 12S/17P,
12S/7P, and 12S/11P, respectively. Based on the initial analysis, 12S/13P was considered for further
analysis and optimized using JMAG built-in Genetic Algorithm (GA). Moreover, thermal analysis
was performed, and the proposed design was compared with the conventional design.

Received: 16 September 2022 Keywords: field-excited machine; flux-switching machine; finite element analysis; genetic algorithm;
Accepted: 9 October 2022 thermal analysis; performance analysis; optimization
Published: 12 October 2022

Publisher’s Note: MDPI stays neutral

with regard to jurisdictional claims in
published maps and institutional affil- 1. Introduction
iations. Flux-switching machines (FSMs) are a class of synchronous machines that attract
researcher interest due to their doubly salient structure, featuring both excitation sources
(permanent magnet (PM) or winding) on the stator and robust rotor structure [1–5]. Since
there are no PMs or windings on the rotor, FSM has a robust rotor structure like switched
Copyright: © 2022 by the authors.
reluctance machines (SRM), making it a strong candidate in many industrial applications
Licensee MDPI, Basel, Switzerland.
where high speed and rigidness are required [6–8]. Thus, due to the inherent feature of
This article is an open access article
high torque and power density, single-phase, three-phase, and multiphase FSM topologies,
distributed under the terms and
such as PM excited, field excited (FE), and hybrid excited (HE), have been researched over
conditions of the Creative Commons
Attribution (CC BY) license (https://
the past few decades [9,10].
creativecommons.org/licenses/by/
Permanent magnet flux-switching machines (PMFSMs) offer high torque density, ef-
4.0/).
ficiency, lower harmonic components in their back-electromotive force (back-EMF), and

Energies 2022, 15, 7531. https://doi.org/10.3390/en15207531 https://www.mdpi.com/journal/energies

Energies 2022, 15, 7531 2 of 17

robust rotor structure [11–13]. Due to the stator’s flux-concentrating effect, PMFSM eas-
ily generates a high magnetic flux density in the air gap and therefore achieves high
torque [14–16]. However, field weakening is a problem for these machines because their
d-axis inductance is lowered due to deep magnetic saturation in the stator, limiting their
constant power speed range. Additionally, the possibility of irreversible demagnetization,
cost of rare earth magnets, uncontrollable flux, and PM’s low mechanical strength are all
critical issues limiting PM machines to some extent. Therefore, reduced PM or no PM
machines are preferred [17,18].
Hybrid excited flux-switching machines (HEFSMs) have the potential to improve
power density, efficiency, and flux weakening performance [19–21], and overall reduce the
volume of PM by adding FE coils (FEC). However, the presence of both excitation sources
(PM and FEC) on the stator makes its geometry complex. To overcome the aforesaid
demerits of HEFSM and PMFSM, field-excited flux-switching machines (FEFSMs) are
introduced that are fully excited by windings and have more robust rotors than other types
of FSM. Despite their lower efficiency, the field current regulation makes FEFSM a suitable
alternative for variable speed applications [4,22].
FEFSM that utilize only field winding as a magnetic excitation are proposed in [23].
The field-weakening capabilities of this machine have been greatly enhanced due to flux
regulation capability with dc excitation current that effectively reduces excitation magnetic
field. However, the field windings are wound around every other stator tooth, causing the
magnetic field to be radially excited and unable to be concentrated, which results in low
torque density. To improve torque-power density, the authors of [23] investigated wound
field-excited FSM with non-overlapped winding and segmented rotor structure. Due to the
advantages of flexible flux weakening capability and short flux path, high-torque density
is achieved. However, FEFSM topologies with segmented rotors suffer from mechanical
constraints, and are not feasible for high-speed applications due to rotor segmentation.
The authors of [24] introduce a partitioned stator FEFSM (PS-FEFSM) with a double
stator, having field and armature winding on a separate stator, so that FEC are arranged in
the inner stator. In contrast, armature windings are placed in outer stators, thus increasing
the torque density by 19% more than conventional FEFSM. However, manufacturing cost
increases, due to the partitioned stator with two air gaps; and with metallic coupling of the
rotor pieces, the cage losses increase, which can be incorporated with non-metallic sticks.
Despite the average torque improvements, the influence of the cogging torque and
torque ripples are dominant in FEFSM. In this regard, the authors of [23] comparatively
investigated FEFSM with and without skew rotor for torque ripple reduction. However,
torque ripple is suppressed at the cost of a reduction in average torque. To further improve
the torque profile, the authors of [25] examined torque characteristics of an OR-FEFSM with
maximum torque control; however, this design suffers from magnetic saturation under
higher currents. Moreover, dual-stator OR-FEFSM for unbalanced shaft magnetic force
in [26] is effectively suppressed at 3% reduction in average torque.
To address the aforementioned problems in various state-of-the-art designs, this
paper proposes a novel OR-FEFSM (as shown in Figure 1) for direct-drive embedded
in-wheel systems, with an improved torque profile utilizing combined semi-closed and
open stators. Furthermore, due to the elimination of mechanical gearboxes, the proposed
OR-FEFSM omits maintenance costs, ultimately improving system operation efficiency.
In contrast, combined semi-closed and open slots dominantly suppress torque ripple
issues and improve overload capability. Therefore, this paper intends a finite element
analysis (FEA)-based detailed investigation of the proposed OR-FESM with possible rotor
pole combinations.
 Section 2 discusses design topology and operating principles, while Section
Section 2 discusses design topology and operating principles, while Section 3 inves-
tigates
tigates FEA-based
FEA-based performance
performance analysis
analysis of of possible
possible rotor rotor poleSection
pole combinations. combinations.
4 de- Sect
scribes
scribes the genetic
the genetic algorithm.
algorithm. Section 5Section
examines5the
examines
best rotorthe best
pole’s rotorFEA-based
detailed pole’s detailed FE
Energies 2022, 15, 7531
performance
performance study. Section 6 depicts a quantitative comparison ofper-
study. Section 6 depicts a quantitative comparison of electromagnetic electromagn
3 of 17
formance, whereas Section 7 investigates a comparison of conventional and proposed de-
formance, whereas Section 7 investigates a comparison of conventional and prop
signs. Section 8 draws a conclusion.
signs. Section 8 draws a conclusion.

(a) (b)
Figure
Figure1.1.Proposed
ProposedOR-FEFSM
OR-FEFSM(a)
(a)cross-sectional
cross-sectionalview
viewand
and(b)(b)
magnetic flux
magnetic density
flux map.
density map.

2. Design Topology
Section and Operating
2 discusses (a) topology
design Principles
and operating principles, while Section(b)3 inves-
tigates
2.1. FEA-based
Design Topology performance analysis of possible rotor pole combinations. Section 4
Figure 1.the
describes Proposed OR-FEFSMSection
genetic algorithm. (a) cross-sectional view
5 examines the bestand (b)pole’s
rotor magnetic fluxFEA-
detailed density map
The proposed 12S/13P OR-FEFSM is shown in Figure 1. It consists of a salient outer
based performance study. Section 6 depicts a quantitative comparison of electromagnetic
rotor and an inner stator having both armature and field winding. The material used for
performance, whereas Section 7Operating
investigates aPrinciples
comparison of conventional and proposed
the2.stator
Design Topology
and rotor and
is 35H210 silicon steel, and copper is used for the winding. Geometric
designs. Section 8 draws a conclusion.
parameters
2.1. Designof the proposed OR-FEFSM are shown in Figure 2 and listed in Table 1, while
Topology
the simplified
2. Design operating
Topology andprinciple
Operatingbased on the flux-switching principle and possible rotor
Principles
The proposed
pole combinations 12S/13P
are explored OR-FEFSM
as follows. is shown in Figure 1. It consists of a sali
2.1. Design Topology
rotor and an inner stator having both armature and field winding. The material
The proposed 12S/13P OR-FEFSM is shown in Figure 1. It consists of a salient outer
the stator and rotor is 35H210 silicon steel, and copper is used for the winding. G
rotor and an inner stator having both armature and field winding. The material used for
parameters
the of the
stator and rotor proposed
is 35H210 OR-FEFSM
silicon are shown
steel, and copper in the
is used for Figure 2 and
winding. listed in Table
Geometric
the simplified
parameters operating
of the proposed principle
OR-FEFSM arebased
shownonin the flux-switching
Figure principle
2 and listed in Table and poss
1, while
the simplified
pole operating are
combinations principle based as
explored on follows.
the flux-switching principle and possible rotor
pole combinations are explored as follows.

Figure 2. Depicted design parameters of proposed OR-FEFSM.

Table 1. Design parameter of proposed OR-FEFSM.

Symbol Value (unit) Symbol Value (unit)

R1 15 mm 𝑊 7.5 mm
R2 20 mm 𝑊 6 mm
R3 35.5 mm 𝐻 7.5 mm
R4 40 mm 𝑟 22 mm
Figure
Figure R
2.5 Depicted
2. Depicted design design44.5 mm
parameters
parameters of proposed
of proposed 𝛼 OR-FEFSM. 10 mm
OR-FEFSM.
R6 45 mm 𝑊 6 mm
2.2. Operating Principles
R 7 52.5 mm
Table 1. Design parameter of proposed OR-FEFSM. 𝐻 15.5 mm
The operating principle of OR-FEFSM is similar to PMFSM, and can be easily explained
by the structure shown in Figure 3a,b. The upper laminated part is a rotor-like SRM, and
Symbol Value (unit) Symbol Value (u
the lower part consists of an armature winding (AW) and FEC labeled in the stator. The
flux generatedRby1 the FEC flows through 15 mm 𝑊
the path with the least reluctance. As shown in 7.5 mm
Figure 3a, whenR2 the rotor poles align with
20 mm the stator teeth, around 𝑊
which the phase coil is 6 mm
wound, the fluxR3 that is linked in the coil leaves
35.5 mm the rotor tooth and 𝐻 the coil. Similarly, 7.5 mm
enters
as the rotor rotates further and another rotor pole aligns with the next stator tooth of the
same coil, fluxRleaves
4 40 the
the coil and enters mmrotor tooth, maintaining𝑟 the same amount of 22 mm
R5 44.5 mm 𝛼 10 mm
R6 45 mm 𝑊 6 mm
R7 52.5 mm 𝐻 15.5 m
Energies 2022, 15, x FOR PEER REVIEW 4 of 18

Energies 2022, 15, 7531 4 of 17

R8 60 mm 𝐻 9 mm
βin 6.76 mm 𝛽 11.67 mm
flux-linkage
W1 but reversing the6polarity,
mm as shown in Figure
W2 3b. The coils flux linkage varies
6 mm
regularlyW3as the rotor rotates, producing
6 mm sinusoidal back-EMF.
Stack length Torque is generated
45 mm when
current is appropriately provided to the coils, which drives the armature forward.
2.2. Operating
Table Principles
1. Design parameter of proposed OR-FEFSM.
The operating principle of OR-FEFSM is similar to PMFSM, and can be easily ex-
plained bySymbol
the structure shown Value
in (unit) Symbol
Figure 3a,b. The upper laminated partValue (unit)
is a rotor-like
SRM, and the R1 lower part consists 15 of
mman armature winding Wry (AW) and FEC labeled
7.5 mm in the
R
stator. The flux 20 mm W
2 generated by the FEC flows through the path with the least reluctance.
rp 6 mm As
R3 35.5 mm Hrp 7.5 mm
shown in Figure 3a, when the rotor poles align with the stator teeth, around which the
R4 40 mm rout 22 mm
phase coil isRwound, the flux that 44.5ismm
linked in the coil leaves
αin the rotor tooth and enters the
10 mm
5
coil. Similarly,
R6 as the rotor rotates45further
mm and another rotor Wsppole aligns with the next stator
6 mm
tooth of the Rsame
7 coil, flux leaves the
52.5 mmcoil and enters the rotor
Hsi tooth, maintaining the same
15.5 mm
R8
amount of flux-linkage but reversing Hso
60 mm the polarity, as shown in Figure 3b. The9 mm
coils flux
βin regularly as the 6.76
linkage varies rotormmrotates, producingβ out
sinusoidal back-EMF.11.67Torque
mm is
W1 6 mm W2 6 mm
generated when current is appropriately provided to the coils, which drives the armature
W3 6 mm Stack length 45 mm
forward.

(a) (b)
Figure
Figure3. 3.
Operating principle
Operating of OR-FEFSM
principle (a) positive
of OR-FEFSM maximum
(a) positive flux and
maximum flux(b)
andnegative maximum
(b) negative maxi-
flux.
mum flux.

2.3.Rotor
2.3. RotorPoles
PolesCombination
Combination
Rotor pole positiondetermines
Rotor pole position determines the
the basic
basic principle
principle ofof FEFSM
FEFSM due
due toto magnetic
magnetic flux
flux link-
link-
ageinin
age armaturewinding
armature windingthat
that will
will bebe positive
positive andand negative,
negative, depending
depending onon the
the rotor
rotor position.
position.
Forinner
For inner and
and outer
outer rotor
rotor FEFSM,
FEFSM, possible
possible rotor
rotor poles
poles combinations
combinations areare computed
computed as: as:
 1 ± 𝑘
𝑁 𝑁 1±k (1)(1)
Nr= Ns 2𝑞
2q
where Nr represents the number of rotor poles, 𝑁 stator slot, 𝐾 is natural number, and
where Nr represents the number of rotor poles, Ns stator slot, K is natural number, and q is
q is number of pairs of phases. For the proposed OR-FEFSM, the values of q and 𝑁 are 6
number of pairs of phases. For the proposed OR-FEFSM, the values of q and Ns are 6 and
and 12, respectively, whereas feasible stator slot and rotor pole obtained are 12S/7P,
12, respectively, whereas feasible stator slot and rotor pole obtained are 12S/7P, 12S/11P,
12S/11P, 12S/13P, and 12S/17P.
12S/13P, and 12S/17P.
Furthermore,
Furthermore,for
forallallpossible
possiblerotor
rotorpoles
polescombinations,
combinations,the thenumber
numberofofturns
turnsper
perphase
phase
(𝑁(N) and number of turns for field winding (𝑁 ) varies based on the armature current
a ) and number of turns for field winding (Ne ) varies based on the armature current
density
density(𝐽(J) and field
) and fieldexcitation
excitationcurrent
currentdensity
density(𝐽(J) as
) asfollows:
follows:
a e
𝐼 𝑁
𝐽 Ia Na (2)
Ja= 𝛼𝑆 (2)
αSa
Ie Ne
Je= (3)
αSe
 Energies 2022, 15, x FOR PEER REVIEW 5 of 18

Energies 2022, 15, 7531 𝐼𝑁 5 of 17

𝐽 (3)
𝛼𝑆
where 𝛼 represent the filling factor while 𝑆 and 𝑆 represent armature and field slot
where α represent the
area, respectively. filling
Thus, factor while
to decide Sa andvalues
an optimal Se represent
for the armature
aforesaid and field slot
associated area,
parame-
respectively. Thus, to decide an optimal values for the aforesaid
ters (𝐽 & 𝐽 ), a detailed electromagnetic performance is investigated. associated parameters
(Je & Ja ), a detailed electromagnetic performance is investigated.
3. Performance Analysis of Feasible Rotor Pole Combination
3. Performance Analysis of Feasible Rotor Pole Combination
In this section, a comparative analysis of the proposed OR-FEFSM is conducted for
In this section, a comparative analysis of the proposed OR-FEFSM is conducted for
the possible rotor pole combinations. The best design based on electromagnetic analysis
the possible rotor pole combinations. The best design based on electromagnetic analysis
will be proceeded for further study.
will be proceeded for further study.
3.1.No
3.1. NoLoad
LoadFlux
FluxLinkage
Linkage
Figure44 shows
Figure shows thethe no-load
no-load fluxfluxlinkage
linkageofofallall
feasible rotor
feasible pole
rotor combinations.
pole It is
combinations.
Itworth noting
is worth thatthat
noting the the
phase fluxflux
phase linkage obtained
linkage obtained under no-load
under operation
no-load 𝐽 =
withwith
operation
Je15=𝐴/mm
15 A/mm . Analysis reveals
2 . Analysis that that
reveals OR-FEFSM
OR-FEFSM withwith
12S-7P design
12S-7P offers
design the the
offers highest flux
highest
linkage of 0.020 Wb, which is 54%, 7.15%, and 37.97% higher than 11,
flux linkage of 0.020 Wb, which is 54%, 7.15%, and 37.97% higher than 11, 13, and 17 rotor13, and 17 rotor
poles,respectively.
poles, respectively.
Flux linkage (Wb)

Figure4.4.No-load
Figure No-loadoperation
operationflux
fluxlinkages
linkagesofofallallfeasible
feasiblecombinations.
combinations.

3.2.
3.2.No
NoLoad
LoadBack-EMF
Back-EMF
Under
Underno-load
no-loadoperation,
operation,phase back-EMF
phase back-EMF forfor
12S/7P,
12S/7P,12S/11P, 12S/13P,
12S/11P, 12S/13P,and
and12S/17P
12S/17P
through
Energies 2022, 15, x FOR PEER REVIEWthrough 2D
2D FEA
FEAis shown
is shownin Figure
in 5.
Figure Analysis
5. Analysisreveals
revealsthat 12S/13P
that 12S/13Pshows
shows the
thehighest
highest
6 of 18
back-EMF
back-EMF of 44.43 V, whereas the phase back-EMF for 12S/7P, 12S/11P, and 12S/17Pare
of 44.43 V, whereas the phase back-EMF for 12S/7P, 12S/11P, and 12S/17P are
2525V,V,18.13
18.13V,V,
and
and38.63 V,V,
38.63 respectively.
respectively.
Back-EMF (V)

Figure 5. No-load operation phase back-EMF.

Figure 5. No-load operation phase back-EMF.

3.3. Cogging Torque

The torque obtained is cogging torque when no current is applied to the armature
winding. This torque is due to slotting effects that cause acoustic noise and vibration, and
is therefore undesirable. Since slotting effects change with the rotor pole number varia-
Energies 2022, 15, 7531 6 of 17
Figure 5. No-load operation phase back-EMF.

3.3.Cogging
3.3. Cogging Torque
Torque
Thetorque
The torqueobtained
obtainedis iscogging
coggingtorque
torquewhen
whennonocurrent
currentis isapplied
appliedtoto the
the armature
armature
winding. This torque is due to slotting effects that cause acoustic noise and
winding. This torque is due to slotting effects that cause acoustic noise and vibration, and vibration, and
is therefore undesirable. Since slotting effects change with the rotor pole
is therefore undesirable. Since slotting effects change with the rotor pole number variation, number varia-
tion,
the the resultant
resultant coggingcogging torquevary.
torque greatly greatly
Thevary.
coggingThetorques
coggingof torques of 12S/7P,12S/13P,
12S/7P, 12S/11P, 12S/11P,
12S/13P, and 12S/17P are illustrated in Figure 6. In regard to peak
and 12S/17P are illustrated in Figure 6. In regard to peak magnitude, analysis reveals magnitude, analysis
that
revealsdesign
12S/17P that 12S/17P designcogging
offers lowest offers lowest
torque cogging
of 0.0508torque of 0.0508
Nm, whereas Nm, whereas
12S/11P 12S/11P
shows highest
shows torque
cogging highestof cogging
0.337 Nm.torque of 0.337 Nm.
Furthermore, theFurthermore,
cogging torquethe cogging torque
profiles for profiles
12S/13P andfor
12S/13P and 12S/7P show
12S/7P show moderate values. moderate values.

Figure
Figure 6. 6. Variation
Variation of of cogging
cogging torque
torque profile.
profile.

3.4.
3.4.Instantenous Torque
Instantenous Torque
Under
Underloaded
loadedand
andrated
ratedoperating conditionsofofJa𝐽 == J𝐽e = 15
operatingconditions A/mm2 and
15 A/mm andspeed
speedof
of1500
1500rpm,
rpm,instantaneous
instantaneous torque
torque profile
profile is shown
is shown in Figure
in Figure 7. Comparative
7. Comparative analysis
analysis of
vari-
Energies 2022, 15, x FOR PEER REVIEW 7 of
of18
ofous
various
rotorrotor
polespoles illustrates
illustrates that 12S/13P
that 12S/13P offers
offers the the highest
highest torque (𝑇
averageaverage ) ofT4.62
torque avg Nm,
4.62 Nm, whereas
whereas 12S/11P 12S/11P
shows the shows
lowestthe lowesttorque
average averageoftorque of 1.68 Nm.
1.68 Nm.
Table 2 represents average torque, cogging torque, torque ripples ratio, flux linkage,
and back-EMF of possible rotor pole combinations. From the above analysis, it is clear that
Instantaneous torque (Nm)

the 12S/13P combination has a more significant torque density and back-EMF at 𝐽 = 𝐽 =
15 A/mm . Therefore, it is considered for further study.

Figure
Figure 7. 7. Instantaneous
Instantaneous torque
torque profile
profile with
with rotor
rotor pole
pole number.
number.

Table 2. Quantitative
Table 2 representsperformance of the OR-FEFSM
average torque, with various
cogging torque, torquerotor poleratio,
ripples numbers 𝐽 =𝐽 =
fluxatlinkage,
15 A/mm
and .
back-EMF of possible rotor pole combinations. From the above analysis, it is clear
that the 12S/13P combination has 12S/7Pa more significant
12S/11Ptorque density
12S/13Pand back-EMF
12S/17P at
Ja = Je = 15 A/mm2 . Therefore, it is considered for further study.
Average Torque (Nm) 2.72 1.68 4.62 3.72
Cogging Torque P-P (Nm) 0.344 0.377 0.246 0.0508
Torque ripples (%) 32 16 31 12
Flux Linkage (Wb) 0.022816 0.010481 0.021185 0.014152
Back-EMF (V) 25.06 18.13 44.84 38.64

4. Genetic Optimization
Energies 2022, 15, 7531 7 of 17

Table 2. Quantitative performance of the OR-FEFSM with various rotor pole numbers at
Ja = Je = 15 A/mm2 .

12S/7P 12S/11P 12S/13P 12S/17P

Energies 2022, 15, x FOR PEER REVIEW
Average Torque (Nm) 2.72 1.68 4.62 3.72 8 of 18
Cogging Torque P-P (Nm) 0.344 0.377 0.246 0.0508
Torque ripples (%) 32 16 31 12
Flux Linkage (Wb) 0.022816 0.010481 0.021185 0.014152
Back-EMF (V) 25.06 3 ≤18.13
𝑊𝑟𝑦 ≤ 12 44.84 38.64
⎧
2 ≤ 𝑊𝑟𝑝 ≤ 5
⎪
4. Genetic Optimization 3 ≤ 𝑊1 ≤ 8
Design variables =
⎨
In this section, optimization using genetic 3 ≤ algorithm
𝑊2 ≤ 8
(GA) is used for the proposed
⎪ 3 ≤ 𝑊3 ≤ 8
OR-FEFSM with 12S/13P. GA is a widely common solution for optimization issues because
⎩12 ≤ 𝑅1 ≤ 15
it is particularly good at solving highly non-linear objective functions. The GA-based
optimization technique is based on the evolution of the world’s animal population. It begins
Table
with a 3. Variables
random for Geneticof
population Optimization.
variables, similar to a pool of chromosomes. A generation
is the term used to describe
Parameters Initialeach
(mm)iteration Boundary
of the algorithm. Only aOptimized
Conditions few chromosomes
(mm)
with the most significant fitness values are carried down into the next generation, known
𝑊 7.5 3 ≤ 𝑊 ≤ 12 6.3505
as exceptional offspring. In addition, the technique generates new children for the next
𝑊 3 1≤𝑊 ≤5 3.3055
generation by simulating crossover and mutation, which are binary and unary acts on
W 1 6 3 ≤ 𝑊 ≤ 8 7.97017
existing chromosomes, respectively. The method is repeated until one of the end conditions
are metW[27].
2 3 ≤ 𝑊 ≤ 8size are defined
The number of6 generations and population 5.9480
to reach an
optimum 6
W3global value of the objective 3 ≤ 𝑊8 illustrates
function. Figure ≤8 6.8116
the GA workflow.

START

Define initial parameters

Import CAD model

Define objective function

Set constrains

Specify population size and

Crossover
number of generation

Number of Execute FEA

childrens

Check No Objective function

constrains converges?

Yes

No Next
Globally optimized?
generation

Yes

STOP

Figure 8. Flow chart of Genetic Optimization.

Figure 8. Flow chart of Genetic Optimization.

JMAG built-in
Geometry editorglobal optimization,
is used which
to develop the utilizes
initial the GA
design, and approach, was used toare
the CAD parameters op-
timize
then the geometrically
imported significant
into the designer. Theparameters. Geometrical
objective function design variables,
and constraints suchbelow,
are given as the
yoke length (Wry), width of rotor tooth (Wrp), the area of the armature slot, area of the FEC
slot, width of stator poles (W1, W2, W3), and the shaft radius (R1), were used to determine
the optimization problem of the design. During optimization, key dimensions, such as the
stator outer radius (R5), air gap, rotor outer radius (R8), stack length, rated field and arma-
Energies 2022, 15, 7531 8 of 17

composed of two sub-objectives. Table 3 describes the ranges and constraints of variable
CAD parameters.

max( Tavg )
Objective function =
min( Tripples )

Tavg > 4.62
Constraints =
Tripples < 0.28

3 ≤ Wry ≤ 12


 2 ≤ Wrp ≤ 5



3 ≤ W1 ≤ 8

Design variables =

 3 ≤ W2 ≤ 8
3 ≤ W3 ≤ 8




12 ≤ R1 ≤ 15


Table 3. Variables for Genetic Optimization.

Parameters Initial (mm) Boundary Conditions Optimized (mm)

Wry 7.5 3 ≤ Wry ≤ 12 6.3505
Wrp 3 1 ≤ Wrp ≤ 5 3.3055
W1 6 3 ≤ W1 ≤ 8 7.97017
W2 6 3 ≤ W2 ≤ 8 5.9480
W3 6 3 ≤ W3 ≤ 8 6.8116

JMAG built-in global optimization, which utilizes the GA approach, was used to
optimize the geometrically significant parameters. Geometrical design variables, such as
the yoke length (Wry ), width of rotor tooth (Wrp ), the area of the armature slot, area of
the FEC slot, width of stator poles (W 1 , W 2 , W 3 ), and the shaft radius (R1 ), were used to
determine the optimization problem of the design. During optimization, key dimensions,
Energies 2022, 15, x FOR PEER REVIEW 9 of
such as the stator outer radius (R5 ), air gap, rotor outer radius (R8 ), stack length, rated
field and armature current densities, and turns, were kept constant to maintain constant
electrical and magnetic loading. In optimization settings, maximum generations were set
at elements
14, population size atof
and nodes 16,each
number
modelof were
children at 17,
12,246 andand stopping
7672, criteria atThe
respectively. 10. mesh
The size w
number of elements and nodes of each model were 12,246 and 7672, respectively.
set to 0.5. After computing 451 case studies with GA, which took nearly 112 hours, the f The mesh
size was set to 0.5. After computing 451 case studies with GA, which took nearly 112 hours,
size was 110 GB, the PC used was a HP core i5, 2.5 GHz, 8 GB RAM, and optimum valu
the file size was 110 GB, the PC used was a HP core i5, 2.5 GHz, 8 GB RAM, and optimum
werewere
values obtained. Optimization
obtained. Optimizationresults, suchas
results, such astorque
torqueandand ripple
ripple ratioratio
w.r.t,w.r.t,
Wry , WW ry, Wrp, W
rp ,
WW 2, W3, and R1, are illustrated in Figure 9 in relation to one another. Table 3 depicts t
1 , W 2 , W 3 , and R1 , are illustrated in Figure 9 in relation to one another. Table 3 depicts
theproposed
proposeddesign’s globaloptimal
design’s global optimal parameters.
parameters.

Figure 9. Cont.
atio (%)
atio (%)
e (Nm)
e (Nm)
Energies 2022, 15, 7531 9 of 17

Torque ripple ratio (%)

Torque ripple ratio (%)
Average torque (Nm)
Average torque (Nm)

Figure
Figure 9. Torque
9. Torque and and ripples
ripples ratiooptimization
ratio w.r.t w.r.t optimization parameters
parameters (a) R1 (b) (a)
WryR1
(c)(b)
W1W ry (c)
(d) W2W1 (d) W2 (e) W
(e) W3
(f) W
(f) Wr . r.

5. FEA Based Electromagnetic Performance of Optimized Design

In this section, a detailed comparative analysis of optimal design is discussed.

5.1. No Load Analysis

The no-load airgap flux density of 12S/13P initial and optimized design substantially
fluctuates, resulting in a significant variation in the magnetic flux density over one periodic
cycle, as shown in Figure 10. The peak value of airgap flux density of optimized design is
30% more than the initial. The variation of the magnetic flux density results in phase flux
linkage variation. Figure 11a,b shows no load flux linkage and harmonic spectra of initial
and optimized design. Analysis shows that the peak value of flux linkage is improved
by 31.15% after optimization. Similarly, odd harmonics and total harmonic distortion
(THD) were improved by 41.88 and 58.33%. The formula for calculating THD is shown in
(4) below.
∑ N ∅i
THD = i=2 (4)
∅1
where ∅1 is fundamental hormonic component and N is natural number
 tra of initial and optimized design. Analysis shows that the peak value of flux linkage is
phase flux linkage variation. Figure 11a,b shows no load flux linkage and harmonic spec-
improved
tra of initialby 31.15%
and after optimization.
optimized Similarly,
design. Analysis oddthe
shows that harmonics andoftotal
peak value fluxharmonic
linkage is dis-
tortion (THD) were improved by 41.88 and 58.33%. The formula for calculating
improved by 31.15% after optimization. Similarly, odd harmonics and total harmonic dis- THD is
shown in (4) below.
tortion (THD) were improved by 41.88 and 58.33%. The formula for calculating THD is
shown in (4) below. ∑ ∅
Energies 2022, 15, 7531 𝑇𝐻𝐷 = (4)
10 of 17
∅
∑ ∅
𝑇𝐻𝐷 = (4)
where ∅ is fundamental hormonic component ∅ and N is natural number

where ∅ is fundamental hormonic component and N is natural number

Figure 10. Variation of the magnetic flux density over one periodic cycle.
Figure10.
Figure 10.Variation
Variationof
ofthe
themagnetic
magnetic flux density over
over one
oneperiodic
periodiccycle.
cycle.
Flux Linkage
Phase Flux
Phase (Wb)
Linkage(Wb)

Figure
Figure 11. Initial and optimized design phase flux linkage (a) (a)
waveform andand
(b) harmonic spectra.
Figure 11.11.Initial
Initialand
andoptimized
optimizeddesign
design phase
phase flux
flux linkage
linkage waveform
(a) waveform and (b)(b) harmonic
harmonic spectra.
spectra.

Energies 2022, 15, x FOR PEER REVIEW Under

Under no-load
Underno-load condition,
condition,phase-back-EMF
no-loadcondition, phase-back-EMF
phase-back-EMF and harmonic
and
and harmonic
harmonic spectra
spectra
spectra forfor
initial
for and
initial
initial opti-
11 opti-
and
and of 18
opti-
mized design
mizeddesign
design through
through 2D FEA is shown in Figure 12a,b. Analysis reveals that back-EMF
mized through 2D2D
FEAFEA is shown
is shown in Figure
in Figure 12a,b.
12a,b. Analysis
Analysis reveals
reveals thatthat back-EMF
back-EMF is
is
is improved
improved by
by28%,
28%, while
whileodd
odd harmonics
harmonics and THD
and THD is improved
is improvedby by
33.2% andand
33.2% 39.63%,
39.63%,
improved by 28%, while odd harmonics and THD is improved by 33.2% and 39.63%, respec-
tively, after optimization.
respectively, Figure 13
after optimization. shows
Figure 13 the comparison
shows betweenbetween
the comparison initial and optimized
initial and op-
cogging
timized torque;
coggingafter optimization,
torque; the cogging
after optimization, thetorque
cogging is torque
reducedis by 54.93%.
reduced by 54.93%.

60
Initial design
50 Optimized design

40

30

20

10

0
1 2 3 4 5 6 7 8 9 10 11 12
Harmonics Order
(b)

Figure 12. Initial and optimized design back-EMF (a) waveform and (b) harmonic spectra.
Figure 12. Initial and optimized design back-EMF (a) waveform and (b) harmonic spectra.
 20

10

0
1 2 3 4 5 6 7 8 9 10 11 12
Energies 2022, 15, 7531 Harmonics Order 11 of 17
(b)

Figure 12. Initial and optimized design back-EMF (a) waveform and (b) harmonic spectra.

Figure 13. Cogging torque analysis of initial and optimized design.

Figure 13. Cogging torque analysis of initial and optimized design.

5.2.Load
5.2. LoadAnalysis
Analysis
Underloaded
Under loadedandandrated
ratedoperating conditionsofofJa𝐽==Je𝐽 ==15
operatingconditions 15A/mm
A/mm2 and andspeed
speedofof
1500rpm,
1500 rpm,instantaneous
instantaneous torque
torque profile
profile of
of 12S/13P
12S/13P is is shown
shown in in Figure
Figure14, 14,which
whichoffers
offersa
a41.99%
41.99%improvement
improvementinintorque torqueand and87.76%
87.76%ininTorque
Torqueripple
rippleratio(𝑇
ratio ( T).rrFurthermore,
). Furthermore, av-
average torqueprofiles
erage torque profileswith
withtorque
torqueripples
ripplesunder
underdifferent
differentelectric
electricloading
loadingare areillustrated
illustratedin
inFigure
Figure 15,15, demonstrating
demonstrating thethe over-load
over-load capability
capability of the
of the proposed
proposed OR-FEFSM
OR-FEFSM before
before and
and
afterafter optimization.
optimization. It be
It can canseen
be seen that under
that under all-electric
all-electric loading
loading conditions,
conditions, average average
torque
torque increases,
increases, whereas whereas the torque
the torque rippleripple
ratio ratio decreases.
decreases. FromFrom the analysis,
the analysis, the torque
the torque and
and ripples
ripples ratios
ratios areare improved
improved byby 41.99%and
41.99% and87.76%,
87.76%,respectively,
respectively, atat the rated conditions.
conditions.
The
Theformula
formulafor forcalculating
calculatingtorque
torqueripple
rippleratio
ratioisisdiscussed
discussedinin(5).(5).
Energies 2022, 15, x FOR PEER REVIEW
𝑇 −−
Tmax 𝑇
Tmin
Trr𝑇= = ×× 100100 (5)
(5)
𝑇
Tavg
In (5) 𝑇 is the maximum value of torque, 𝑇 is the minimum value of torque,
and 𝑇 is the average value of torque. Figure 16 shows electromagnetic torque perfor-
mance of initial and optimized design w.r.t different electrical degrees.

Figure
Figure 14. 14. Instantaneous
Instantaneous toqueofprofile
toque profile of initial
initial and anddesign.
optimized optimized design.
e ripple ratio (%)
ge torque (Nm)
Energies 2022, 15, 7531 12 of 17

Figure 14. Instantaneous toque profile of initial and optimized design.

Torque ripple ratio (%)

Average torque (Nm)

Torque ripple ratio (%)

Average torque (Nm)

Figure15.
Figure 15.Overload
Overloadcapability
capabilityofofproposed
proposeddesign
designbefore
beforeand
andafter
afteroptimization.
optimization.(a)
(a)Torque
Torquew.r.t
w.r.tJa Ja
and Je before optimization; (b) ripples w.r.t Ja and Je before optimization; (c) torque w.r.t Ja and Je after
and Je before optimization; (b) ripples w.r.t Ja and Je before optimization; (c) torque w.r.t Ja and Je
optimization; and (d) ripples w.r.t Ja and Je after optimization.
after optimization; and (d) ripples w.r.t Ja and Je after optimization.

In (5) Tmax is the maximum value of torque, Tmin is the minimum value of torque,
Energies 2022, 15, x FOR PEER REVIEW 13 of 18
and Tavg is the average value of torque. Figure 16 shows electromagnetic torque perfor-
mance of initial and optimized design w.r.t different electrical degrees.
Average torque (Nm)

Figure 16.16.Torque
Figure Torquew.r.t
w.r.tdifferent electrical
different degrees
electrical before
degrees and
before after
and optimization.
after optimization.

5.3. Dynamic Analysis

In designing an electric machine, dynamic characteristics analysis is one of the key
studies that provides wide operational capability. In this regard, torque and output
power, characteristics curve versus speed and power factor, give a detailed illustration of
its behavior under low-speed, high-speed, constant torque, and constant output power
regions, which is a prerequisite for direct-drive in-wheel-embedded systems.
Energies 2022, 15, 7531 13 of 17

5.3. Dynamic Analysis

Energies 2022, 15, x FOR PEER REVIEW In designing an electric machine, dynamic characteristics analysis is one of the 14 key
of 18
studies that provides wide operational capability. In this regard, torque and output power,
characteristics curve versus speed and power factor, give a detailed illustration of its
behavior under low-speed, high-speed, constant torque, 𝑃 and constant output power regions,
𝑃𝐹 = 𝑐𝑜𝑠∅ =
which is a prerequisite for direct-drive in-wheel-embedded systems. (15)
𝑆
The output power of the proposed machine is calculated by multiplying Tavg with
The power factor
the corresponding speed.of Input
initialpower
and optimized design is of
is the summation shown in power
output Figure and
18. total losses,
A comparison of the initial and optimized OR-FEFSM designs is performed in order
including core and copper losses. Core losses can be calculated from 2D-FEA at the specified
to verify
points, andand
(6) examine the
is used for performance
copper of the GA optimization technique. Table 4 illus-
losses calculation.
trates a quantitative assessment of the proposed machine.
PCopper = PCopper ( AW ) + PCopper ( FEC ) (6)
Table 4. Performance comparison of initial and optimized design.
where,
Parameters Initial Value Optimized Value Improvement
PCopper ( AW ) = Irms ρJL( NQ)(1000) (7)
Torque (𝑁𝑚) 4.877 6.925 41.99%
Torque Ripples (%) PCopper ( FEC) =0.30
IρJL( NQ)(1000) 0.037 87.76% (8)
FluxInlinkage (𝑊𝑏)
(8), I represent 0.021 J represents current
current, ρ is resistivity, 0.0276density, N 23.91%
and Q are
Airgap flux density (𝑇)
number of turns and slot pair, respectively. 1.3886 1.9276 27.94%
Back-emf peak(𝑉) 41.07 55.362 25.91%
Cogging torque(𝑁𝑚) = Tavg ∗ ω
Pout0.3160 0.1424 54.93% (9)
Odd harmonics in flux (Wb) 0.000573 0.000333 41.88%
Pinput = Pout + PCopper + PCore (10)
THD in flux (%) 2.88 1.2 58.33%
OddFinally, efficiency
harmonics is given as:
in back-EMF
6.1206 4.09209 33.2%
(V)
Pout
THD in back emf (%) η = 16.4 100 % 9.9 39.63%(11)
Pinput
Output power (𝐾𝑤) 2.27 3.223 29.56%
Power
Thefactor (𝑃𝑓)
characteristics curve of torque and 0.84
power versus speed0.97 13.4%
on initial and optimized
Efficiency (%) 83.4
designs is depicted in Figures 17a and 17b, respectively. 87.70 4.9%
Average torque (Nm)

Output power (kW)

Figure 17. (a) Torque vs speed curve; (b) power vs speed curve.
Figure 17. (a) Torque vs. speed curve; (b) power vs. speed curve.

Similarly, the power factor can be calculated as:

3

P= V Iq + Vq Id (12)
2 d
3

Q= V Iq + Vq Id (13)
2 d
p
S = P2 + Q2 (14)
Energies 2022, 15, 7531 14 of 17

Energies 2022, 15, x FOR PEER REVIEW P 15 of 18

PF = cos∅ = (15)
S
The power factor of initial and optimized design is shown in Figure 18.

Figure18.
Figure 18.Power
Powerfactor
factorofofinitial
initialand
andoptimized
optimizeddesign
designw.r.t
w.r.t different
different speed.
speed.

A comparison
6. Thermal of the initial and optimized OR-FEFSM designs is performed in order to
Analysis
verify Thermal
and examine the performance
analysis of theimportant
is one of the most GA optimization technique.
analyses in electricTable 4 illustrates
machine design, aas
quantitative assessment of the proposed machine.
heat dissipation limits the long-term functionality of the machine. Therefore, it helps to
determine the operating limits and the insulating class of the machine. The electromag-
Table 4. Performance comparison of initial and optimized design.
netic performance reduces when the temperature exceeds a certain permissible range, re-
sulting inParameters
an inter-turn short circuit problem
Initial Value [28]. Optimized
Current flowValuethrough the conductor
Improvement
produces
Torque (Nm) heat dissipation and power loss,
4.877 which raises the temperature
6.925 of the whole ma-
41.99%
chine. The machine
Torque Ripples (%) has core losses in addition
0.30 to the copper
0.037 losses associated
87.76%with the
windings,
Flux linkagewhich
(Wb) are due to the eddy current 0.021 and hysteresis losses in the core.
0.0276 All of these
23.91%
Airgap flux density ( T ) 1.3886 1.9276
losses behave as a heat source and raise temperatures. Therefore, the magnetic loss study 27.94%
Back
is used− emf peak (V)
to determine 41.07 analysis is used
the losses, and thermal 55.362to analyze the25.91%
temperature
Cogging torque(Nm) 0.3160 0.1424 54.93%
distribution.
Odd harmonics in flux (Wb) 0.000573 0.000333 41.88%
THD For the(%)
in flux proposed machine, initial 2.88 losses were determined 1.2 using a 3D FEA 58.33%followed
by thermal analysis in
Odd harmonics in back-EMF (V) 3D thermal studies,
6.1206 as 3D analysis is more
4.09209 precise than 33.2%analysis.
2D
To
THD calculate the(%)
in back emf temperature distribution 16.4 of the complete9.9 machine, the 3D39.63%loss study is
Output power
combined (Kw
with the) 3D heat study. The2.27 3D thermal study3.223 29.56%
reveals that the stator tempera-
Power
ture factor (Pf) rises due to the presence
significantly 0.84 0.97 on the stator;13.4%
of all excitation sources however, the
Efficiency (%) 83.4 87.70 4.9%
temperature of the rotor minimally rises, as the proposed machine has a salient rotor. Figure
19 depicts a contour plot of the temperature distribution. In the stator, the temperature dis-
6. Thermal Analysis
tribution in the windings reaches a maximum value of 46 °C. At the rotor, the highest tem-
Thermal
perature analysis
observed is 24is one of the most important analyses in electric machine design, as
°C degrees.
heat dissipation limits the long-term functionality of the machine. Therefore, it helps to
determine the operating limits and the insulating class of the machine. The electromagnetic
performance reduces when the temperature exceeds a certain permissible range, resulting
in an inter-turn short circuit problem [28]. Current flow through the conductor produces
heat dissipation and power loss, which raises the temperature of the whole machine. The
machine has core losses in addition to the copper losses associated with the windings, which
are due to the eddy current and hysteresis losses in the core. All of these losses behave as a
heat source and raise temperatures. Therefore, the magnetic loss study is used to determine
the losses, and thermal analysis is used to analyze the temperature distribution.
For the proposed machine, initial losses were determined using a 3D FEA followed
by thermal analysis in 3D thermal studies, as 3D analysis is more precise than 2D analy-
sis. To calculate the temperature distribution of the complete machine, the 3D loss study
Energies 2022, 15, 7531 15 of 17

is combined with the 3D heat study. The 3D thermal study reveals that the stator tem-
perature significantly rises due to the presence of all excitation sources on the stator;
however, the temperature of the rotor minimally rises, as the proposed machine has a
Energies 2022, 15, x FOR PEER REVIEW
salient rotor. Figure 19 depicts a contour plot of the temperature distribution. In the 16 of 18
stator,
◦
the temperature distribution in the windings reaches a maximum value of 46 C. At the
rotor, the highest temperature observed is 24 ◦ C degrees.

Figure19.
Figure 19.Thermal
Thermalanalysis
analysisofofthe
theproposed
proposedOR-FEFSM.
OR-FEFSM.

7.7.Comparison
Comparisonof ofProposed
Proposedand andConventional
ConventionalDesign Design
In
In this section, a detailed comparison of theconventional
this section, a detailed comparison of the conventionaland andproposed
proposeddesigndesignisis
performed. For fair comparison with [23], the proposed model is redesigned
performed. For fair comparison with [23], the proposed model is redesigned and opti- and optimized
under
mizedthe same
under theouter
samedimensions, stack length,
outer dimensions, currentcurrent
stack length, density, and airand
density, gap airlength. The
gap length.
average torque of scaled 12S/13P design is 47.95 Nm which is 36.6%,
The average torque of scaled 12S/13P design is 47.95 Nm which is 36.6%, 57.1%, 48.77%, 57.1%, 48.77%,
and
and96.52%
96.52% more than the
more than the conventional
conventionalstraight
straightWFSM,
WFSM,skewed
skewed WFSM,
WFSM, 12S/8P
12S/8P SRM,SRM,
and
and 12S/10P SRM, respectively. Based on the key performance matrix listed
12S/10P SRM, respectively. Based on the key performance matrix listed in Table 5, without in Table 5,
without any influence
any influence on efficiency,
on efficiency, it is that
it is clear clearthe
thatproposed
the proposed design
design offers
offers 9.4%,9.4%, 65.8%,
65.8%, and
and 63.7% less ripples than the conventional straight WFSM, 12S/8P SRM, and 12S/10P
63.7% less ripples than the conventional straight WFSM, 12S/8P SRM, and 12S/10P SRM,
SRM, respectively. Furthermore, individual machine part weight and overall weight
respectively. Furthermore, individual machine part weight and overall weight of the pro-
of the proposed machine are the lowest, resulting in a better torque density than the
posed machine are the lowest, resulting in a better torque density than the aforementioned
aforementioned state-of-the-art model.
state-of-the-art model.
Table 5. Performance comparison of conventional and proposed design.
Table 5. Performance comparison of conventional and proposed design.
Straight Skewed 12/8 12/10 Proposed
Parameters Straight WFSMSkewed SRM12/8 12/10 Proposed
WFSM SRM Machine
Parameters
Average Torque (Nm ) 35.1WFSM 30.52 WFSM 32.32SRM 28.4SRM Machine
47.95
Average Torque
Torque Ripple (%)(𝑁𝑚) 27.6 35.1 6.8530.52 73.132.32 68.928.4 47.95
25
Efficiency ( % )
Torque Ripple (%) 89.4 27.6 88 6.85 89.9 73.1 90.568.9 89.325
Weight of iron
Efficiency (%)core (kg) 19.87
89.4 88 19.5489.9 19.7490.5 16.19
89.3
Weight of copper (kg) 17.86 13.74 12.63 3.121
Weight
Weight inoftotal
iron(kgcore
) (𝑘𝑔) 37.73 19.87 34.3319.54 32.3719.74 16.19
19.311
Weight of copper
Torque Density (𝑘𝑔) )
(Nm/kg 0.930 17.86
0.808 0.94113.74 0.87712.63 3.121
2.483
Weight in total (𝑘𝑔) 37.73 34.33 32.37 19.311
8.Torque Density (𝑁𝑚/𝑘𝑔)
Conclusions 0.930 0.808 0.941 0.877 2.483
This paper proposed a novel outer rotor field-excited flux-switching machine with a
8. Conclusions
combined semi-closed and open slots stator to improve power factor and torque profile.
This paper proposed a novel outer rotor field-excited flux-switching machine with a
combined semi-closed and open slots stator to improve power factor and torque profile.
Semi-closed slots utilization improves power factor, whereas open slots suppress slot
leakages that effectively suppress leakage reluctance that causes flux pulsation, improv-
Energies 2022, 15, 7531 16 of 17

Semi-closed slots utilization improves power factor, whereas open slots suppress slot
leakages that effectively suppress leakage reluctance that causes flux pulsation, improving
torque ripples, and mean torque. The effectiveness of the proposed OR- FEFSM was
investigated through FEA with different stator slot and rotor pole combinations. Initial
FEA-based comprehensive electromagnetic performance reveals that 12S/13P offers the
highest average torque of 4.62 Nm, which was considered for further study and optimized
through GA optimization. The maximum output power obtained is 3.22 kW with an
efficiency of 87.7%. Finally, the proposed design was compared with the state-of-the-art
model, and the results showed there was significant improvement in average torque, torque
ripples, weight of machine, and torque density.

Author Contributions: Conceptualization, S.A., W.U. and F.K.; methodology, S.A. and B.U.; software,
S.A. and B.U.; validation, S.A. and W.U.; resources, F.K., A.H.M. and A.A.A.; formal analysis, B.U.,
A.H.M. and A.A.A.; original draft preparation, S.A. and W.U.; visualization, B.U.; review and editing,
B.U. and F.K.; supervision, F.K.; project administration, F.K., A.H.M. and A.A.A.; funding acquisition,
A.H.M., A.A.A. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Ullah, W.; Khan, F.; Sulaiman, E.; Umair, M.; Ullah, N.; Khan, B. Analytical validation of novel consequent pole E-core stator
permanent magnet flux switching machine. IET Electr. Power Appl. 2020, 14, 789–796. [CrossRef]
2. Jiang, W.; Huang, W.; Lin, X.; Zhao, Y.; Zhu, S. Analysis of Rotor Poles and Armature Winding Configurations Combinations of
Wound Field Flux Switching Machines. IEEE Trans. Ind. Electron. 2021, 68, 7838–7849. [CrossRef]
3. Jiang, W.; Huang, W.; Lin, X.; Zhao, Y.; Wu, X.; Zhao, Y.; Dong, D.; Jiang, X. A Novel Stator Wound Field Flux Switching Machine
with the Combination of Overlapping Armature Winding and Asymmetric Stator Poles. IEEE Trans. Ind. Electron. 2022, 69,
2737–2748. [CrossRef]
4. Zulu, A.; Mecrow, B.; Armstrong, M. A Wound-Field Three-Phase Flux-Switching Synchronous Motor with All Excitation Sources
on the Stator. IEEE Trans. Ind. Appl. 2010, 46, 2363–2371. [CrossRef]
5. Tang, Y.; Paulides, J.; Motoasca, T.; Lomonova, E. Flux-Switching Machine with DC Excitation. IEEE Trans. Magn. 2012, 48,
3583–3586. [CrossRef]
6. Wu, Z.; Zhu, Z.; Wang, C. Reduction of On-Load DC Winding-Induced Voltage in Partitioned Stator Wound Field Switched Flux
Machines by Dual Three-Phase Armature Winding. IEEE Trans. Ind. Electron. 2022, 69, 5409–5420. [CrossRef]
7. Ullah, W.; Khan, F.; Hussain, S. A Comparative Study of Dual Stator with Novel Dual Rotor Permanent Magnet Flux Switching
Generator for Counter Rotating Wind Turbine Applications. IEEE Access 2022, 10, 8243–8261. [CrossRef]
8. Tang, Y.; Motoasca, E.; Paulides, J.; Lomonova, E. Comparison of flux-switching machines and permanent magnet synchronous
machines in an in-wheel traction application. COMPEL-Int. J. Comput. Math. Electr. Electron. Eng. 2012, 32, 153–165. [CrossRef]
9. Othman, S.; Ahmad, M.; Rahim, J.; Bahrim, F.; Sulaiman, E. Design Improvement of Three Phase 12Slot-14Pole Outer Rotor Field
Excitation Flux Switching Motor. Int. J. Power Electron. Drive Syst. (IJPEDS) 2017, 8, 239. [CrossRef]
10. Shen, Y.; Hu, P.; Jin, S.; Wei, Y.; Lan, R.; Zhuang, S.; Zhu, H.; Cheng, S.; Chen, J.; Wang, D.; et al. Design of Novel Shaftless
Pump-Jet Propulsor for Multi-Purpose Long-Range and High-Speed Autonomous Underwater Vehicle. IEEE Trans. Magn. 2016,
52, 1–4. [CrossRef]
11. Su, P.; Hua, W.; Wu, Z.; Chen, Z.; Zhang, G.; Cheng, M. Comprehensive Comparison of Rotor Permanent Magnet and Stator
Permanent Magnet Flux-Switching Machines. IEEE Trans. Ind. Electron. 2019, 66, 5862–5871. [CrossRef]
12. Hua, W.; Cheng, M.; Zhu, Z.; Howe, D. Analysis and Optimization of Back EMF Waveform of a Flux-Switching Permanent
Magnet Motor. IEEE Trans. Energy Convers. 2008, 23, 727–733. [CrossRef]
13. Du, Y.; Zhang, C.; Zhu, X.; Xiao, F.; Sun, Y.; Zuo, Y.; Quan, L. Principle and analysis of doubly salient PM motor with Π-shaped
stator iron core segments. IEEE Trans. Ind. Electron. 2018, 66, 1962–1972. [CrossRef]
14. Ullah, W.; Khan, F.; Hussain, S. A Novel Dual Rotor Permanent Magnet Flux Switching Generator for Counter Rotating Wind
Turbine Applications. IEEE Access 2022, 10, 16456–16467. [CrossRef]
15. Chau, K.; Chan, C.C.; Liu, C. Overview of permanent-magnet brushless drives for electric and hybrid electric vehicles. IEEE Trans.
Ind. Electron. 2008, 55, 2246–2257. [CrossRef]
16. Du, Z.S.; Lipo, T.A. Efficient utilization of rare earth permanent-magnet materials and torque ripple reduction in interior
permanent-magnet machines. IEEE Trans. Ind. Appl. 2017, 53, 3485–3495. [CrossRef]
17. Dorrell, D.; Parsa, L.; Boldea, I. Automotive Electric Motors, Generators, and Actuator Drive Systems with Reduced or No
Permanent Magnets and Innovative Design Concepts. IEEE Trans. Ind. Electron. 2014, 61, 5693–5695. [CrossRef]
Energies 2022, 15, 7531 17 of 17

18. Boldea, I.; Tutelea, L.N.; Parsa, L.; Dorrell, D. Automotive Electric Propulsion Systems with Reduced or No Permanent Magnets:
An Overview. IEEE Trans. Ind. Electron. 2014, 61, 5696–5711. [CrossRef]
19. Yang, G.; Lin, M.; Li, N.; Tan, G.; Zhang, B. Flux-Weakening Control Combined with Magnetization State Manipulation of Hybrid
Permanent Magnet Axial Field Flux-Switching Memory Machine. IEEE Trans. Energy Convers. 2018, 33, 2210–2219. [CrossRef]
20. Wardach, M.; Palka, R.; Paplicki, P.; Prajzendanc, P.; Zarebski, T. Modern Hybrid Excited Electric Machines. Energies 2020, 13, 5910.
[CrossRef]
21. Zhang, G.; Hua, W.; Cheng, M.; Liao, J.; Wang, K.; Zhang, J. Investigation of an improved hybrid-excitation flux-switching
brushless machine for HEV/EV applications. IEEE Trans. Ind. Appl. 2015, 51, 3791–3799. [CrossRef]
22. Pollock, C.; Pollock, H.; Barron, R.; Coles, J.R.; Moule, D.; Court, A.; Sutton, R. Flux-switching motors for automotive applications.
IEEE Trans. Ind. Appl. 2006, 42, 1177–1184. [CrossRef]
23. Zhao, G.; Hua, W.; Qi, J. Comparative Study of Wound-Field Flux-Switching Machines and Switched Reluctance Machines.
IEEE Trans. Ind. Appl. 2019, 55, 2581–2591. [CrossRef]
24. Zhu, Z.Q.; Wu, Z.Z.; Evans, D.J.; Chu, W.Q. A Wound Field Switched Flux Machine with Field and Armature Windings Separately
Wound in Double Stators. IEEE Trans. Energy Convers. 2015, 30, 772–783. [CrossRef]
25. Yang, S.; Zhang, J.; Jiang, J. Modeling Torque Characteristics and Maximum Torque Control of a Three-Phase, DC-Excited
Flux-Switching Machine. IEEE Trans. Magn. 2016, 52, 1–4. [CrossRef]
26. Nguyen, H.Q.; Jiang, J.Y.; Yang, S.M. Design of a wound-field flux switching machine with dual-stator to reduce unbalanced shaft
magnetic force. J. Chin. Inst. Eng. 2017, 40, 441–448. [CrossRef]
27. Parappagoudar, M.B.; Pratihar, D.K.; Datta, G.L. Forward and reverse mappings in green sand mould system using neural
networks. Appl. Soft Comput. 2008, 8, 239–260. [CrossRef]
28. Zhao, J.; Guan, X.; Li, C.; Mou, Q.; Chen, Z. Comprehensive evaluation of inter-turn short circuit faults in pmsm used for electric
vehicles. IEEE Trans. Intell. Transp. Syst. 2020, 22, 611–621. [CrossRef]