Energies 11 00274 PDF

energies
Article
Development of Automotive Permanent Magnet
Alternator with Fully Controlled AC/DC Converter
Tareq S. El-Hasan ID
Electrical Engineering Department, Zarqa University, Zqrqa 13132, Jordan; tareq.elhasan@zu.edu.jo;

Tel.: +962-777-747-230
Received: 18 November 2017; Accepted: 9 January 2018; Published: 24 January 2018
Abstract: This paper proposes the design of a three-phase axial flux permanent magnet alternator
(AFPMA) that is characterized with an air-cored stator and two-rotor (ACSTR) configuration.
The AFPMA is harnessed with fully controlled AC/DC converter using six bridge Insulated Gate
Bipolar Transistor (IGBTs) capable to deliver a constant DC output power as an attempt to replace
the Lundell alternator for automotive applications. First, the design methodology and analysis of
the AFPMA is introduced. The most effective parameters, such as rotor diameter, magnet thickness,
number of turns, and winding thickness are determined. A smart digital control which facilitates
the comparison between the magnitudes of the three-phase input signals instead of finding the zero
crossing points is developed. Moreover, custom design comparators are specially designed and
developed to generate adaptive signals that are fed into an Arduino Uno microcontroller. Accordingly,
the Arduino generates the timely precise pulses that are necessary to maintain the appropriate
triggering of the IGBTs. This technique allows the IGBTs to conduct in an adaptive manner to
overcome the problem of asymmetrical voltage outputs from the AFPM alternator. The system is
also capable of handling the variation in the speed of the AFPMA via the rigor code in Arduino that
detects the change in the supply frequency and voltages in a real time process. The system is first
analyzed via simulations using MATLAB/Simulink and then experimentally validated at certain
speed and loading conditions. The preliminary tests results indicate that such system is capable to
provide an efficient solution to satisfy automotive electric power demands.
Keywords: automotive alternator; axial flux permanent magnet alternator; air cored stator alternator;
fully controlled AC/DC converter
1. Introduction
When implementing electrical technology or components into vehicles, the size, weight, cost,
performance, and efficiency are the major factors that need to be considered for successful integration.
Steadily but surely, automobile manufacturers are looking to improve vehicle component performance
and increasing their power density. Among these components is the on-board electrical power charging
system which normally consists of the alternator, power converter, and battery. Such electrical systems
are pushed to their limits as more and more vehicle’s components are replaced by electrically powered
or electrically assisted items. In addition, the evolving customer demands, increasing car functionalities,
and developing vehicle regulations and legislations have necessitated the development of the electrical
power generating system to satisfy the increasing demands of electrical power in automobiles.
For the last 50 years, the three-phase claw-pole alternators (or Lundell alternators) have been used
for the generation of electric power in automotive applications. Until very recent years, this alternator
was the most economic choice in vehicle technology due to its low manufacturing cost. However,
Lundell alternators suffer from major drawbacks such as reduced efficiency, limited output power,
and the need for continuous maintenance. This in turn has limited the utilization of such machines in
Energies 2018, 11, 274; doi:10.3390/en11020274 www.mdpi.com/journal/energies

Energies 2018, 11, 274 2 of 28
modern vehicles due to the challenges initiated by increased power demands. Hence, other alternatives
for electrical power generating systems have been considered to meet such challenges in motor
vehicles [1–4].
To meet the increasing power demands in modern vehicles, automobile manufacturers have
exploited the distinguished merits offered by the rare earth, high power, permanent magnets (PM) in
the development of electrical power generator to replace the Lundell alternators.
Alternators utilizing PM offer numerous attractive features over other conventional types of
electrical machines, including higher efficiency, higher power density, no maintenance, and better
reliability [5,6]. PMs can also improve the dynamic performance, and output quality of the machine.
PMs are also becoming less expensive, which makes PM machines more prevalent. The recent advances
in power electronic devices have led to better, easier, and cost effective control of PM machines, with the
tendency of operating the machine over a large range of speeds while maintaining a reasonable
efficiency. Several types of PM machines have been constructed based on their magnetic field direction.
For example, PM machines with axial, radial, or transverse configurations have found their way into
many applications, such as home appliances, industrial use, renewable energy, and automobiles [7–10].
Other benefits such as short length, ease of manufacturing and assembly, and modularity are realized
by axial flux PM (AFPM) machines when compared to other configurations. Such machines have
demonstrated their relative fitness in applications that are limited with confined spaces or those which
require direct coupling of the machine with short overhang [11]. AFPM machines can be used for
both low speed [12,13] and high speed applications [14–16]. Moreover, AFPM machines with ACSTR
configurations are characterized with extra advantages such as lower cogging torque, reduced eddy
current losses, better output wave (sinusoidal-like) generation, and less harmonics [15] as compared to
other types of PM machines with cored stator configurations. Hence they have potential for use in
wind energy applications [17–19] as well as automotive applications [20–23].
This paper introduces the design and optimization methodology for an AFPM alternator (AFPMA)
as part of the integrated battery charging system in automobiles. The proposed AFPMA, which is
intended to work as a replacement for the Lundell alternator diverges from the high speed version that
has been previously developed for high speed applications [14–16]. The objective is to rebuild that
AFPMA with minimal design effort to make it suitable for vehicle applications. This is achieved by
redesigning the stator windings within the existing geometrical constraints and without jeopardizing
the machine’s efficiency. To do so, the sizing equations for the machine are developed, and the
machine performance is analysed using MATLAB/Simulink modelling. Moreover, a three-phase fully
controlled AC/DC rectifier is developed and integrated with the AFPMA. The complete system is
experimentally tested and simulation results are validated and results obtained are presented in detail
in this paper.
The ultimate goal is to design a suitable on-board electrical power system that can satisfy the
full requirement of increasing demand of modern automotive industry. The machine shall be capable
of providing the car with the necessary power to charge the battery and to turn all electrical and
electronics loads whether the engine is running at idle speed or at full speed.
2. Description and Design Requirements of the AFPMA

The machine under investigation has an air-cored stator and two-rotor (ACSTR) system with no
iron core in the stator coils. Originally, the machine was designed for high speed applications [14].
Hence, the rotors were carefully designed to allow the machine to withstand high levels of centrifugal
forces. The proposed AFPMA in this paper is depicted in Figure 1. To redesign the machine for low to
moderate speed applications, the rotor configuration and dimensions are kept the same except the
retainment ring which encapsulates the rotor magnets are selected from lower cost aluminium alloy
instead of expensive, high strength maraging alloy. In addition, simpler manufacturing and assembly
processes are performed which in turn reduces the machine cost. However, the stator is completely
redesigned in order to achieve the automobile requirements. Still, the stator is air cored (ironless)
Energies 2018, 11, x FOR PEER REVIEW 3 of 30
redesigned in order to achieve the automobile requirements. Still, the stator is air cored (ironless) in
orderEnergies
to minimize weight, losses, and cogging torque. Figure 2 shows the main parameters3 ofthat
2018, 11, 274 28 are
considered in the redesign process. These are: the outer radius of the coil which is similar to the
redesigned in order to achieve the automobile requirements. Still, the stator is air cored (ironless) in
outerinradius of the PM disc, ( R olosses,
order to minimize weight,
), inner radius
losses, and
of coil
and cogging
which
cogging torque.
is similar
torque. Figure
Figure22shows
tothe
showsthe
the innerparameters
main radius ofthat
mainparameters
PM disc
thatare
( Ri ),areconsidered
considered
number in
inthe
the redesign
of winding redesign ( Nt ), axial
turns process. These are:
airare: the
gabthe outer(stator
outer
length radiusof
radius ofthe
thecoil
coilwhich
thickness), which is similar
andismechanical
similar to to
the the
speed
outer radius of the PM disc, (Ro ),Rinner
outer radius of the PM disc, ( ), inner radius of coil which is similar to the inner radius of
radius of coil which is similar to the inner radius of PM disc PM disc
( N m(R
o
). ),The target is to reconstruct a mathematical
air gab lengthmodel that can be andused to design
speedthe
(N stator
i( Ri ), number of winding turns
number of winding t ( N ), axial air gab length (stator thickness), and mechanical speed
turns (N ), axial
t
(stator thickness), mechanical
m ).
winding to fitisfor
The target the AFPMA
to reconstruct to achieve the
a mathematical machine
model that canrequirements presented
be used to design the statorinwinding
Table 1towithin
fit for the
existing( N m
the AFPMA). The target is to reconstruct a mathematical model that can be used to design the
to achieve the machine requirements presented in Table 1 within the existing constraints.
constraints. stator
winding to fit for the AFPMA to achieve the machine requirements presented in Table 1 within the
existing constraints.
(a) (b)
(a) (b)
Figure 1. AFPMA
Figure 1. AFPMA
Figure
under
1. AFPMA
consideration:
under consideration:
under
(a)
(a)3D
consideration:
exploded
3D3D
(a) exploded view
view
exploded ofofACSTR;
view
ACSTR;(b)(b)
of ACSTR;
assembled
assembled
(b)
AFPMA.
AFPMA.
assembled AFPMA.
Figure
Figure 2. AFPMA
2. AFPMA rotor
rotor andand stator
stator main
main parameters.
parameters.
Figure 2. AFPMA rotor and stator main parameters.
Table
Table 1. AFPMA
1. AFPMA Requirements.
Requirements.
Table 1.Set
AFPMA AFPMA Requirements.
Parameters Value Unit
AFPMA Set Parameters Value
AFPMA RMSSet
line–line voltage Value
Parameters 17 Unit V
Unit
RMS line–line voltage
Rated current 17 40 V A
RMS Rated
line–line voltage
current 40 17 V
Rated speed 5000 A RPM
Ratedcurrent
Rated speed 500040 RPMA
Rated
Rated power
power @ rated
@ rated speedspeed1.2 1.2 kVAkVA
Rated speed 5000 RPM
Rated
3. Design Consideration powerof@the
and Sizing rated speed
AFPMA 1.2 kVA
3. Design Consideration and Sizing of the AFPMA
In order to facilitate the AFPMA for automobile applications, the machine shall be designed to
3. DesignInConsideration
order to facilitate
and the AFPMA
Sizing for automobile
of the AFPMA applications, the machine shall be designed
fit for the gasoline engine speed. However, coupling the alternator directly to the automobile engine
to fit for the gasoline engine speed. However, coupling the alternator directly to the automobile
In will
order
engine
require an increased
willtorequire
facilitate number of
the AFPMA
an increased
stator
for
number
winding turns
automobile
of
to compensate
applications,
stator winding
for the low
the machine
turns to compensate
operating
shall
for the
speed, to
lowbeoperating
designed
thus
fit forspeed, increasing
the gasoline overall machine
engineoverall volume.
speed.machine
However, Lundell
coupling alternators are
thealternators therefore
alternatorare directlycommonly
to commonly run
the automobilewith aengine
belt
thus increasing volume. Lundell therefore run with
ratio of
will require an 3:1 to obtain
increased a higher
number rotational
of stator speed.turns
winding For the proposed AFPMA,
to compensate for the alowbeltoperating
ratio of 5:1 is
speed,
thus increasing overall machine volume. Lundell alternators are therefore commonly run with a belt
ratio of 3:1 to obtain a higher rotational speed. For the proposed AFPMA, a belt ratio of 5:1 is
Energies 2018, 11, 274 4 of 28
a belt ratio of 3:1 to obtain a higher rotational speed. For the proposed AFPMA, a belt ratio of 5:1 is
adopted to achieve higher rotational speed hence increasing the overall power density of the system.
Accordingly the AFPMA shall be capable to operate in the range of 3750–15,000 rpm when coupled
to an engine running in the range of idle speed of 750 rpm to the cruise speed of 3000 rpm. As far
as the mechanical integrity is concerned, the AFPMA has its PMs encapsulated by a high strength
retainment ring that is shrunk fit on the rotor disc. Mechanical integrity tests showed safe operation of
the machine at rotational speeds in the excess 25,000 [16]. However, in this paper, and for the purpose
of laboratory testing, the proposed machine is driven by a three-phase electric motor with a belt ratio
of 1.7:1 which will limit the machine maximum speed of the AFPMA to 5100 rpm.
Once the range of operating speed is known, it is possible to develop the necessary sizing
equations and mathematical models that can be utilized for optimization purposes in order to
determine the machine’s parameters within the competing demands to meet the system requirements.
To derive the necessary equations, the induced voltage formula will be applied on the stationary
conductors in the presence of a rotating magnetic field. The rotating magnetic field in the proposed
machine is secured by the axially magnetized PM semicircular discs that are located alternately on
both rotor sides. The induced voltage in the conductor relies on the flux density, conductor length,
and relative velocity. Calculation of the flux density in the middle of the air gap relies on several
parameters such as: magnet type/grade, magnet axial length Lm , magnet shape and air gap length
Xg between the two PM rotors. In automotive applications, components in the engine compartment
are fitted in confined spaces and often exposed to high ambient temperatures and insufficient cooling.
Mounted near the engine, the alternator is exposed to high temperatures, thereby reducing its
performance and lifetime. Therefore, PM alternators have to be designed with great care to avoid
the risk of PM demagnetization, especially for the NdFeB which has a lower Curie temperature and
lower working temperature. In automotive applications, the under hood temperature may vary in
the range from 50 ◦ C to 80 ◦ C, adding more thermal stress on the PM alternator as it may increase
the risk of PM demagnetization. Therefore, selecting the type and the grade of PM is a vital issue in
the design process as one has to compromise between achieving higher power density and higher
temperature operability. Among the recently developed PMs, comes the neodymium iron boron
(NdFeB) magnet which has high remanence Br , high coercive force Hc , high energy production,
and a high performance/cost ratio. As can be shown in Figure 3, NdFeB magnets with grade N52 have
the highest normalized energy product (with average remanence value of 1.45 T) which makes them
the most powerful magnets among other grades. However, such magnets have a limited operating
temperature of 80 ◦ C which may degrade the performance of the proposed AFPM alternator when
used for automotive applications. In this stage of research, it was decided to choose the type and grade
of PM that provides the best magnetic characteristics in order to achieve the maximum possible power
density of the AFPM alternator to perform as an advanced development model (ADM). The objective
is to prove that the AFPMA would perform according to the requirements and interface well with other
systems in terms of form, fit, and function (F3). Thanks to its simple shape and design, the ACSTR
configuration of the proposed AFPMA allow for natural ventilation and heat removal with minimal
effort. The NdFeB magnets are held by an Alumec carrier which in turn plays a crucial role in
dissipating the heat that may transfer from the stator windings to the rotor surface. In addition,
the rotating PM discs work as natural ventilators since they continuously expel the hot air surrounding
their surfaces during rotation. Consequently, once the performance of the machine is experimentally
validated, the engineering development model (EDM) of the AFPMA will consider the intended
final materials and/or deployment of the necessary cooling arrangement to account for reliability,
availability, and maintainability (RAM).
Energies 2018, 11, 274 5 of 28
(a) (b)
Figure 3. Normalized energy product vs. different grades of NdFeB at their corresponding maximum
operating temperatures: (a) calculated at 20 °C; (b) calculated at 80 °C. Based on available information
from China YY Magnetics (Zhejiang, China) [24].
To account for temperature

(a) effect on the energy product of the
(b) PMs hence on the machine
performance, B r 3.3.and
Figure
Figure
H c are
Normalized
Normalized
calculated
energy
energy productvs.
product vs.respectivelygradesas
differentgrades
different
[25]:
ofofNdFeB
NdFeBatattheir
theircorresponding
correspondingmaximum
maximum
operatingtemperatures:
operating temperatures:(a)(a)calculated
calculatedatat2020◦ C;
°C;(b)
(b)calculated
calculatedatat8080◦ C.
°C.Based
Basedononavailable
availableinformation
information
from China YY Magnetics (Zhejiang, China)  
[24]. 
from China YY Magnetics (Zhejiang, China) [24].
Br  Br 20 1  TPM  20
B
(1)
To account for  100 
temperature effect on the energy product of
the PMs hence on the machine
To account for temperature effect on the energy product of the PMs hence on the machine
performance,Br B
performance, r and
and H ccalculated
Hc are are calculated
 as [25]:as [25]: 
respectively
respectively
H c  H c 20h 1  H TPM  20 
i  (2)
 α100
 ) 


B 1 +1  ( TPM
B B
Br B
=Br20 T −2020 (1)
r r 20  100 PM  (1)
 100 
where Br 20 and H c 20 are the NdFeB gradeh52 remanence αH and
i coercive force respectively given at
Hc = Hc20 1 +  100 H( TPM − 20)  (2)
20 °C,  B and  H are the temperature H c coefficients
 H c 20 1  TPM B 20and
for  H and they are set to(2)
−0.12 and
where Br20 and Hc20 are the NdFeB grade 52 remanence 100 r 
and coercive
c
force respectively given at
−0.6 respectively,
20where BB
and
◦ C, α and THPMtheare
α H are
and
istemperature
the PM temperature
the NdFeB coefficients for Bwhich
grade 52 remanence r and
is set to are
andHcoercive
c and they
thesetmaximum
to −0.12
force respectively
operating
and
given at
r 20 c 20
temperature and TPM is the PM
of 80 °C. Accordingly, thetemperature
remanence which
value is set
andto the maximum
energy operating
product are temperature
calculated for N52
of20 ◦ C. 
80°C, B and  H the areremanence
the temperature coefficients for B r are and H c andfor
they
N52are
at set
theto −0.12 and
at the maximum operating temperature which showed a decrease in the normalizedmaximum
Accordingly, value and energy product calculated
energy product
operating temperature and TPMshowed
which is thea PM temperature
decrease which isenergy
in the normalized set toproduct
the maximum operating
by approximately
by approximately
8%.
8%.
The effect of
The
of 80
effect
temperature
of temperature
rise on the
the remanence
rise
normalizedvalue
on
energy
the normalized
product
energy product for several
temperature °C. Accordingly, and energyfor severalare
product grades of NdFeB
calculated as
for N52
grades ofcompared
NdFeB
at asthe
to
the maximum compared
energy totemperature
product
operating theofenergy product
N52 based 20 ◦ CofisN52
onshowed
which based
apresented
decrease in on 20
Figure
the °C
4. is presented
normalized in Figure 4.
energy product
by approximately 8%. The effect of temperature rise on the normalized energy product for several
grades of NdFeB as compared to the energy product of N52 based on 20 °C is presented in Figure 4.
Figure
Figure 4.Figure 4. Degrading
Degrading of theofof the normalized
normalized energy product as
energy as aa function of of
thethe
temperature increase for
4. Degrading the normalized energy product function
product as a function of the temperature
temperature increaseincrease
for for
several grades of NdFeB. Based on available information from e-magnets UK (Hertfordshire, UK) [26].
several grades
several of NdFeB.
grades BasedBased
of NdFeB. on available information
on available informationfrom e-magnetsUK
from e-magnets UK (Hertfordshire,
(Hertfordshire, UK) [26].
UK) [26].
For simplification purposes, circular magnet shapes are considered to determine the magnetic
For simplification
For simplification purposes,
purposes, circularmagnet
circular magnet shapes
shapes are considered to determine the magnetic
flux density in the air gap between the two PM rotors. Theare
full considered to determine
derivation of the theand
sizing equation magnetic
flux density in the air gap between the two PM rotors. The full derivation of the sizing equation and
flux density innecessary
other the air gap between
equations the
for the two PMparameter
AFPMA’s rotors. The full derivation
determination of the sizing
and optimization equation
purposes are and
other necessary equations for the AFPMA’s parameter determination and optimization purposes are
shown
other necessary in Appendix A.
shown in equations
Appendix A.for the AFPMA’s parameter determination and optimization purposes are
shown in Appendix A.
Energies 2018, 11, 274 6 of 28
4. Optimization of the AFPMA

In order to optimize the machine design, two main characteristics will be investigated. These are
machine efficiency and machine internal power factor, as will be described in detail in the
following sections.
4.1. Machine Efficiency

To predict the machine efficiency, the losses that are affected by the redesign process of the stator
will be investigated. The two main losses that are seen to have major impact on the efficiency of the
redesigned AFPMA these are the stator copper losses and the windage losses. The copper losses, Pcu ,
are calculated at the designed armature current, I A = 40 A as:
2
Pcu = 3I A RW (3)
The windage losses of the machine is highly dependent on the rotor diameter and the rotational
speed. If the rotor diameter is kept constant, then the only parameter that affects the windage losses is
the rotor speed as:
3
PW = 0.5NR C f ρ air ωm R5o − R5sh (4)
3.87
Cf = √ (5)
Re
ρ air ωm R2o
Re = (6)
µ air
where C f is the coefficient drag for turbulent flow, Re is Reynolds number, NR is number of rotating
discs, ρ air is the specific density of the surrounding air = 1.18 kg/m3 , µ air is the coefficient of viscosity
of the air = 1.98 × 10−5 kg/m·s, and Rsh is shaft radius.
Moreover, the amount of saving in windage losses is a critical aspect which may diversely affect
the overall performance of the machine if precautions are not considered during the design. Hence,
for fixed rotor geometry, the windage losses will only vary as a function of the rotor speed as:
2.5
PW = Kwdg ωm (7)
The maximum windage losses PW,max are obtained at the maximum calculated speed,
and minimum number of winding turns. i.e., at Nt = 2.
Accordingly, the saving in the windage losses PW,sav can be calculated as:
PW,max − PW
PW,sav = (8)
PW,max
Assuming unity load power factor, and rated armature current of 40 A, the machine output power
can be calculated as:
Pout = 3Vϕ, f l I A (9)
Assuming the machine is operating at unity power factor, the full load phase voltage Vϕ, f l is
calculated as: q
Vϕ, f l = E2ϕ,rms − ( I A XW )2 − I A RW (10)
Other losses, such as rotor eddy current losses, are not considered in the investigation since the
rotor geometry is not changed. Hence, the machine efficiency, η at the rated armature current and
unity load power factor is calculated as:
Pout
η= × 100% (11)
Pout + Pcu + PW
unity load power factor is calculated as:
Pout
 100% (11)
Pout  Pcu  PW
Energies 2018, 11, 274 7 of 28
4.2. Machine Internal Power Factor (cos φ)
The machine internal power factor (PFm) can be calculated as PFm  cos , where
4.2. Machine Internal Power Factor (cos ϕ)
1  X W  . Also, the voltage regulation of the machine can be calculated as:
  tan  

The machine R internal power factor (PFm) can be calculated as PFm = cos ϕ, where ϕ = tan−1 X W
RW .
 W 
Also, the voltage regulation of the machine can be calculated as:
V ,nl  V , fl
VR  100% (12)
Vϕ,nlV− V
 , fl ϕ, f l
VR = · 100% (12)
Vϕ, f l
The no load voltage Vnl equals the per phase internal generated voltage of the machine E ,rms .
The
Hence,no load 17
voltage VnlVequals
. the per phase internal generated voltage of the machine E ϕ,rms .
E , rms   9.815
17 3
Hence, E ϕ,rms = √ = 9.815 V.
3
To perform
To perform the theoptimization,
optimization,the
thenumber
number of ofturns
turnsperper pole
pole perper phase
phase N t Nist varied
is varied in steps
in steps in in
the range
the range from two turns up to nine turns. The maximum number of turns is limited with the spacespace
from two turns up to nine turns. The maximum number of turns is limited with the
between the shaft
between and magnet
the shaft inner inner
and magnet radius. For a For
radius. specific number
a specific of turns,
number the corresponding
of turns, rotational
the corresponding
speed,rotational
power speed,factor, power
savingfactor, saving in
in windage windage
losses, losses,regulation,
voltage voltage regulation, and efficiency
and efficiency are
are calculated.
Usingcalculated. Using the set
the set parameters parameters
found in Tablefound in Table
2, the 2, the first
first group group is
of results of graphically
results is graphically
presented to
showpresented
the machine to show the machine efficiency and the internal power factor against the change in
efficiency and the internal power factor against the change in number of turns
number of turns (eventually the speed) as shown in Figure 5a. The second group of results are also
(eventually the speed) as shown in Figure 5a. The second group of results are also arranged to show
arranged to show the saving in windage losses, voltage regulation, and rotational speed against the
the saving in windage losses, voltage regulation, and rotational speed against the change in number of
change in number of turns as shown in Figure 5b. Such graphs are utilized to assist in selecting the
turnsbest
as shown in Figure
combinations of 5b. Such performance
machine graphs are utilized to assist
characteristics in selecting
hence, the best
the calculated combinations
parameters as of
machine
summarized in Table 3. It can be seen from the results that the optimum design is obtained at N t = 4It can
performance characteristics hence, the calculated parameters as summarized in Table 3.
be seen from
turns the PF
where results thatand
= 0.94 theefficiency
optimumofdesign
94% is is obtained
achieved. at Nt = 4atturns
Moreover, where
this point thePFsaving
m = 0.94
in and
m
efficiency of 94% is achieved. Moreover, at this point the saving in normalized
normalized windage losses is 93% and the design speed ( N m ) of 4814 rpm is achieved. The line–line
windage losses is 93%
and the design speed (Nm ) of 4814 rpm is achieved. The line–line RMS output voltage is dependent
RMS output voltage is dependent on the AFPMA speed. Hence the machine voltage/speed constant
on the AFPMA speed. Hence the machine voltage/speed constant mc which correlates the AFPMA
mc which correlates the AFPMA line–line RMS output voltage to its rotational speed N m can be
line–line RMS output voltage to its rotational speed Nm can be found as mc = VLL /Nm .
found as mc  VLL / N m .
(a) (b)
Figure 5. Optimization results for the AFPMA: (a) Machine efficiency and the internal power factor vs.
Figure 5. Optimization results for the AFPMA: (a) Machine efficiency and the internal power factor vs. number
number of turns; (b) Saving in windage losses, voltage regulation, and rotational speed vs. number
of turns; (b) Saving in windage losses, voltage regulation, and rotational speed vs. number of turns.
of turns.
Table 2. AFPMA Set Parameters.
AFPMA Set Parameters Value Unit

Desired line–line RMS output voltage 17.6 V
Number of phases (star connected) 3 -
Air gap half-distance, Xg 10 mm
Number of poles, p 8 -
Axial thickness of PM, Lm 17 mm
Inner radius of PM, Ri 27 mm
Outer radius of PM, Ro 48 mm
Remanence for NdFeB Grade 52, Br 1.45 T
Wire diameter, dW 2.2 mm
Energies 2018, 11, 274 8 of 28
Table 3. AFPMA Calculated Parameters.
AFPMA Calculated Parameters Value Unit

Number of winding turns per pole per phase, Nt 4 -
Winding resistance per phase, RW 22.8 mΩ
Winding inductance per phase, LW 4.14 µH
Design speed, Nm 4814 RPM
No load machine voltage/speed constant, mc 3.537 × 10−3 V/RPM
Generated frequency, f 320 Hz
Air gap flux density, Bg 0.546 T
Voltage regulation, VR % 10.3 -
Machine internal power factor (cos ϕ), PFm 0.94 -
Machine power angle, δ 2 Deg
Efficiency, η (%) 94 -
5. Modelling and Simulation of the AFPMA

The AFPMA under consideration has its magnet topology designed to produce a sinusoidal flux in
its ironless stator. This implies that the induced voltages will be almost sinusoidal. MATLAB/Simulink
R2012a (Mathworks, Galway, Ireland) [27], has a block that is already defined in the SimPowerSystems
which is called a permanent magnet synchronous machine (PMSM). This block is used to simulate the
performance of the AFPMA under consideration after setting the corresponding parameters as shown
in Figure 6.
Figure 6. AFPMA MATLAB/Simulink model and parameter configuration.

The AFPMA equivalent circuit is depicted in Figure 7, whereas the corresponding mathematical
TheofAFPMA
model equivalent
the machine in rotorcircuit is depicted
reference in Figure
frame (dq-axis) 7, whereas the
is summarized corresponding
in Appendix B. mathematical
model of the machine in rotor reference frame (dq-axis) is summarized in Appendix B.
d p m d
Ld id Lq i q Ld id Rwid
dt Rwid dt id
id 2
        
p m pm
Lq iq Vd r Vq
2 2
 
(a) (b)
Figure 7. AFPMA equivalent circuit: (a) d-axis model; (b) q-axis model.
Once the parameters of the AFPMA are defined, MATLAB/Simulink model can be used to
The11,
Energies 2018, AFPMA
274 equivalent circuit is depicted in Figure 7, whereas the corresponding mathematical 9 of 28
model of the machine in rotor reference frame (dq-axis) is summarized in Appendix B.
d p m d
Ld id Lq i q Ld id Rwid
dt Rwid dt id
id 2
        
p m pm
Lq iq Vd r Vq
2 2
 
(a) (b)
Once the parameters of the AFPMA are defined, MATLAB/Simulink model can be used to
Once thethe
predict parameters of the AFPMA
machine performance at are defined,
several MATLAB/Simulink
conditions which can bemodel cananalyzed
further be used after
to predict
experimental
the machine validation.atThe
performance expected
several line–line which
conditions rms voltages
can beoffurther
the machine at noafter
analyzed loadexperimental
and for
several The
validation. rotational speeds
expected are represented
line–line in Figure
rms voltages of 8.
the machine at no load and for several rotational
speeds are represented
Energies 2018, 11, x FOR PEERinREVIEW
Figure 8. 10 of 30
Figure 8. Expected AFPMA line–line rms voltages at no load for several operating speeds.
Figure 8. Expected AFPMA line–line rms voltages at no load for several operating speeds.
6. Development of the Controlled AC/DC Rectifier for the AFPMA

6. Development of the Controlled AC/DC Rectifier for the AFPMA
Regulation of the output voltage is a key issue in automotive alternator. The output voltage of
Regulation of the output voltage is a key issue in automotive alternator. The output voltage
the alternator shall be maintained at about 14 V DC, which is the nominal charging voltage of a 12 V
of the alternator shall be maintained at about 14 V DC, which is the nominal charging voltage of a
lead–acid battery. The Lundell alternator regulates its output voltage by controlling the current
12 V lead–acid battery. The Lundell alternator regulates its output voltage by controlling the current
supplied to the filed windings hence the magnetic flux inside the alternator. In PM machines where
supplied to the filed windings hence the magnetic flux inside the alternator. In PM machines where
the flux is relatively constant, the back-emf is only proportional to the alternator’s speed. Therefore,
the flux is relatively constant, the back-emf is only proportional to the alternator’s speed. Therefore,
driving the machine above the rated speed will generate a back-emf that is exceeding the rated
driving the machine above the rated speed will generate a back-emf that is exceeding the rated output
output voltage. Accordingly, at top speed the power electronic components in the converter shall be
voltage. Accordingly, at top speed the power electronic components in the converter shall be capable
capable to withstand both rated current and the considerable machine back-emf which may involve
to withstand both rated current and the considerable machine back-emf which may involve increased
increased power electronics cost and complexity. This could be solved by employing a DC-to-DC or
power electronics cost and complexity. This could be solved by employing a DC-to-DC or controlled
controlled AC-to-DC converter to maintain a constant output voltage. The separate converter can be
AC-to-DC converter to maintain a constant output voltage. The separate converter can be mounted in
mounted in a location where the temperatures and vibration levels are lower. When used together
a location where the temperatures and vibration levels are lower. When used together with an IGBT
with an IGBT based AC/DC converter, the alternator can respond in a fast manner to provide the
based AC/DC converter, the alternator can respond in a fast manner to provide the required charging
required charging current, at a stable voltage, over a large range of rotational speeds.
current, at a stable voltage, over a large range of rotational speeds.
The regulation and control of alternator’s output has been described in a number of studies [28–31].
Most of this work relies on using the Pulse Width Modulation (PWM) and boost rectifier technique.
Published research on the integration and implementation of fully controlled AC/DC converter to
PM generators is still scarce. In this paper, the description of the development of three-phase full
wave fully controlled rectifier system using six IGBT bridge will be demonstrated to regulate the
output of the AFPMA. As shown in Figure 9, the proposed system will utilize smart digital control
Energies 2018, 11, 274 10 of 28
The regulation and control of alternator’s output has been described in a number of studies [28–31].
Most of this work relies on using the Pulse Width Modulation (PWM) and boost rectifier technique.
Published research on the integration and implementation of fully controlled AC/DC converter to PM
generators is still scarce. In this paper, the description of the development of three-phase full wave
fully controlled rectifier system using six IGBT bridge will be demonstrated to regulate the output
of the AFPMA. As shown in Figure 9, the proposed system will utilize smart digital control using
specially designed comparators interfaced with Arduino Uno to generate the required pulses to the
gate of the IGBTs. The proposed system shall be capable to operate under several conditions such as
variable input voltage, variable input frequency, various loading conditions. The objective is to supply
aEnergies
regulated DCx FOR
2018, 11, voltage
PEERto a vehicle’s battery and loads under varying alternator speeds.
REVIEW 11 of 30
Figure 9.9.Proposed
Figure automotive
Proposed electrical
automotive powering
electrical systemsystem
powering using AFPMA harnessed
using AFPMA with fullywith
harnessed controlled
fully
AC/DC rectifier.
controlled AC/DC rectifier.
The following main components form the corner stones of the fully controlled AC/DC converter:
The following main components form the corner stones of the fully controlled AC/DC converter:
 IGBT six-pack module (CM100TF-12H)
• IGBT six-pack module
Microcontroller (Arduino (CM100TF-12H)
Uno)
• Microcontroller
Comparator circuit (Arduino
(voltageUno)
sensing).
• Comparator
LC filter. circuit (voltage sensing).
• LC filter.
6.1. Control Scheme
6.1. Control Scheme
To obtain the required and regulated DC voltage, the IGBT switches have to be triggered at a
properTo obtain
time and the sequence.
required and Theregulated
voltages DCof thevoltage,
threethe IGBTfrom
phases switches have to be
the AFPMA aretriggered
measured at
a proper time and sequence. The voltages of the three phases from the AFPMA
instantaneously to determine the most positive phase and the most negative phase using comparator are measured
instantaneously to determine
circuitry that is shown in Figurethe 10.
most positive phase
It compares and the(AB,
two phases mostAC,
negative phase using
BC) amplitudes and comparator
generates
circuitry that is shown in Figure 10. It compares two phases (AB, AC, BC) amplitudes
a pulse at the output if the phase at the non-inverting input is higher than the phase at the inverting and generates
ainput.
pulseThe
at the output if
generated the phase
pulses at the
are fed non-inverting
to Arduino input is
via a special higher thancircuit
conditioning the phase
and at theainverting
then decision
input.
is made The
to generated
trigger two pulses are switches
specific fed to Arduino
of the via a special
IGBT. conditioning
The switching circuit
scheme is and then
based on aadecision
specific
is made to trigger two specific switches of the IGBT. The switching scheme is
combination sequence as shown in Figure 11. To keep the output voltage at a specific (preset value) based on a specific
combination
voltage, the DC sequence
outputas shown
from in Figure
the filter 11. compared
is then To keep the output
with voltageset
the voltage at apoint
specific
and(preset value)
if the output
voltage,
voltage isthe DC output
lower than thefrom the filter
set point, the is then compared
rectifier with theaccordingly.
will be activated voltage set point and if the output
voltage is lower than the set point, the rectifier will be activated accordingly.
Energies 2018, 11, 274 11 of 28
Energies2018,
Energies 2018,11,
11,xxFOR
FORPEER
PEERREVIEW
REVIEW 12of
12 of30
30
VCC
VCC
77 11 55
Inputvoltage
Input voltagefrom
fromalternator
alternator
33
3PH
3PH 66
22
44 741
741
VCC
VCC
77 11 55
33
66
22
44 741
741
VCC
VCC
77 11 55
33
66
22
44 741
741
Figure10.
Figure 10.Voltage
Voltagemeasurement
Voltage measurementand
measurement andcomparator
and comparatorcircuit.
comparator circuit.
circuit.
Figure11.
Figure 11.IGBT
IGBTswitching
switchingsequence
sequencefor
forcontrolled
controlledrectifier
controlled rectifieroperation.
rectifier operation.
operation.
6.2.Controlled
6.2.
6.2. ControlledRectifier
Controlled RectifierDesign
Rectifier DesignCalculations
Design Calculations
Calculations
Thethree
The
The threephase
three phasefull
phase fullwave
full wavefully
wave fullycontrolled
fully controlled rectifier
rectifierisisisimplemented
controlledrectifier implemented
implementedusingusing six
six
using IGBTs
IGBTs
six IGBTs connected
connected
connected in
in
the
the form
form of
of aa full
full wave
wave bridge
bridge configuration.
configuration. All
All the
the six
six IGBTs
IGBTs are
are controlled
controlled switches
switches
in the form of a full wave bridge configuration. All the six IGBTs are controlled switches which are which
which are
are
turnedon
turned
turned onat
on atappropriate
at appropriatetimes
appropriate timesby
times byapplying
by applyingsuitable
applying suitablegate
suitable gatetrigger
gate triggersignals.
trigger signals.The
signals. Theaverage
The averageoutput
average outputvoltage
output voltage
voltage
VVdcdcdcofofofthe
V the
the rectifier
rectifier
rectifier can
can
can be
bebe calculated
calculated
calculated as
asas [32]:
[32]:
[32]:
√
33 22V
2VVLL
LL cos α
Vdc =
V
V 
dc
dc
LL
cos
cos (13)
(13)

π
where VLL is the rms line voltage from the three-phase AFPMA and α is the firing angle applied to the
where V
where VLL isisthe
IGBT gate.LL
therms
rmsline
linevoltage
voltagefrom fromthe
thethree-phase
three-phaseAFPMA
AFPMAand and  isisthe
thefiring
firingangle
angleapplied
applied
tothe
to the IGBT
IGBT
The gate. output voltage is regulated by controlling the firing angle of the IGBTs. To otbain
gate.
AFPMA
The
The AFPMA
AFPMA outputvoltage
output
a constant output voltage, avoltage
feedbackisisregulated
regulated bycontrolling
by
signal must controlling
be the
the
sent to the firingangle
firing
controllerangle ofthe
of theIGBTs.
to compare IGBTs. To
it withTothe obtain
presetaa
obtain
constant
constant
value output
output
of 14.3 voltage,
V tovoltage, a feedback
allow fora controlled
feedback signalsignal must
batterymust be sent
be sent
charging to the controller
to the controller
operation. to compare
to compare
When the battery it with
it with
is fully charged, the
the
preset
preset
the valueof
value
controller of 14.3
keeps14.3the
VV to
toallow
allow
firing forat
for
angle controlled
controlled
90 degrees. battery
battery charging
Whencharging operation.
operation.
the battery When
When
charge goes below thethe
the battery
battery
preset isislevel,
fully
fully
charged,the
charged, thecontroller
controllerkeeps
keepsthe thefiring
firingangle
angleat at90
90degrees.
degrees.When
Whenthe thebattery
batterycharge
chargegoes
goesbelow
belowthe the
preset level,
preset level, then
then the
the controller
controller triggers
triggers the
the IGBT
IGBT with
with aa firing
firing angle
angle below
below 9090 degrees
degrees which
which isis
Energies 2018, 11, 274 12 of 28
then the controller triggers the IGBT with a firing angle below 90 degrees which is proportional to the
voltage difference. The higher the voltage difference, the lower the firing angle will be.
The minim line–line RMS voltage VLL,min that is required to achieve the preset value of 14.3 V for
the battery charger at α = 0 degree can be determined from (13) as:
√
3 2VLL,min
14.3 = (14)
π
Rearranging for VLL,min
14.3π
VLL,min = √ = 10.58 V (15)
3 2
The minim speed of the AFPMA, Nm,min that keeps the output voltage at the preset value while
considering the machine constant value can be calculated as
VLL,min 10.58
Nm,min = = = 2997 RPM (16)
mc 3.537 × 10−3
At the design speed, Nm = 4814RPM, the line–line RMS output voltage will be 17.6 V and the
line-line peak output voltage, VLL,pk will be 24.89 V. Hence the maximum rectifier output voltage,
Vdc at α = 0o will be 23.7 V. The predicted results for AFPMA voltage and rectifier voltage with respect
to speed are summarized in Table 4.
Table 4. AFPMA and rectifier output voltages at no load vs. speed at α = 0o .
AFPMA Speed, Nm VLL,rms VLL,pk Vdc

(RPM) (V) (V) (V)
Design speed = 4814 17.6 24.89 23.7
Minimum speed = 2997 10.6 14.9 14.3
Maximum speed 5000 18.29 25.8 24.7
7. Overall System Simulation

The complete system that includes the AFPMA, integrated with the three-phase fully controlled
AC/DC converter, LC filter, battery, load, and controller is modeled using MATLAB/Simulink as
shown in Figure 12. The controller design relies on the measurement of the AFPMA output voltages
where both most positive and negative phases are determined at a certain instant. Based on the
feedback signal from the battery side, the controller generates the necessary pulses to trigger the IGBT
switches accordingly. For simulation purposes, logic OR gates are used to implement the switching
sequence that was presented in Figure 8. The algorithm that is used to enable the controller is depicted
in Figure 13. Whereas the controller output zero-one pulses are depicted in Figure 14.
Energies 2018, 11, 274 13 of 28
Figure 12. Complete system simulation using MATLAB/Simulink model.

Figure 12. Complete
Figure 12. Complete system
system simulation
simulation using
using MATLAB/Simulink model.
MATLAB/Simulink model.
Figure 13. Algorithm used to enable the controller for simulation purposes.
Figure
Figure 13.
13. Algorithm
Algorithm used
used to
to enable
enable the
the controller for simulation
controller for simulation purposes.
purposes.
In practice, charging system in an automobile shall be capable to operate at various loads and
In practice, charging system in an automobile shall be capable to operate at various loads and
various engine/alternator
In practice, speed while
charging system in an supplying
automobilea shall
regulated/fixed
be capable voltage to battery
to operate terminals
at various and
loads and
various engine/alternator speed while supplying a regulated/fixed voltage to battery terminals and
any loads connected across the terminals. In the following section, simulations for different
various engine/alternator speed while supplying a regulated/fixed voltage to battery terminals and scenarios
any loads connected across the terminals. In the following section, simulations for different scenarios
are developed
any in order
loads connected to predict
across the system
the terminals. behavior
In the at various
following section,operating conditions.
simulations for different scenarios
are developed in order to predict the system behavior at various operating conditions.
are developed in order to predict the system behavior at various operating conditions.
Energies 2018, 11, 274 14 of 28
Figure 14. Controller output pulses (zero–one).
7.1. System
7.1. Simulation with
System Simulation with Open
Open Circuit
Circuit
In this
In this scenario,
scenario, the
the AFPMA
AFPMA speed
speed isis set
set to
to 5000
5000 RPM,
RPM, thethe load
load and
and the
the battery are disconnected
battery are disconnected
from the
from the rectifier
rectifier terminals,
terminals,and
andthe
thefiring angleα isis set
firingangle set to
to zero.
zero. The
Thesix
sixpulse
pulserectified
rectifiedoutput
outputasasa
aresult
resultfrom
fromthe
thesimulation
simulationisisshown
shownininFigure
Figure15.
15.The
TheAFPMA
AFPMA line–line
line–line peak
peak voltage
voltage isis found to be
found to be
V, hence
24.8 V, hencethe
therectifier
rectifieroutput
outputvoltage Vdcis found
voltageVdc is found to be
to be 23.723.7 V. Such
V. Such initial
initial simulation
simulation results
results are
validated by comparing them to the analytical results that are presented earlier in Table 4. Hence
are validated by comparing them to the analytical results that are presented earlier in Table 4. Hence other
simulation scenarios
other simulation can becan
scenarios conducted with more
be conducted with confidence.
more confidence.
Energies 2018, 11, 274 15 of 28
Figure 15.
Figure 15. Simulation
Simulationresults
resultsfor
foropen
opencircuit
circuitrectifier
rectifieroutput
outputvoltage
voltageatat5000
5000rpm
rpmatatfiring angleα == 0.
firingangle 0.
7.2. System
Figure 15. Simulation results for open circuit rectifier output voltage at 5000 rpm at firing angle  = 0.
7.2. System Simulation
Simulation with
with Constant Load and
Constant Load and Variable
Variable Speed
Speed
7.2. System Simulation
In this with Constant Loadload and Variable Ω Speed
In this scenario,
scenario, aaconstant
constantelectrical
electrical loadofof0.70.7 Ωis fitted across
is fitted thethe
across battery terminals
battery terminals while the
while
speed
the of AFPMA
speed ofInAFPMA is ramped
this scenario,
is ramped upup
a constant from 3800
electrical
from loadRPM
3800 of 0.7 to
RPM Ωtois5100
5100RPM
fitted RPMthrough
across the batterythe
through simulation
terminals
the while the
simulation runtime.
runtime.
speedresults
Simulation of AFPMAare is rampedin
depicted upFigure
from 380016 RPM tobattery
where 5100 RPM throughvoltage
terminal the simulation
(after runtime.rectifier
filter),
Simulation results are depicted in Figure 16 where battery terminal voltage (after filter), rectifier
voltage Simulation
voltage (before
resultsand
(before filter),
filter), are load
and and
depicted
loadload
in Figure
current
current
16 where battery
are captured
are captured terminal
in response
in response tovoltage
speed (after filter),with
variation rectifier
respect to
to
voltage (before filter), current are captured in response to to speed
speed variation
variation with respect
with respect to
time. The
time. The charging
charging limit for the
limit limit
for the battery
battery is is set
set to 14.3 V. V.
time. The charging for the battery is settoto14.3
14.3 V.
FigureFigure 16. Simulation

16. Simulation results
results forfor constant electrical
constant electrical load
loadand variable
and AFPMA
variable speed.speed.
AFPMA
Figure 16. Simulation results for constant electrical load and variable AFPMA speed.
Energies 2018, 11, 274 16 of 28
7.3.System
7.3. SystemSimulation
Simulationwith
withConstant
ConstantSpeed
Speedand
andVariable
VariableLoad
Load
Inthis
In thisscenario,
scenario,thethespeed
speedof ofAFPMA
AFPMAisisset setto
to5100
5100RPM
RPMwhilewhilethe
theisiselectrical
electricalload
loadconnected
connected
acrossthe
across thebattery
battery terminal
terminal is decreased from 3 Ω to 0.7 Ω
is decreased Ω of
of through
throughthe thesimulation
simulationruntime
runtimefrom
from0
to 1.5 s. Simulation results are depicted in
0 to 1.5 s. Simulation results are depicted in Figure Figure 17 where battery terminal voltage (after filter),
terminal voltage (after filter),
rectifiervoltage
rectifier voltage(before
(beforefilter),
filter),and
andload
loadcurrent
currentarearecaptured
capturedininresponse
responsetotoelectrical
electricalload
loadvariation
variation
withrespect
with respecttototime.
time.The
Thecharging
charginglimit
limitfor
forthe
thebattery
batteryisisset
settoto14.3
14.3V.V.
Figure 17.Simulation
Figure17. Simulationresults
resultsfor
forconstant
constantspeed
speedand
andvariable
variableelectrical
electricalload.
load.
7.4.System
7.4. SystemSimulation
Simulationwith
withVariable
VariableSpeed
Speedand
andVariable
VariableLoad
Load
Inthis
In thisscenario,
scenario,both
bothspeed
speedofofAFPMA
AFPMAand andelectrical
electrical load
load areare varied
varied simultaneously.
simultaneously. TheThe speed
speed is
is increased
increased fromfrom
38003800
RPMRPM toRPM
to 5100 5100while
RPMthewhile the electrical
electrical load acrossloadtheacross
batterythe batteryisterminal
terminal decreasedis
from 3 Ω tofrom
decreased 0.7 Ω3of
Ω through
to 0.7 Ω the
of through the runtime
simulation simulation runtime
from 0 to 1.5from 0 to 1.5 s.results
s. Simulation Simulation results are
are depicted in
depicted in Figure 18 where battery terminal voltage (after filter), rectifier voltage (before
Figure 18 where battery terminal voltage (after filter), rectifier voltage (before filter), and load currentfilter), and
load current are captured in response to electrical load variation with respect to time. The charging
limit for the battery is set to 14.3 V.
Energies 2018, 11, 274 17 of 28
are captured in response to electrical load variation with respect to time. The charging limit for the
Energies 2018,
battery is set11,tox FOR
14.3 PEER
V. REVIEW 18 of 30
Figure 18.
Figure 18. Simulation
Simulation results
results for
for variable
variable speed
speed and
and variable
variable electrical
electrical load.
load.
8. Testing
8. Testing and
and Experimentation
Experimentation of
ofthe
theAFPMA
AFPMA
The AFPMA
The AFPMAalongalongwith
withthethe controlled
controlled AC/DC
AC/DC converter
converter are are experimentally
experimentally testedtested to obtain
to obtain their
their performance characteristics as shown in Figure 19. As an alternative prime
performance characteristics as shown in Figure 19. As an alternative prime mover to the car engine, mover to the car
aengine, a three-phase
three-phase electricalelectrical
motor thatmotor
is fedthat is fed
from from three-phase
three-phase variable
variable power power
supply is supply
used to is usedthe
drive to
drive the AFPMA at several speeds through pulleys and belt. A dummy resistive load
AFPMA at several speeds through pulleys and belt. A dummy resistive load is used to test the system is used to test
the system
under under
various various
loading loadingThe
conditions. conditions.
three-phaseTheoutput
three-phase
from theoutput
AFPMAfromisthe
fedAFPMA is fed toload
to the electrical the
electricalthe
through load through
IGBT whichthe IGBTcontrolled
is fully which is fully controlled
via the via
controller the controller
circuit. circuit.circuit
The controller The controller
contains
circuit contains the comparator circuit, Arduino Uno, and the IGBT
the comparator circuit, Arduino Uno, and the IGBT gate driver circuit. To demonstrate the gate driver circuit.
systemTo
demonstrate atthe
performance system
several performance
operating at several
conditions, operating
the following tests conditions,
are performed theasfollowing
described tests
below.are
performed as described below.
Energies 2018, 11, 274 18 of 28
Figure 19. Experimental test rig; (1) Three-phase driving motor, (2) AFPMA, (3) controller circuit,
Figure(4)19. Experimental
IGBT six-pack moduletest CM100TF-12H,
rig; (1) Three-phase drivingpower
(5) three-phase motor, (2) AFPMA,
supply, (3) resistive
(6) dummy controller circuit,
load,
(4) IGBT six-pack module CM100TF-12H, (5) three-phase power supply,
(3a) comparator circuit, (3b) Arduino Uno, (3c) IGBT gate driver circuit. (6) dummy resistive load,
(3a) comparator circuit, (3b) Arduino Uno, (3c) IGBT gate driver circuit.
Figure 19. Experimental test rig; (1) Three-phase driving motor, (2) AFPMA, (3) controller circuit,
8.1. No
(4) Load
IGBTTests
six-pack module CM100TF-12H, (5) three-phase power supply, (6) dummy resistive load,
(3a)Tests
8.1. No Load comparator circuit, (3b) Arduino Uno, (3c) IGBT gate driver circuit.
In this section, the system is tested under no load condition with AFPMA initially is running at
rated speed of 5000 RPM. The line–line output voltage from the AFPMA is captured and compared
In
8.1.this section,
No Load Teststhe system is tested under no load condition with AFPMA initially is running at
with the simulation results obtained from MATLAB as shown in Figure 20. Results obtained show a
rated speed of 5000 RPM. The line–line output voltage from the AFPMA is captured and compared with
goodIn this section,
match betweenthe thesystem
simulation is tested under no load
and experimental condition
results. with AFPMA
The results initially is line-neutral
of the three-phase running at
the simulation
rated
voltagesspeed results
are of 5000obtained
captured RPM.
using The from MATLAB
line–line
PicoScope output
3000
as shown
voltage
oscilloscope
in Figure
from
(Pico the AFPMA 20. Results
Technology,
obtained
is captured show
and compared
Cambridgeshire, UK) as
a good
match between
with
depicted the
the simulation simulation
in Figure 21. results and
The obtained
quality ofexperimental
fromthe MATLAB results.
AFPMA voltageas shown The results
in Figure
signal is alsoof the three-phase
20.analyzed
Results obtained
where the line-neutral
show
totala
voltages
good are
harmonic captured
match betweenusing
distortion the
(THD) PicoScope
simulation
is found and 3000
to oscilloscope
experimental
be close to 2% as (Pico
results. Technology,
The results
shown in Figureof the22.Cambridgeshire,
three-phase
The output line-neutral
from the UK) as
voltages
depicted
rectifier arezero
in Figure
at captured
21.
firing using
The angle PicoScope
quality
is also the3000
of captured oscilloscope
AFPMA voltage
in Figure (Pico ItTechnology,
23. signalis shownis also Cambridgeshire,
analyzed
from where
the results UK)
that as total
the
the
depicted
rectified
harmonic in Figure
output
distortion 21. The
voltage
(THD) quality
isisvery
found of to
close
to the
bethe AFPMA
output
close to 2%voltage
obtained signal
as shown fromin isthealso analyzed
simulation
Figure 22. The where
results. theis total
It
output also the
from
harmonic
shown
rectifier thatdistortion
at zero phase angle
firing (THD)
C has its is found
peak
is also to bevoltage
line–line
captured close
in toless
Figure 2% asItshown
than
23. theshown
is in Figure
other two
from 22.
the The
phases by output
thatfrom
1.5 volts.
results This
the the
is
rectified
rectifier
explained atby zero
thefiring
fact angle
that the iscoil
also captured
that is in Figure
responsible for 23. It is
phase C shown
is from located
physically the results that the
between
output voltage is very close to the output obtained from the simulation results. It is also shown that
rectified
other twooutput
phasesvoltage is very of
in the middle close to the
the air gap.output obtained
Therefore, from the flux
less magnetic simulation
density results. It is also
arrives from PM
phase C has its peak line–line voltage less than the other two phases by 1.5 volts. This is explained by
shown
to that that phasehence
location, C haslessits peak
induced line–line
voltage voltage less than
is produced in the
thatother
coil. two
Thisphases
makes by the1.5 volts.voltage
output This is
the fact that thebycoil
explained the
that isthatresponsible for phase C is physically located between the other two phases
asymmetrical andfactspecial the coil that ishave
precautions responsible
to be takenfor phase
into C is physically when
consideration locateddesigning
between the
in theother
middle
scheme two of the
forphases air gap.
in the middle
the controller Therefore,
of the of the air gap. Therefore, less magnetic flux density arrives from location,
converter.less magnetic flux density arrives from PM to that PM
hencetoless
thatinduced
location,voltage
hence less is produced
induced voltage in thatis coil. This makes
produced the output
in that coil. This makes voltage asymmetrical
the output voltage and
special precautionsand
asymmetrical have to beprecautions
special taken into consideration
have to be taken when intodesigning
consideration the scheme for the controller
when designing the
of thescheme for the controller of the converter.
converter.
Figure 20. AFPMA no load voltage (line-neutral) experimental results vs. simulation results at
N m = 5000 RPM.
Figure
Figure 20. AFPMA
20. AFPMA no no load
load voltage(line-neutral)
voltage (line-neutral) experimental
experimentalresults vs.vs.
results simulation results
simulation at
results at
N = 5000
Nm = 5000RPM.
m RPM.
Energies 2018, 11, 274 19 of 28
Figure 21. AFPMA no load three-phase line-neutral voltages tested at N m = 5000 RPM.
Figure 21. AFPMA no load three-phase line-neutral voltages tested at Nm = 5000RPM.
Figure 21. AFPMA no load three-phase line-neutral voltages tested at Nm = 5000 RPM.
Figure 21. AFPMA no load three-phase line-neutral voltages tested at Nm = 5000 RPM.
Figure 22. AFPMA phase voltage THD % experimental result at Nm = 5000 RPM.
Figure 22. AFPMA phase voltage THD % experimental result at N m = 5000 RPM.
Figure 22. AFPMA phase voltage THD % experimental result at N= 5000
Figure 22. AFPMA phase voltage THD % experimental result at N m m
= 5000RPM.
RPM.
Figure 23. Rectifier output voltage at zero firing angle at Nm = 5000 RPM.
Figure 23. Rectifier output voltage at zero firing angle at Nm = 5000RPM.
Energies 2018, 11, 274 20 of 28
The controller
Energies of the
2018, 11, x FOR PEER AC/DC
REVIEW converter is designed to deal with the voltages under asymmetrical 21 of 30
conditions.
Energies 2018, 11, x FOR PEER REVIEWthat gives the controller the adaptively feature as it compares
It is the comparator 21 of 30 the
The controller of the AC/DC converter is designed to
most positive signal with the most negative one. Therefore, an adaptive pulse is generated from the deal with the voltages under
comparator Thethat
asymmetrical controller
isconditions.of the
proportional It isAC/DC
the converterthat
comparator
to the difference is designed
gives
between thethe totwodeal
controller with
the the phases.
adaptively
consecutive voltages
feature under
as itin an
Even
asymmetrical
compares the conditions.
most positiveIt issignal
the comparator
with the that gives
most negative the controller
one. the adaptively
Therefore, an adaptive feature
pulseas isit
asymmetrical situation where the three phase voltages are not balanced, the comparator is capable to
compares from
generated the most positive signal
the comparator that with the most negative
is proportional one. Therefore,
to the difference betweenan theadaptive pulse is
two consecutive
create the necessary pulse and send it to Arduino Uno to generate the necessary pulses for triggering
generated
phases. Evenfrom in the
an comparator
asymmetrical that is proportional
situation where the to the
three difference betweenare
phase voltages thenot
two consecutive
balanced, the
the appropriate
phases. Evenis switches
in in the IGBT
an asymmetrical module.
situation For the
where example,
three Figure
phase 24 showsnot the output from the
comparator capable to create the necessary pulse and send it to voltages
Arduinoare Uno tobalanced,
generate the
comparator
comparator
necessary forpulses
Cisand phases
capable
for A. When
to create
triggering thethe Cnecessary
phase (blue
appropriate pulsecolor)
switches and is most
send
in the positive
it to
IGBT Arduino
module. than
ForUno phase AFigure
(green
to generate
example, 24color)
the
the comparator
necessary generates
pulses for the
triggeringappropriate
the pulse
appropriate (yellow
switches incolor)
the throughout
IGBT
shows the output from the comparator for C and phases A. When C phase (blue color) is most module. that
For interval.
example, The
Figure results
24
also show
shows that
the the
output peak fromvalue
the of phase
comparator A is
for higher
C and than
phases that
A. for
When phase
positive than phase A (green color) the comparator generates the appropriate pulse (yellow color) C C.
phase As a
(blue result,
color) an
is adaptive
most
positive
pulsethroughout than
is generated thatphase
from Athe
interval. (green
The color) the
comparator
results also comparator
witnessed
show thatby generates
the an increase
peak the of
value appropriate
inphase A ispulse
the pulses width
higher (yellow
than color)
(yellow
that forcolor).
throughout
phase C.from
The output that
As a the interval.
result, The results
an adaptive are
comparators also
pulse show that
is generated
sent to Arduino the
from peak value
the comparator
Uno of phase A is higher
witnessed by an
where the appropriate than that
increase
train for
of pulesin are
phase
the
generated C. As
pulses
and a result,
width
sent an adaptive
to(yellow
IGBT color).
gate pulsecircuit
The
driver is generated
output from
from which
from the the
thecomparator
comparators
IGBT are witnessed
sent
switches toare by anUno
Arduino
triggered increase
where in
accordingly.
the pulses
appropriatewidthtrain (yellow
of color).are
pules The output from
generated and the
sent comparators
to IGBT gateare sent to
driver Arduino
circuit from Uno
whichwhere
the
As can be shown from Figure 25, the six Arduino pulses have different widths to allow for adaptive
the appropriate
IGBT switches are train of pules
triggered are generated
accordingly. andbe
As can sent to IGBT
shown fromgate driver
Figure circuit
25, the six from
Arduino which the
pulses
control of the AC/DC converter.
IGBTdifferent
have switcheswidths
are triggered
to allowaccordingly.
for adaptiveAs can be
control ofshown fromconverter.
the AC/DC Figure 25, the six Arduino pulses
have different widths to allow for adaptive control of the AC/DC converter.
Figure 24. Phase C and A with adaptive pulse generated from the comparator at Nm = 5000 RPM.
Figure Phase
24. 24.
Figure CC
Phase and
andAAwith
withadaptive pulsegenerated
adaptive pulse generatedfrom
from
thethe at N
comparator
comparator = 5000RPM.
at N=m5000 RPM. m
Figure 25. Arduino pulses with appropriate pulse widths to allow for adaptive control of AC/DC
Figure 25. Arduino pulses with appropriate pulse widths to allow for adaptive control of AC/DC
converter.
Figure 25. Arduino pulses with appropriate pulse widths to allow for adaptive control of
converter.
AC/DC
8.2. Loadconverter.
Tests
8.2. Load Tests
In this section, the AFPMA and the AC/DC converter are tested under load conditions as shown
8.2. Load Tests
In this
in Figure 26.section, theofAFPMA
First set results and the AC/DC
is obtained converter
when are tested
the system under load
is connected conditions
to the as shown
battery only. The
in Figure
In this set
second 26.
section,Firstthe
of resultsset of results when
is AFPMA
obtained is
andobtained whenisthe
the system
AC/DC system to
converter
connected is are
connected
the tested to
battery the
under
and battery only.
loadhead
a single The as
conditions
lamp.
shown second
The setset
inthird
Figure of of
results
26. is is
results
First obtained
obtained
set when
when
of results the
is system
the iswhen
system
obtained connected
is connected tobattery
the battery
the system isand and head
two a single
connected head
lamps.
to the lamp. only.
Results
battery
The third
The second setsetofofresults
results is
is obtained
obtainedwhenwhenthethe
system is connected
system battery
is connected toand
thetwo head and
battery lamps. Resultshead
a single
Energies 2018, 11, 274 21 of 28
lamp. The third set of results is obtained when the system is connected battery and two head lamps.
Results obtained for11,the
Energies 2018, three
x FOR PEERloading
REVIEW conditions as compared to the no load condition are 22 ofsummarized
30
in Table 5. It is shown from the results that the controller was successful in maintaining almost
obtained for the three loading conditions as compared to the no load condition are summarized in Table
a constantEnergies
output voltage close to 14 V DC with an average decrease of 2.7% for a 30% increase in
5. It is2018,
shown 11, x FOR PEER
from the REVIEW
results that the controller was successful in maintaining almost a constant22 of 30
load current and 6.6% decrease in average output voltage for 60% increase in
output voltage close to 14 V DC with an average decrease of 2.7% for a 30% increase in load current load current. Also,
obtained for the three loading conditions as compared to the no load condition are summarized in Table
when analyzing
and 6.6%the alternator’s
decrease in averageoutput
output waveforms,
voltage for 60%it is noticed
increase in loadthat the Also,
current. line–line
when voltage
analyzinggets more
5. It is shown from the results that the controller was successful in maintaining almost a constant
thethe
distorted as alternator’s
controlled output waveforms,
converter is it is noticedto
processed that the line–line
provide more voltage gets more
charging distorted
current while asthe
the alternator
output voltage close to 14 V DC with an average decrease of 2.7% for a 30% increase in load current
controlled converter is processed to provide more charging current while the alternator running is
running isand
also decreased
6.6% decrease inasaverage
can beoutput
shown in Figure
voltage for 60%27.increase in load current. Also, when analyzing
also decreased as can be shown in Figure 27.
the alternator’s output waveforms, it is noticed that the line–line voltage gets more distorted as the
controlled converter is processed to provide more charging current while the alternator running is
also decreased as can be shown in Figure 27.
Figure 26. Testing of AFPMA and AC/DC converter under charging process (battery and load connected).
Figure 26. Testing of AFPMA and AC/DC converter under charging process (battery and load
connected).
Figure 26. Testing of AFPMA and AC/DC converter under charging process (battery and load connected).
(a) (b) (c)
Figure 27. Distorted line–line output voltages due to increased load current; (a) without rectifier and
no load at 5085(a)rpm; (b) with rectifier and no load(b)
at 5032 rpm; (c) with rectifier and charging
(c) current
of 11 A at 4740 rpm.
Figure 27. Distorted line–line output voltages due to increased load current; (a) without rectifier and
Figure 27.noDistorted
loadTable line–line
at 5085 output voltages dueatto5032
increased loadrectifier
current; (a) without rectifier and
5. rpm;
AFPMA(b) with rectifier
and AC/DC and no load
converter rpm;conditions
under loading (c) with vs. no and
loadcharging
conditioncurrent
no load atof5085
11 Arpm;
at 4740(b) with rectifier and no load at 5032 rpm; (c) with rectifier and charging current
rpm.
of 11 A at 4740 rpm. Speed VLL ,rms Vdc I dc
Table 5. AFPMATest Condition
and AC/DC converter under loading conditions vs. no load condition
(RPM) (V) (V) (A)
No load (rectifier
Table 5. AFPMA and AC/DC converter under only) 5032
Speed VLL ,conditions
loading18.41
rms Vdc vs.I 0dcno load condition.
24.15
Test Condition
Battery only 5020 18.01 14.53 2.85
(RPM) (V) (V) (A)
Battery and one head lamp 4876 16.89 13.57 7.62
No load (rectifier only) Speed 5032 V18.41
LL,rms 24.15Vdc 0 Idc
Test Condition
Battery and two head lamps 4740 16.05 13.21 11.04
Battery only 5020 18.01 14.53 2.85
(RPM) (V) (V) (A)
Battery and one head lamp 4876 16.89 13.57 7.62
9. Conclusions and Future Recommendations
No load (rectifier
Battery and twoonly)
head lamps 5032 4740 18.41 13.2124.15
16.05 11.04 0
Batterythe
This paper exhibited only
design methodology 5020 and optimization
18.01 14.53
of AFPMA2.85 having ACSTR
9. Battery
configuration
Conclusions and and
andits one headwith
integration
Future lamp 4876fully controlled
three phase
Recommendations 16.89 AC/DC 13.57 7.62
converter as part of the
Battery
onboard battery and two
charging headfor
system lamps 4740 applications.
use in automobile 16.05 The13.21 necessary11.04
sizing equations
This paper exhibited the design methodology and optimization of AFPMA having ACSTR
were developed, and optimization of the machine was performed based on parametric analysis. The
configuration and its integration with three phase fully controlled AC/DC converter as part of the
number of winding turns, hence machine design speed, were optimized against the machine
onboard
9. Conclusions battery
and
efficiency.
charging
Future
Analytical
system for use in automobile applications. The necessary sizing equations
Recommendations
results showed that the AFPMA having four turns per coil per phase is suitable
were developed, and optimization of the machine was performed based on parametric analysis. The
This number
paper exhibited
of winding theturns,design methodology
hence machine and optimization
design speed, were optimized of AFPMA
against having ACSTR
the machine
efficiency.
configuration and Analytical results showed
its integration that thephase
with three AFPMA having
fully four turns per
controlled AC/DCcoil perconverter
phase is suitable
as part of the
onboard battery charging system for use in automobile applications. The necessary sizing equations
were developed, and optimization of the machine was performed based on parametric analysis.
The number of winding turns, hence machine design speed, were optimized against the machine
efficiency. Analytical results showed that the AFPMA having four turns per coil per phase is suitable to
Energies 2018, 11, 274 22 of 28
run at nearly 4800 rpm to achieve an power output of 1.2 kW at rated line voltage of 17 and rated current
of 40 A at an expected efficiency of 94%. The AFPMA was modeled using MATLAB/Simulink and the
results obtained were close to analytical solutions. The response of the AFPMA while operating with
a three-phase controlled rectifier using six IGBT bridge is modeled and experimentally tested, forming
the basis for the design of a closed-loop controller to regulate the output voltage. The experimental
performance of the closed-loop system is examined and evaluated. A custom design comparator circuit
was developed to produce the necessary pulses for Arduino Uno and was capable of dealing with
asymmetrical output voltage conditions from the AFPMA. The Arduino was used as a controller that
generates the required pulses for triggering the IGBTs sequentially in a timely manner. The Arduino
custom code enabled the production of a regulated DC output voltage across the battery via controlling
the firing angle of the IGBT model in proportion to the amount of voltage difference between the
battery voltage preset value and the measured value. The overall system was experimentally tested
under partial loading conditions and results obtained showed the validity of the analytical as well as
the simulation results within acceptable level. Experimental results demonstrated that the closed-loop
controller was successful in regulating the output voltage close to 14 V DC while the alternator was
operating in the range of the rated speed. Measurements showed a 2.7% decrease in average output
voltage for a 30% increase in load current. It is worth mentioning that some amount of the voltage drop
can be attributed to the drop in the alternator speed (approximately 2.9%) due to the increased load
torque on the electrical motor that is driving the alternator. For accurate evaluation and assessment
of the system performance, it is recommended to use alternative prime movers that can provide
constant speed during load variations. Further investigations of the output power for the system
using the controlled rectifier are also required at higher speed operation. Moreover, it was noticed
that during the loading tests, that the machine voltage/speed constant (mc ) tends to change under
high current loading. This can be explained by the flux weakening that occurs in the middle of the
air gap due to armature reaction. More experimental results need to be conducted to determine the
exact relationship and to account for this change in the MATLAB/Simulink model. Also, preliminary
laboratory tests showed a rise of 50 ◦ C in the temperature of the stator winding above ambient
temperature of 25 ◦ C when the full load current of 40 A is drawn from the machine for a period
of two hours without using integral fan. It is anticipated that the stator winding temperature may
rise in the range of 100–130 ◦ C when the AFPMA is tested in an engine compartment due to the
elevated ambient temperature. This in turn may cause a temperature increase in the nearby magnets
which may lead to demagnetization of the PMs. Therefore, the cooling aspect of the AFPMA needs
to be considered in order to minimize the temperature rise by using an integral fan on the shaft or
other methods of forced ventilation. More experimental tests are required to test the system under
several full loading conditions. Using alternative grades of NdFeB can be decided once the machine
is tested on-board of the vehicle at several loading conditions. In general, the preliminary results
showed the potential capabilities of the AFPMA for meeting increasing power demands in automotive
applications. In addition, a scaled up system can be also made in the future to satisfy higher power
ratings of automobiles.
Acknowledgments: The author acknowledges that this research is funded by the Deanship of Research and
Graduate Studies at Zarqa University/Jordan. The author also would like to extend his gratitude to the laboratory
supervisor Eng. Mohammad Mosleh as well as to Eng. Basil Al Jamal and Eng. Yazan Shihab for their help in
conducting the experimental work at the electrical engineering department of Zarqa University.
Conflicts of Interest: The author declares no conflict of interest.
Energies 2018, 11, 274 23 of 28
Nomenclature
Amg Air gap/magnet interface area
Aw Winding cross sectional area
Bg Air gap magnetic flux density
Br Magnetic remanence
Cf Coefficient drag for turbulent flow
Cl Coil mean path length
dW Wire diameter
Eind Induced voltage in a conductor
Per phase internal generated
E ϕ,rms
voltage
f Electrical frequency of the voltage
Hc Magnetic coercive force
IA Armature current
Id d-axis current
Iq q-axis current
Kg Machine geometrical constant
Km Machine magnetic constant
Lc Conductor length
Ld d-axis inductance
LL Per phase leakage inductance
Lm Magnet axial length
Lmg Total magnet air gap length
Lq q-axis inductance
Lw Winding Inductance
LZ Magnetization inductance
mc Machine voltage/speed constant
Nm Mechanical speed
Minim acceptable speed of the
Nm,min
AFPMA
NR Number of rotating discs
Nt Number of winding turns
p Rotor number of poles
Pcu Copper losses
pf Wire packing factor
Internal power factor for the
PFm
machine
Pout Machine output power
PW Windage losses
PW,max Maximum windage losses
PW,sav Saving in the windage losses
r Average radius of conductor
R Centre point of the PM disc
Re Reynolds number
Reluctance of the path of the
Req
magnetising flux
Ri Inner radius of the coil
Ro Outer radius of the coil
Rsh Shaft radius
RW Winding resistance
Te Electromagnetic torque
→
v Linear velocity vector
Vd d-axis voltage
Vdc Rectifier average output voltage
VLL RMS line voltage
Energies 2018, 11, 274 24 of 28
VLL,min Minim line-line RMS voltage

VR Voltage regulation of the machine
Vq q-axis voltage
Vϕ, f l Full load phase voltage
Vϕ,nl No load phase voltage
Wl Total winding length
Xg Air gap half-distance
XW Winding reactance
α Rectifier firing angle
αB Temperature coefficient for Br
αH Temperature coefficient for Hc
δ Machine power angel
η Efficiency of the machine
λr Peak flux reaching the stator windings
µ air Coefficient of viscosity of the air
ρ air Specific density of the surrounding air
ωm Mechanical angular velocity
Appendix A. Derivation of the Sizing Equations for the AFPMA

The magnetic flux density in the air gap between the two PM rotors (assuming circular PMs) is calculated
as [33–36]:  
X g + Lm Xg
Bg = Br  q −q  (A1)
R2 + ( X g + L m )2 R2 + Xg2
where R is center point of the PM disc which can be determined as:
R o − Ri
R= (A2)
2
Hence, the induced voltage in the conductors is given by:
→ →
Eind = ( v × B g ) · Lc (A3)
→ →
Since v is perpendicular to B g , the magnitude of the induced voltage reduces to:
Eind = v · Bg · Lc (A4)
The tangential velocity of the conductor can be represented in terms of angular velocity ωm as:
v = ωm · r (A5)
where r is the average radius of the conductor (m), also called the magnet centerline and can be found as:
R o + Ri
r= (A6)
2
The angular velocity ωm can be expressed in terms of mechanical speed of the rotor Nm as:
2π
ωm = Nm (A7)
60
The average effective length of the winding conductor, Lc can be calculated as
L c = 2( R o − R i ) (A8)
Substituting (A7) into (A5) and then into (A4)
π · Nm
Eind = r · Bg · L c (A9)
30
Energies 2018, 11, 274 25 of 28
For number of turns Nt per phase and for number of poles p, substituting (A6) and (A8) into (A9) can be
represented as:
π 2
Eind = ( R − R2i ) · Bg · Nm · Nt · p (A10)
30 o
For three phase stator, the line-line rms induced voltage can be expressed as:
√
π 3 2
ELL = √ ( Ro − R2i ) · Bg · Nm · Nt · p (A11)
30 2
Hence, the number of turns can be found as:

√
30 2 · ELL
Nt = √ (A12)
π 3 · ( R2o − R2i ) · Bg · Nm · p
 q q 
( X + L ) R 2 + X2 − X R 2 + ( X + L )2
g m g g g m
Bg = Br  (A13)
 
q q 
2 2 2 2
R + ( X g + Lm ) · R + X g
 q q 
ELL 7.796  R2 + ( Xg + Lm )2 · R2 + Xg2
Nt = (A14)

Nm p · Br
 q q 
R · r · [( Xg + Lm ) R2 + Xg2 − Xg R2 + ( Xg + Lm )2 ]
The above equation can be reduced to:
ELL
Nt = · Km · K g (A15)
Nm
" q
2 √ #
7.796 R2 +( Xg + Lm ) ·
R2 + Xg2
where Km = p· Br , and K g = √ 2 2 q
2
are the machine magnetic constant and
R·r ·[( Xg + Lm ) R + Xg − Xg R2 +( Xg + Lm ) ]
geometric constant, respectively.
The frequency of the generated voltage is determined by:
p · Nm
f = (A16)
120
The AFPMA under consideration has a non-salient rotor topography. Also the stator has an air-cored
structure which enables sinusoidal flux distribution. Hence, the inductance of the machine can be calculated
as [37]:
Lw = L L + L Z (A17)
The per phase leakage inductance can be calculated as:
pNt3 Cl
L L = µo (A18)
12
where Cl is the coil mean path length which can be determined based on the geometry of the coil as:

2π Nt dw N dw
Cl = p Ro + + Ri − t + 2( Ro − Ri + Nt dw ) · p f (A19)
p f 2 2

π π
Cl = 2Ro + 1 + 2Ri − 1 + 2Nt dw · p f (A20)
p p
Wl = p · Nt · Cl (A21)
The winding resistance can be calculated as:
ρ · Wl
RW = (A22)
Aw
where Wl is the total winding length, Aw is the winding cross sectional area, dw is the wire diameter, and p f is
the wire packing factor and it can be defined in the range between 1.2 and 1.3 depending on the wire diameter.
Energies 2018, 11, 274 26 of 28
By performing magnetic circuit analysis, the reluctance of the PM and the coil can be calculated, hence the
magnetizing inductance can be determined as:
pNt2
LZ = (A23)
Req
where Req is total reluctance of the path of the magnetising flux and can be calculated as:
Lmg
Req = (A24)
µo Amg
Lmg = 2Lm + 2Xg (A25)
R2o − R2i

Amg = π (A26)
p
where Xg is the air gap (axial length) half-distance, Lmg is the total magnet air gap length, and Amg is the air
gap/magnet interface area. The per-phase synchronous reactance of the stator windings are calculated as:
XW = 2π · f · L (A27)
Appendix B. Simulink Mathematical Model of the AFPMA in Rotor Reference Frame (dq-Axis)
pωm d
Vd = − Rw id + Lq iq − Ld id (A28)
2 dt
pωm pωm d
Vq = − Rw iq − Ld id + λr − L q i q (A29)
2 2 dt
The dq-axis currents can be described as
d −Vd − Rw id + 0.5pωm Lq iq
id = (A30)
dt Ld
d −Vq − Rw iq − 0.5pωm Ld id + 0.5pωm λr

iq = (A31)
dt Lq
The coreless (ironless) stator of the AFPMA implies that the stator inductance as referred to the dq-axis does
not vary with rotor position, which makes Ld = Lq = Lw . Hence the electromagnetic torque of the machine can be
expressed as
3
Te = p iq λr (A32)
4
References
1. Ivankovic, R.; Cros, J.; Kakhki, M.T.; Martins, C.A.; Viarouge, P. Power Electronic Solutions to Improve the
Performance of Lundell Automotive Alternators. In New Advances in Vehicular Technology and Automotive
Engineering; InTech: London, UK, 2012; Chapter 6. [CrossRef]
2. Whaley, D.M.; Soong, W.L.; Ertugrul, N. Extracting More Power from the Lundell Car Alternator. In Proceedings of the
Australasian Universities Power Engineering Conference (AUPEC 2004), Brisbane, Australia, 26–29 September 2004.
3. Tang, S.C.; Otten, D.M.; Keim, T.A.; Perreault, D.J. Design and Evaluation of a 42-V Automotive Alternator
with Integrated Switched-Mode Rectifier. IEEE Trans. Energy Convers. 2010, 25, 983–992. [CrossRef]
4. Perreault, D.J.; Caliskan, V. Automotive Power Generation and Control. IEEE Trans. Power Electron. 2004, 19,
618–630. [CrossRef]
5. Wang, Q.; Niu, S.; Yang, L. Design Optimization and Comparative Study of Novel Dual-PM Excited Machines.
IEEE Trans. Ind. Electron. 2017, 64, 9924–9933. [CrossRef]
6. Pop, A.C.; Gyselinck, J.J.; Pinto, D.E.; Vintiloiu, I. Optimization of Low-Power Brushless PM-Machines for
Automotive Applications with Focus on High-Volume Mass Production. IEEE Trans. Ind. Electron. 2017, 64,
9767–9775. [CrossRef]
Energies 2018, 11, 274 27 of 28
7. Chen, A.; Nilssen, R.; Nysveen, A. Performance Comparisons among Radial-Flux, Multistage Axial-Flux,
and Three-Phase Transverse-Flux PM Machines for Downhole Applications. IEEE Trans. Ind. Appl. 2010, 46,
779–789. [CrossRef]
8. Cavagnino, A.; Lazzari, M.; Profumo, F.; Tenconi, A. A comparison between the axial flux and the radial flux
structures for PM synchronous motors. IEEE Trans. Ind. Appl. 2002, 38, 1517–1524. [CrossRef]
9. Ishikawa, T.; Amada, S.; Segawa, K.; Kurita, N. Proposal of a Radial- and Axial-Flux Permanent-Magnet
Synchronous Generator. IEEE Trans. Magn. 2017, 53. [CrossRef]
10. Pop, A.A.; Jurca, F.; Oprea, C.; Chirca, M.; Breban, S.; Radulescu, M.M. Axial-flux vs. radial-flux permanent-
magnet synchronous generators for micro-wind turbine application. In Proceedings of the 15th Conference
on Power Electronics and Applications (EPE’13 ECCE Europe), Lille, France, 3–5 September 2013.
11. Kahourzade, S.; Mahmoudi, A.; Ping, H.W.; Uddin, M.N. A Comprehensive Review of Axial-Flux
Permanent-Magnet Machines. Can. J. Electr. Comput. Eng. 2014, 37, 19–33. [CrossRef]
12. Nguyen, T.D.; Tseng, K.J.; Zhang, S.; Nguyen, H.T. A Novel Axial Flux Permanent-Magnet Machine for
Flywheel Energy Storage System: Design and Analysis. IEEE Trans. Ind. Electron. 2010, 58, 3784–3794.
[CrossRef]
13. Chan, T.F.; Lai, L.L. An axial-flux permanent-magnet synchronous generator for a direct-coupled
wind-turbine system. IEEE Trans. Energy Convers. 2007, 22, 86–94. [CrossRef]
14. El-Hasan, T.S.; Luk, P.C.; Bhinder, F.S.; Ebaid, M.S. Modular design of high-speed permanent-magnet
axial-flux generators. IEEE Trans. Magn. 2000, 36 Pt 1, 3558–3561. [CrossRef]
15. El-Hasan, T.S.; Luk, P.C. Magnet topology optimization to reduce harmonics in high-speed axial flux
Generators. IEEE Trans. Magn. 2003, 39 Pt 2, 3340–3342. [CrossRef]
16. Fei, W.; Luk, P.C.K.; El-Hasan, T.S. Rotor Integrity Design for a High-Speed Modular Air-Cored Axial-Flux
Permanent-Magnet Generator. IEEE Trans. Ind. Electron. 2011, 58, 3848–3858. [CrossRef]
17. Bumby, J.R.; Martin, R. Axial-Flux Permanent-Magnet Air-Cored Generator for Small-Scale Wind Turbines.
IEE Proc. Electr. Power Appl. 2005, 152, 1065–1075. [CrossRef]
18. Ferreira, Â.P.; Costa, A.F. Direct Driven Axial Flux Permanent Magnet Generator for Small-Scale Wind Power
Applications. In Proceedings of the International Conference on Renewable Energies and Power Quality
(ICREPQ’11), Las Palmas de Gran Canaria, Spain, 13–15 April 2010.
19. El-Hasan, T.S. Development of axial flux permanent magnet generator for direct driven micro wind turbine.
In Proceedings of the IEEE International Conference on Renewable Energy Research and Applications
(ICRERA 2016), Birmingham, UK, 20–23 November 2016.
20. Di Gerlando, A.; Foglia, G.M.; Iacchetti, M.F.; Perini, R. Sizing comparison of axial flux PM motors, for
automotive application. In Proceedings of the 2014 International Conference on Electrical Machines (ICEM),
Berlin, Germany, 2–5 September 2014.
21. Bastiaens, K.; Jansen, J.W.; Jumayev, S.; Lomonova, E.A. Design of an axial-flux permanent magnet machine
for an in-wheel direct drive application. In Proceedings of the 2017 IEEE International Electric Machines and
Drives Conference (IEMDC), Miami, FL, USA, 21–24 May 2017.
22. Javadi, S.; Mirsalim, M. Design and Analysis of 42-V Coreless Axial-Flux Permanent-Magnet Generators for
Automotive Applications. IEEE Trans. Magn. 2010, 46, 1015–1023. [CrossRef]
23. Javadi, S.; Mirsalim, M. A Coreless Axial-Flux Permanent-Magnet Generator for Automotive Applications.
IEEE Trans. Magn. 2008, 44, 4591–4598. [CrossRef]
24. China YY Magnetics. Available online: http://www.smi-mag.com/Neodymiummagnets.html?gclid=
EAIaIQobChMIhZ2C5qa32AIVJyjTCh0OUAYgEAAYASAAEgI7-_D_BwE (accessed on 6 January 2018).
25. Deeb, R.; Janda, M.; Makki, Z. Design of Cooling System of Permanent Magnet Servo Motor. In Proceedings
of the 14th International Scientific Conference Electric Power Engineering 2013; VSB—Technical University of
Ostrava: Ostrava, Czech Republic, 2013; pp. 205–209. ISBN 978-80-248-2988-3.
26. E-Magnets UK Company. Available online: http://e-magnetsuk.com/neodymium_magnets/temperature_
ratings.aspx (accessed on 6 January 2018).
27. MathWorks. Available online: https://www.mathworks.com/help/physmod/sps/powersys/ref/
permanentmagnetsynchronousmachine.html (accessed on 5 March 2017).
28. Krah, J.O.; Holtz, J. High-Performance Current Regulation and Efficient PWM Implementation for Low
Inductance Servo Motors. IEEE Trans. Ind. Appl. 1999, 35, 1039–1049. [CrossRef]
Energies 2018, 11, 274 28 of 28
29. Krahenbuhl, D.; Zwyssig, C.; Kolar, J.W. Half-Controlled Boost Rectifier for Low-Power High-Speed
Permanent-Magnet Generators. IEEE Trans. Ind. Electron. 2011, 58, 5066–5075. [CrossRef]
30. Bai, H. Study of PWM Rectifier/Inverter for a High Speed Generator. In Proceedings of the Power and
Energy Engineering Conference, Chengdu, China, 28–31 March 2010.
31. El-Hasan, T.; Elnasser, E.M. Development of a Low Cost Sensored Control for High-Speed Axial Flux
Permanent Magnet Machines. J. Innov. Syst. Des. Eng. 2014, 5, 7–16.
32. Hart, D.W. Power Electronics, 1st ed.; McGraw-Hill Education: New York, NY, USA, 2010; ISBN 10 0073380679.
33. El-Hasan, T.S. Modelling, Simulation and Experimentation of PM Spring. Int. Rev. Model. Simul. 2016, 9,
56–63. [CrossRef]
34. El-Hasan, T.S. Application of Permanent Magnets in Suspension and Recoil Buffer Systems. J. Innov. Syst. Des. Eng.
2015, 6, 1–14.
35. Camacho, J.M.; Sosa, V. Alternative method to calculate the magnetic field of permanent magnets with
azimuthal symmetry. Rev. Mex. Fis. E 2013, 59, 8–17.
36. Babic, S.I. Improvement in the Analytical Calculation of the Magnetic Field Produced by Permanent Magnets
ring. Prog. Electromagn. Res. C 2008, 5, 71–82.
37. El-Hasan, T.S. Inductance Determination for Air-Cored Permanent Magnet Axial Flux Machine. In Proceedings of
the Global Summit on Electronics and Electrical Engineering, Valencia, Spain, 3–5 November 2015.
© 2018 by the author. 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/).

Energies 11 00274 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Energies 11 00274 PDF

Uploaded by

Copyright:

Available Formats

energies

Electrical Engineering Department, Zarqa University, Zqrqa 13132, Jordan; tareq.elhasan@zu.edu.jo;

Received: 18 November 2017; Accepted: 9 January 2018; Published: 24 January 2018

Energies 2018, 11, 274; doi:10.3390/en11020274 www.mdpi.com/journal/energies

2. Description and Design Requirements of the AFPMA

To account for temperature

4. Optimization of the AFPMA

4.1. Machine Efficiency

Table 2. AFPMA Set Parameters.

AFPMA Set Parameters Value Unit

Table 3. AFPMA Calculated Parameters.

AFPMA Calculated Parameters Value Unit

5. Modelling and Simulation of the AFPMA

Figure 6. AFPMA MATLAB/Simulink model and parameter configuration.

6. Development of the Controlled AC/DC Rectifier for the AFPMA

Table 4. AFPMA and rectifier output voltages at no load vs. speed at α = 0o .

AFPMA Speed, Nm VLL,rms VLL,pk Vdc

7. Overall System Simulation

Figure 12. Complete system simulation using MATLAB/Simulink model.

Figure 14. Controller output pulses (zero–one).

Energies 2018, 11, x FOR PEER REVIEW 16 of 30

FigureFigure 16. Simulation

Energies 2018, 11, x FOR PEER REVIEW 19 of 30

(a) (b) (c)

VLL,min Minim line-line RMS voltage

Appendix A. Derivation of the Sizing Equations for the AFPMA

where R is center point of the PM disc which can be determined as:

Hence, the induced voltage in the conductors is given by:

The average effective length of the winding conductor, Lc can be calculated as

Substituting (A7) into (A5) and then into (A4)

Hence, the number of turns can be found as:

The above equation can be reduced to:

Lmg = 2Lm + 2Xg (A25)

d −Vq − Rw iq − 0.5pωm Ld id + 0.5pωm λr

You might also like