First Insights on the Electromagnetic Design of

Axial-Flux Synchronous-Reluctance Maschine

1st Adrian-Cornel Pop 2nd Florin Pop Piglesan 3rd Radu Martis
Advanced Development Group Electrical Machines and Drives Electrical Machines and Drives
Brose Fahrzeugteile GmbH & Co. KG., Technical University of Cluj-Napoca Technical University of Cluj-Napoca
Würzburg, Germany Cluj-Napoca, Romania Cluj-Napoca, Romania
adriancornel.pop@brose.com florin.pop@emd.utcluj.ro radu.martis@emd.utcluj.ro

4th Ioana Vintiloiu 5th Claudia Martis

Advanced Development Group Electrical Machines and Drives
Brose Fahrzeugteile GmbH & Co. KG., Technical University of Cluj-Napoca
Würzburg, Germany Cluj-Napoca, Romania
ioana.vintiloiu@brose.com claudia.martis@emd.utcluj.ro

Abstract—This paper presents a new topology of synchronous brushed-DC Machines (employed in applications demanding
reluctance machine, based on the axial-flux concept (AF-SynRM). low cost, power and dynamic such as window lifters or
As case study, an electric drive used in a cooling fan module various seat adjusters) and the most common type, Permanent
for vehicles is considered, but the applications field can be
much wider. At first, the existing/reference radial flux Interior Magnet Synchronous Machines (PMSMs). Their choice is
Permanent Magnet Synchronous Motor (IPMSM) has been governed by three important factors: cost, weight and size
modeled by means of Finite Element Method (FEM) to check [3], [4], at given torque-speed requirements and input power
the electromagnetic behavior, (i.e. flux linkage and torque). The specifications. When high performance criteria are demanded,
next step consists in the design and optimization of a concentrated the PM-based synchronous machine is first choice, thanks
winding, magnet-free, radial flux Synchronous Reluctance Ma-
chine (SynRM). Following that, the newly proposed AF-SynRM to its higher torque density. In order to reduce the motor
is benchmarked against the both magnet-based and magnet- size/mass (at given torque), PMs are typically added on the
free radial flux (RF) counterparts. As the reference machines rotor side (for both radial or axial flux designs) to increase
are employing a fractional slot concentrated winding, the same the flux density, without increasing the size of the magnetic
is done for the AF-SynRM, even though superior behavior or electrical circuit [3].
is expected when considering its distributed variant. Results
obtained during the 3D FE-based optimization are reported However, despite their attractiveness, the limited availability
highlighting the main challenges needed to be addressed (e.g. and challenges associated with the mining and processing of
higher torque ripple and q-axis inductance). Being a motor rare earth materials, inherently imply increased costs of the
topology that is rather complex regarding its manufacturability, PM-based machines [4], [5]. For these reasons, and thanks to
possible directions and their problems are discussed. the development of new materials and concepts, PM-less motor
Index Terms—automotive applications, axial-flux machines,
cooling-fan applications, electric drives, electromagnetic design, topologies such as Synchronous Reluctance Motors (SynRM)
synchronous reluctance motors are perceived as a promising alternative.
Typical usage of Axial Flux Motors (AFMs - often named
I. I NTRODUCTION disk-type motors [6]) in automotive can be observed especially
Technological innovations in the mobility sector perma- for applications where there are stringent limitations regarding
nently trigger new developments in electric machinery tech- the axial length (e.g. fans, in-wheel traction motors). In this
nology. Many electric drive systems have been developed for context AFMs can compete against the commonly used RF-
automotive applications [1], [2], traditionally for auxiliaries machines, thanks to their compact geometrical shape and
and more recently for traction. In a quest to reduce the excellent power/torque density [7]. Despite the fact that nu-
fuel consumption and the harmful emissions, the deployed merous motor topologies and geometries for different applica-
electrical machines should fulfill various conflicting require- tions are reported as an alternative against RF-machines, they
ments (i.e. high torque and power density, wide speed range, didnt succeed to establish themselves as a mature technology
enhanced torque quality, low NVH and increased robustness). because of their drawbacks which are mainly related to the
Depending on the application, several types of electric manufacturing difficulties as well as the strong axial magnetic
motors, the vast majority in radial flux configuration, have attraction forces [8].
found their way in this field: Induction Machines (for traction, This work aims to report the first results on the design and
successfully used in Tesla vehicles), Switched Reluctance Ma- optimization of an axial flux SynRM. For that, the proposed
chines (e.g. for high-speed applications such as turbocharging), motor topology is benchmarked against a ferrite spoke PMSM

l-))) 
(the reference machine) and a PM-free SynRM, both of radial Finally, one may notice in Table I that all three motors
flux type. Their cross-sections along with the requirements have a different airgap width. The spoke-type PMSM has
and specifications of a 300W drive used for a car’s cooling a rather thin web - the small part connecting the hub (the
fan are provided in Section II. In Section III, a comparative concentric cylinder around the shaft) with the iron pole - see
study between all three topologies is performed by means also Fig. 2a. As the latter along with the magnets has to be held
of FE-based electromagnetic analysis in order to evaluate entirely by this thin part against the acting of the centrifugal
their key-performance indicators, i.e. power to weight ratio, force, the rotor is molded in plastic, meaning that a larger
average torque and torque ripple as well as their speed-torque airgap is needed. On the other hand, for the RF-SynRM, the
(and -efficiency, respectively) characteristics, while keeping FEA-based structural analysis (Fig. 1), performed in ANSYS
the outer diameter fixed (installation space constraint), and Mechanical, shows that the maximum von-Mises stress is
assuming concentrated windings for all three designs. In Sec- below the used steel’s yield point (i.e. 300 N/mm2 ), when
tion IV, the results obtained during the optimization process considering a minimum thickness of the bridges of 0.8mm.
of the proposed AF-SynRM are provided. Therefore no additional fixation measures are needed, thus the
To the knowledge of the authors, there are no previous airgap can be reduced. Furthermore, unlike the PMSM where
attempts neither to design nor to manufacture such a machine, decreasing the airgap would inherently mean an increase in
therefore following the numerical and optimization sections, cogging torque, for the PM-free SynRM topologies it is not
Section V is dedicated to the more challenging production the case.
process and possible materials that could be used. In this For improved visualization purposes, the model consists of
respect, possibilities, ideas and challenges are highlighted three laminations (3 × 0.5mm), whereas in order to increase
along with suggestions for future works to be addressed by the accuracy of the computation, the mesh is refined around
other researchers interested in this topic. Finally, conclusions the thinnest parts of the rotor (Fig. 1a). Consequently, in these
are drawn in the last section. sensitive regions, a maximum von-Mises stress around 210
N/mm2 is obtained at rated speed (well below the maximum
II. C ONSIDERED MOTORS , REQUIREMENTS AND allowed point). For safety reasons, computations have been
SPECIFICATIONS
performed and it is shown that the rotor can withstand forces
As design constraints, both the RF- and AF-SynRM are corresponding to speed values up to 5000rpm, while ensuring
sized having the same outer diameter Dout , and in the first full integrity.
step, the same axial length LstkS as the reference PMSM.
Furthermore, the input power has to be the same, i.e. battery
voltage (13V), RMS phase current (17A) at given DC and
AC resistances of the electrical circuit between the battery
and the motor (power electronic converter, cables). Another
rather noteworthy constraint is that all three topologies are
employing two-layer concentrated winding [9]. Finally, the
newly designed motors should have a rated speed Ωb of
2700rpm and a rated torque Tb of 0.85Nm. An overview of
the motor topologies key parameters along with those of the
electrical circuit are listed in Table I.
(a) (b)
TABLE I: Configurations, magnetic materials and electrical
circuit parameters of the analyzed motors.

RF-IPMSM RF-SynRM AF-SynRM

Number of stator slots Ns 12 12 12

Number of rotor pole pairs Npp 5 2 2

PM material type ferrite - -

PM remanence Br [T] 0.4 - -

(c) (d)
Radial air-gap length g [mm] 0.6 0.4 0.5
Fig. 1: Results of the static structural analysis at (b) rated
Winding connection delta star star speed (2700rpm) and (c, d) at a higher speed for checking the
Wire diameter [mm] 1.18 1.8 1.8 over-speed capability (5000rpm).
Number of turns 22 18 18
As for the AF-SynRM, in comparison with its RF counter-
Number of parallel paths 2 1 1 part, a slightly larger airgap value is assumed as reserve in
Phase resistance [mΩ] 8.8 27.6 21.5 case of problems caused by axial unbalance (typical for axial
flux machines). However, this value is far from being neither


numerically nor experimentally proven, as a AF-SynRM’s
detailed mechanical analysis is out of the scope of this paper.

A. Spoke-type ferrite-based RF Synchronous Motor

The considered 12 slots 8 pole, ferrite-based, spoke-type
reference in RF-IPMSM topology is widely used in automotive
applications for relatively low-power density, low-cost appli-
cations. Its cross-section being shown in Fig. 2a.

Fig. 3: 3D-view of the 12/4 Axial Flux Synchronous Reluc-

tance Machine.

A. FE-Electromagnetic analysis

The field analysis is conducted using the commercial FEA

software Ansys Electronics (formerly Maxwell). The variation
(a) (b) of the magnetic parameters is obtained using a current-fed FE-
Fig. 2: Cross-sections of the analyzed RF machines. Model.
1) Radial Flux Synchronous Machines: At first the RF-
SynRM is designed while considering (as constraints) that
B. RF-Synchronous Reluctance Motor the motor has to fit into the same installation space as the
Nowadays, the design and development of synchronous PMSM (i.e. fixed outer diameter and stack length), and to
reluctance machines is a common research topic [10], [11]. be supplied with the same input power. Obviously, this is
However, there are only few works dealing with the fractional a though requirement with the lack of the magnets. Another
slot concentrated SynRMs [12]. Furthermore, since most of strong limitation which hinders the possibility to build up a
the activities are being focused on high power electric drives, reasonable saliency is the concentrated winding. The reluc-
there are even less reported results when it comes to low tance torque produced by the SynRM depends on the saliency
power applications as the ones met in automotive auxiliaries. effect and is proportional to the difference between the d- and
For this study, a 12 slots 4 pole four flux-barriers per pole q-axis inductances (Ld and Lq ). Under these circumstances,
SynRM with concentrated winding is considered, its cross- the resulted design has a average electromagnetic torque, that
section being shown in Fig. 2b. is -17% below the target and a higher torque ripple (30%).
In a second step, the analysis is done by allowing both
C. AF-Synchronous Reluctance Motor the current and the stack length to exceed the values of the
When it comes to axial flux concepts, similar to the radial reference machine. This (resized) design is further denoted as
case, the most common one is the Permanent Magnet Syn- RF-SynRM-Re. In Fig. 4 the torque of the first SynRM design
chronous type (AF-PMSM), with a large variety of designs (RF-SynRM-ref) and that of the resized one are shown in red
in various configurations [13], [14]. At least at research and blue, respectively. It can also be seen that if the current
level, there are also reported works based on the switched is increased with 41% and the axial length with 30%, the RF-
reluctance concept (AF-SRM). The 12/4 variant with four- SynRM meets the same performances as the reference PMSM.
barriers designed for the RF case, is also considered in the Obviously, higher drive costs are expected as higher current
AF one, a view with one of the designs considered in this cause higher VA-requirements of the inverter components.
paper being shown in Fig. 3. The two RF-SynRMs are being prototyped at the moment of
writing this paper, side and top views of the rotor stacks being
III. FE- BASED COMPARATIVE ANALYSIS OF THE shown in Fig. 5 and Fig. 6, respectively. Furthermore, in the
MACHINES UNDER STUDY same figure, the rotor of the reference spoke-design PMSM
In this section the comparison between the three types of can be seen as well.
synchronous motors, namely RF-IPMSM, RF-SynRM and AF- 2) Axial Flux Synchronous Machines: As the axial flux
SynRM is performed, highlighting various key performance machines are belonging to class 3D topologies, the most
indicators both at electromagnetic (e.g. torque density and accurate way for modeling these machine is by means of three-
quality) and electromechanical level (e.g. torque/efficiency- dimensional finite element method (3D-FEM). For one of
speed characteristic). the best designs, extracted from the pareto-front optimization,


 B. Best Design Assessment
The resized RF-SynRM is further used in comparison with
the reference PMSM and the AF-SynRM which is briefly
described in Section III-A2.
1) Electromagnetic torque waveforms: For the three con-
sidered motors, the waveforms of the electromagnetic torque
as a function of normalized electrical angle (half of an
electrical period) are shown in Fig. 8. At first, one can see
in blue dashed line the FE-obtained electromagnetic torque
for the PM-based topology at rated load conditions (i.e. 17A
RMS phase current). The average torque is 0.85 Nm, while its
relative ripple is around 9%. As it is also reported in Section
Fig. 4: Electromagnetic torque vs. angle (in electric cycles) III-A1, the resized RF-SynRM produces the same torque (0.85
for the RF-SynRMs. Nm) but the torque ripple is larger (34%), when supplied with
≈ 40% higher current - as shown in black in the same figure.

Fig. 5: Side-view with the rotor stacks of (from the left to the
right) SynRM-ref, PMSM, SynRM-re.

Fig. 8: Electromagnetic torque vs. angle (in electric cycles).

Finally, the best AF-design obtained after conducting the

first optimization loop, develops an electromagnetic torque
as the one shown in red in Fig. 8, when supplied with
17A RMS phase current. It can be seen that the AF-design
Fig. 6: Side-view with the rotor stacks of (from the left to the
develops the target average torque at given RMS current and
right) SynRM-ref, PMSM, SynRM-re.
installation space limitation, thus being superior to its radial
flux counterpart. In terms of torque density it might even
challenge the PMSM. However, regarding the torque quality
presented in Section IV the magnitude of the magnetic flux there is still some work to be done at optimization level as a
density can be seen in Fig. 7. value of ≈ 55% is reported for the peak-to-peak torque ripple.
2) Torque/Efficiency-Speed Characteristics: Fig. 9 shows
the torque-speed characteristics for all motor topologies dis-
cussed in this paper. The PMSM generates the highest torque
at the constant torque region. Furthermore, it can be seen that
the AF-SynRM topology exhibits better torque performances
against the RF-SynRM when both topologies are using the
reference machine input parameters. Finally, it can be seen
that for the SynRM-re it is quite challenging to reach the
base point without over-sizing (more torque than needed below
the base point). This happens mainly because, unlike in the
PMSM case where the base point can be easily reached by
adjusting the number of ampere-turns and the stack length,
in the SynRM case by employing these degrees of freedom
Fig. 7: AF-SynRM: Magnitude of the magnetic flux density
(DoFs) the saliency can be undesirably altered. Thus designing
at rated current.
a motor with proper saliency becomes a non-trivial task due
to the non-linearities of the magnetic circuit.


 TABLE II: Active material weights and their costs.

RF-IPMSM RF-SynRM-Ref RF-SynRM-Re AF-SynRM

Steel [g] 528 953 1237 853

Copper [g] 164 300 339 318

Magnet [g] 147 - - -

Total mass [g] 839 1253 1576 1171

Total costs [e] 2.14 2.15 2.59 2.12

Δ costs [%] 0 +0.63 +21.2 -0.7

Fig. 9: Torque vs. speed characteristics for the considered

motors.

Finally, the superiority of the newly proposed axial flux

motor can also be seen from the efficiency vs. speed char-
acteristics shown in Fig. 10, where one can observe that the
AF-SynRM has the highest efficiency in the class of PM-free
motors.

Fig. 11: AF-SynRM: Magnetic flux density streamlines on the

partial FE-model used in the optimizations at rated current

In Fig. 12 Rsin and Rsout are the inner and the outer stator
radius, respectively, Yt is the stator yoke thickness, Th is the
tooth height (excluding the tooth tip), Sow is the slot opening,
St is the tooth tip height and Twfront and Twback are the inner
Fig. 10: Efficiency vs. speed characteristics for the considered and the outer stator tooth tip width, respectively.
motors.

3) Costs and active materials weight: Following the per-

formance analysis, the cost calculation was performed and
reported in Table II, where a net cost of 6.5e/kg, 4e/kg,
1e/kg, is considered for magnet, copper and steel, respectively.
Due to the very small size and power of the compared motors
as well as due to the fact that the employed magnet in the
PMSM is a very cheap one, the PM-free machines are not
necessarily more attractive in terms of cost. There is a slight
decrease when comparing with the AF-SynRM, however this
Fig. 12: Parametrization of the Stator.
number can be considered as negligible.
In Fig. 13 Rrin and Rrout are the inner and the outer
IV. 3D FEA- ELECTROMAGNETIC O PTIMIZATION OF THE
stator radius, respectively, and hfci is the thickness of the i-
AF-S YN RM
th flux-carrier (in q-axis). Normally, if n, is the number of
In order to reduce the motor weight (and thus the costs), and flux carriers, there are n − 1 flux barriers, as shown in Fig.
torque ripple while maximizing the average torque, hereafter 13 their thicknesses on q-axis being denoted with hfbj . In this
a FE-based optimization is performed. The partial 3D FE- particular representation, 3 flux barriers and thus 4 flux carriers
electromagnetic model (Fig. 11) is parametrized as shown in have been considered. Wfci are the widths of the flux carriers
Fig. 12 for the stator and in Fig. 13 for the rotor. on d-axis, whereas W1in and W1out are the inner and the outer


lengths of the linear part of the flux carriers, respectively. though is a known fact that usually this can contribute to the
Finally Rfillet is the curvature radius of the nonlinear portion power factor improvement, can be understood after analyzing
of the flux carriers (highlighted in green). how these designs perform in terms of torque ripple (Fig.
14c). Again, the best designs pareto-front is poorly populated,
pointing out anew, how difficult it is to find designs with high
average torque and low torque ripple (at least for this particular
motor topology).
Finally, the mass of the iron parts in function of the average
toque is represented in Fig. 14d, along with the best designs
pareto-front (the red points).

V. AF-S YN RM: M ANUFACTURING POSSIBILITIES

Fig. 13: Parametrization of the Rotor. The general requirements of soft magnetic materials
(SMMs) are high saturation flux density and permeability (in
all directions or specifically in one direction, i.e. isotropy
At first, having the aforementioned DoFs, a multi-objective
versus anisotropy), and low hysteresis and eddy currents
optimization is performed having both the average value of
losses, but also ease of construction and low cost.
torque (max(Tav )) as well as its ripple (min(Tp2p )) as
objective functions. The obtained Tp2p = f (Tav ) dependence A. The Stator
is shown in Fig.14a, based on which one can already draw
In terms of working principle and concepts, there is nothing
the first conclusion that indeed the chosen topology exhibits
new about the stator, with the manufacturing still being the bot-
significant torque ripple. It has to be also stated that since
tleneck. Therefore, many attempts are being made to simplify
this is a new motor concept, no stringent boundary conditions
the rather complicated production process. One such example
were set on torque ripple (i.e. ≈ 50%). In the same figure, best
is shown in Fig.15a to Fig.15d, where at first the tooth seg-
designs within the pareto-front are highlighted in red, showing
ments are obtained Fig.15a and then they are mounted together
a rather limited number of designs with an average torque
to form a stator, but for simplification purpose the resultant
Tav ≥ 0.85Nm and which fulfill the torque ripple constraint
structure has an alternate type of winding, i.e. one tooth is
(i.e. Tp2p ≤ 0.45Nm).
winded whereas its 2 neighbors are left unwound, being uses
only for flux guidance (Fig. 15b). Furthermore, for fractional
concentrated windings, employing coil-band type, instead of
the conventional wired-coils, contributes significantly to the
filling factor improvement. Furthermore, the coil-band can be
manufactured either out of copper or aluminum as it can be
seen on the prototype in Fig. 15c and Fig. 15d, respectively.
Particularly for the AF-SynRM, its stator can be built in a
similar manner as for the PM-based AF-Syncronous machines,
either using Si-Fe laminations or powder based materials.
(a) (b)
B. The Rotor
1) Si-Fe Laminations: A rather straightforward approach
would be to built the rotor core by applying the principle of
the axially laminated (ALA) RF-rotor presented in [16], to the
axial case. Thus, outstanding saliency values can be achieved
as the transversal, out-of-plane flux is limited by the very
low permeability in that direction, due to the non-magnetic
(c) (d) insulation between the laminations (which coincides with the
Fig. 14: Pareto-front obtained while maximizing average rotor’s q-axis), but leads to in-plane induced currents that can
torque and minimizing torque ripple. flow freely in the laminations, without being hindered by the
insulation, and can lead to a substantial increase of the losses
In Fig. 14b it is shown that the average torque and the motor [17]. Moreover, from the process point of view, some extra
saliency are inter-linked. The points found on the pareto-front logistics steps are to be performed as each lamination (per
are represented in Fig. 14a in red. No clear trend can be pole) has its own width.
observed, the best designs are spread all over the map, however 2) Soft magnetic composites (SMCs): When it comes to
it is noteworthy that there are many designs with saliencies finding substitutes for the more conventional Fe-Si lamina-
exceeding that of the chosen design (i.e. 4.8). The reason tions, compacted insulated iron powder, commonly known as
why designs with higher saliency cannot be considered, even soft magnetic composite (SMC), were primarily investigated


 to be used in the FMC materials have the thickness and
width of order of tens of micrometres and the length varying
from few centimeters up to the order of meters, the magnetic
characteristic of this new family of materials (FMC) can be
a combination of magnetic properties of SMC’s and Fe-Si
laminated sheets.
Obviously, the final physical properties of the material
depend on the ”recipe” that is used, i.e. the type of lubricants
(a) (b) or binders, but equally on the process, e.g. higher tempera-
ture treatment combined with optimal additive will provide
improved magnetic properties. Moreover, the insulating layer
that covers the entire fibre is of major importance in order
to minimize the dissipative character of the magnetic core at
high frequency.
Due to their particular anisotropic rotor topology, SynRM
makes full use of the FMC magnetic properties. That leads to
higher saliency ratio and thus, to better performances of the
machine. Such a rotor stack is envisaged in Fig. 16, showing
how such fibers can define the rotor stack, with the mention
that only the fibers appearing on the outer surface are shown,
(c) (d) thus the apparent interruption is caused by their non-uniformed
distribution in the XY-plane at constant Z coordinate (they are
Fig. 15: Laminated stator of an axial flux synchronous motor.
leaving the viewable plane around q-axis).

because of their isotropic nature and unique shaping possi-

bilities, which facilitates the implementation of 3D topologies
with an additional degree of freedom in terms of flux guidance
[18]. Fundamentally, SMC is comprised of Fe powder mate-
rials coated with an electrically insulating layer. The design
and production of the SMC should be robust with regard
to mechanical tolerances, leading to additional costs [19].
Compared to Fe-Si laminations, with in-plane flux, SMC have Fig. 16: FMSCs principle.
a lower permeability and saturation flux density, and higher
hysteresis losses. As they are isotropic, with low eddy current However, there are quite a few technical challenges related
losses, irrespective of the flux direction, they are particularly with the material physical properties, including good magnetic
suited for 3D flux applications, with easy segmentation of properties with low power loss and resilience to aging factors.
the machine, improved winding configuration (with higher fill Furthermore, in terms of modelling, it is clearly very difficult
factor), and a lower part number and thus lower production to consider all the fibres in a direct and explicit way in a FE
time and costs [20]. However, it is exactly their isotropic model, as this requires a detailed 3D modelling, including the
character which increases the complexity of the production fine and explicit discretisation of more such fibres. Therefore,
process when attempting to create a controlled anisotropy as similarly as for Fe-Si lamination case [22]- [24], dedicated
desired in an SynRM. numerical techniques are to be developed for FSMCs so as
3) Fiber Soft magnetic composites (FMCs): A more radical to consider different magnetic effects (e.g. eddy currents) at a
approach, would be the introduction of a new type of mate- reasonable computational cost. The fibers are to be replaced
rials from SMC materials class, namely Fibre Soft Magnetic by a homogeneous, coarsely meshed volume in which ad-
Composites (FMCs) [21]. The FMC is an anisotropic iron- hoc material properties and equations are taken into account.
fibre based material which includes various material addi- For highly saturated machines where the assumption of linear
tives/impurities as fibre coatings. material behavior would imply many inaccuracies, saturation
From the fundamental point of view, FMCs are situated has to be considered, thus the homogenization of the core
in between the two important classes of materials: SMC and becomes much more challenging. Extensions regarding hys-
Fe-Si laminated sheets. If in the case of powders, the size teresis modeling brings further challenges.
of particles (in all 3 directions) is of tens of micrometres
order, in the case of laminated sheets, the thickness is of CONCLUSIONS
order of hundreds of micrometres and the length and width is In this paper, a new concept of AF synchronous machines
dependent on application. Considering that the fibres intended is briefly introduced. Only the fractional slot concentrated


winding configuration was addressed as this was also used
[3] A.-C. Pop and J. J. C. Gyselinck and D. E. Pinto and I. Vintiloiu,
in the reference machine. ”Optimization of Low-Power Brushless PM-Machines for Automotive
The newly introduced motor has shown its superiority both Applications With Focus on High-Volume Mass Production”, IEEE
in terms of torque/power density as well as efficiency in Transactions on Industrial Electronics, vol. 64, pp. 9767–9775, 2017
[4] S. Taghavi, ”Design of Synchronous Reluctance Machines for Automo-
comparison with its RF variant. It is however to acknowledge tive Applications”, Doctoral dissertation, Concordia University, 2015.
its inferiority when comparing it to the PMSM. Moreover, due [5] S. Ruoho, Demagnetisation of permanent magnets in electrical machines,
to the higher q-axis axis inductance (longer flux’s paths) and OH presentation, Helsinki University, 2007.
[6] F. Daldaban, & E. Cetin, ”Prototyping of Axial Flux Permanent Magnet
resistance, it exhibits a lower base speed, despite the fact that Motors,” 3rd International Symposium on Innovative Technologies in
it is able to develop the same torque in the same installation Engineering and Science, vol. 13, 2015.
space and assuming the same input power. The resistance [7] A. Cavagnino, M. Lazzari, F. Profumo, & A. Tenconi, ”A comparison
between the axial flux and the radial flux structures for PM synchronous
can be further decreased when increasing the filling factor by motors,” IEEE transactions on industry applications, vol. 38(6), pp.
means of using band-coils (briefly introduced in this paper). 1517-1524, 2002.
One of the main design challenges is the torque ripple [8] W. E. Jara Montecinos, ”Axial Flux Permanent Magnet MachinesDe-
velopment of Optimal Design Strategies”, Acta Universitatis Lappeen-
mitigation, problem addressed in this paper, however with rantaensis, 2016.
rather modest results. Another challenging issues, which are [9] Y. K. Chin, ”A permanent magnet synchronous motor for an electric
typical for all machines in axial flux configuration, are related vehicle-design analysis”, Doctoral dissertation, 2004.
[10] R. R. Moghaddam, ”Synchronous reluctance machine (SynRM) design”,
to the manufacturing process. In this context, a laminated Royal Institute of Technology (KTH), 2007
stator concept was shown, that has features, that can be rather [11] T. Seyedmorteza, ”A Core Analysis of the Synchronous Reluctance
straightforward automatized in view of large volume mass Motor for Automotive Applications”,” Electrical Machines (ICEM),
2014 International Conference, pp. 961-967, 2014.
production (e.g. segmented teeth, low number of coils to be [12] C. M. Spargo, B. C. Mecrow, & J. D. Widmer, ”Application of fractional
winded). As for the rotor, a new material concept was briefly slot concentrated windings to synchronous reluctance machines”, IEEE
introduced which from the properties point of view can be International Electric Machines & Drives Conference (IEMDC), pp.
618–625 2013.
placed between SiFe-laminations and SMC, but (theoretically) [13] F. Daldaban, & E. Cetin, E., ”Prototyping of Axial Flux Permanent Mag-
has the main advantage that an anysotropic structure can be net Motors,” in 3rd International Symposium on Innovative Technologies
created, which is desirable in general for the rotors of SynRMs in Engineering and Science, ISITES A (Vol. 13), 2015.
and in particular for the AF-SynRM introduced here. [14] W. E. Jara Montecinos, ”Axial Flux Permanent Magnet MachinesDe-
velopment of Optimal Design Strategies,” Acta Universitatis Lappeen-
It has to be stated that there are significant challenges to be rantaensis, 2016.
addressed, before such a material can be used at industrial [15] R. Madhavan, & B. G. Fernandes, ”Comparative analysis of axial flux
scale, but its theoretical properties are very promising and SRM topologies for electric vehicle application,” in IEEE International
Conference on Power Electronics, Drives and Energy Systems (PEDES),
the preparation, characterization and production at industrial 2012 (pp. 1-6), 2012.
level of such materials can lead to important steps far beyond [16] I. Boldea, Z. X. Fu, & S. A. Nasar, ”Performance evaluation of axially-
the state of the art in the field of magnetic materials used in laminated anisotropic (ALA) rotor reluctance synchronous motors”,
IEEE Industry Applications Society Annual Meeting, vol. 1, pp. 212–
electrical machines. 218 1992.
Also at modeling level and incorporating of various effects [17] B. Scheerlinck, H. De Gersem, & P. Sergeant, ”3-D Eddy Current
and material properties in the FE-model, there are major and Fringing-Flux Distribution in an Axial-Flux Permanent-Magnet
Synchronous Machine With Stator in Laminated Iron or SMC.”, IEEE
hurdles to overcome. Therefore in this paper the conventional Transactions on Magnetics, vol. 51(11), pp. 1–4 2015.
SiFe were used when designing the AF-SynRM, however, [18] L. O. Hultman, A. G. Jack, ”Soft magnetic composites-materials and
when thinking at an ideal high saliency value which can be applications,” IEEE International Electric Machines and Drives Confer-
ence, (IEMDC’03), vol.1, pp. 516-522, 2003
obtained by simply imposing the right directions of the fibers, [19] G. Lei, J. G. Zhu, Y. G. Guo, J. F. Hu, W. Xu & K. R. Shao,
the potential is absolutely huge. ”Robust Design Optimization of PM-SMC Motors for Six Sigma Quality
Manufacturing,” in IEEE Transactions on Magnetics, vol. 49, no. 7, pp.
ACKNOWLEDGMENT 3953-3956, July 2013.
[20] O. Andersson, P. Hofecker, ”Advances in Soft Magnetic Composites
This work was supported by a grant of Strengthening Materials and Applications”, PowderMet2009, 2009.
the Research potential of CAREESD in the field of Elec- [21] P. Pszola, M. Pszola, ”Stack System of an Electric Machine, Electric
tromechanical Systems and Power Electronics for Sustain- Machine, and Method for Producing the Stack System,” patent applica-
tion number PCT/EP2017/054942, 2017.
able Applications, ESPESA, 692224/2015 H2020-TWINNing [22] J. Gyselinck, C. Geuzaine, & R. V. Sabariego, ”Considering laminated
-2015. The authors equally acknowledge the support of Brose cores and eddy currents in 2D and 3D finite element simulation of
Fahrzeugteile GmbH & Co. KG, Würzburg, Germany. electrical machines”, in Proceedings of the 18th Conference on the
Computation of Electromagnetic Fields (COMPUMAG2011), Sydney,
R EFERENCES Australia. 2011.
[23] J. Gyselinck, R. V. Sabariego, & P. Dular, ”A nonlinear time-domain
[1] B. Lequesne, “Automotive electrification: The nonhybrid story,” IEEE homogenization technique for laminated iron cores in three-dimensional
Transactions on Transportation Electrification, pp. 40–53, 2015. finite-element models”, IEEE transactions on magnetics vol 42(4) pp.
[2] J. De Santiago, H. Bernhoff, B. Ekergard, S. Eriksson, S. Ferhatovic, 763–766, 2006.
R. Waters, & M. Leijon, ”Electrical motor drivelines in commercial all- [24] R. V. Sabariego, I. Niyonzima, J. Gyselinck & C. Geuzaine, ”Com-
electric vehicles: A review,” IEEE Transactions on vehicular technology, parison of two nonlinear finite-element homogenization methods for
vol. 61(2), 475-484, 2012. laminated iron cores”, 2013.