Outer Rotor Permanent Magnet Brushless AC Synchronous

Motor Design

Alexandre Cruz
[email protected]

Instituto Superior Técnico, Universidade de Lisboa, Portugal

July 2019

Note - This extended abstract is a brief summary of the full document, Introduction
and because of that, misinterpretations of some subjects in this In a brief statement, the focus of this work is the design of an
document may occur. The reading of the full document is Outer Rotor Permanent Magnet Brushless AC Synchronous Motor.
recommended to get the full scope of the subject. Electric machines with a high power density like the one referred
in this document have an important relevance in the aerospace
industry, offerings the possibility to create systems with low mass,
Abstract resilient to failure and simple to operate.
The goal of this work is to design an outer rotor brushless The electric machine on this document was designed for the Z1
permanent magnet synchronous electric machine, to be applied in a Model of the Darxzeus Electric Wheel, which in a brief description, is a
project of an electrical motorized direct drive wheel. The main wheel with incorporated electric power train, self-refrigerated, without
requirements of the project are to obtain a high power density a gear box (direct-drive), which includes all the main systems required
machine, high torque, low angular speed in a thin ring form factor. to propel and break a passenger or service vehicle.
The electric machine design was executed using an Despite the specificity of this application which the machine was
electromagnetic study and a thermal study. Both studies were designed and optimized, this document can be consulted as a guide or a
performed using finite element methods. The electromagnetic study tool for help in the process of designing an Outer Rotor Permanent
consists of a magnetostatic analysis, dynamic analysis and a D-Q model Magnet Brushless AC Synchronous Motor. The document approaches
was build. The Thermal Study was performed by coupling an thermal the main subjects of the theoretical design for this type of electrical
analysis with an magnetostatic analysis to analyze the machine thermal machine.
behavior and to evaluate cooling systems strategies.
The results of the work give us a preliminary design that can be We start be designing a pre-dimension of the machine according
used to help produce a physical prototype of the electrical machine as to the pre-requirements of the application. The pre-dimension includes
well guide lines for the cooling system strategies to be used and the preliminary geometry, electric configuration, type of winding,
components grades. dimension of the coils conductors, evaluating the types of lamination
that best suited in the stator and rotor cores of this machine, and the
Keywords: Electric Machine, Permanent Magnets, Synchronous type of magnet to be used.
Machine, Outer Rotor, Electromagnetic Study, Thermal The next step was to conduct and electromagnetic study, using
Study. finite element methods to help complete the electromagnetic sizing of
the machine. This includes geometrical and electrical improvements/
Index optimizations with the goal of increase the energy efficiency of the
 Introduction machine and other aspects relevant to the electric wheel project.
 I - Pre-Dimension of the Electrical Machine To complete the theoretical study, a thermal analysis using finite
 II - Electromagnetic Study element method was conducted, to inspect the thermal behavior of the
 III - Thermal Study electrical machine. This study also includes evaluation and creation of
 Conclusions cooling strategies, and evaluation and choice of the materials that have
critical properties influenced by the temperature.
 References

1
 I - Pre-Dimension of the Electrical Machine – Geometry Slot Tooth (Length across ) [
– Geometry Slot Tooth (Length across ) [
]
]
– Geometry Slot Tooth Tip (Spacing between Tips) [ ]
Application Pre-requirements – Geometry Slot Tooth Tip (Tip Depth) [ ]
The requirements for the electric machine are the follow, – Geometry Rotor Diameter 1 (Outer Diameter) [ ]
– Geometry Rotor Diameter 2 (Inner Diameter) [ ]
1) Wheel without gear Box.
– Geometry Rotor Back-Iron [ ]
2) Machine with the lowest mass possible, without compromise the – Geometry Rotor Front-Iron [ ]
machine integrity. – Geometry Rotor Magnet (Length across or Wide Angle) [ ] or [ ] or [ ]
– Geometry Rotor Magnet (Length across or Depth) [ ]
3) Neglecting the aerodynamic effects, the wheel need to be able to
– Geometry Rotor Space between Magnets [ ] or [ ]
move a body (1000 Kg) on a slop with with a constant velocity – Air Gap Length [ ]
of ⁄ ( ⁄ ), assuming that the wheel supports – Lamination Length [ ]
tires/rims with radius of 0.5m, 0.4m e 0.375m.
4) The machine need to be able to propel the wheel to achieve a tip The preliminary geometry obtained was,
velocity of ( ).
5) The volume available to fit the electrical machine without chassis, is
equivalent of the volume shown in the figure below,

6) The machine need to be controlled with a three-phase inverter

( = ).
7) The relations below need to be respected. The parameters are the
number of Slots in the Stator ( ), the number of Poles ( )
and the number of phases ( ),

(1.1) 𝑚𝑚
𝐿𝑎𝑖𝑟𝑔𝑎𝑝
, se Double Layer Winding 𝐿𝑙𝑎𝑚𝑖𝑛𝑎𝑡𝑖𝑜𝑛
(1.2)

, se Single Layer Winding

(1.3)

8) The machine needs to be wired in star (Wye Type) configuration,

with the winding preferential in Single Layer Winding (Lap Type,
Parallel Paths = 1, Slot Pitch = 3) configuration.
9) The Slots Fill Factor ( ) needs to be aprox. 65%.
10) Preferential using NdFeB permanent magnets.
Figura 2 - Descrição Geométrica da Geometria Preliminar da Maquina Eléctrica (Pré-
Dimensionamento)
RPM, Torque and Power Estimation
Assuming the requirements of the application, the electrical machine Winding
need to be able to deliver = [ ] at a mechanical angular The winding that was chosen was the Single Layer Winding (Lap Type,
frequency of = [ ]. The mechanical power needed is Parallel Paths = 1, Slot Pitch = 3)
= , and first estimation for the eletrical power needed,
assuming an efficiency of = , is = .

Preliminary Geometry
The next step was creating the preliminary geometry of the machine,
and for that a template below describing the geometrical parameters
was created.

Figure 3 - Single Layer Winding (Lap Type, Parallel Paths = 1, Slot Pitch = 3) in Star
configuration

Coils Conductors
For helping with the conductors configuration design, was conduted a
Figure 1 - Parâmetros Descritivos de Geometria da Máquina Eléctrica PM de Rotor Exterior skin depth calculation, that gives us an penetration depth of 3.53mm.
With that value, we design a conductor with filaments with a diameter
– Angle between Slots [ ] or [ ]
lower than the penetration depth. The result was a coil with 20
– Angle between Poles [ ] or [ ]
– Geometry Slot Diameter 1 (Outer Diameter) [ ] conductors per slot, and 10 strands in each conducter. The strands had
– Geometry Slot Diameter 2 (Inner Diameter) [ ] a wire diameter of 1.024mm (AWG 18)
– Geometry Slot Back-Iron [ ]

2
 Lamination First Approximation
The preliminary material chosen for the core of the stator and rotor, The main goal of the first approximation was to get a more realistic
was the Nonoriented Electrical Steel M-15 Gauge 29. Below is the B-H estimation of the input power needed to satisfy the application
curve of this material, requirements, and the levels of the phase current required to meet this
requirements ( = @ = ). The results
obtained are shown in the table below,

RMS Stator Total Efficiency Input RMS RMS Phase

Phase Current Torque power Phase Back-EMF
Voltage
[A] [N.m] [%] [W] [V] [V]
1 12 81 1723 580 574
10 125 97 17272 582 574
Figura 4 - B-H Curve of Electric Steel M15-29G
25 310 98 42824 590 572
50 595 97 83202 614 576
Magnets
100 1046 96 151105 665 587
The preliminary material chosen for the magnets, was the Sintered
133 1266 94 188141 700 585
NdFeB N42xx. The exact variant was left to be chosen in the chapter 3
150 1363 93 205204 716 581
(Thermal Study).
200 1603 91 251141 764 562
Nota: O Discretization Error da análise foi em média 3%
II – Electromagnetic Study
Table 1 – General Results of the First Approximation

On this study was used the Maxwell equations to analyze the behavior
of the electrical machine. The calculation was executed using the finite The phase current required to meet the requirements is 133A RMS.
element method (FEM). For helping in the meshing and pre/post-
processing, was used the computation tool MotorAnalysis-PM® The table below show more broth results for the 133A RMS case,
Discretization Error [ ] 3 Average d-axis Current ( ) [ ] 0
integrated in the mathematical computation tool MATLAB®. Rotor Speed ( )[ ] 1300 Average q-axis Current ( ) [ ] 188
Two types of analyses were used. Magnetostatic analysis, Advance Angle ( ) [ ] 0 Average d-axis Voltage ( ) [ ] -724
Supple Frequency ( ) [Hz] 346.667 Average q-axis Voltage ( ) [ ] 666
assuming the Maxwell equations simplification, Total Torque ( ) [ ] 1266 Input Electrical Power ( ) [ ] 188141
Reluctance Torque ( ) [ ] -382 Output Mechanical Power ( )[ ] 172419
Electromagnetic Torque ( ) [ ] 1649 Efficiency ( ) [ ] 94
(2.2.3) Torque Ripple (%) 48 Power factor ( ) 0.65
RMS Phase Current ( )[ ] 133 Stator Winding Loss ( )[ ] 8775
(2.2.4)
RMS Phase Voltage ( )[ ] 700 Total Iron Core Loss ( )[ ] 416
RMS Phase Back-EMF ( )[ ] 585 Eddy Current Iron Core Loss ( ) [ ] 111
RMS Current Density ( )[ ⁄ ] 16 Hysteresis Iron Core Loss ( )[ ] 305
And Dynamic analysis assuming the Maxwell equations, Max. Demagnetization Field [ ⁄ ] 450031 Magnet Eddy Current Loss ( ) [ ] 404
Max. Demagnetization Field (% of )[ ] 47 Other Eddy Current Loss [ ] 0

Table 2 – Additional results of the First Approximation

( ) (2.2.1)
The figure below show the distribution of the Magnetic Flux Density (B),

( ) (2.2.2)

The torque calculation was performing using the Maxwell Stress

Tensor,

( )
∫ (2.2.6)

Meshing
The meshing parameters chosen was, air gap with 3 layers od elements
and a element growth ratio of 1.4 (40%), the mesh control of the rest of
the 2D model was automaticly meshed with the optimized mesh
algorithem that the software had. The result was,

Figure 6 –Flux Density (B) Distribution of the First Approximation

The A,B,C,D zones, show location that can be geometrically improved.

Figure 5 - Mesh of the last optimized geometry (1.5 Pole and 5 Slots)

3
 Geometric Optimization Analise
Analise Dinâmica
Magnetostática
The strategy of the geometric optimization, is to start from the Computing time 20 minutes 3 hours
Discretization Error [ ] 0.25 0.7
geometry of the first approximation and try to improve/optimize eash Total Mechanical Torque ( )[ ] 1427 1413
geometrical parameter, that are described in the Figure 1. Reluctance Torque ( ) [ ] -244 -240
Electromagnetic Torque ( )[ ] 1671 1653
Torque Ripple (%) 37 45
Gs3 Optimization (Slot Back-Iron) RMS Phase Current ( )[ ] 133 132
RMS Phase Voltage ( )[ ] 678 676
The First approximation RMS Phase Back-EMF ( )[ ] 577 576
RMS Current Density ( )[ ⁄ ] 16.14 16.01
geometry had =
Average d-axis Current ( ) [ ] 0 (aprox.) 0.54
Average q-axis Current ( ) [ ] 188 187
Average d-axis Voltage ( ) [ ] -625 -624
Average q-axis Voltage ( ) [ ] 722 724
Max. Demagnetization Field [ ⁄ ] 525424 529015
Max. Demagnetization Field (% of )[ ] 55 55
Input Electrical Power ( ) [ ] 203688 203176
Output Mechanical Power ( )[ ] 194301 192303
Efficiency ( ) [ ] 95.0 94.4
Power factor ( ) 0.75 0.76
Total Power Losses ( )[ ] 9703 10672
 Stator Winding Loss ( )[ ] 8876 9391
 Total Iron Core Loss ( )[ ] 414 641
Rotor Iron Core Loss [ ] - 0.005
Stator Iron Core Loss [ ] - 641
The results obtained was,  Total Eddy Current Iron Core Loss ( )
112 206
[ ]
Stator Mass Total Torque Eddy Current Rotor Iron Core Loss [ ] - 0.002
Efficiency ( ) Eddy Current Stator Iron Core Loss [ ] - 206
( ) ( )  Total Hysteresis Iron Core Loss (
[ ] )[ ] 301 434
[ ] [ ] [ ] Hysteresis Rotor Iron Core Loss [ ] - 0.003
Hysteresis Stator Iron Core Loss [ ] - 434
19 mm
14.7 1270 94.9 Nota - Na análise dinâmica, o cálculo das perdas de potência no núcleo ( ) foi efectuado para uma gama de
(Ref.) frequências de fluxo magnético compreendida entre ]0Hz; 10kHz]. Na secção “Apêndice A”, pode ser consultado as
curvas de Iron Loss do Electrical Steel (M-15 29Ga) para várias frequências.
9 mm 11.2 1269 94.9
Table 4 – Dynamic and Magnetostatic Analysis Results of the Optimized Geometry
7 mm 10.5 1258 94.8
3 mm 9 1134 94.3
Electrical Optimization
Table 3 - Gs3 Optimization (Slot Back-Iron)
The main goal of this topic was to get a understanding of the behavior
of the machine for a wide range of Torque ( ) and Speed ( )
The chose was = because give as the best results of ( ) e ( ),
situations. The eletrical optimization that was made, was the
and offer a significant reduction of the total stator lamination mass.
computation of the optimizer Advance Angle ( ) for a range of
The same strategy was made for some key geometrical parameters, and ( ) combinations, assuming that the inverter (supply) is able to
the result was the follow geometry, excite each phase with the required current and voltage.

Figure 8 – Optimizer Advance Angle ( )

The efficiency map was computed as well,

𝑚𝑚

𝐿𝑎𝑖𝑟𝑔𝑎𝑝
𝐿𝑙𝑎𝑚𝑖𝑛𝑎𝑡𝑖𝑜𝑛

Figure 7 – Final Geometry of the Electric Machine

A magnetostatic analysis and a dynamic analysis was conducted do

made a deeper evaluation of the optimized geometry, the may results
are shown in the table below, Figure 9 - Efficiency Map (Efficiency vs Torque vs RPM)

4
 III – Thermal Study Magnets
The chosen variant was N42SH( ), N42UH( ) or
The thermal analysis using finite element method was conducted, to N42EH( ). With a NiCuNi + Black Epoxy isolation.
examine the thermal behavior of the electrical machine. This study also
includes evaluation and creation of cooling strategies, and evaluation Magnetic Wire isolation
and choice of the materials that have critical properties influenced by The choice was IEC 60317-13.
the temperature
Electrical Steel Isolation
Cooling Strategy The choice was C-3, C-4 e C-5.
The cooling strategy that was suggested is a close and forced
refrigeration system with a water Jacket fitted in the stator back-iron. Conclusions
The choice for the Jacket material was aluminum, and a water
The final results of the magnetostatic study suggest that is
base fluid for the coolant. In order to safely cool down the machine, the
possible to create an electric machine to satisfy the requirements of the
combination of the coolant velocity in the jacket and the radiator
application ( ), with a theoretical energy
performance to extract the heat away from the coolant, need to ensure
efficiency of 95%.
a Convective Heat Transfer Coefficient ( = 1000 [ ( )]) in
The final results of the thermal study suggest that is possible to
the walls of the jacket, and an income coolant temperature of 300K.
cool the electric machine with a 2.5 factor of safety.

Bibliographical References
Books
[1] Dinâmica de Maquinas Eléctricas. Prof. Gil Marques (2002) [IST]
[2] Controlo de Motores Eléctricos, Prof. Gil Marques (1999) [IST]
[3] Brushless Permanent Magnet Motor Design, 2nd Ed, Duane Hanselman
(2006) [McGraw-Hill]
[4] Permanent Magnet Motor Technology – Design and Applications, 3rd Ed,
Jacek F.Gieras (2010) [CRC]
[5] Electric Machinery, 6th Ed, Fitzgerald, Kingsley, Umans (2003) [McGraw-Hill]
[6] Fundamentals of Engineering Thermodynamics, 5th Ed, M. J. Moran and H.
N. Shapiro (2006) [Wiley]
[7] Electromagnetic Modeling by Finite Element Methods, J. Bastos,
N.Sadowski (2003) [CRC Press]

Scientific Papers
[8] Interior PM Synchronous Motors for Adjustable-Speed Drives, Jahns,
Kliman, Neumann (1986) [IEEE]
[9] Dynamic Model of PM Synchronous Motors (using D-Q Model), Dal Ohm
(2000) [Drivetech]

Encyclopedia
[10] McGraw-Hill Encyclopedia of Science & Technology, [http://MHEST.com]
[www.accessscience.com]
[11] IEC 60050, International Electrotechnical Vocabulary by IEC,
[www.electropedia.org]

Standards Organizations
Figure 10 - Distribuição de Temperatura (10, 1000) [W/m2K]
[12] ASTM International (American Society for Testing and Materials),
[www.astm.org]
[13] IEC (International Electrotechnical Commission), [www.iec.ch]
[14] ISO (International Organization for Standardization), [www.iso.org]

Software Tools
[15] EMWorks EMS, Documentation [www.emworks.com]
[16] MotorAnalysis-PM, Documentation. Dr. Vladimir Kuptsov
[www.motoranalysis.com]
[17] FEMM, Documentation. David Meeker [[email protected]]
[www.femm.info]
[18] SolidWorks, [www.solidworks.com] [www.3ds.com]
[19] MATLAB, [www.mathworks.com]
[20] Multisim, [www.ni.com]
[21] Eagle PCB Design, [www.autodesk.com]
[22] Atmel Studio, [www.microchip.com]
[23] Microchip MPLAB, [www.microchip.com]
[24] Arduino IDE, [www.arduino.cc]

Technical Documents
[25] Brushless DC (BLDC) Motor Fundamentals (2003), Application Note AN885
[Microchip].
[26] AKSteel Non-Oriented Electrical Steels, Datasheets [www.aksteel.com]
[27] NdFeB Reference Guide, Datasheet Rev-181031 [ARNOLD Magnetics]
[www.arnoldmagnetics.com]
[28] NdFeB Datasheet, Datasheet Rev-1 [Eclipse Magnets]
Water Jacket Tipo A Water Jacket Tipo B [www.eclipsemagnetics.com]