Design of a Permanent Magnet Synchronous

Motor for an Electric Traction Application

by
Hewa Gamage Nandun Senevirathna

Submitted to the Department of Electrical Engineering, Electronics,

Computers and Systems
in partial fulfillment of the requirements for the degree of
Erasmus Mundus Master Course in Sustainable Transportation and
Electrical Power Systems
at the
UNIVERSIDAD DE OVIEDO
August 2020

c Universidad de Oviedo 2020. All rights reserved.

Author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Nandun Senevirathna

Certified by . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Dr. Gaurang Vakil
Associate Professor
Thesis Supervisor
2
 Design of a Permanent Magnet Synchronous Motor for an
Electric Traction Application
by
Hewa Gamage Nandun Senevirathna

Submitted to the Department of Electrical Engineering, Electronics, Computers and

Systems
on August 31, 2020, in partial fulfillment of the
requirements for the degree of
Erasmus Mundus Master Course in Sustainable Transportation and Electrical
Power Systems

Abstract
This thesis details a design of an exterior - rotor, fractional - slot, double layer per-
manent magnet synchronous motor for a two-wheeler application.The cooling method
used was natural cooling. The initial design procedure was adopted from previously
published works by other authors [1, 2] . ANSYS MotorCAD software was used in
design of the machine,optimization and verification of the machine’s performance.A
review of different electric motor types was carried out to choose a suitable ma-
chine type and configuration. Comparative analysis of BLAC and BLDC motors was
performed and considering the application requirements Sinusoidally fed permanent
magnet synchronous motor was chosen (commonly referred as BLAC motor) . An
extensive sensitivity analysis was done using the tools available in MotorCAD simu-
lation platform. Further the optimization of the selected design in terms of electro-
magnetic performance was carried out using MotorCAD. Finally, the electromagnetic
and thermal performance validation was done using the same software.
Keywords : Outer-Rotor, in-wheel motor, Permanent Magnet Synchronous Mo-
tor, FEA, Concentrated winding, fractional-slot windings

Thesis Supervisor: Dr. Gaurang Vakil

Title: Associate Professor

3
4
Acknowledgments
I bestow my sincere gratitude to the supervisor Dr. Gaurang Vakil for his continuous
guidance and support provided for carrying out this work. Erasmus Mundus STEPS
(Sustainable Transportation and Electrical Power Systems) master program was a
milestone in my career as an engineer.I’m very much thankful to Prof. Pablo Arboleya,
Prof. Jorge Garcia and to to all of my professors at the University of Oviedo,Spain,
Sapienza University of Rome,Italy and the University of Nottingham,UK for giving
me this wonderful opportunity to participate in this excellent academic program. Last
but not least I’m very much grateful to my loving parents who supported me from
the distance throughout this journey.

5
6
Contents

1 Introduction 19
1.1 Aim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.3 The Structure of the Thesis . . . . . . . . . . . . . . . . . . . . . . . 20
1.4 Direct Drive In-Wheel Motors . . . . . . . . . . . . . . . . . . . . . . 21
1.5 Electrical Machine Classifications . . . . . . . . . . . . . . . . . . . . 21
1.6 Fundamental Relationships in Permanent Magnet Synchronous Motors 22
1.6.1 Speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.6.2 Air gap magnetic flux density . . . . . . . . . . . . . . . . . . 23
1.6.3 Voltage induced (EMF) . . . . . . . . . . . . . . . . . . . . . 24
1.6.4 Armature line current density and current density . . . . . . . 24
1.6.5 Electromagnetic Power . . . . . . . . . . . . . . . . . . . . . . 25
1.6.6 Electromagnetic Torque . . . . . . . . . . . . . . . . . . . . . 25
1.7 Magnetic Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2 Study of Machine Topologies and Motor Windings 29

2.1 Conventional Motor Types . . . . . . . . . . . . . . . . . . . . . . . . 29
2.1.1 Synchronous Motor (SM) . . . . . . . . . . . . . . . . . . . . . 29
2.1.2 Interior permanent magnet synchronous machine (IPMSM) . . 32
2.1.3 Brushless direct current machines (BLDCM) . . . . . . . . . . 33
2.1.4 Induction machine (IM) . . . . . . . . . . . . . . . . . . . . . 33
2.1.5 Switched reluctance machine (SRM) . . . . . . . . . . . . . . 34
2.2 Unconventional flux direction machines . . . . . . . . . . . . . . . . . 35

7
 2.2.1 Axial flux machines . . . . . . . . . . . . . . . . . . . . . . . . 35
2.2.2 Transverse flux machines (TFM) . . . . . . . . . . . . . . . . 36
2.3 Unconventional machine Topologies . . . . . . . . . . . . . . . . . . . 37
2.3.1 Flux switching machine (FSM) . . . . . . . . . . . . . . . . . 37
2.3.2 Synchronous Reluctance Machine (SynRelM) . . . . . . . . . . 39
2.3.3 Permanent magnet assisted synchronous reluctance machine
(PMa-SynRelM) . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4 Brief comparison of the motor type chosen for the application . . . . 40
2.4.1 BLDC/SMPM . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.4.2 BLDC/SMPM - Limitations . . . . . . . . . . . . . . . . . . . 41
2.5 Rotor Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.5.1 Interior Rotor Configuration . . . . . . . . . . . . . . . . . . . 42
2.5.2 Exterior Rotor Configuration . . . . . . . . . . . . . . . . . . 43
2.5.3 Exterior Rotor Vs Interior Rotor . . . . . . . . . . . . . . . . 44
2.6 Winding options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.6.1 Concentrated Winding Versus Distributed Windings . . . . . . 46

3 Developing Initial Design Approach 49

3.1 Understanding the Application requirements . . . . . . . . . . . . . . 49
3.2 Preliminary Design Considerations . . . . . . . . . . . . . . . . . . . 50
3.3 Generalized Design procedure . . . . . . . . . . . . . . . . . . . . . . 50
3.3.1 Exterior Rotor Configuration . . . . . . . . . . . . . . . . . . 51
3.4 Determination of Motor Dimensions . . . . . . . . . . . . . . . . . . . 52
3.5 Magnet Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.6 Selection of the Number of Poles and Slots . . . . . . . . . . . . . . . 54
3.7 Concentrated Windings . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.7.1 Winding Factors and Winding Feasibility . . . . . . . . . . . . 57
3.8 Design Approach with Analytical Formulae . . . . . . . . . . . . . . . 58
3.8.1 Per-Unit Machine Model . . . . . . . . . . . . . . . . . . . . . 59
3.8.2 Analysis of Inductance . . . . . . . . . . . . . . . . . . . . . . 61

8
 3.8.3 Power factor Maximization . . . . . . . . . . . . . . . . . . . . 63
3.8.4 Design Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.8.5 Summary of Equations Applicable to Radial Flux Permanent
Magnet Motors . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.9 Simulation Results of Sinusoidally fed PMSM . . . . . . . . . . . . . 68
3.9.1 Simulation Results of the initial Design BLDC Motor fed with
Square-wave Drive . . . . . . . . . . . . . . . . . . . . . . . . 70
3.10 Choice between 20P18Q motor and 12P36Q Motor . . . . . . . . . . 71

4 Sensitivity Analysis and Design Optimization 79

4.1 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.1.1 Electromagnetic Model Sensitivity Analysis . . . . . . . . . . 79
4.1.2 Thermal Model Output Data . . . . . . . . . . . . . . . . . . 91
4.2 Optimization of Motor Models . . . . . . . . . . . . . . . . . . . . . . 91
4.2.1 Electromagnetic Model Optimization . . . . . . . . . . . . . . 93
4.2.2 Multi-Physics Model optimization . . . . . . . . . . . . . . . . 93

5 Validation of the Motor Model by Finite Element Analysis 97

5.1 Electromagnetic Validation (FEA of optimized design) . . . . . . . . 97
5.2 Lab Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
5.3 Thermal Model Analysis and Validation (FEA of optimized design) . 99
5.3.1 Thermal Validation . . . . . . . . . . . . . . . . . . . . . . . . 99
5.3.2 Mechanical model (FEA of optimized design) . . . . . . . . . 100

6 Conclusion 109
6.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
6.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

A Tables 111

B Figures 121

9
10
List of Figures

1-1 Conventional Electrical Machine Classification [3] . . . . . . . . . . . 22

1-2 Typical torque versus speed characteristic.[4] . . . . . . . . . . . . . . 23
1-3 Second quadrant normal B-H curves of different types of permanent
magnet materials. Source: Magnequench [5] . . . . . . . . . . . . . . 27
1-4 Impact of increasing dysprosium content on the coercivity Hcj and re-
manent flux density Br of NdFeB magnets. Source : Arnold Magnetics
[6] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

2-1 Cross Section of a Wound Field SM [7]. . . . . . . . . . . . . . . . . . 30

2-2 PM machines with different rotor configurations [8] . . . . . . . . . . 30
2-3 Surface Mounted PM motor . . . . . . . . . . . . . . . . . . . . . . . 31
2-4 Interior Permanent Magnet Synchronous Motor . . . . . . . . . . . . 32
2-5 Induction Motor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2-6 Switched Reluctance Motor . . . . . . . . . . . . . . . . . . . . . . . 34
2-7 Double Rotor Axial Flux IPMSM [9] . . . . . . . . . . . . . . . . . . 36
2-8 Transverse Flux Machine . . . . . . . . . . . . . . . . . . . . . . . . . 37
2-9 Outer Rotor Flux Switching SM . . . . . . . . . . . . . . . . . . . . . 38
2-10 Synchronous Reluctance Motors . . . . . . . . . . . . . . . . . . . . . 39
2-11 Interior-rotor brushless permanent-magnet motor [10] . . . . . . . . . 42
2-12 Examples of interior-rotor brushless permanent-magnet rotors [10]. . . 43
2-13 Exterior Rotor Configuration [10]. . . . . . . . . . . . . . . . . . . . . 44
2-14 Exterior and Interior Rotor Configurations [11] . . . . . . . . . . . . . 45

11
2-15 Typical stator winding configurations (four pole) [12]. (a) 24-slot,
overlapping (distributed). (b) Twelve slot, overlapping (concentrated).
(c)6-slot, nonoverlapping, all teeth wound. (d) 6-slot, non-overlapping,
alternate teeth wound. . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3-1 Generalized Design Procedure. . . . . . . . . . . . . . . . . . . . . . . 51

3-2 Typical cross section of the exterior-rotor surface PM motor. courtesy:[13] 52
3-3 Definition of the geometrical parameters for the outer-rotor SMPM
motors [14] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3-4 Demagnetization Curves for NdFeB magnets. . . . . . . . . . . . . . . 54
3-5 Fundamental winding factors ξ1 for concentrated two-layer windings (q
≤ 0.5). [2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3-6 Elementary block of an SPM machine [1] . . . . . . . . . . . . . . . . 58
3-7 Definition of Power Factor [2] . . . . . . . . . . . . . . . . . . . . . . 61
3-8 Motor CAD model of Initial Design -Radial cross section 20P18Q Ma-
chine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3-9 Winding Layout of 20P18Q Machine. . . . . . . . . . . . . . . . . . . 70
3-10 Torque-Speed Curve Initial Design Fractional Slot 20P18Q Motor . . 71
3-11 Output waveforms Fractional Slot Motor Initial Design . . . . . . . . 74
3-12 Motor CAD model of Initial Design Integer slot Motor -Radial cross
section (Square-Wave drive). . . . . . . . . . . . . . . . . . . . . . . 76
3-13 Winding Layout of Integer Slot 12P36Q Machine. . . . . . . . . . . . 76
3-14 Torque-Speed Curve Initial Design Integer Slot Motor . . . . . . . . . 77
3-15 Output Waveforms Integer Slot Motor Initial Design : Square-wave drive 78

4-1 Effect of Magnet Thickness . . . . . . . . . . . . . . . . . . . . . . . . 80

4-2 Airgap:Sensitivity Analaysis . . . . . . . . . . . . . . . . . . . . . . . 81
4-3 Armature Diameter:Sensitivity Analysis . . . . . . . . . . . . . . . . 82
4-4 Magnet Arc:Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . 83
4-5 Slot Depth:Sensitivity Analaysis . . . . . . . . . . . . . . . . . . . . 84
4-6 Slot Opening:Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . 85

12
4-7 Tooth Tip Angle:Sensitivity Analysis . . . . . . . . . . . . . . . . . . 86
4-8 Tooth Width:Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . 87
4-9 Turns Number:Sensitivity Analysis . . . . . . . . . . . . . . . . . . . 88
4-10 Copper Slot Fill:Sensitivity Analysis . . . . . . . . . . . . . . . . . . 89
4-11 Peak Current:Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . 90
4-12 Thermal Sensitivity : Ambient Temperature . . . . . . . . . . . . . . 91

5-1 Electromagnetic FEA Window :Optimized design . . . . . . . . . . . 98

5-2 Output graphs of Optimized machine . . . . . . . . . . . . . . . . . . 101
5-3 Torque-Speed Characteristic of the optimized 20P18Q Machine . . . . 102
5-4 Contour graph of efficiency and shaft torque versus speed . . . . . . . 102
5-5 Torque, Efficiency - Speed curves . . . . . . . . . . . . . . . . . . . . 103
5-6 Schematic view of equivalent thermal network (steady state) model
optimized 20P18Q motor . . . . . . . . . . . . . . . . . . . . . . . . . 103
5-7 Radial cross section FEA evaluated temperature distribution of Opti-
mized 20P18Q motor . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5-8 FEA evaluated temperature distribution of rotor and stator off opti-
mized 20P18Q motor . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5-9 Thermal validation Data of the Optimized 20P18Q Machine . . . . . 106
5-10 Thermal validation Graph . . . . . . . . . . . . . . . . . . . . . . . . 106
5-11 Temperature surface graphs of the optimized 20P18Q machine . . . . 107
5-12 mechanical stress distribution on rotor . . . . . . . . . . . . . . . . . 107

B-1 Torque-angle characteristics of a salient-pole synchronous machine with

Xsd ¿ Xsq : 1 — synchronous torque Tdsyn , 2 — reluctance torque Tdrel ,
3 — resultant torque Td [15] . . . . . . . . . . . . . . . . . . . . . . . 122
B-2 Flux Density Vs Magnetic Field for M350 50A Electrical steel . . . . 122
B-3 Torque, speed and efficiency surface graph of optimized 20P18Q Machine122
B-4 Axial cross Section and Motor CAD 3D Model . . . . . . . . . . . . . 123

13
14
List of Tables

2.1 Performance comparison of Motor Topologies . . . . . . . . . . . . . 41

3.1 Basic Specifications of the Motor. . . . . . . . . . . . . . . . . . . . . 50

3.2 Typical values of TRV, K and Sigma. Reference:[10] . . . . . . . . . . 52
3.3 Comparison of Magnets. [16] . . . . . . . . . . . . . . . . . . . . . . . 54
3.4 Preliminary Design Data of Elementary Block . . . . . . . . . . . . . 66
3.5 PMSM Preliminary Design Data Fractional Slot 20P18Q Machine . . 70
3.6 Preliminary Design Drive Output Data of Fractional-Slot Machine . . 72
3.7 Preliminary Design ElectroMagnetic Output Data of Fractional-Slot
Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.8 Flux Densities PMSM Preliminary Design Fractional Slot Motor . . . 73
3.9 Winding Data Initial Design 20P18Q Motor . . . . . . . . . . . . . . 75
3.10 PMSM Preliminary Design Data Integer Slot Machine . . . . . . . . . 75
3.11 Output Data of Integer-Slot Motor :Square-Wave Drive . . . . . . . . 77

4.1 Fractional Slot Motor Thermal Model Output Data . . . . . . . . . . 92

4.2 Comparison Output Data of the Initial Design and the Initial Opti-
mized Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.3 Final Optimization Results . . . . . . . . . . . . . . . . . . . . . . . . 95

5.1 Analytical and Simulated Values of Output variables . . . . . . . . . 99

5.2 Operating Point Data Under Maximum Torque Per Ampere Control
Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

A.1 Initial Design Calculations 20P18Q Machine with Sinusoidal Excitation 111

15
A.2 Integer - Slot Motors Initial Designs. . . . . . . . . . . . . . . . . . . 115
A.3 Performance Comparison of Initial Designs . . . . . . . . . . . . . . . 119

16
Acronyms

AC - Alternating Current

BLAC - Brushless Alternating Current

BLDC - Brushless Direct Current

DC - Direct Current

EV - Electric Vehicle

EMF - Electro Motive Force

FEA - Finite Element Analysis

FSCW - Fractional-Slot Concentrate Winding

FSM - Flux Switching Machine/Motor

GCD - Greatest Common Divisor

HEV - Hybrid Electric Vehicle

IM - Induction Machine/Motor

IPSM - Interior Permanent Magnet Machine/Motor

LCM - Lowest Common Multiple

MMF - Magneto Motive Force

PMFSM - Permanent Magnet Flux Switching Machine/Motor

PMSM - Permanent Magnet Synchronous Machine/Motor

PMa-SynRelM - Permanent Magnet assisted Synchronous Reluctance Machine/Mo-

tor

17
SPM/SMPM - Surface mounted Permanent Magnet Machine/Motor

SRM - Switched Reluctance Machine/Motor

SynRelM - Synchronous Reluctance Machine/Motor

TFM - Transverse Flux Machine/Motor

WFFSM - Wound Field Flux Switching Machine/Motor

18
Chapter 1

Introduction

Improvement of the air quality and reducing the environmental impacts is a primary
reason for revolutionizing the vehicles through electrification. Introduction of electri-
fied two wheelers such as electric scooters for daily short distance commutations in
urban transport systems can help reduce traffic congestion and hence to reduce pol-
lution. Therefore researching on technologies more suitable for low speed low torque
applications such applications is important.

1.1 Aim
The aim of this work is to design a permanent magnet synchronous motor for a low-
speed,low-power electrical traction application (Two-wheeler) with acceptable perfor-
mance and size.

1.2 Objectives
The project would focus on developing a solution for an electrified two wheeler and
the project objectives include;

1. Understanding the application requirements.

2. Defining the specifications of the machine required for the application.

19
 3. Analysing different typologies used for the electrical machines used in auto-
motive applications.BLDC, BLAC and SynRelM are mainly suggested in liter-
ature.

4. Developing initial design approach

5. Modelling the selected designs using MotorCAD software, performing the

sensitivity analysis and optimization.

6. Electromagnetic,thermal and mechanical Validation through finite element

analysis.

1.3 The Structure of the Thesis

This thesis is organized in the following manner.

Chapter one introduces the project and details the scope of the project and the
structure of the thesis. The chapter also includes the equations governing perma-
nent magnet synchronous machines and some background information on permanent
magnets.

Chapter two presents the literature review of electrical motors used in traction appli-
cations and describes winding options.

Chapter three describes the basic design flow and motivations for the design decisions
made.

Chapter four provides the sensitivity analysis results of the selected initial design and
the optimization outcomes.

Chapter five describes the validation through finite element analysis. It also includes
the conclusion and suggestions for future work.

Chapter six includes the conclusion of the work. Some suggestion for future work are
also included at the end.

Appendices and Bibliography are provided at the end of the thesis.

20
1.4 Direct Drive In-Wheel Motors
The motor can be positioned in different locations of the vehicles using electrical
traction. Locating the motor directly inside the wheel is one possibility which reduces
the transmission path between the motor and the whee, hence reducing the losses in
transmission. The use of direct drives in low-speed electric machines offers benefits
attributed to the removal of the gearbox [17] such as:

• Improved efficiency: Removal of the gearbox eliminates losses due to friction in

mechanical gears.

• Reduced maintenance: Gearboxes are main sources of mechanical system failures.

The necessity of frequent lubrication is also minimized with direct drive systems.

• Improved reliability: Having less number of mechanical components such as gears

reduces the probability of failures improving the overall reliability.

• Reduced vibration and noise: Audible noise and vibrations due the gear teeth are
eliminated.

• Simpler design: The number of components on the power train is reduced hence
simplifying the overall system.
However, there are some drawbacks of in-wheel motors. The unsprung mass (
mass components not supported by the suspension system) is increased in in-wheel
motor configuration. This results in issues with the suspension and steering and loss
of comfort.

1.5 Electrical Machine Classifications

There are various electrical machine topologies and technologies. Therefore several
different classifications exist. Some of these classifications are:

Supply for the armature: - AC, DC or Pulsed

Field flux production: - Windings (whether externally supplied or through induc-

tance), Permanent magnets(PM) or no field flux at all in the case of pure reluctance

21
Position of the stator and rotor: - inner rotor radial, outer rotor radial, axial, trans-
verse, linear or multiple rotors or stators

Rotation speed: Synchronous or asynchronous relative to the armature frequency. A

classification of a selection of conventional machines is shown in figure 1-1.A Typical
torque speed characteristics of an electrical traction motor is shown in figure 1-2.

Figure 1-1: Conventional Electrical Machine Classification [3]

1.6 Fundamental Relationships in Permanent Mag-

net Synchronous Motors

This section presents the basic mathematical equations governing the operation of
permanent magnet synchronous motors.

22
 Figure 1-2: Typical torque versus speed characteristic.[4]

1.6.1 Speed

The steady-state rotor speed is given by,

f
ns = (1.1)
p

where f is the input frequency and p number of pole pairs.ns is equal to the syn-
chronous speed of the rotating magnetic field produced by the stator.

1.6.2 Air gap magnetic flux density

The fundament air gap magnetic flux density is [15],

Z 0.5αi π
2 4 αi π
Bmg1 = Bmg cos αdα = sin (1.2)
π −0.5αi π π 2

The coefficient αi is the ratio of the average–to–maximum value of the normal com-
ponent of the air gap magnetic flux density (αi = Bavg /Bmg ).

23
1.6.3 Voltage induced (EMF)

The no-load rms voltage induced in one phase of the stator winding (EMF) by the
d.c. magnetic excitation flux Φf of the rotor is

√
Ef = π 2f N1 kw1 Φf (1.3)

where N1 is the number of the stator turns per phase, kw1 is the stator winding
coefficient, and the fundamental harmonic Φf 1 of the excitation magnetic flux density
Φf without armature reaction is
Z τ
π 2
Φf 1 = L i Bmg1 sin xdx = τ Li Bmg1 (1.4)
0 τ π

where τ is the pole pitch.

1.6.4 Armature line current density and current density

The peak value of the stator (armature) line current density (A/m) or specific electric
loading is defined as the number of conductors in all phases 2m1 N1 times the peak
√
armature current 2Ia divided by the armature circumference πD1in ,
√ √ √
2m1 2N1 Ia m1 2N1 Ia m1 2N1 Ja sa
Am = = = (1.5)
πD1in pτ pτ

where Ja is the current density (A/m2) in the stator (armature) conductors and sa
is the cross section of armature conductors including parallel wires. For air cooling
systems Ja ≤ 7.5 A/mm2 (sometimes up to 10 A/mm2) and for liquid cooling systems
10 ≤ Ja ≤ 28 A/mm2. The top value is for very intensive oil spray cooling systems
[15].

24
1.6.5 Electromagnetic Power

Neglecting stator winding resistance, for an m-phase salient pole synchronous motor,
the electromagnetic power is given by[15],

V1 2
   
V1 Ef 1 1
Pem =m sin δ + − sin 2δ (1.6)
Xsd 2 Xsq Xsd

where V1 is the input (terminal) phase voltage, Ef is the EMF induced by the
rotor excitation flux (without armature reaction), δ is the power angle which is the
angle between V1 and Ef , Xsd is the synchronous reactance in the direct axis (d-axis
synchronous reactance), and Xsq is the synchronous reactance in the quadrature axis
(q-axis synchronous reactance).

Synchronous Reactance

The d-axis and q-axis synchronous reactances of a salient pole synchronous motor are
given by,
Xsd = X1 + Xad (1.7)

Xsq = X1 + Xaq (1.8)

where X1 = 2πf L1 is the stator leakage reactance, Xad is the d-axis armature reaction
reactance( d-axis mutual reactance) and Xaq is the q-axis armature reaction reactance
(q-axis mutual reactance). The reactance Xad is sensitive to the saturation of the
magnetic circuit whilst the influence of the magnetic saturation on the reactance Xaq
depends on the rotor construction [15]. The leakage reactance X1 consists of the slot,
end-connection differential and tooth–top leakage reactances.

1.6.6 Electromagnetic Torque

The electromagnetic torque developed by the synchronous motor is determined by

the electromagnetic power Pem and synchronous speed ωs = 2π ns (equal to the
mechanical speed of the rotor), and neglecting the stator winding resistance, it is

25
given by,
V1 2
   
m1 V1 Ef 1 1
Td = sin δ + − sin 2δ (1.9)
2πns Xsd 2 Xsq Xsd
Therefore Td can be expressed as,

Td = Tdsyn + Tdrel (1.10)

where the fundamental synchronous torque

m V1 Ef
Tdsyn = sin δ (1.11)
2πns Xsd

and the reluctance torque,

mV1 2
 
1 1
Tdrel = − sin 2δ (1.12)
4πns Xsq Xsd

For a cylindrical rotor synchronous motor where Xsd =Xsq , the reluctance component
is zero and
m V1 Ef
Td = Tdsyn = sin δ (1.13)
2πns Xsd

1.7 Magnetic Materials

The excitation field of an electrical machine can be provided either by permanent
magnets or by electrically energized windings. However the former method is pop-
ular where high efficiency and smaller size of the machines are desirable. Among
different magnet grades, Neodymium-Iron-Boron magnets was a main advancement
in permanent magnet technology.B-H curves for different types of permanent mag-
nets shown in figure 1-3. The remanent flux density Br and coercivity Hc of sintered
NdFeB magnets are higher than those of other magnet grades shown in the figure
1-3, including those of samarium-cobalt (Sm2 Co17 ) magnets. Even though NdFeB
magnets have attractive performance they have a relatively low Curie temperature
compared to other types of magnets, including samarium-cobalt magnets. This is
disadvantageous for use in demanding electric machine applications that often push

26
Figure 1-3: Second quadrant normal B-H curves of different types of permanent
magnet materials. Source: Magnequench [5]

the thermal limits of their wire insulation systems. Adding small amounts of another
rare-earth element, dysprosium (Dy) is one way of increasing the maximum temper-
ature range of NdFeB magnets.This effect is depicted in figure 1-4. Figure 1-4 shows
that the maximum operating temperature of NdFeB magnets increases monotonically
as the percentage of dysprosium by mass is increased from 0 % to greater than 10 %.
This pushes up the maximum operating temperature. However the impact of adding
dysprosium on magnet cost is significant as Dy is rarer and more expensive than Nd.

27
Figure 1-4: Impact of increasing dysprosium content on the coercivity Hcj and rema-
nent flux density Br of NdFeB magnets. Source : Arnold Magnetics [6]

28
Chapter 2

Study of Machine Topologies and

Motor Windings

Various technologies and topologies of machines were considered to achieve the design
targets. Each type of machine has its advantages and challenges. The comparison is
mainly focused on the ability to meet both the peak torque and speed requirements
while offering satisfactory performance in terms of torque ripple,efficiency.

2.1 Conventional Motor Types

Conventional machine technologies have existed for a long time. Most of them have
been used in traction applications and those technologies are relatively more mature.

2.1.1 Synchronous Motor (SM)

Wound field synchronous machines have windings on both the stator and the rotor,
as shown in Figure 2-1. The flux produced by the stator and rotor current interact
to produce a torque that is only present when the machine is running at synchronous
speed. A convenient and accurate field flux control is offered by field winding which
is useful for both field weakening and strengthening. Supplying the dc field to the
rotor poses the main challenge. This will result in additional copper losses, which

29
 Figure 2-1: Cross Section of a Wound Field SM [7].

are difficult to remove thermally and results in reduction of the overall efficiency.
Supplying the current to the rotating conductors is another issue. Usually this is
done by implementing brushes or exciters. However, they have significant drawbacks,
particularly at high speeds such as sparking, brush wearing etc. PM machines with
different rotor magnet configurations are shown in 2-2.

Figure 2-2: PM machines with different rotor configurations [8]

PMSM operational concepts

A PMSM is an electrical machine in which the rotor excitation field is provided by

permanent magnets unlike in induction machines where excitation is being induced by
feeding current through windings. The stator construction similar to those of in AC
induction machines.They require less maintenance since they do not have a system
to provide a rotor current.Moreover, PMSM can be designed to operate without
a gearbox.The PMSM has a constant magnetic field. Therefore the frequency of
induced voltage is equal to the electrical speed of the machine. The stator currents

30
 Figure 2-3: Surface Mounted PM motor

must also have the same frequency as the electrical speed of the machine to produce
torque. Elimination of gear box means the gearing has to be done electrically. This is
done by selecting an appropriate number of poles that the machine is equipped with
allowing it to operate at its optimum for a certain speed ranges.

Surface mount permanent magnet synchronous machine (SMPMSM)

SMPMSM machines have PMs on the rotor surface (shown in 2-3 ). They provide the
field flux. When the PMs as close to the air gap as possible ,the maximum magnet
utilisation factor resulting in higher torque densities is achieved . However, there are
some drawbacks of this. The absence of rotor saliency and so no reluctance torque
is produced, this means that obtaining torque in the field weakening region, while
maintaining the VA limits is difficult [[18],[19]]. Also, the PMs need a retaining sleeve
to withstand the high centrifugal forces present at high speeds [[20]].This type of
machine gives the high power and torque density required.But it is unlikely to meet
the high speed voltage and efficiency requirements [[20]]. However, the application
considered in this thesis does not require high speed operation or operating in the
field weakening region.Also high-performance PMs add considerable cost to machines.
SMPMSMs are commonly used for low speed high torque traction applications.

31
2.1.2 Interior permanent magnet synchronous machine (IPMSM)

The IPMSM machine operates on the same principle as the SMPMSM but the PMs
are buried under the surface of the rotor, shown in Figure 2-4. This gives an almost as
good power density and magnet alignment torque density as the SPM [18].This also
improves rotor robustness, even at higher speeds, because the PMs are buried under
the rotor surface. The interior magnets also add reluctance torque. It helps improve
torque density and high speed performance.However this reluctance value is small
compared to some of the topologies mentioned later [21]. The drawbacks are the need

Figure 2-4: Interior Permanent Magnet Synchronous Motor

for expensive PMs and while it can reach higher speeds than the SMPMSM machine,
the performance (in terms of efficiency and torque per amp) is not optimal at these
higher speeds.IPSM has high torque density and acceptable high speed capabilities.
They are used in many current hybrid vehicle applications [22] [23]. They have been
heavily investigated for these targets because of the torque density and moderate
speed range of the IPMSM design.Major example of this is the General Electric (GE)
machine [24].It uses radial spoke PMs inset in the rotor. Another example is the V-
shape IPMSMs in [25], performances of these designs are little worse than the radial
spoke machine. However the V-shape design used smaller magnets.

32
2.1.3 Brushless direct current machines (BLDCM)

BLDCMs have a polyphase stator and a rotor with PMs, similar to PM synchronous
machines. But in a BLDCM the PMs are designed to produce a square wave flux
density, which results in a trapezoidal back EMF and requires a square wave armature
current, in contrast to the sinusoidal counterparts found in the PMSM. This machine
has similar advantages to the synchronous machines, in that it provides high torque
with high efficiency [26]. The control of this BLDC is simpler and easy. In general
BLDC produces slightly higher torque than the PM synchronous machines.However
this simpler control and increased torque come at a cost which is the much higher
torque ripple in the machine than that of a synchronous machine.

2.1.4 Induction machine (IM)

IMs (shown in Figure 2-5) have existed for over a century and have been used regularly
in electric vehicles, especially larger vehicles. IMs exploits electromagnetic induction

Figure 2-5: Induction Motor

to generate a field flux from electrically conductive rotor bars, shown in Figure 2-5.
Therefore only the stator needs to be supplied with current.They are cheap, reliable

33
and rugged and can produce traction adequately [27]. A key benefit of IMs is that
the rotor flux can be controlled through field orientated control [28] making them
very competitive over PM machines in the high speed region. Current flowing in the
rotor causes some drawbacks, such as high losses and rotor cooling complexity [29].
While IMs are comparable to PM motors at high speed, it is unlikely the same design
would be able to meet the peak torque requirement with the current and temperature
limits.

2.1.5 Switched reluctance machine (SRM)

SRMs produce torque solely based on the reluctance of the rotor. SRMs have a simple,
robust, low cost rotor structure and capability for high-speed rotation [30],[31].This is
a consequence of the rotor being purely made out of laminated steel (with a toothed
shape to produce the reluctance) as shown in Figure 2-6. Compared to other machines

Figure 2-6: Switched Reluctance Motor

the motion of SRM is more discretised and it results in higher vibrations causing
acoustic noise and torque ripple [20, 32]. The simple rotor also means lower power
and torque density.So high torque SRMs need to be very large.
SRMs do not use a standard 3 phase AC stator winding as many of the other
machines do. So they require dedicated power electronics which increases cost and

34
reduces flexibility [32, 33].High torque low speed machines for full electric vehicles
are shown in [34] , though the only machines with high enough torque density for
the targets have low efficiency. A similar machine with high torque density but
with a lower efficiency SRM for EVs is mentioned in [32].The SRMs are optimal for
high speed applications, they are also candidates for applications where robustness
is important and the vibrations they produce is acceptable such as in off-highway
vehicles mentioned in [22].

2.2 Unconventional flux direction machines

These machines use conventional methods of producing the flux fields in the machine.
But unlike in conventional machines ,the flux of these machines travels in directions
other than radial direction.

2.2.1 Axial flux machines

Axial flux machines have flux travelling in the axial direction. Both the rotor and the
stator are annular shaped. Conventional machine types could be made axially, but
in generall a surface mount or interior PM rotor is implemented, with a synchronous
wound stator.
Axial machines generally have a smaller volume for the same power than their
radial counterparts. The extremely high torque density values achievable in axial flux
machines is the main benefit. However, this is quite dependent on the dimensions
and the power node being investigated. If the specifications favour the size of the
axial flux machine it would be very difficult for a different topology to outperform
it. However, the range of dimensions and power nodes where this superiority would
occur is very narrow [35]. The shape of axial machines makes them very suited to
applications where the radial space is limited, such as the in-wheel motor shown in
[36] . Because of the shape of axial machines,they can easily have double stator
or double rotor machines. This is very common with axial machines [35, 37] [9] ,
as shown in Figure 2-7. These machines are difficult to scale than radial machines

35
 Figure 2-7: Double Rotor Axial Flux IPMSM [9]

because the torque is proportional to the cube of the diameter and independent of
the stack length, so to increase the torque produced, the diameter must be increased
and so different size laminations must be manufactured. This lack of easy scalability
is a big issue when considering traction machines, if they are to be produced for a
range of different vehicles which is usually the case.

2.2.2 Transverse flux machines (TFM)

Flux in TFMs travels the transverse direction. This machine uses a homopolar phase
winding, a stator core formed by laminated C-cores and a rotor provided with het-
eropolar PMs[38]. The flux path associated with TFM is shown in the cross-section in
figure 2-8(a). Each TFM is single phased, so for multiple phases, multiple machines
are needed, a TFM with two phases is shown in Figure 2-8(b).The main advantages of
TFM are the potential high torque density and high electric loading [28]. Increasing
the pole number increases the power and reduces the speed of the machine for a fixed
electrical frequency. This means the torque increases significantly [38]. Each phase is
an independent system, giving modularity and improving fault tolerance capability

36
(a) Cross section of a transverse flux machine [38](b) Two phase outer rotor transverse flux machine
[39]

Figure 2-8: Transverse Flux Machine

[40].

2.3 Unconventional machine Topologies

These machines have the conventional radial shape of a machine but their methods
of producing or using the flux are different. Many of them are a variation of or a
hybrid between two of the conventional technologies.

2.3.1 Flux switching machine (FSM)

FSMs can be developed to produce the flux purely in the stator. The stator consists
of both the armature windings and the field producing arrangement. The field pro-
duction arrangement can either be PMs (PMFSM) or with a wound field (WFFSM)
to produce the field flux or combination of both methods. The rotor usually have
a similar design to an SRM rotor. So It has the same benefits of robustness and

37
 Figure 2-9: Outer Rotor Flux Switching SM

low cost [41]. More Control of the field flux is possible by using field-winding, but
it requires a DC current excitation .An exterior rotor machine using this is shown in
Figure 2-9.No field current input is required if PMs are used and so losses produced
are lower.Though, field weakening is required at higher speeds to reduce the field flux
to permissible levels. When both PMs and field windings are present in combination
the PMs can produce the most of the field flux and the windings can control this
allowing to strengthen or weaken the field.

The PMFSM has high air gap flux density similar to a conventional PM machine[42].
The benefit of this is that the design is less dependent on a small air gap, as SRM are,
making the construction simpler.The power density of PMFS machines is comparable
to that of a PM synchronous machine while the PM utilisation ratio is lower than
that of IPMSM machine [43]. Therefore larger PM mass is needed to get a similar
power density to an IPMSM design. However since PMs would be in the stator, they
would be closer to the cooling system.Hence less dysprosium (needed to improve the
maximum operating temperature of NdFeB) would be required. This can possibly
reduce the overall cost of PM material[44].Having fewer issues with magnet retention
is another advantage of the PMs in the stator.

38
2.3.2 Synchronous Reluctance Machine (SynRelM)

A section of synchronous reluctance motor (without PMs) is shown in figure 2-10(a).

The stators of SynRelMs is simlilar to those of synchronous machines but only produce

(a) An Outer Rotor SynRel Motor [45] (b) Permanent Magnet Assisted SynRel Motor [46]

Figure 2-10: Synchronous Reluctance Motors

reluctance torque Production of reluctance torque is based on the saliency of the

machine. The rotor which has flux barriers placed within it and it is made of laminated
steel,depicted in Figure 2-10(a), these flux barriers create different reluctances along
the d and q axis.As there’s no PMs or windings, the rotor of a SynRelM offers the
benefits of the SRM rotor, i.e it is cheap and robust and can be designed to operate
at high speeds [47].An acceptable level of torque density can be achieved [47, 48]
(having no PMs) with SynRelMs. But its performance cannot be compared to that
of a PM machine and thus it is unlikely to reach the requirements of the peak and
rated torque of the application being considered while respecting the size restrictions
and torque ripple.The required small air-gap to maintain the required performance
is an additional complexity of SynRelMs, making constructing and maintaining this
air gap more difficult than PM machines.The torque ripple and the power factor are
other challenges with SynRelMs. These can be minimized but are still worse than an

39
IPMSM machine in general [21].

2.3.3 Permanent magnet assisted synchronous reluctance ma-

chine (PMa-SynRelM)

PM-assisted synchronous reluctance machine (PMa-SynRelM) has PMs to improve

the torque density of the machine. However PMs increase the cost and mechani-
cal issues at high speeds [49] .These machines have an increased reluctance torque
compared to an IPMSM, which can be used to reduce supply current or magnet
mass requirements for a given torque requirement [50]. Figure 2-10(b) shows a PMa-
SynRelM machine where the rotor barriers are filled with ferrite PMs. This design has
high saliency while producing a moderate amount of magnet alignment torque.When
cheaper PMs or less NdFeB can be used than in an IPMSM the costs can be lowered
whilst achieving a high torque density.

The SynRel motor in particular with PM assistance, has been suggested for HEV
and EV applications. Using the space within the barriers, by implementing ferrite
PMs instead of rare earth is a common approach. In [46],it shows that this machine
can have high efficiencies at high speeds. However if large quantities of ferrite PMs
are used the mechanical issues caused by the centrifugal forces at high speeds of these
machines become even more critical[21].

2.4 Brief comparison of the motor type chosen for

the application

Considering the application requirements,the opportunities and limitations posed by

each motor type studies in the literature review it was decided to choose BLDC/BLAC
surface mounted synchronous motor for the application. Table 2.1 briefly summarizes
the comparison among different motor types discussed before.

40
 Machine Type Permanent Magnet Mass Torque Density Field Weakening Range
SMPMSM High High Very Poor
IPMSM High High Poor
IM None Medium Medium
SRM None Poor Very Wide
SynRelM None Medium Wide
FS Medium Medium Wide

Table 2.1: Performance comparison of Motor Topologies

2.4.1 BLDC/SMPM - Strengths

• High torque/power density

• High efficiency

• Torque linearly proportional to Iq , simplifying control

• High compatibility with FSCW stators

• Good candidates for achieving high CPSR using FSCW stators

2.4.2 BLDC/SMPM - Limitations

• Requires magnet containment means for high-speed operation.

• Vulnerable to high rotor losses at high speeds with FSCW stators.

• Magnets may need to be segmented to avoid high losses at high speeds.

• Short-circuit fault currents cannot be extinguished for any non-zero speed.

• Complete dependence on magnet torque requires high magnet content and cost.

2.5 Rotor Configurations

The motor type choice is the most basic design decision,this is due to relatively high
cost of magnets and issues related to packaging, magnet retention, and winding [10].
Various configurations of brushless motors which use rotating permanent magnets and

41
stationary phase coils exist. The utilization of different magnet grades in addition to
the wide range of applications are the main reasons for having many different varia-
tions. For example, for an application needing rapid acceleration and deceleration of
the load the torque/inertia ratio should be as high as possible. This call for the use
of an interior-rotor motor with high-energy magnets.However,for an application re-
quires constant speed at medium to high speed ,using an exterior-rotor configuration
(where rotor is outside of the wound stator) may be the preferred choice.Due to its
high inertia an exterior-rotor motor may be a good candidate for applications where
a very uniform and constant speed is required.

2.5.1 Interior Rotor Configuration

The interior-rotor motor (shown in figure 2-11 ) has a similar configuration to that of
the traditional SM or the IM.The stator is similar to that of the three-phase induction
motor. This design has a high torque/inertia ratio.But considering the manufacturing

Figure 2-11: Interior-rotor brushless permanent-magnet motor [10]

concerns ,it has the following two shortcomings :

1. Magnet retention must be carefully implemented so that the rotor does not fly
apart.

2. Exterior stators are expensive to wind without automatic equipment even though
cooling is easier.Examples of interior rotor brushless permanent magnet motors are
shown in figure 2-12.

42
 Figure 2-12: Examples of interior-rotor brushless permanent-magnet rotors [10].

2.5.2 Exterior Rotor Configuration

Exterior rotor configuration (shown in figure 2-13) can result in the most cost effective
use of ferrite magnets in brushless DC motors [10]. This type of stator is simple to
wind using DC-motor fly-winding machines. Some designs are produced on winding
machines that wind all three phases simultaneously. The rotor comprises a cup made
of soft iron mounted on the shaft with magnet arcs or a molded or bonded ring magnet
fixed inside the steel rotor cup with epoxy or Loctite[10]. The balancing is crucial
due to the large rotating mass.Magnet retention can be achieved by the rotor cup
on the outside of the magnets.The use of a single bearing support of aluminum or
die-cast zinc is an advantage of this configuration. Most interior-rotor motors need
a bearing at each end,hence requiring two bearing end-bells and higher cost [10]. In
general ,exterior-rotor brushless motors are used for continuous- speed applications.

43
 Figure 2-13: Exterior Rotor Configuration [10].

The magnet grades are usually the lower-cost versions of bonded rare-earth, bonded
or sintered ferrite grades.Outer rotor motor is popular due to low cost and ease of
manufacture.

2.5.3 Exterior Rotor Vs Interior Rotor

The external rotor design has some performance advantages.Outer rotor motor has
higher inertia as it has a rotor larger than that of a conventional DC motor. Higher
inertia helps to mitigate the torque ripple and hence provides smooth, stable operation
even at low speeds.
Generally,external rotor motors can produce higher torque than their similar sized
internal rotor counterparts. Torque is proportional to product of the magnetic force
and the air gap radius. Outer rotor motors have a larger air gap area than inner rotor
counterparts for a fixed motor diameter. The larger air gap permits a higher force
to build. They also have a larger air gap radius, which increases the lever arm for
torque production. The larger diameter (and, therefore, circumference 2-14) of the
rotor in external rotor designs allows the rotor to accommodate more poles further
increasing the magnetic flux. In [11], it presents a general comparison of permanent
magnet machines of interior and exterior rotor type based on scaling laws. Even
though the exterior rotor machine could theoretically produce a much higher torque
than achievable with the optimal interior rotor setup. But in real conditions the
torque is reduced due to the lowered air gap field in case of magnetic saturation [11].
Outer-rotor motors are shorter in axial direction than inner rotor motors with sim-

44
 Figure 2-14: Exterior and Interior Rotor Configurations [11]

ilar performance.They have compact size and high torque production. The external
rotor can serve as the hub of the fan or blower impeller providing a compact package
and allows the impeller to act as a large, rotating heat sink and assist with motor
cooling [51]. In [52] a comparison of Outer-Rotor-Type BLDC Motor and BLAC Mo-
tor based on numerical analysis is presented where they concluded the BLAC motor
to have a reduced of torque ripple and motor size,under identical current density con-
ditions, considering the driving method. In [53] design and optimization of an outer
rotor BLDC machine is presented.

2.6 Winding options

Ubiquitously used Winding configurations in three-phase radial-field permanent mag-

net (PM) brushless machines, can be categorized as [12]:

1) overlapping, either distributed 2-15(a) (two slots/ pole/phase) or concentrated

2-15(b) (one slot/pole/ phase);

2) non-overlapping, that is , concentrated, with either all teeth wound 2-15(c) or

alternate teeth wound 2-15(d). Distributed overlapping winding usually results in a
more sinusoidal magnetomotive force (MMF) distribution and EMF waveform, and
therefore it is ubiquitously used in PM brushless ac (BLAC) machines [54].
Fractional-slot concentrated-windings (FSCW) provide several benefits in syn-

45
chronous permanent magnet (PM) machines including high-power density, high effi-
ciency, short end turns, high slot fill factor particularly when coupled with segmented
stator structures, low cogging torque, flux-weakening capability, and fault tolerance
[55][56],[57]. In [58] method for directly calculating the winding factor, without do-
ing a winding layout first has been proposed. Fractional-slot winding machines are
good candidates for applications with certain requirements.For example they perform
well for low speed applications and the machine can be designed to have a higher
number of poles [59]. Fractional -slot winding can minimize the manufacturing costs
[60] while reducing end-winding lengths [61]. Further these fractional slot windings
can be designed to have a low torque ripple and to reduce the periodicity between
Q(the number of slots) and p (the number of pole pairs) [62]. These motors with
fractional-slot windings can be used in direct-drive applications.More over they can
be candidates to use design machines with fault-tolerance capability [63, 64].

2.6.1 Concentrated Winding Versus Distributed Windings

Concentrated and distributed windings are two common winding configurations.A coil
is wound around each tooth in the case of Concentrated windings. In a distributed
winding a coil can span multiple teeth.Concentrated windings usually have shorter
end windings and it’s possible to use segmented stators allowing to achieve better slot
fill factors. This can result in reduced copper losses and reduced machine sizes [65].
Distributed windings can produce higher reluctance torque as they can have higher
saliency between the d and q axis. This is a consequence of both the d and q axis
reluctances being proportional to both the winding factor and ratio of number of
coils and square of pole number [10]. Each of these can be equal for concentrated and
distributed designs. As distributed winding machines usually have higher number
of slots and integral slots per pole per phase typically these values are larger for
distributed designs [65, 66]. These distributed sinusoidal windings can result in lower
leakage flux and hence lower losses [10, 67].Concentrated winding designs are usually
more suitable to achieve a higher torque density as they permits higher slot fill factors
[68]. However, since distributed winding machines have a higher reluctance torque

46
they can perform better in field weakening region.

47
Figure 2-15: Typical stator winding configurations (four pole) [12]. (a) 24-slot, over-
lapping (distributed). (b) Twelve slot, overlapping (concentrated). (c)6-slot, nonover-
lapping, all teeth wound. (d) 6-slot, non-overlapping, alternate teeth wound.

48
Chapter 3

Developing Initial Design

Approach

This chapter details the design approach (mainly adopted from [1]) and the calcula-
tions related to basic sizing used in the initial design. Two main initial motor designs
viz one with fractional-slot winding fed by a sinusoidal drive and the other one with
integer- slot winding fed by a square wave drive are presented for the comparison
purpose. The BLDC motor fed by a square-wave drive allows to use a simpler control
strategy while offering slightly higher torque per volume. However, BLAC motor fed
by sinusoidal drive has much superior performance in terms of torque ripple provid-
ing a smooth torque.Therefore eventually BLAC motor fed by sinusoidally drive was
chosen for the application and further optimization.

3.1 Understanding the Application requirements

The specifications for the motor are tabulated in 3.1. The motor is an in-wheel motor
whose radial and axial dimensions are restricted by the available space which depends
on the dimensions of the rim. The maximum speed is as low as 400 RPM and the
speed range is narrow which does not require to operate the motor in flux weakening
region. The peak torque requirement is approximately 19-20 Nm and the motor peak
output power is 800 W(±10%). The transmission is direct drive where there’s no

49
gear box. The cooling method is natural air cooling therefore it’s required to ensure
temperature levela are maintained within the desired limits.

Parameter value
Voltage Rating 72V/48V
Motor Speed 400/250 RPM
Wheel Radius 0.3 m
Peak Power Rating 800 W/ 500 W ±10%
Torque 20 Nm/ 10 Nm ±10%
Continuous Power 500 W/ 250 W
Max Frequency 400 Hz
Transmission Direct drive
Winding Type Double layer concentrated
Temperature limits 100 C/ 150 C

Table 3.1: Basic Specifications of the Motor.

3.2 Preliminary Design Considerations

The aim was to design a permanent magnet motor for an electrical traction application
which is able to able to satisfy the following features:

high efficiency;

smaller size and light weight

suitable for direct-drive;

variable speed;

low noise and vibration;

easy to manufacture and assemble;

moderate cost.

3.3 Generalized Design procedure

The basic design algorithm and motor sizing procedure was adopted from previously
published works and textbooks by other authors [69, 1, 2]. The general design flow

50
is shown in figure 3-1

Figure 3-1: Generalized Design Procedure.

3.3.1 Exterior Rotor Configuration

Exterior-rotor configuration was selected to achieve a higher torque while keeping the
motor dimensions sufficiently smaller. The torque is produced at a radius that is quite
large relative to the outside diameter of the machine.This implies the same torque
can be obtained with lower electric and magnetic loading than would be required in
a traditional interior rotor motor [13].Permanent magnets are retained inside rotor
cup. It also provides a return flux path. It offers simplicity in manufacturing. This
solution does not require to apply any banding for the magnet retention. Further it
permits reducing the air-gap length compared the interior rotor counterpart. Cross
section of a typical exterior rotor is shows in figure 3-2.

51
Figure 3-2: Typical cross section of the exterior-rotor surface PM motor. courtesy:[13]

3.4 Determination of Motor Dimensions

The determination of the main dimensions of the motor (outer stator diameter De
and the stack length Lstk ) are chosen based on the available space and other phys-
ical restrictions imposed by the application itself. Being the rated torque Tn of the
machine related to the active volume, the adopted procedure is based on the ratio
between the torque and the outer volume of the machine itself. For this purpose
factor kT V is defined, which corresponds to this ratio. Typical kT V (TRV) values are
given in following table 3.2.

Cooling Type Typical values of TRV (kN/m3 ) Sigma lbf/in2 K lbf-in/in3

Small totally enclosed motors (Ferrite) 7- 14 0.5 -1 0.8-1.6
Totally enclosed motors (sintered rare earth or NdFeB) 14-42 1-3 1.6 -4.7
Totally enclosed motors (Bonded NdFeB) 21 typ. 1.5 typ. 2.4 typ.
Integral-hp industrial motors 7-30 0.5 -2 0.8-3
High-performance servo motors 15 -50 1-3 1.5 -5
Aerospace machines 30-75 2-5 3-7.5
large liquid-cooled machines 100 -250 10-15 15 -200

Table 3.2: Typical values of TRV, K and Sigma. Reference:[10]

The air gap volume is typically related to the rated motor torque, since the tan-
gential (shear) stress due to the interaction between the electric and magnetic fields
occurs at the air gap surface. However, the volume considered here refers to the outer

52
diameter. The outer volume has a more direct relation with the machine size.Since
the ratio between outer and inner stator diameters, De and Di respectively, changes
with the machine dimensions (small size machines have a De -to-Di ratio larger than
large size machines).This means the factor kT V also depends on the rated torque. For
example rated torque in the range between 5 and 50 Nm, a proper value (expressed
in Nm/litre) is, [70],
kT V ' 10N m/l (3.1)

It is noted that the value of kT V is within the range between 8 and 12 N m/l. This
range depends on the cooling effectiveness.Once the outer volume is fixed, De and
Lstk are segregated on the basis of further practical needs, such as maximum outer
space, maximum available length, existing stator lamination, and so on [70]. The
geometrical parameters of the outer-rotor designs are defined in 3-3.

Figure 3-3: Definition of the geometrical parameters for the outer-rotor SMPM motors
[14]

53
3.5 Magnet Selection
A comparison of different magnet grades are tabulated in 3.3.

Material of Magnet BHmax Br T Hc kA/m

NdFeB 200-500 0.97-1.45 740-1000
SmCo 120-400 0.85-1.10 620-840
Ferrite 7-42 0.20-0.48 120-360
AlNiCo 10-35 0.60-1.16 40-120

Table 3.3: Comparison of Magnets. [16]

Ferrite magnets have lower electrical conductivity than rare-earth magnets, so

the eddy current losses are much lower, hence the temperature rise due to eddy cur-
rent losses are lower. Ferrite magnets have a lower remanent magnetic flux density
compared to neodymium magnets, however ferrite magnets have a higher Curie tem-
perature, So ferrite magnets are suitable to use in high-temperature environments,
such as electric vehicles, thus offering improved reliability compared to the use of rare-
earth permanent magnets [71]. The figure 3-4 depicts the demagnetization curves for
NdFeB magnets.

Figure 3-4: Demagnetization Curves for NdFeB magnets.

3.6 Selection of the Number of Poles and Slots

The selection of pole and slot combinations is based on some rules when designing an
electrical machine. These are also applicable to concentrated windings.

54
The pole number is an even number.

the number of pole pairs, Pp , in a section of the machine can not be a multiple of the
phase number. This would lead to unbalanced windings [72, 73].

The number of poles can not be equal to the number of slots. This could results in
an undesired cogging torque in the machine in addition to the machine being a single
phase machine. The number of slots must be a multiple of Nph .

3.7 Concentrated Windings

Concentrated windings are usually preferred in industry especially due to the low cost
production technology.The motor length is main constraint for in-wheel motors,and
therefore concentrated winding is preferred. Moreover, the reduction of copper loss
and weight , higher fill factor and the lowered production cost are other advantages.
The main drawback is higher torque ripple of concentrated winding compared to
distributed winding . So modifying the geometrical parameters such as the stator
teeth to reduce the torque ripple is crucial and the most important method is the
combination between number of poles and slots [74, 75].

Fractional slot Windings

The chart in figure 3-5 shows the winding factors corresponding to different slot-pole
combinations.
A multipole PM motor can be a good solution for low-speed high-torque applica-
tions.Low iron mass per rated torque due to rather low flux per pole is the benefit
here. A high pole number in conventional distributed winding motors result a high
slot number increasing costs and, in the worst cases it can result in a low copper
fill factor. The fractional-slot concentrated-winding solution does not require many
slots although the pole number is high. This reduces both the iron and copper mass
in the motor. The fractional-slot winding allows a longer stator stack in the same
frame length than conventional windings as the axial length of end winding is typ-
ically shorter. It is possible to have a larger airgap diameter in a certain limited

55
Figure 3-5: Fundamental winding factors ξ1 for concentrated two-layer windings (q
≤ 0.5). [2]

stator outer diameter as the stator yoke can be manufactured very thin. This of-
fers a great potential to increase the torque density. Thus ,the multipole PM motor
with fractional-slot concentrated windings is selected as the direct-drive motor.In a
fractional-slot winding motor, various possible slot and pole number combinations ex-
ist. It is crucial to select the slot and pole number combinations that can achieve the
best machine performance. The selection criteria for fractional-slot winding motors
under study are as follows [75].

The fundamental winding factor should be as high as possible. This alllows to produce
the highest possible torque.

The lowest common multiple (LCM) between the number of poles and slots should
be as high as possible. The LCM order value increases the cogging torque frequency
and lowers its magnitude.

An unbalanced magnetic pull should be avoided. The slot/pole combinations (Qs /2p)
giving winding layouts without any symmetry such as combinations with Qs = 9+6m,
where m = 0, 1, 2,..., and 2p = Qs ± 1 should be avoided [76].
In addition to these criteria, a two-layer winding is used in order to achieve the

56
shortest possible end windings.

3.7.1 Winding Factors and Winding Feasibility

The machine periodicity t, the greatest common divisor (GCD) between Q and p is,

t = GCD(Q, p) (3.2)

A winding is feasible when the number of spokes per phase are equal,that is the ratio
qph = Q/mt should be an integer [77].

Distribution Factor

The ratio between the geometrical and the arithmetic sum of the phasors of the same
phase is defined as the distribution factor. The fundamental winding factor can be
calculated by [77],if qph is even,

qph αph
sin 2 2
kd = qph α (3.3)
2
sin 2ph

and if qph is odd kd i,s given by

α
sin qph 4ph
kd = α (3.4)
qph sin 4ph

where
2π
αph = (3.5)
( Qt )

Pitch Factor

The pitch factor is computed from the coil throw. The slot pitch yq is approximated
by,
Q
yq = round
2p

57
The lowest possible value for yq is unity. The coil span angle is ,

(2πpyq )
σw =
Q

The fundamental pitch factor is given by [77],

σw
kp = sin (3.6)
2

The winding factor for an motor (without skewing) can be expressed by

kw = kd .kp (3.7)

3.8 Design Approach with Analytical Formulae

The elementary block of a SMPM illustrating geometrical parameters is shown in

figure 3-6.

Figure 3-6: Elementary block of an SPM machine [1]

Main design parameters including shear stress (related to torque density), power
factor (PF), and Joule loss per exterior surface unit is expressed as a function of q and
the geometric quantities. The PF can be maximized by design, given the shear stress,
or vice versa. It’s beneficial to maximize PF specially for fractional slot machines due
to the following facts[1].

58
• A low PF negatively affects the size of the power converter.

• A low PF indicates that the machine can be prone to load dependent core saturation,
leading to a torque reduction.

• The machine inductance is the key design parameter of fractional slot SPM ma-
chines. [1]

3.8.1 Per-Unit Machine Model

Magnetic Loading

Magnetic loading B (peak fundamental flux density) at no load

Br
B = B̂gap,f und = kb . g (3.8)
1 + kc . lm

Where Br is the remanence flux of permanent magnets, kc is carter coefficient and kb

is a shape factor accounting for quantifying fundamental and is given by [78]

4 π
kb = . sin km (3.9)
π 2

where km is magnet’s pole arc

For a fixed air-gap length, the no-load magnetic loading B depends on rotor pa-
rameters and kc only. It does not depend on rotor pole pitch a.

The magnetic loading and the resistance to demagnetization of the machine is

eventually determined by the normalized PM thickness lm /g. In [1] , It says that
Over certain values, such as lm /g = 6, it is not convenient to further increase lm /g to
improve B,unless it is required by special overload needs and related demagnetization
issues.

59
Electric Loading and Shear Stress

The electrical loading is given by

3 N
A = . .kw .Iq (3.10)
2 a

where kw is the winding factor, N is the number of conductors in series per pole
per phase, and Iq is the phase current amplitude.To satisfy the maximum torque per
ampere scenario the current vector is aligned with the quadrature axis. The average
shear stress is given by,
σ = BA[N/m2 ] (3.11)

For a cylindrical machine σ is proportional to its torque per rotor volume density.
The electromagnet torque is given by,

π 2
Tem = D Lσ = 2Vr σ (3.12)
2

Once B is determined then the shear stress will depend on the electric loading
only.The upper limit of electrical loading related to thermal limit or efficiency target
or to demagnetization.

Specific Joule Loss

Joule loss factor kj is given by [1],

2ρCu kend A 2 l
kj = .( ) . (3.13)
kC u(1 − kt . BBf e ) kw lt

where ρCu is the electric resistivity of copper, kCu is the slot filling factor and kend is
the length of the conductors, including end connections, divided by their active length.
Bf e is the peak flux density in the stator back iron. Bf e is inversely proportional to
the cross section of the stator yoke, as defined in figure 3-6. kt is the tooth scaling
factor which is proportional to the tooth width.kj (in W/m2 ) can be calculated by
dividing the copper loss of the elementary block by its outer surface area.kj does not

60
 Figure 3-7: Definition of Power Factor [2]

depend on pole pitch and air-gap length. It mainly depends on tooth length hence at
continuous conditions kj determines tooth length value.

Power factor

The current vector is in time quadrature with the PM flux linkage λm,pole , and the
stator resistance voltage drop is neglected. The PF angle Φ can be calculated in
normalized quantities by,
4µ0 A
tan Φ ∼
= .Lpole,pu . (3.14)
3π B
where Lpole,pu is base inductance and the base inductance is given by,

2
µ0 .l 2
Lbase = ( .kw .N )
2 π

Here l is the stack length. If rare earth magnets are used then factor B has very
little variations when changing from one machine to another,and then the per-unit
inductance directly relates the PF to the shear stress (torque density)[1].

3.8.2 Analysis of Inductance

In this section , the minimization of the pole per-unit inductance is described. The
criteria for a best compromise between shear stress and PF is presented.The induc-
tance of the elementary block in Figure 3-6 is the sum of the slot leakage and the

61
air-gap inductances
Lpole,pu = Lg,pu + Lslot,pu (3.15)

In per unit values, components of Lpole,pu depend on the geometric variables defined
in Figure 3-6, with distinct expressions for distributed (integer q) and concentrated
(fractional q) windings.

Distributed Winding Machines

The magnetization inductance is given by[1],

π2 1 a
Lg,pu ' 2 .( lm
). (3.16)
6.kw kc + g
g

The slot inductance is,

lt
π2 g a −1
Lslot,pu = . .( ) (3.17)
2kw 2 1 − B
k g
Bf e t

Equations 3.16 and 3.17 are valid for q = 1.

It can be shown that the minimum inductance condition is when Lg = Lslot
[1].Setting Lg = Lslot , the polepitch- to-air-gap ratio that minimizes the inductance
can be calculated by, v
u lt
a u 3. g .(1 + lm
g
)
( ) =t
g Lmin 1 − BBf e kt

(3.18)

The minimum inductance, corresponding to 3.8.2 is [1]

v
2
u lt
π u g
(Lpole,pu )min =√ .t
lm B
(3.19)
3kw 2 (1 + g
).(1 − K)
Bf e t

Typical values of kt are 0.8–0.9 for distributed windings [1].

62
Fractional Slot Machines

The slot inductance for nl number of layers is,

lt
π2
 
g 3.(nl − 1) a −1
Lslot,p = . . 1 − .( ) (3.20)
2kw 2 1 − B
k 4.Q0 g
Bf e t

Q0 is the number of slots corresponding to half the electrical periodicity of the machine
[77], for those q where antiperiodic symmetry conditions apply, or corresponding to
the full electrical period, when they do not. Lg,pu is given by [1],
!
1 π2 1 a
Lg,pu = . . lm
. (3.21)
nl 12(qkw ) kc + g
g

the minimum inductance condition and the minimum inductance value are given
by[1], v    
u 6nl 1 − 3(nl −1) lt 1 + lm
  u
a 4.Q0 g g
= qt (3.22)
u
 
g Lmin 1− B k Bf e t

v  
u lt 3(nl −1)
2
1 π u
u
g
1 − 4Q0
(Lpole,pu )min = .√    
q 6kw 2 n 1 + lm . 1 − B k
t
l g Bf e t

(3.23)

The minimum inductance pole pitch in equation (3.28) is proportional to the

fractional q, while it was independent of integer q.

3.8.3 Power factor Maximization

The following conclusions about machines having minimized inductance have been
made in [1]

• Fractional slot machines tend to have a smaller (a/g)Lmin and then a higher number

63
of poles when the minimum inductance criterion is satisfied.

• In such conditions, double-layer machines can be close to integral slot ones for values
such as q = 1/2 or 2/5.

• Low-q machines and single-layer machines tend to have a high number of poles (low
a/g) for keeping the inductance low.

• Integral slot machines are insensitive to q, whereas fractional slot ones are very
sensitive to q.

• The minimum inductance is inversely proportional to the fractional q and becomes

very large for little values of q such as 1/8 or 1/10.

3.8.4 Design Algorithm

Machine’s torque T and number of pole pairs are given by,

T = σ.2π.r2 .l (3.24)

π.r
p= (3.25)
a
where σ is shear stress, r is rotor radius, l is stack length and a is pole pitch. Initial
designing of the elementary block , the reference value for Power factor is set and
designs that do not satisfy the requirements are rejected and the model is to be
redesigned with new input parameter values.An spread sheet was prepared for the
calculation and the results are provided in Appendix A.

Preliminary Input data

• Air-gap length g.

• q and type of winding.

• PM grade Br and thickness lm /g.

• Steel exploitation Bf e (peak).

• Cooling and thermal constraint, represented by target specific loss kj0 .

64
• Target shear stress σ0 , with reference to typical figures of machines having the same
type of cooling and the same size.

Design of the Elementary Block

1) Magnetic loading B is calculated via 3.8.

2) The electric loading is calculated from B and the σ0 target shear stress, according
to the equation 3.11.

3) The tooth length is chosen appropriately according to the loss target kj0 , according
to the equation 3.13. The end connection factor kend is also an approximate value.
So it would be recalculated once the active length and the pole pitch are determined.
This step might require iterative calculations.

4) Reference pole pitch (a/g)Lmin is calculated according to the minimum inductance

condition (3.8.2 or 3.19) respectively.

5) The minimized PF is evaluated and compared with the threshold.

i)If the PF is satisfactory, then the block is completely defined.

ii) If the PF is lower than the satisfactory threshold , then lt is reduced and the
flowchart is restarted from point 3. One of the two targets σ0 and kj0 must be
relaxed.

iii). If there is a PF margin much greater than the reference PF , then the pole pitch
is reduced with respect to (a/g)Lmin for increasing the number of poles. The outputs
of this stage are:

• pole pitch a/g and tooth length lt /g;

• the shear stress, the PF, and the Joule loss factor.

Sizing of the Motor

The additional input data are,

• target torque T0 and the rated speed

65
• maximum outer radius R0 and stack length l0 . The following six conditions are met
when the shear stress is provided.

1) The product r2 l is evaluated by 3.24, according to the target torque.

2) The rotor radius and stack length can be determined from r2 l within the allowable
length limit.

3) The number of pole pairs 3.25 is calculated and approximated to the closest feasible
number. Because only certain integers are feasible with fractional slots.

4) The end connection length and the specific loss are re-calculated accordingly.

5) Additionally the machine inductance and the PF are recomputed once the pole-pair
number is determined.

6) The stator outer radius is determined. If dimensions are within the permissible
limits then the flowchart is completed, the final design is modelled in MotorCAD and
FEA evaluated.
Preliminary data for the elementary block are tabulated in table 3.4.
Parameter value
Air-gap length 1 mm
q 4/11, 1
Winding Type Concentrated , Double layer
Br 1.45T
lm/g 10
Bf e (peak) 1.7T
Target kj0 600 W/m2
Lstack 5cm
σ0 7kN/m2
Target Torque 19.1 ±10% N at 400 RPM
maximum outer radius R0 10 cm

Table 3.4: Preliminary Design Data of Elementary Block

3.8.5 Summary of Equations Applicable to Radial Flux Per-

manent Magnet Motors

Back emf and Torque Equations: Sinusoidal Excitation

It is assumed the phase currents sinusoidal and all three phases are conducting during
the commutation period. Considering only fundamental back emf component , the

66
electromagnetic torque is given by,

3erms Irms
Te m = (3.26)
ωm

Peak back emf is given by,

dλ
epeak = −ωelectrical (3.27)
dθ

dλ
= Nph Ag Bav (3.28)
dθ
π
epeak = ωm Nph Bgav Di L (3.29)
2

epeak
er ms = (3.30)
2
Torque expressed in geometrical parameters,

3π
Te m = √ Nph Bgav Di LIrms (3.31)
2 2

Electrical loading A is defined by,

2mNph Irms
A= (3.32)
πDi

For a 3-phase motor where m = 3,

6Nph Irms
A= (3.33)
πDi

Therefore finally the torque under sine-wave excitation Tem,sine becomes,

π2
Tem,sine = √ Bgav ADi 2 L (3.34)
4 2

67
Back emf and Torque Equations : Square-wave Excitation

In this case, it is assumed that two phases are conducting simultaneously in commu-
tation period. The peak value of trapezoidal current is the dc link current IDC . For
a flattop value of the phase back emf Eph , the electromagnetic power is given by,

Pem = 2Eph IDC = Tem ωmech (3.35)

The rms value of a DC trapezoidal current with 1200 electrical angle is given by,
r
3
Irms = IDC (3.36)
2

The following set of equations are used to calculate flattop value of back emf Eph .

λ = Nph Φg (3.37)

dλ dλ dθelec dλ
e=− =− = −ωelec (3.38)
dt dθelec dt dθelec

2
φg = Bg Ag = Bg θmech Ri L = Bg θmech (3.39)
p
P 2Nph Φg
Eph = ωmech = Nph Bg LDi ωmech (3.40)
2 θelec
The torque equation under square-wave excitation becomes,

π
Tem,square = √ Bg ADi 2 L (3.41)
6

3.9 Simulation Results of Sinusoidally fed PMSM

For comparison purposes ,two initial designs viz one with fractional slot winding
with sinusoidal excitation and another one with integer slot (q=1) with square-wave
excitation were considered. An excel spread sheet was used for the initial calculations
done by using analytical equations mentioned in previous sections. The resulting

68
values are provided as tables in appendix A. Using those calculated parameter values
the initial design was simulated in MotorCAD software.

A wide range of slot-pole combinations and q values were considered and slot
number Q = 18 and pole number 2p = 20 was chosen for the fractional slot machine
with sinusoidal excitation.For the integer slot machine (q =1) , 12 poles and 36 slots
were chosen accordingly.The former machine and the latter machine will be referred
as 20P18Q motor and 12P36Q motor for the rest of this document.The radial cross
section of initial design of 20P18Q machine (with sinusoidal excitation) is shown in
3-8(a).The magnetic flux density distribution of the radial cross section is shown in
figure 3-8(b). The preliminary design data of the sinusoidal excited machine are

(a) Initial Design -Radial cross section (Sinusoidal) (b) Electromagnetics Window

Figure 3-8: Motor CAD model of Initial Design -Radial cross section 20P18Q Ma-
chine.

tabulated in table 3.5. The winding layout of 20P18Q machine is shown in figure 3-9.
The torque-speed characteristic of the 20P18Q machine’s initial design is shown in
figure 3-10. The output waveforms fractional Slot 20P18Q motor initial Design are
shown in figure 3-11. The drive output data of initial 20P18Q machine are tabulated
in the table 3.6. Electromagnetic output data of the initial 20P18Q machine are
tabulated the table 3.7. The flux densities of preliminary design 20P18Q Motor are
tabulated in table 3.8. The winding data of initial 20P18Q motor are shown in tabel
3.9.

69
 Stack length mm 5
Outer rotor diameter mm 164
Outer stator diameter mm 147
PM thickness mm 10
N.turns per phase 20
Wire size mm 2.28
Copper Slot Fill 0.4
Phase current Amax 14.98
Average torque @ 400 rpm Nm 19.1±10%
Output power W 800
Temperature: Winding C 0 140
PM C 0 120
Magnetization Radial
Peak Current 15.2A
RMS Current 10.75A
RMS Current Density 2.8A/mm2

Table 3.5: PMSM Preliminary Design Data Fractional Slot 20P18Q Machine

(a) Winding Layout Radial view (b) Linear Coil Pattern.

Figure 3-9: Winding Layout of 20P18Q Machine.

3.9.1 Simulation Results of the initial Design BLDC Motor

fed with Square-wave Drive

Preliminary design data for the integer slot motor are tabulated in table 3.10. The
radial cross section of initial design of integer motor is shown in 3-12(a).The magnetic
flux distribution of 12P36Q motor is shown in figure 3-12(b). The winding layout
(radial view) of the 12P36Q motor is shown in figure 3-13(a).The linear winding
pattern of 12P36Q motor is shown in figure 3-13(b). The torque- speed characteristics
of 12P36Q integer slot motor is shown in figure 3-14. The output waveforms of
12P36Q motor is shown in figure 3-15. The output data of integer slot motor fed by

70
 Figure 3-10: Torque-Speed Curve Initial Design Fractional Slot 20P18Q Motor

square-wave drive is tabulated in table 3.11.

3.10 Choice between 20P18Q motor and 12P36Q

Motor
According to the simulation results 12P36Q motor with square-wave excitation has
a higher torque for the same peak drive current than that of 20P18Q motor with
sinusoidal excitation. This is mainly due to higher the RMS current in 12P36Q
machine, hence 12P36Q machine has a higher torque per rotor volume compared to
20P18Q machine. 12P36Q machine can have a simpler control strategy compared to
the complicated control scheme to be used in 20P18Q machine with sinusoidal drive.
However the efficiency of 20P18Q motor is little higher than that of 12P36Q machine.
Most importantly the torque ripple in 12P36Q machine with square-wave excitation
exhibits a extremely high torque ripple due to higher harmonic content which is at
an acceptable level. In contrast 20P18Q machine with sinusoidall excitation has a
very little torque ripple as low as 6%. The torque ripple of both of the machines can

71
Variable Value Units Variable Value Units
DC Bus Voltage 72 Volts D Axis Inductance 0.3024 mH
Line-Line Supply Voltage (rms) 50.91 Volts Q Axis Inductance 0.488 mH
Phase Supply Voltage (rms) 29.39 Volts Line-Line Inductance (DQ) 0.7904 mH
Line-Line Terminal Voltage (peak) 67.32 Volts Self Inductance 1.059 mH
Line-Line Terminal Voltage (rms) 47.18 Volts Mutual Inductance -0.07318 mH
Phase Terminal Voltage (rms) 27.3 Volts Line-Line Inductance 2.265 mH
Harmonic Distortion Line-Line Terminal Voltage 1.585 % ”Armature End Winding Inductance
Harmonic Distortion Phase Terminal Voltage 6.557 % (Rosa and Grover)” 0.02921 mH
Back EMF Line-Line Voltage (peak) 64.91 Volts D Axis Current (rms) 0 Amps
Back EMF Line-Line Voltage (peak) (fundamental) 64.2 Volts Q Axis Current (rms) 10.75 Amps
Back EMF Phase Voltage (peak) 38.65 Volts Torque Constant (Kt) 1.317 Nm/A
Back EMF Line-Line Voltage (rms) 45.4 Volts Motor Constant (Km) 3.548 Nm/(Wattsˆ0.5)
Back EMF Phase Voltage (rms) 26.26 Volts Back EMF Constant (Ke) 1.55 Vs/Rad
Harmonic Distortion Back EMF Line-Line Voltage 1.682 % Back EMF Constant (Ke) (fundamental) 1.533 Vs/Rad
Harmonic Distortion Back EMF Phase Voltage 6.374 % Electrical Constant 4.299 msec
Max Line-Line / Phase Voltage Ratio 1.732 Mechanical Constant 0.8813 msec
DC Supply Current (mean) 12.09 Amps Electrical Loading 1.502E004 Amps/m
Line Current (peak) 15.2 Amps Stall Current 391.6 Amps
Line Current (rms) 10.75 Amps Stall Torque 515.9 Nm
Phase Current (peak) 15.2 Amps Short Circuit Line Current (peak) 242.7 Amps
Phase Current (rms) 10.75 Amps Short Circuit Current Density (peak) 63.58 Amps/mm
Phase Advance 0 EDeg Short Circuit Current Density (rms) 44.96 Amps/mm
Drive Offset Angle (Open Circuit) 130 EDeg Short Circuit Braking Torque -193.9 Nm
Drive Offset Angle (On load) 130 EDeg Short Circuit Max Braking Torque -206.3 Nm
Phase Advance to give maximum torque 1.818 EDeg Short Circuit Max Braking Torque Speed 283.9 rpm
Phasor Offset Angle 30 EDeg Short Circuit Max Demagnetizing Current -587.6 Amps
Phasor Angle (Ph1) 0 EDeg Fundamental Frequency 66.67 Hz
Phasor Angle (Ph2) 120 EDeg Current Shaft Speed RPM 400 rpm
Phasor Angle (Ph3) 240 EDeg Max Angle Between Phasors 120 EDeg

Table 3.6: Preliminary Design Drive Output Data of Fractional-Slot Machine

be reduced by skewing the rotor or the stator. However this would complicate the
manufacturing of the machine and on the other hand skewing would cause reduction
in winding factor hence reducing the torque. Therefore considering all these facts
20P18Q machine with sinusoidal excitation was chosen for the application and for
further optimization.

72
Variable Value Unit Variable Value Unit
Maximum torque possible (DQ) 72 Volts Flux Linkage D (Q axis current) 88.7622 mVs
(For Phase Advance of 1.818 EDeg) 20.248 Nm Flux Linkage Q (Q axis current) 7.37758 mVs
Average torque (virtual work) 20.099 Nm Flux linkage D (On load) 88.518 mVs
Average torque (loop torque) 20.035 Nm Flux linkage Q (On load) 7.26165 mVs
Torque Ripple (MsVw) 1.3636 Nm —–
Torque Ripple (MsVw) [%] 6.8094 % Torque Constant (Kt) 1.31741 Nm/A
Cogging Torque Ripple (Ce) 21.634 Nm Motor Constant (Km) 3.54777 Nm/(Wattsˆ0.5)
Cogging Torque Ripple (Vw) 0.075363 Nm Back EMF Constant (Ke) 1.54969 Vs/Rad
Speed limit for constant torque 64.91 Volts Back EMF Constant (Ke) (fundamental) 1.53256 Vs/Rad
(For Phase Advance of 0 EDeg) 430.73 rpm —–
No load speed 443.67 rpm Stall Current 391.617 Amps
Speed limit for zero current 1E009 rpm Stall Torque 515.921 Nm
Electromagnetic Power 838.79 Watts —–
Input Power 870.65 Watts Cogging Period 2 MDeg
Output Power 807.92 Watts Cogging Frequency 1200 Hz
Total Losses (on load) 62.728 Watts Fundamental Frequency 66.6667 Hz
System Efficiency 92.795 % Mechanical Frequency 6.66667 Hz
Shaft Torque 19.288 Nm Optimum Skewing Angle 2 MDeg
Power Factor [Waveform] (leading) 0.99689 Amps —–
Power Factor Angle [Waveform] 4.5208 EDeg Magnetic symmetry factor 2
Power Factor [Phasor] (leading) 0.99619 EDeg Magnetic Axial Length (Slice1) 50 mm
Power Factor Angle [Phasor] 5 EDeg Magnetic Axial Length Multiplier 1
Load Angle [Phasor] 4.6096 EDeg Number of Force Points 180
Phase Terminal Voltage (rms) [Phasor] 27.337 Volts X Force (On Load) 0 kN
Rotor Inertia 0.019573 kg.m2 Y Force (On Load) 0 kN
Shaft Inertia 0 kg.m2 Unbalanced Magnetic Pull (Open Circuit) 0 kN
Total Inertia 0.019573 kg.m2 Unbalanced Magnetic Pull Angle (Open Circuit) 0 MDeg
Torque per rotor volume 44.264 kNm/m Tangential Force (Open Circuit) -0.0135626 kN
Rotor peripheral velocity (on load) 3.4348 m/s Radial Force (Open Circuit) -8.07932 kN
Unbalanced Magnetic Pull (On Load) 0 kN X Force (Open Circuit) 0 kN
Unbalanced Magnetic Pull Angle (On Load) 0 MDeg Y Force (Open Circuit) 0 kN
Tangential Force (On Load) 0.225589 kN Radial Force Ripple (Open Circuit) (Rotor) 0.0393789 kN
Radial Force (On Load) -8.06805 kN Radial Force Ripple (On Load) (Rotor) 0.0416875 kN

Table 3.7: Preliminary Design ElectroMagnetic Output Data of Fractional-Slot Ma-

chine

Variable Value Units

Airgap flux density (mean) 0.9017 Tesla
Airgap Flux Density (peak) 1.182 Tesla
Stator Tooth Flux Density (peak) 2.022 Tesla
Stator Tooth Tip Flux Density (peak) 2.243 Tesla
Stator Back Iron Flux Density (peak) 0.04306 Tesla
Rotor Back Iron Flux Density (peak) 2.374 Tesla

Table 3.8: Flux Densities PMSM Preliminary Design Fractional Slot Motor

73
 (a) Back EMF (b) On-Load Flux Linkage

(c) Drive Currents (d) Air Gap Flux Density

(e) Terminal Voltages (f) Torque

(g) Harmonics: currents (h) Harmonics:BackEMF

Figure 3-11: Output waveforms Fractional Slot Motor Initial Design

74
Variable Value Units Variable Value Units
Armature Conductor CSA 3.817 mm2 Wire Slot Fill (Wdg Area) 0.8699
Armature Conductor Current Density 2.816 Amps/mm2 Copper Slot Fill (Wdg Area) 0.8082
Armature Conductor MLT 169.6 mm Wire Slot Fill (Slot Area) 0.592
Armature Turns per Phase 120 Copper Slot Fill (Slot Area) 0.55
Armature Turns per Coil 20 Heavy Build Slot Fill 1.108
Length of phase 2.035E004 mm Slot Area 277.6 mm2
Phase Resistance 0.09193 Ohms Winding Area (+Liner) 199.1 mm2
Line-Line Resistance 0.1839 Ohms Slot Area (FEA) 223.9 mm2
Armature Conductor Temperature 20 C Wedge Area 53.69 mm2
Mean Coil Pitch (Calculated) 15.78 mm Slot Opening Area 5.127 mm2
Mean Coil Pitch (Used) 15.78 mm Liner-Lam Imp Area 0 mm2
Fundamental Winding Factor 0.9452 Impreg Area 24.57 mm2
Winding Factor Sum 0.009002 Liner Area 10.17 mm2
Armature End Winding MLT (Calculated) 69.59 mm Coil Divider Area 24.81 mm2
Armature End Winding MLT (User adjustment) 1 Volume Copper EWdg Front 4.351E004 mm3
Armature End Winding MLT (Used) 69.59 mm Volume Copper Active 1.374E005 mm3
Wire Ins Thickness 0.04134 mm Volume Copper EWdg Rear 4.351E004 mm3
Copper Diameter 2.204 mm Conductors/Slot 40

Table 3.9: Winding Data Initial Design 20P18Q Motor

Stack length mm 5
Outer rotor diameter mm 164
Outer stator diameter mm 147
PM thickness mm 10
N.turns per phase 20
Wire size mm 1.48
Copper Slot Fill 0.55
Peak Phase current Amax 10.91
Average torque @ 400 rpm Nm 19.1
Output power W 800
Temperature: Winding C 0 140
PM C 0 60
Magnetization Radial
Peak drive Current 15.2A
RMS drive Current 12.41A
RMA Current Density 4.6A/mm2

Table 3.10: PMSM Preliminary Design Data Integer Slot Machine

75
(a) Motor CAD model of Initial Design Integer slot(b) Electromagnetics Window Integer Slot Ma-
12P36Q Motor -Radial cross section chine

Figure 3-12: Motor CAD model of Initial Design Integer slot Motor -Radial cross
section (Square-Wave drive).

(a) Winding Layout of Integer Slot 12P36Q Machine Radial (b) Linear Coil Pattern
view

Figure 3-13: Winding Layout of Integer Slot 12P36Q Machine.

76
 Figure 3-14: Torque-Speed Curve Initial Design Integer Slot Motor

Parameter value Units Parameter Value Units

Average torque (virtual work) 22.121 Nm Torque Constant (Kt) 1.44971 Nm/A
Average torque (loop torque) 21.789 Nm Motor Constant (Km) 2.41449 Nm/(Wattsˆ0.5)
Torque Ripple (MsVw) 59.492 Nm Back EMF Constant (Ke) 1.58832 Vs/Rad
Torque Ripple (MsVw) [%] 269.55 % Back EMF Constant (Ke) (fundamental) 1.62977 Vs/Rad
Electromagnetic Power 924.48 Watts Stall Current 198.515 Amps
Input Power 1006.1 Watts Stall Torque 287.789 Nm
Output Power 905.56 Watts Cogging Period 10 MDeg
Total Losses (on load) 100.58 Watts Cogging Frequency 240 Hz
System Efficiency 90.004 % Fundamental Frequency 40 Hz
Shaft Torque 21.619 Nm Mechanical Frequency 6.66667 Hz
Power Factor [Waveform] (leading) 0.94436 Optimum Skewing Angle 10 MDeg
Power Factor Angle [Waveform] -19.203 EDeg DC Bus Voltage 72 Volts
Rotor Inertia 0.030365 kg.m2 Line-Line Supply Voltage (rms) 50.91 Volts
Armature Conductor Current Density 4.619 Amps/mm2 Phase Supply Voltage (rms) 50.91 Volts
Total Inertia 0.030365 kg.m2 Line-Line Terminal Voltage (peak) 71.74 Volts
Torque per rotor volume 48.786 kNm/m3 Line-Line Terminal Voltage (rms) 51.74 Volts
Rotor peripheral velocity (on load) 3.4348 m/s Phase Terminal Voltage (rms) 51.74 Volts
Airgap flux density (mean) 0.6951 Tesla Harmonic Distortion Line-Line Terminal Voltage 16.51 %
Airgap Flux Density (peak) 1.078 Tesla DC Supply Current (mean) 13.98 Amps
Stator Tooth Flux Density (peak) 1.788 Tesla Drive Current Limit 15.2 Amps
Stator Tooth Tip Flux Density (peak) 1.831 Tesla Drive Current (peak) 15.2 Amps
Stator Back Iron Flux Density (peak) 0.1565 Tesla Line Current (rms) 12.32 Amps
Rotor Back Iron Flux Density (peak) 2.682 Tesla Phase Current (peak) 10.89 Amps
Electrical Constant 2.784 msec Phase Current (rms) 7.144 Amps
Mechanical Constant 7.174 msec Harmonic Distortion Line Current 30.48 %
Electrical Loading 3.444E004 Amps/m Harmonic Distortion Phase Current 31.96 %

Table 3.11: Output Data of Integer-Slot Motor :Square-Wave Drive

77
 (a) Back EMF (b) On-Load Flux Linkage

(c) Drive Currents (d) Air Gap Flux Density

(e) Terminal Voltages (f) Torque

(g) Harmonics: currents (h) Harmonics:BackEMF

Figure 3-15: Output Waveforms Integer Slot Motor Initial Design : Square-wave drive

78
Chapter 4

Sensitivity Analysis and Design

Optimization

4.1 Sensitivity Analysis

This section presents the results obtained from MotorCAD sensitivity analysis tool.

4.1.1 Electromagnetic Model Sensitivity Analysis

The parameters affecting the electromagnetic performance of the motor were analyzed
using Motor-CAD’s built-in sensitivity analysis tool.

Sensitivity analysis was performed by varying number of different design param-

eters including geometrical parameters, winding parameters and other parameters
such as peak current, magnet temperatures. Some of these parameters mainly affects
the electromagnetic performance of the machine while some others mainly affects the
thermal performance of the machine.The resulting graphs illustrating the variation
motor important output variables according to the variation of design parameters are
depicted in the following sections.

79
Magnet Thickness

Resulting graphs from sensitivity analysis with respect to magnet thickness are de-
picted in figure 4-1.

(a) efficiency (b) Harmonic Distortion

(c) Power (d) Torque

(e) Torque Ripple Percentage (f) torque per Rotor Volume

Figure 4-1: Effect of Magnet Thickness

80
Air Gap Length

Resulting graphs from sensitivity analysis with respect to air-gap length are depicted
in figure 4-2.

(a) Efficiency (b) Harmonic Distortion

(c) Power Factor (d) Torque

(e) Torque Ripple Percentage (f) Torque per Rotor Volume

Figure 4-2: Airgap:Sensitivity Analaysis

81
Armature Diameter

Resulting graphs from sensitivity analysis with respect to armature diameter are
depicted in figure 4-3.

(a) Efficiency (b) Harmonic Distortion

(c) Power Factor (d) Torque

(e) Torque Ripple Percentage (f) Torque per Rotor Volume

Figure 4-3: Armature Diameter:Sensitivity Analysis

82
Magnet Arc

Resulting graphs from sensitivity analysis with respect to magnet arc are depicted in
figure 4-4.

(a) Efficiency (b) Harmonic Distortion

(c) Power Factor (d) Torque

(e) Torque Ripple Percentage (f) Torque per Rotor Volume

Figure 4-4: Magnet Arc:Sensitivity Analysis

83
Slot Depth

Resulting graphs from sensitivity analysis with respect to slot depth are depicted in
figure 4-5

(a) Back EMF (b) Torque

(c) Efficiency (d) Harmonic Distortion

(e) Torque Ripple Percentage (f) Torque per Rotor Volume

Figure 4-5: Slot Depth:Sensitivity Analaysis

84
Slot Opening

Resulting graphs from sensitivity analysis with respect to slot opening are depicted
in figure 4-6.

(a) Efficiency (b) Harmonic Distortion

(c) Power Factor (d) Torque

(e) Torque Ripple Percentage (f) Torque per Rotor Volume

Figure 4-6: Slot Opening:Sensitivity Analysis

85
Tooth Tip Angle

Resulting graphs from sensitivity analysis with respect to tooth tip angle are depicted
in figure 4-7.

(a) Efficiency (b) Harmonic Distortion

(c) Power Factor (d) Torque

(e) Torque Ripple Percentage (f) Torque per Rotor Volume

Figure 4-7: Tooth Tip Angle:Sensitivity Analysis

86
Tooth Width

Resulting graphs from sensitivity analysis with respect to tooth width are depicted
in figure 4-8.

(a) Efficiency (b) Harmonic Distortion

(c) Power Factor (d) Torque

(e) Torque Ripple Percentage (f) Torque per Rotor Volume

Figure 4-8: Tooth Width:Sensitivity Analysis

87
Turns Number

Resulting graphs from sensitivity analysis with respect to turns number are depicted
in figure 4-9

(a) Efficiency (b) Electrical Loading

(c) DQ and End Winding Inductances (d) Torque

(e) Torque Ripple Percentage (f) Torque per Rotor Volume

Figure 4-9: Turns Number:Sensitivity Analysis

88
Copper Slot Fill

Resulting graphs from sensitivity analysis with respect to copper slot fill are depicted
in figure 4-10.

(a) Efficiency (b) Power

(c) Torque (d) Torque Ripple Percentage

(e) Torque per Rotor Volume (f) DQ Inductances

Figure 4-10: Copper Slot Fill:Sensitivity Analysis

89
Peak Current

Resulting graphs from sensitivity analysis with respect to peak drive current are
depicted in figure 4-11.

(a) Efficiency (b) Magnetic Flux Density

(c) Harmonic Distortion (d) Torque

(e) Torque Ripple Percentage (f) Torque per Rotor Volume

Figure 4-11: Peak Current:Sensitivity Analysis

90
Thermal Sensitivity to variations in Ambient Temperature

Sensitivity analysis graphs with respect to ambient temperature affecting thermal

performance of the motor are illustrated in 4-12.

(a) Magnet and Winding temperatures (b) Surface Temperatures

(c) Armature Losses (d) Total Losses

Figure 4-12: Thermal Sensitivity : Ambient Temperature

4.1.2 Thermal Model Output Data

Thermal model output data of fractional slot motor fed by sinusoidal drive is tabulated
in table 4.1.

4.2 Optimization of Motor Models

The optimization tool in MotorCAD allows to input a wide range of different de-
sign parameters including geometrical parameters, winding parameters, calculation

91
Temperature Value Temperature Value Temperature value
T [Housing - Overhang (F)] 59.236 T [Housing - Active] 59.322 T [Housing - Overhang (R)] 59.23
T [Housing - Front] 59.137 T [Stator Lam (back iron)] 99.159 T [Housing - Rear] 59.13
T [Endcap - Front] 58.508 T [Stator Surface] 98.558 T [Endcap - Rear] 58.46
T [Bearing - Front] 77.311 T [Rotor Surface] 62.087 T [Bearing - Rear] 77.29
T [Axle Ohang - Front] 95.678 T [Magnet] 61.118 T [Axle Ohang - Rear] 95.68
T [Axle - Front] 95.68 T [Rotor Lamination] 59.621 T [Axle - Rear] 95.68
T [End Space (F)] 72.252 T [Axle - Center] 95.68 T [End Space (R)] 72.23
T [Rotor (F)] 59.672 T [Active Winding Maximum] 99.668 T [Rotor (R)] 59.67
T [EWdg (F) Maximum] 99.767 T [Active Winding Average] 99.623 T [EWdg (R) Maximum] 99.77
T [EWdg (F) Average] 99.614 T [Active Winding Minimum] 99.437 T [EWdg (R) Average] 99.61
T [EWdg (F) Minimum] 99.195 T [Winding Maximum] 99.767 T [EWdg (R) Minimum] 99.19
T [Winding Average] 99.619 T [End Winding Average] 99.614 T [Model Minimum] 58.462
T [Winding Minimum] 99.195 T [Model Maximum] 99.767
Winding Temperatures ————————— ———————————— ———————————– ——————————— ————————————–
Temperature Cuboid1 Value C 0 Cuboid2 C 0 Temperature Cuboid1 Value C 0 Cuboid2 C 0
T [EWdg (F) Maximum] 99.767 99.767 T [Active Winding Minimum] 99.437 99.449
T [EWdg (F) Average] 99.651 99.578 T [EWdg (R) Maximum] 99.766 99.766
T [EWdg (F) Minimum] 99.423 99.195 T [EWdg (R) Average] 99.65 99.578
T [Active Winding Maximum] 99.668 99.634 T [EWdg (R) Minimum] 99.423 99.195
T [Active Winding Average] 99.638 99.608 T [Tooth] 99.39 99.024
Losses —————————– ———————————– ——————————— ——————————– ————————————–
Variable Value Unit Variable Value Unit
Loss [Armature Copper] 41.64 Watts Main Winding (Copper Loss Multiplier) 1
Loss [Armature Copper] (Active) 24.55 Watts Stall Operation (Copper Loss Multiplier) 1
Loss [Armature Copper] (EWdg Front) 8.544 Watts Fault Operation (Copper Loss Multiplier) 1
Loss [Armature Copper] (EWdg Rear) 8.544 Watts Loss[Stator Back Iron] Schematic Addition 0 Watts
Loss [Stator Back Iron] 2.578 Watts Dissipation - Housing - Active [Con] 11.61 Watts
Loss [Stator Tooth] 25.16 Watts Dissipation - Housing - Active [Rad] 7.128 Watts
Loss [Magnet] 1.778 Watts Dissipation - Front Housing OH [Con] 0 Watts
Loss [Rotor Back Iron] 0.2564 Watts Dissipation - Front Housing OH [Rad] 0 Watts
Loss [Airgap Banding] 0 Watts Dissipation - Rear Housing OH [Con] 0 Watts
Loss [Winding Sleeve] 0 Watts Dissipation - Rear Housing OH [Rad] 0 Watts
Loss [Windage] 0 Watts Dissipation - Front Endcap [Con] 17.71 Watts
Loss [Windage] (Ext Fan) 0 Watts Dissipation - Front Endcap [Rad] 8.319 Watts
Loss [Friction - F Bearing] 0 Watts Dissipation - Rear Endcap [Con] 18.14 Watts
Loss [Friction - R Bearing] 0 Watts Dissipation - Rear Endcap [Rad] 8.511 Watts
Loss [Total] 71.4097 Watts Total Dissipation to model ambient node 71.4097 Watts

Table 4.1: Fractional Slot Motor Thermal Model Output Data

parameters and material as well. The optimization criteria are two-fold viz. error
criteria and validity criteria.
Error Criteria: Error criteria usually work on outputs, for example flux densities.
These are evaluated after an optimisation model is solved, and the error result is the
sum of a multiple selection. The optimisation objective is to minimise the error sum,
therefore lower values are considered as improvements.
Validity Criteria: Validity criteria work on inputs, for example geometry of the
model, and they are Boolean expressions. If multiple validity criteria are selected and
if at least one of the selected validity criteria is false the combined criteria will yield
false.A false value is not considered a valid design to test. Therefore validity criteria
are evaluated before an optimisation model is solved. MotorCAD default criteria are
given below.

• Outer stator validity – a validity criterion

92
• Maximise Motor Constant – an error criterion

• Maximise Average Torque – an error criterion

• Maximise Load Point Torque – an error criterion

4.2.1 Electromagnetic Model Optimization

Initially the motor optimization was done solely considering the electromagnetic
model without taking thermal aspects into account.The output data for the initial
design and for an acceptable optimization solution are tabulated in table for compar-
ison. 4.2.

4.2.2 Multi-Physics Model optimization

A suitable design solution was chosen after the initial optimization stage. Then
the selected design was optimized considering both electromagnetic and thermal as-
pects.MotorCAD offers an iterative calculation option which takes electromagnetic
and thermal coupling into account converging to a electromagnetic-thermal coupled
solution. The final optimization results are tabulated in table 4.3.
According to the optimization results design 4 has the highest magnet weight re-
duction and highest torque per rotor volume(TRV). Design 6 has next best magnet
weight reduction and TRV value. Design 6 also has a better efficiency and a lower
torque ripple as well as lower magnet and armature winding temperatures. Consid-
ering these facts design 6 was chosen for the application.

93
Parameter (Initial Design) Value Unit Parameter (Optimized Design) Value Unit
Maximum torque possible DQ Maximum torque possible DQ
19.678 Nm 19.923 Nm
(For Phase Advance of 2.033 EDeg) (For Phase Advance of 2.627 EDeg)
Average torque (virtual work) 19.535 Nm Average torque (virtual work) 19.8 Nm
Average torque (loop torque) 19.454 Nm Average torque (loop torque) 19.709 Nm
Torque Ripple (MsVw) 1.1199 Nm Torque Ripple (MsVw) 1.3628 Nm
Torque Ripple (MsVw) [%] 5.7544 % Torque Ripple (MsVw) [%] 6.917 %
(For Phase Advance of 0 EDeg)” 428.49 rpm (For Phase Advance of 0 EDeg)” 460 rpm
No load speed 453.82 rpm No load speed 475.09 rpm
—– —–
Electromagnetic Power 815.23 Watts Electromagnetic Power 825.26 Watts
Input Power 877.01 Watts Input Power 852.54 Watts
Output Power 786.61 Watts Output Power 795.48 Watts
Total Losses (on load) 90.399 Watts Total Losses (on load) 57.059 Watts
System Efficiency 89.692 % System Efficiency 93.307 %
—– —–
Shaft Torque 18.779 Nm Shaft Torque 18.991 Nm
—– —–
Power Factor [Waveform] (leading) 0.99639 Power Factor [Waveform] (leading) 0.99434
Power Factor Angle [Waveform] 4.8696 EDeg Power Factor Angle [Waveform] 6.0963 EDeg
Power Factor [Phasor] (leading) 0.99619 Power Factor [Phasor] (leading) 0.99452
Power Factor Angle [Phasor] 5 EDeg Power Factor Angle [Phasor] 6 EDeg
Load Angle [Phasor] 4.9836 EDeg Load Angle [Phasor] 6.0563 EDeg
Phase Terminal Voltage (rms) [Phasor] 27.506 Volts Phase Terminal Voltage (rms) [Phasor] 22.738 Volts
—– —–
Rotor Inertia 0.030365 kg.m Rotor Inertia 0.021061 kg.m
Total Inertia 0.030365 kg.m Total Inertia 0.021061 kg.m
Torque per rotor volume 43.021 kNm/m Torque per rotor volume 59.219 kNm/m
Rotor peripheral velocity (on load) 3.4348 m/s Rotor peripheral velocity (on load) 3.4111 m/s
Magnet Weight 1.762 kg Magnet Weight 1.055 kg
—–
DC Supply Current (mean) 12.15 Amps DC Supply Current (mean) 11.84 Amps
Line Current (peak) 15.2 Amps Line Current (peak) 18 Amps
Line Current (rms) 10.75 Amps Line Current (rms) 12.73 Amps
Phase Current (peak) 15.2 Amps Phase Current (peak) 18 Amps
Phase Current (rms) 10.75 Amps Phase Current (rms) 12.73 Amps
—–
DC Bus Voltage 72 Volts DC Bus Voltage 72 Volts
Line-Line Supply Voltage (rms) 50.91 Volts Line-Line Supply Voltage (rms) 50.91 Volts
Phase Supply Voltage (rms) 29.39 Volts Phase Supply Voltage (rms) 29.39 Volts
Line-Line Terminal Voltage (peak) 67.84 Volts Line-Line Terminal Voltage (peak) 55.78 Volts

Table 4.2: Comparison Output Data of the Initial Design and the Initial Optimized
Design

94
Parameter Initial Value Min Max Design 1 Design 2 Design 3 Design 4 Design 5 Design 6 Design 7
Cup Radial Thickness(mm) 5 3 5 4.631239301 4.323260358 3.666975 4.359372 4.38799 4.957881 4.860439
Magnet Thickness(mm) 10 5 10 7.619230835 5.33763369 9.016304 6.412512 7.5926 6.163726 7.613213
Tooth Width(mm) 7 4 8 5.773974263 5.858767205 7.11223 7.96436 6.395715 6.610801 6.767478
Slot Depth(mm) 18 15 30 21.34443973 22.89401301 19.47368 15.7339 25.19741 27.2623 29.00687
Slot Opening(mm) 5 2 8 5.501536212 5.562118865 6.659138 6.741229 3.116827 4.74794 4.316576
Tooth Tip Angle(MDeg) 30 5 30 20.41676281 27.49306554 17.3236 20.65164 26.493 12.94273 26.9106
Armature Diameter(mm) 164 150 170 152.5414892 161.2547386 157.7696 152.3687 151.8595 162.8699 162.1341
Airgap(mm) 1 0.5 1 0.586364517 0.946017166 0.676639 0.637564 0.859044 0.982833 0.636688
Turns: 20 15 25 20 17 22 16 25 19 19
Copper Slot Fill 0.4 0.4 0.55 0.42376882 0.465240048 0.501002 0.52679 0.442017 0.526041 0.541373
Coil Divider(mm) 2 1 2 1.804407181 1.9901891 1.114667 1.547005 1.352052 1.193622 1.182279
Liner Thickness(mm) 0.25 0.25 1 0.443218018 0.695263786 0.810241 0.654167 0.823047 0.871555 0.791715
Total Weighted Error : 0.023328833 0.023328833 0.033855 0.033855 0.033855 0.028828 0.028828
Peak Current (A) 15.2 13 20 18 20 15 18 14 18 18
RMS Current(A) 10.75 12.73 14.4 10.61 12.73 9.89 12.73 12.73
RMS current Density(A/mm2 ) 3.87 3.67 2.8 3.11 3.5 3.23 2.136 2.09
Torque(Nm) 18.78 18.46 17.33 19.07 18.09 18.54 18.518 19.42
Torque Ripple % 5.7 5.8 4.5 7.1 6.05 5.89 5.6 7.69
Output Power(W) 786.6 773.4 725 798.87 758.7 776.7 775.7 813.7
Efficiency(%) 89.69 88.9 90.77 90.8 90.9 90.09 92.36 92.12
Loss(W) 90.39 96 73.79 80.6 75.3 85.4 64.1 69.6
Magnet weight improvement % -31.8181818 -48.8636364 -14.7727 -42.0455 -30.6818 -40.3409 -26.1364
TRV (kNm/m3 ) 43.02 57.28 72.67 54.95 63.09 57.8 57.7 56.12
TRV Improvement 33.14737331 68.92143189 27.73129 46.65272 34.35611 34.12 30.45095
Armature Winding Temperature(C 0 ) 126.1 25 150 137.5 115.8 121.7 123.4 129.2 104.3 107.4
Magnet Temperature(C 0 ) 102.4 25 120 116.8 96.95 103.2 102.2 106.9 88.19 93.63
Shaft Temperature(C 0 ) 103.8 25 150 110.6 94.63 98.3 98.57 104.2 86.79 89.68

Table 4.3: Final Optimization Results

95
96
Chapter 5

Validation of the Motor Model by

Finite Element Analysis

This chapter presents the MotorCAD simulation results of the selected design (Design
6 in the table 4.3 ). The results obtained by MotorCAD show that the designed motor
meets the specified requirements of application. The simulation results closely comply
with the analytical values.

5.1 Electromagnetic Validation (FEA of optimized

design)

The flux density distribution and current density of radial cross -section of the op-
timized design is shown in figure 5-1. Some variables of the design were calculated
analytically using the equations mentioned in chapter 3. Some of those analytical val-
ues and corresponding values yielded by simulations are tabulated in table 5.1. The
output graphs corresponding to the optimized 20P18Q machine model are shown in
figure 5-2. The torque- speed characteristic of the optimized 20P18Q motor is shown
in figure 5-3.

97
 (a) Flux Density (b) Current Density

Figure 5-1: Electromagnetic FEA Window :Optimized design

5.2 Lab Model

Motor-CAD Lab is a combined electromagnetic and thermal modelling toolbox.It can

be used for calculations facilitating the modelling and optimisation of a motor design
over its entire operating range.Motor-CAD Lab can be used create efficiency maps,
plot torque/speed characteristics, study the continuous and peak thermally con-
strained operational envelope and analyse performance over driving cycles.However
analysis of motor performance over a particular drive cycle is beyond the scope of
this thesis.

The optimized 20P18Q was modelled in Motor-CAD Lab and the results maximum
torque per ampere control strategy are provided in following sections.

Efficiency Map

The contour graph of efficiency and shaft torque versus speed of the final design is
shown in 5-4. Torque, efficiency -speed curves are shown in figure 5-5.

98
 Parameter Analytical Value Simulated Value Units
Torque 19.1 18.52 Nm
Output Power 800 775.74 W
Phase Current 11.79 12.73 A
RMS Current Density 2 2.13 A/mm2
Airgap Flux Density 1.52 1.03 T
Peak Current 16.67 18 A
Peak flux density in stator teeth 1.7 1.8 T
Losses 42 64 Watts
System Efficiency 95 92 %
Power Factor 0.998 0.993
Torque per rotor volume 59.9 57 kNm/m3
Electrical Loading 1.65E004 1.701E004 Amps/m

Table 5.1: Analytical and Simulated Values of Output variables

Operating Point under Maximum Torque per Amp Control Strategy

Operating point data under maximum torque per Ampere control strategy are tabu-
lated in table 5.2.

5.3 Thermal Model Analysis and Validation (FEA

of optimized design)

Schematic of thermal network steady state model of the optimized 20P18Q motor
design is illustrated in 5-6. FEA evaluated temperature distribution of radial cross
section is shown in 5-7. FEA evaluated temperature distributions of rotor and stator
are shown in 5-8.

5.3.1 Thermal Validation

The temperatures and deviation from validation temperatures are shown in 5-9 and in
5-10. It can be seen that the simulated temperature values comply with the validation

99
 Operating Point Parameter Value Unit Operating Point Parameter Value Unit
Shaft Speed 400 rpm Total Loss 66.05 Watts
Shaft Torque 19 Nm Stator Copper Loss 37.7 Watts
Shaft Power 795.9 Watts Iron Loss 28.2 Watts
Efficiency 92.34 Magnet Loss 0.1501 Watts
Stator Phase Current (peak) 18.31 Amps Mechanical Loss 0 Watts
Stator Line Current (peak) 18.31 Amps Electromagnetic Power 824.2 Watts
DC Terminal Current 11.97 Amps Electromagnetic Torque 19.68 Nm
Phase Voltage (peak) 31.44 Volts Magnet Torque 19.63 Nm
Line Voltage (peak) 54.46 Volts Reluctance Torque 0.05154 Nm
Phase Advance 3.103 EDeg Power Factor 0.998
Flux Linkage D 71.27 mVs Stator Winding Temperature (average) 113.5 C
Flux Linkage Q 8.637 mVs Stator Winding Temperature (max) 113.8 C
Magnet Flux Linkage 71.55 mVs Magnet Temperature 93.4 C
D axis Inductance 0.2827 mH Q axis Inductance 0.4723 mH

Table 5.2: Operating Point Data Under Maximum Torque Per Ampere Control Strat-
egy

temperatures. Temperature surface graphs of stator winding and magnets are shown
in figure 5-11.

5.3.2 Mechanical model (FEA of optimized design)

Motor-CAD mechanical tool facilitates calculating the mechanical stresses using finite
element analysis. The resulting mehanical stress distribution is shown in figure 5-12.
It is suggested to fabricate a sample rotor cut and perform a spin test to evaluate the
realistic centrifugal stresses and to validate the structural integrity,and durability of
the rotor. It can be seen that the maximum stress on rotor cup (about 304kPa) is
much less than typical yield stress 69MPa of aluminum.Also the stress on magnets is
also at an acceptable level.

100
 (a) Back EMF (b) Back EMF Harmonics

(c) Air-gap Flux (d) Flux Linkage(on-load)

(e) Torque (f) Torque Harmonics

(g) Terminal Voltage Harmonics (h) Current Harmonics

Figure 5-2: Output graphs of Optimized machine

101
Figure 5-3: Torque-Speed Characteristic of the optimized 20P18Q Machine

Figure 5-4: Contour graph of efficiency and shaft torque versus speed

102
 (a) Torque -Speed curves for different current levels (b) Efficiency Vs Speed

Figure 5-5: Torque, Efficiency - Speed curves

Figure 5-6: Schematic view of equivalent thermal network (steady state) model opti-
mized 20P18Q motor

103
 (a) Temperature distribution (FEA):Radial (b) Temperatures:Radial cross section

Figure 5-7: Radial cross section FEA evaluated temperature distribution of Optimized
20P18Q motor

104
 (a) Rotor Temperature Distribution (b) Stator Temperature Distribution

Figure 5-8: FEA evaluated temperature distribution of rotor and stator off optimized
20P18Q motor

105
Figure 5-9: Thermal validation Data of the Optimized 20P18Q Machine

Figure 5-10: Thermal validation Graph

106
 (a) Stator winding temperature (b) Magnet temperature

Figure 5-11: Temperature surface graphs of the optimized 20P18Q machine

Figure 5-12: mechanical stress distribution on rotor

107
108
Chapter 6

Conclusion

6.1 Conclusion

In this project, a permanent magnet synchronous motor was designed for a two
wheeler application. A review on available motor types/topologies was performed
and based on that study two main permanent magnet synchronous motor(surface
mounted) designs (A motor with Square-Wave shaped Back EMF waveform and a
motor with Sinusoidal shaped back EMF waveform ) were chosen for modelling in
MotorCAD software. Several possible winding configurations were considered and
fractional- slot double layer winding and a integer slot double layer winding were
used in initial simulations.Initial design calculations were performed using analyti-
cal equations and design algorithms which are published by different authors [1, 2].
Performance evaluation of two main motor designs was carried out using MotorCAD
software.The torque density of square-wave motor was higher than that of sinusoidal
-wave motor for a given peak drive current. However Considering the favorable per-
formance(in terms of torque ripple and efficency) of the motor with sinusoidal back
EMF waveform fed by a sinusoidal-wave drive was chosen for further analysis and
optimization. A sensitivity analysis was performed using tools available in Motor-
CAD focusing on main design parameters affecting the electromagnetic and thermal
performance of the machine.

109
 After that,in order to improve the performance and cost effectiveness of the motor
a general optimization was performed using the optimization tool provided by Mo-
torCAD. The final design was validated with respect to electromagnetic performance
using MotorCAD.Further thermal performance validation was done using MotorCAD
thermal model analysis to ensure the temperature levels of the motor are satisfactory.
According to simulation results it can be concluded that the designed motor meets
the requirements specified in the application in terms of electromagnetic and thermal
performance.

6.2 Future Work

The performance of proposed motor design with respect to electromagnetic and ther-
mal aspects was validated using MotorCAD. However the mechanical losses of the
machine were not considered in simulations.In the calculations mechanical losses were
neglected.Therefore modelling the design taking into mechanical losses is suggested
as a future work. Moreover the author suggests to fabricate a sample stator and
rotor cuts and to conduct spin tests to evaluate the realistic mechanical stresses and
performance of the machine.
Even though the designed motor exhibits good performance and meets the basic
requirements of the application there’s still some room for further improvement of the
design. Particularly, the proposed design uses stranded windings in the stator where
it’s practically very difficult have a higher copper slot fill.Therefore the author suggests
to investigate the possibility using hairpin type windings (which allows to achieve a
higher copper slot fill factor) in any future attempt to improve the performance of
the motor.Also the investigation of suitability of the motor for a particular driving
cycle is yet to be done as a future work. Moreover , the scope of this project was
limited to the design of the motor and hence the aspects related to motor drive and
control wasn’t considered. The circular tracking sine drive modulation method which
is the default option in MotorCAD was used in simulations.Therefore, it is suggested
to investigate the appropriate motor drive technologies as a future work.

110
Appendix A

Tables

Table A.1: Initial Design Calculations 20P18Q Machine with Sinusoidal Excitation
Parameter Design Design Design Design Design Design
1 2 3 4 5 6
L stk 0.05 0.05 0.05 0.05 0.05 0.05
g Air gap 0.001 0.001 0.001 0.001 0.001 0.001
m 3 3 3 3 3 3
q 1/4 2/5 1/3 2/7 2/7 0.5
Kc Carter Coefficient 1.085 1.085 1.085 1.085 1.085 1.085
n l Number of layers 2 2 2 2 2 2
Br T 1.45 1.45 1.45 1.45 1.45 1.45
lm/g 10 10 10 10 10 10
Bfe peak 1.7 1.7 1.7 1.7 1.7 1.7
kjo (W/mˆ2) target 700 700 700 700 700 700
Sigma 0 target (N/mˆ2) 7000 7000 7000 7000 7000 7000
km 5/6 5/6 5/6 5/6 5/6 0.83
Design of Elementary
Block
kb 1.23 1.23 1.23 1.23 1.23 1.23
B (T) 1.61 1.61 1.61 1.61 1.61 1.61

111
A (A/m) 4351.23 4351.23 4351.23 4351.23 4351.23 4351.23

Area 0.08 0.08 0.08 0.08 0.06 0.08

Loss 58.69 58.69 58.69 58.69 40.08 58.69
Kcu Fill factor 0.4 0.4 0.4 0.4 0.4 0.4
Temperature C 40 40 40 40 40 40
Rho Cu 1.8E-08 1.8E-08 1.8E-08 1.8E-08 1.8E-08 1.8E-08
kw Winding factor 0.95 0.95 0.95 0.95 0.95 0.87
kt 0.5 0.5 0.5 0.5 0.5 0.5
k end 0.3 0.3 0.3 0.3 0.3 0.3
l t (m) 0.009 0.009 0.009 0.009 0.009 0.009
lt/g 9 9 9 9 9 9
kj 120.33 121.41 121.41 121.41 121.41 144.63
Q0 10 10 10 9 9 6
a/g |L min 11.42 18.27 16.61 13.64 13.64 22.21
L pole,pu min 15.15 9.56 10.51 12.68 12.68 8.86
n s (turns per slot per 20 20 20 20 20 20
phase)
N (turns per pole per 5 8 7 6 6 10
phase)
L base H 4.6E-06 4.5E-06 4.5E-06 4.5E-06 4.5E-06 3.8E-06
L pole min 6.9E-05 4.3E-05 4.7E-05 5.7E-05 5.7E-05 3.4E-05
Tan Phi 0.04 0.04 0.04 0.04 0.04 0.04
Phi 0.04 0.04 0.04 0.04 0.04 0.04
PF 1.00 1.00 1.00 1.00 1.00 1.00
Power W 800 800 800 800 500 800
T target Nm 19.10 19.10 19.10 19.10 11.94 19.10
Speed RPM 400 400 400 400 400 400

112
rˆ2L stk 0.000 0.000 0.000 0.000 0.000 0.000
r 0.093 0.093 0.093 0.093 0.074 0.093
D rotor (Outer) 0.186 0.186 0.186 0.186 0.147 0.186
Stator Diameter 0.164 0.164 0.164 0.164 0.125 0.164
TRV kN/mˆ3 14 14 14 14 14 14
a |L min 0.01 0.02 0.02 0.01 0.01 0.02
p Pole pairs 25.64 16.03 17.63 21.47 16.97 13.18
2p Approximated 26 16 20 20 20 8
Slots Z 24 18 18 18 18 12
q calculated 0.31 0.38 0.30 0.30 0.30 0.50
Pole pitch (a) 0.02 0.04 0.03 0.03 0.02 0.07
n (turns per pole per phase) 6.15 7.50 6.00 6.00 6.00 10.00
DC bus Voltage V 72 72 72 72 72 72
Phase Supply voltage V 29.39 29.39 29.39 29.39 29.39 29.39
Efficiency 0.9 0.9 0.9 0.9 0.9 0.9
Iq Current 11.18 14.97 14.97 14.97 11.84 24.52
I dc DC mean current 12.35 12.35 12.35 12.35 7.72 12.35
LCM(2p,Z) 312 144 180 180 180 24
Cogging torque Periodicity 13 8 10 10 10 2
n
Is (Slot current) 335.43 449.25 449.25 449.25 355.16 735.49
Iph 7.91 10.59 10.59 10.59 8.37 17.34
Ipk 11.18 14.97 14.97 14.97 11.84 24.52
t periodicity 1 2 2 2 2 4
q ph= Z/mt (an integer) 8 3 3 3 3 1
Alpha ph 0.26 0.70 0.70 0.70 0.70 2.09
Kd if q ph is even 0.96 0.97 0.97 0.97 0.97 1.15
Kd if q ph is odd 0.96 0.96 0.96 0.96 0.96 1.00

113
y q Slot pitch max 1 1 1 1 1 1
Epsilon (remainder) 0.92 0.13 0.90 0.90 0.90 0.50
Cph coils per phase 8.00 6.00 6.00 6.00 6.00 4.00
HCF(Cph,p) number of 1.00 2.00 2.00 2.00 2.00 4.00
sections
Nss Slots per Section 24.00 9.00 9.00 9.00 9.00 3.00
Sf progressive ,epsilon=0.5 2.00 2.00 2.00 2.00 2.00 2.00
Sf retrogressive ,epsilon0.5 23.00 8.00 8.00 8.00 8.00 2.00
Sigma w Coil span Angle 3.40 2.79 3.49 3.49 3.49 2.09
Kp Pitch factor 0.99 0.98 0.98 0.98 0.98 0.87
Kp 0.99 0.98 0.98 0.98 0.98 0.87
Kw winding factor calcu- 0.95 0.95 0.95 0.95 0.95 0.87
lated
Q0 new 10 10 10 9 9 9
kjo Recalculated 477.07 477.07 477.07 477.07 698.69 477.07
k end Recalculated 1.19 1.18 1.18 1.18 1.73 0.99
L slot pu 7.58 4.78 5.26 6.34 6.34 4.64
L g pu 1.45 1.91 2.17 1.78 1.78 1.90
L pole pu new 12.31 10.19 12.74 12.68 12.68 9.07
L base new 0.00 0.00 0.00 0.00 0.00 0.00
L pole new H 0.00 0.00 0.00 0.00 0.00 0.00
tan Phi new 0.02 0.01 0.02 0.02 0.02 0.01
Phi new 0.02 0.01 0.02 0.02 0.02 0.01
PF new 1.00 1.00 1.00 1.00 1.00 1.00
minimum Wt Tooth width 0.00 0.01 0.01 0.01 0.01 0.01
Back iron thickness 0.00 0.01 0.01 0.00 0.00 0.01
Rotor Back iron 0.01 0.01 0.01 0.01 0.01 0.01
Stator diameter Ds 0.16 0.16 0.16 0.16 0.13 0.16

114
np 13 8 10 10 10 2
Magnet arc 150 150 150 150 150 150
Throw 1 1 1 1 1 1

Table A.2: Integer - Slot Motors Initial Designs.

Parameter Design 1 Design 2 Design 3 Design 4 Design 5
L stk 0.05 0.05 0.05 0.05 0.05
g Air gap 0.001 0.001 0.001 0.001 0.001
m 3 3 3 3 3
q 1 2 3 1 1
Kc Carter Coefficient 1.05 1.05 1.05 1.05 1.05
n l Number of layers 2 2 2 2 2
Br T 1.45 1.45 1.45 1.45 1.45
lm/g 10 10 10 10 10
Bfe peak 1.7 1.7 1.7 1.7 1.7
kjo (W/m2 ) target 450 450 450 450 450
Sigma 0 target (N/m2 ) 7000 7000 7000 7000 7000
km 5/6 1 1 1 2/3
Design of Elementary Block
kb 1.23 1.27 1.27 1.27 1.102658
B (T) 1.61 1.67 1.67 1.67 1.446927
A (A/m) 4337.489326 4189.693 4189.693 4189.693 4837.841
Kcu Fill factor 0.4 0.4 0.4 0.4 0.4
Temperature C 40 40 40 40 40
Rho Cu 1.81104E-08 1.81E-08 1.81E-08 1.81E-08 1.81E-08
kw Winding factor 0.866025404 0.933013 0.945214 0.866025 0.866025
kt 0.5 0.5 0.5 0.5 0.5
k end 0.3 0.3 0.3 0.3 0.3
Continued on next page

115
 Table A.2 – continued from previous page
Parameter Design 1 Design 2 Design 3 Design 4 Design 5

l t (m) 0.009 0.009 0.009 0.009 0.009

lt/g 9 9 9 9 9
kj 144.1285948 119.6717 116.6022 138.901 163.9755
a/g —L min 23.77700327 24.16523 24.16523 24.16523 22.73832
L pole,pu min 9.48 8.30 8.09 9.64 9.067403
n s (turms per slot per phase) 20 20 20 20 20
L base H 3.81972E-06 4.43E-06 4.55E-06 3.82E-06 3.82E-06
L pole min 3.6217E-05 3.68E-05 3.68E-05 3.68E-05 3.46E-05
Tan Phi 0.041569569 0.038785 0.038785 0.038785 0.051713
Phi 0.04154565 0.038766 0.038766 0.038766 0.051667
PF 0.999137104 0.999249 0.999249 0.999249 0.998666
Power W 800 800 800 800 800
T target Nm 19.09859317 19.09859 19.09859 19.09859 19.09859
Speed RPM 400 400 400 400 400
Efficiency 0.9 0.9 0.9 0.9 0.9
DC bus Voltage V DC V 72 72 72 72 72
Supply Voltage v s V 29.39387691 29.39388 29.39388 29.39388 29.39388
Phase Current I ph A 10.08891039 10.08778 10.08778 10.08778 10.09367
Peak Current I peak 12.35634125 12.35496 12.35496 12.35496 12.36218
Slot Current Is A 244.3898373 109.557 72.09519 236.0625 272.5814
r2 L stk 0.000434234 0.000434 0.000434 0.000434 0.000434
r 0.093191592 0.093192 0.093192 0.093192 0.093192
D rotor 0.186383185 0.186383 0.186383 0.186383 0.186383
a —L min 0.023777003 0.024165 0.024165 0.024165 0.022738
p Pole pairs 12.31315901 12.11534 12.11534 12.11534 12.87562
2p Approximated 12 12 12 12 12
Continued on next page

116
 Table A.2 – continued from previous page
Parameter Design 1 Design 2 Design 3 Design 4 Design 5

Slots 36.00 72.00 108.00 36.00 36

q calculated 1 2 3 1 1
Z/2p 3 6 9 3 3
t periodicity 6 6 6 6 6
q ph= Z/mt 2 4 6 2 2
Alpha ph 1.047197551 0.523599 0.349066 1.047198 1.047198
Kd even q ph 1 0.965926 0.959795 1 1
Kd odd q ph 0.965925826 0.957662 0.956143 0.965926 0.965926
y q Slot pitch 2 5 8 2 2
Sigma w Coil span Angle 2.09 2.62 2.79 2.09 2.09
Kp 0.866025404 0.965926 0.984808 0.866025 0.866025
Kw calculated 0.866025404 0.933013 0.945214 0.866025 0.866025
kjo Recalculated 477.0741353 477.0741 477.0741 477.0741 477.0741
k end Recalculated 0.993017664 1.195957 1.227441 1.03039 0.872827
L slot pu 4.740800325 4.15118 4.044704 4.818207 4.533702
L g pu 4.71934874 4.132397 4.026402 4.796405 4.513187
L pole pu new 9.48160065 8.302361 8.089408 9.636414 9.067403
L base new 1.90986E-07 2.22E-07 2.28E-07 1.91E-07 1.91E-07
L pole new H 1.81085E-06 1.84E-06 1.84E-06 1.84E-06 1.73E-06
tan Phi new 0.013627615 0.011133 0.010848 0.012922 0.016212
Phi new 0.013626772 0.011133 0.010847 0.012922 0.016211
PF new 0.999907157 0.999938 0.999941 0.999917 0.999869
minimum Wt Tooth width 0.002866889 0.001508 0.001005 0.003016 0.002458
Back iron thickness 0.007184854 0.00756 0.00756 0.00756 0.00616
Rotor Back iron 0.005 0.005 0.005 0.005 0.005
Stator diameter Ds 0.164383185 0.164383 0.164383 0.164383 0.164383
Continued on next page

117
 Table A.2 – continued from previous page
Parameter Design 1 Design 2 Design 3 Design 4 Design 5

np 1 1 1 1 1
Magnet arc 150 180 180 180 120
Throw 2 5 8 2 2

118
 12P36Q
Motor
20P18QMotor
150Deg 12P36Q
Sinusoidally fed by
Parameter MagnetArc 120Deg
fed 20P18Q Square wave
fed by MagnetArc
drive
Sq-Wave
Drive
pole number 20 20 12 12
slots 18 18 36 36
Winding factor Kw 0.945 0.945 0.866 0.866
Current desisty J A/mmˆ2 4.2 2.77 6 5.99
Tooth width (mm) 7 7 6 5
Amature dia (mm) 164 164 164 164
stack lenghth (mm) 50 50 50 50
Turns 20 20 20 20
Throw 1 1 2 2
Average Torque (Nm) 20.8 11.83 20.4 20.39
Shaft Torque Nm 20.04 11.08 19.93 19.9
Torque ripple 7% 25% 121% 139%
Input Power W 932.27 521 960.5 971.59
EM power W 869.57 493.8 853.95 852.35
Output Power W 839.44 464.2 834.8 833.78
total Loss (on load) W 92.83 56.8 125.7 137.8
Efficiency 90.04% 89.10% 86.90% 85.81%
Power factor 0.99 0.91 0.94 0.95
PF angle Edeg 5.49 -23.5 -19.8 -17.44
torque per Volume kNm/m3 45.88 26.06 45.06 44.98
Torque const Kt Nm/A 1.29 0.736 1.27 1.27
BEMF const Ke Vs/Rad 1.54 0.92 1.71 1.52
BEMF const fundamental Ke fund Vs/Rad 1.518 0.87 1.65 1.45
Motor Const Km Nm/Wˆ0.5 2.62 2.26 1.97 1.86
Cogging Period Mdeg 2 2 10 10
Cogging Frequency Hz 1200 1200 240 240
Fundamental Freq Hz 66.67 66.67 40 40
Mechanical freq Hz 6.67 6.67 6.67 6.67
Drive peak Line current A 16 16 16 16
peak phase current A 16 12.47 10.72 10.67
RMS Line current A 11.31 12.72 11.47 7.49

Table A.3: Performance Comparison of Initial Designs

119
120
Appendix B

Figures

121
Figure B-1: Torque-angle characteristics of a salient-pole synchronous machine with
Xsd ¿ Xsq : 1 — synchronous torque Tdsyn , 2 — reluctance torque Tdrel , 3 — resultant
torque Td [15]

Figure B-2: Flux Density Vs Magnetic Field for M350 50A Electrical steel

Figure B-3: Torque, speed and efficiency surface graph of optimized 20P18Q Machine

122
(a) Axial Section of 20P18Q(b) 3D Model without Rotor Cup 20P18Q Machine
Machine

Figure B-4: Axial cross Section and Motor CAD 3D Model

123
124
Bibliography

[1] B. Boazzo, G. Pellegrino, and A. Vagati, “Multipolar SPM machines for direct-
drive application: A general design approach,” IEEE Transactions on Industry
Applications, vol. 50, no. 1, pp. 327–337, 2014.

[2] P. Salminen, Fractional Slot Permanent Magnet Synchronous, 2004.

[3] A. E. Machines, “ADVANCED ELECTRICAL MACHINES.”

[4] N. Bianchi and G. Berardi, “Analytical Approach to Design Hairpin Windings

in High Performance Electric Vehicle Motors,” 2018 IEEE Energy Conversion
Congress and Exposition, ECCE 2018, pp. 4398–4405, 2018.

[5] “Homepage - Magnequench.” [Online]. Available: https://mqitechnology.com/

[6] “Permanent Magnet Division — Arnold Magnetic Technologies.” [Online].

Available: https://www.arnoldmagnetics.com/permanent-magnets/

[7] J. Pyrhonen, T. Jokinen, and V. Hrabovcova, Design of Rotating Electrical

Machines- Wiley 2009,0470695161.pdf, 2008. [Online]. Available: www.wiley.
com

[8] K. Wang and Z. Q. Zhu, Third harmonic utilization in permanent magnet ma-
chines, 2018.

[9] A. Mebarki, D. Gerada, and N. L. Brown, “Analysis of an axial PM flux machine

with field weakening capability for engine integration,” 7th IET International
Conference on Power Electronics, Machines and Drives, PEMD 2014, 2014.

[10] J. R. Hendershot, T. J. E. Miller, D. C. Hanselman, B. Boukais, H. Zeroug, A. N.

Patel, Y. Kim, C. Lee, and A. N. Patel, Design of Brushless PM motors, 2016,
vol. 46, no. 11.

[11] T. Reichert, T. Nussbaumer, and J. Kolar, “Torque scaling laws for interior
and exterior rotor permanent magnet machines,” Aa, vol. 1, pp. 1–4, 2009.
[Online]. Available: http://www.pes.ee.ethz.ch/uploads/tx ethpublications/
Reichert Transmag09 TorqueScaling Final.pdf

125
[12] A. M. EL-Refaie, Z. Q. Zhu, T. M. Jahns, and D. Howe, “Winding inductances of
fractional slot surface-mounted permanent magnet brushless machines,” COM-
PEL - The International Journal for Computation and Mathematics in Electrical
and Electronic Engineering, vol. 28, no. 6, pp. 1590–1606, 2009.

[13] F. Parasiliti, M. Villani, and M. Castello, “PM Brushless DC Motor with exterior
rotor for high efficiency household appliances,” Proceedings - 2014 International
Conference on Electrical Machines, ICEM 2014, pp. 623–628, 2014.

[14] F. Libert, Design, optimization and comparison of permanent magnet motors

for a low-speed direct-driven mixer, 2004. [Online]. Available: http://faculty.mu.
edu.sa/public/uploads/1338064112.6994pmsg.pdf

[15] J. F. Gieras, Permanent magnet motor technology : design and applications,

3rd ed. Boca Raton, Fla.: CRC Press, 2010.

[16] Y. B. Apatya, A. Subiantoro, and F. Yusivar, “Design and prototyping of 3-phase

BLDC motor,” QiR 2017 - 2017 15th International Conference on Quality in Re-
search (QiR): International Symposium on Electrical and Computer Engineering,
vol. 2017-Decem, pp. 209–214, 2017.

[17] F. Meier, Permanent-Magnet Synchronous Machines with Non-Overlapping

Concentrated Windings for, 2008. [Online]. Available: http://www.ee.kth.se/
php/modules/publications/reports/2008/TRITA-EE 2008 041.pdf

[18] A. Vagati, G. Pellegrino, and P. Guglielmi, “Comparison between SPM and IPM
motor drives for EV application,” 19th International Conference on Electrical
Machines, ICEM 2010, 2010.

[19] K. T. Chau, C. C. Chan, and C. Liu, “Overview of permanent-magnet brushless

drives for electric and hybrid electric vehicles,” IEEE Transactions on Industrial
Electronics, vol. 55, no. 6, pp. 2246–2257, 2008.

[20] J. B. Bartolo, H. Zhang, D. Gerada, L. De Lillo, and C. Gerada, “High speed elec-
trical generators, application, materials and design,” Proceedings - 2013 IEEE
Workshop on Electrical Machines Design, Control and Diagnosis, WEMDCD
2013, pp. 47–59, 2013.

[21] Y. Jeong, K. Kim, Y. Kim, B. Park, and S. Jung, “Design Characteristics of

PMa-SynRM and Numerical Analysis,” pp. 164–170, 2012.

[22] A. El-Refaie, “Motors / Generators for Traction / Propulsion Applications,”

IEEE Vehicular Technology Magazine, no. march, pp. 90–99, 2013.

[23] R. Cao, C. Mi, and M. Cheng, “Quantitative comparison of flux-switching

permanent-magnet motors with interior permanent magnet motor for EV, HEV,
and PHEV applications,” IEEE Transactions on Magnetics, vol. 48, no. 8, pp.
2374–2384, 2012.

126
[24] A. M. El-Refaie, J. P. Alexander, S. Galioto, P. B. Reddy, K. K. Huh, P. De Bock,
and X. Shen, “Advanced high-power-density interior permanent magnet motor
for traction applications,” IEEE Transactions on Industry Applications, vol. 50,
no. 5, pp. 3235–3248, 2014.

[25] J. K. Tangudu and T. M. Jahns, “Comparison of interior PM machines with

concentrated and distributed stator windings for traction applications,” 2011
IEEE Vehicle Power and Propulsion Conference, VPPC 2011, 2011.

[26] M. Norhisam, A. Nazifah, I. Aris, H. Wakiwaka, and M. Nirei, “Effect of mag-

net size on torque characteristic of three phase permanent magnet brushless DC
motor,” Proceeding, 2010 IEEE Student Conference on Research and Develop-
ment - Engineering: Innovation and Beyond, SCOReD 2010, no. SCOReD, pp.
293–296, 2010.

[27] M. Jalalifar, A. F. Payam, B. Mirzaeian, and S. M. Nezhad, “Dynamic mod-

eling and simulation of an induction motor with adaptive backstepping design
of an input-output feedback linearization controller in series hybrid electric ve-
hicle,” 2006 International Conference on Power Electronics, Drives and Energy
Systems, PEDES ’06, 2006.

[28] W. Xu, J. Zhu, Y. Guo, S. Wang, Y. Wang, and Z. Shi, “Survey on electrical
machines in electrical vehicles,” 2009 International Conference on Applied Su-
perconductivity and Electromagnetic Devices, ASEMD 2009, no. c, pp. 167–170,
2009.

[29] D. Gerada, D. Borg-Bartolo, A. Mebarki, C. Micallef, N. L. Brown, and C. Ger-

ada, “Electrical machines for high speed applications with a wide constant-power
region requirement,” 2011 International Conference on Electrical Machines and
Systems, ICEMS 2011, pp. 3–8, 2011.

[30] K. Nakamura, K. Murota, and O. Ichinokura, “Characteristics of a novel switched

reluctance motor having permanent magnets between the stator pole-tips,” 2007
European Conference on Power Electronics and Applications, EPE, pp. 10–14,
2007.

[31] J. B. Bartolo and C. Gerada, “The electromagnetic design of a high speed, 45kW,
switched reluctance machine having a novel rotor geometry for aerospace appli-
cation,” Proceedings - 2014 International Conference on Electrical Machines,
ICEM 2014, pp. 2513–2519, 2014.

[32] K. Watanabe, S. Aida, A. Komatsuzaki, and I. Miki, “Driving force charac-

teristics of 40kW switched reluctance motor for electric vehicle,” Proceeding of
International Conference on Electrical Machines and Systems, ICEMS 2007, pp.
1894–1898, 2007.

127
[33] B. Bilgin, A. Emadi, S. Member, and M. Krishnamurthy, “Comprehensive Eval-
uation of the Dynamic Performance of a 6 / 10 SRM for Traction Application in
PHEVs,” vol. 60, no. 7, pp. 2564–2575, 2013.

[34] W. Wang and B. Fahimi, “Comparative study of electric drives for EV/HEV
propulsion system,” Electrical Systems for Aircraft, Railway and Ship Propul-
sion, ESARS, 2012.

[35] P. Giangrande, D. Ronchetto, G. Pellegrino, F. Cupertino, C. Gerada, and

M. Sumner, “Hybrid sensorless control of axial flux permanent magnet motor
drives, including zero speed,” Proceedings of the 2011 14th European Conference
on Power Electronics and Applications, EPE 2011, pp. 1–8, 2011.

[36] C. Versèle, Z. De Grève, F. Vallée, R. Hanuise, O. Deblecker, M. Delhaye, and

J. Lobry, “Analytical design of an axial flux permanent magnet in-wheel syn-
chronous motor for electric vehicle,” 2009 13th European Conference on Power
Electronics and Applications, EPE ’09, no. Im, 2009.

[37] T. Arakawa, M. Takemoto, S. Ogasawara, K. Inoue, O. Ozaki, H. Hojo, and

H. Mitani, “Examination of an interior permanent magnet type axial gap motor
for the hybrid electric vehicle,” IEEE Transactions on Magnetics, vol. 47, no. 10,
pp. 3602–3605, 2011.

[38] K. Lu and E. Ritchie, “Preliminary comparison study of drive motor for electric
vehicle application,” ICEMS 2001 - Proceedings of the 5th International Confer-
ence on Electrical Machines and Systems, vol. 2, pp. 995–998, 2001.

[39] M. Gärtner, P. Seibold, and N. Parspour, “Laminated circumferential transverse

flux machines - Lamination concept and applicability to electrical vehicles,” 2011
IEEE International Electric Machines and Drives Conference, IEMDC 2011, pp.
831–837, 2011.

[40] E. Padurariu, L. E. Somesan, I. A. Viorel, and L. Szabo, “Large power permanent

magnet transverse flux motor, steady-state and dynamic behavior,” Proceedings
of 9th International Conference, ELEKTRO 2012, pp. 221–224, 2012.

[41] F.-s. P.-m. Machine, T. Raminosoa, C. Gerada, and M. Galea, “Design Consid-
erations for a Fault-Tolerant,” Design, vol. 58, no. 7, pp. 2818–2825, 2011.

[42] W. Hua, M. Cheng, and G. Zhang, “A novel hybrid excitation flux-switching

motor for hybrid vehicles,” IEEE Transactions on Magnetics, vol. 45, no. 10, pp.
4728–4731, 2009.

[43] R. Cao, C. Mi, and M. Cheng, “Quantitative comparison of flux-switching

permanent-magnet motors with interior permanent magnet motor for EV, HEV,
and PHEV applications,” IEEE Transactions on Magnetics, vol. 48, no. 8, pp.
2374–2384, 2012.

128
[44] T. Raminosoa, A. M. EL-Refaie, D. Pan, K. K. Huh, J. P. Alexander, K. Grace,
S. Grubic, S. Galioto, P. B. Reddy, and X. Shen, “Reduced Rare-Earth Flux-
Switching Machines for Traction Applications,” IEEE Transactions on Industry
Applications, vol. 51, no. 4, pp. 2959–2971, 2015.

[45] F. N. Jurca and C. Martis, “Analysis of outer rotor synchronous reluctance

motor for low-speed applications,” International Conference on Electical Drives
and Power Electronics, vol. 2017-Octob, pp. 242–247, 2017.

[46] M. Obata, S. Morimoto, M. Sanada, and Y. Inoue, “Characteristic of

PMASynRM with ferrite magnets for EV/HEV applications,” ICEMS 2012 -
Proceedings: 15th International Conference on Electrical Machines and Systems,
2012.

[47] J. Malan, M. J. Kamper, and P. N. Williams, “Reluctance synchronous machine

drive for hybrid electric vehicle,” IEEE International Symposium on Industrial
Electronics, vol. 2, pp. 367–372, 1998.

[48] M. M. Ansari, W. A. Cronje, and A. Meyer, “Evaluation of a reluctance syn-

chronous motor: For use in an Electric Mine Shuttle Vehicle (EMSV),” 2012
IEEE International Electric Vehicle Conference, IEVC 2012, 2012.

[49] F. Cupertino, G. M. Pellegrino, E. Armando, and C. Gerada, “A SyR and IPM

machine design methodology assisted by optimization algorithms,” 2012 IEEE
Energy Conversion Congress and Exposition, ECCE 2012, pp. 3686–3691, 2012.

[50] G. Pellegrino, F. Cupertino, and C. Gerada, “Automatic Design of Synchronous

Reluctance Motors Focusing on Barrier Shape Optimization,” IEEE Transac-
tions on Industry Applications, vol. 51, no. 2, pp. 1465–1474, 2015.

[51] “External rotor motor basics: Design and applica-

tions.” [Online]. Available: https://www.motioncontroltips.com/
external-rotor-motor-basics-design-applications/

[52] T. Y. Lee, M. K. Seo, Y. J. Kim, and S. Y. Jung, “Motor Design and Characteris-
tics Comparison of Outer-Rotor-Type BLDC Motor and BLAC Motor Based on
Numerical Analysis,” IEEE Transactions on Applied Superconductivity, vol. 26,
no. 4, pp. 5–10, 2016.

[53] P. D. I. Milano, “DESIGN AND OPTIMIZATION OF AN OUTER ROTOR

PERMANENT MAGNETS BRUSHLESS MOTOR DESIGN AND OPTIMIZA-
TION OF AN OUTER ROTOR PERMANENT,” 2018.

[54] A. M. EL-Refaie, “Fractional-slot concentrated-windings synchronous permanent

magnet machines: Opportunities and challenges,” IEEE Transactions on Indus-
trial Electronics, vol. 57, no. 1, pp. 107–121, 2010.

129
[55] A. M. El-Refaie, T. M. Jahns, and D. W. Novotny, “Analysis of surface perma-
nent magnet machines with fractional-slot concentrated windings,” IEEE Trans-
actions on Energy Conversion, vol. 21, no. 1, pp. 34–43, 2006.

[56] J. Cros and P. Viarouge, “Synthesis of high performance PM motors with concen-
trated windings,” IEEE International Electric Machines and Drives Conference,
IEMDC 1999 - Proceedings, vol. 17, no. 2, pp. 725–727, 1999.

[57] F. Magnussen and C. Sadarangani, “Winding Factors and,” 2006.

[58] S. Skaar, O. Krovel, and R. Nilssen, “Distribution, coil-span and winding

factors for PM machines with concentrated windings,” XVII International
Conference on Electrical Machines, ICEM 2006, p. 346, 2006. [Online]. Available:
http://www.elkraft.ntnu.no/eno/Papers2006/icem-skaar-krovel-nilssen06.pdf

[59] A. D. Gerlando, G. M. Foglia, R. Perini, M. Ubaldini, E. Politecnico, D. O. Ayer,

and C. O. W. Indings, “Synchronous Machines with Concentrated Armature
Windings,” Leonardo, pp. 1165–1172, 2005.

[60] J. Cros, P. Viarouge, and C. Gelinas, “Design of PM brushless motors using iron-
resin composites for automotive applications,” Conference Record - IAS Annual
Meeting (IEEE Industry Applications Society), vol. 1, pp. 5–11, 1998.

[61] A. G. Jack, B. C. Mecrow, P. G. Dickinson, D. Stephenson, J. S. Burdess,

N. Fawcett, and J. T. Evans, “Permanent-magnet machines with powdered iron
cores and prepressed windings,” IEEE Transactions on Industry Applications,
vol. 36, no. 4, pp. 1077–1084, 2000.

[62] N. Bianchi and S. Bolognani, “Design techniques for reducing the cogging torque
in surface-mounted PM motors,” IEEE Transactions on Industry Applications,
vol. 38, no. 5, pp. 1259–1265, 2002.

[63] A. G. Jack, B. C. Mecrow, and J. A. Haylock, “A comparative study of perma-

nent magnet and switched reluctance motors for high-performance fault-tolerant
applications,” IEEE Transactions on Industry Applications, vol. 32, no. 4, pp.
889–895, 1996.

[64] N. Bianchi, M. D. Pré, G. Grezzani, and S. Bolognani, “Design considerations on

fractional - Slot fault - Tolerant synchronous motors,” 2005 IEEE International
Conference on Electric Machines and Drives, pp. 902–909, 2005.

[65] S. O. Kwon, S. I. Kim, P. Zhang, and J. P. Hong, “Performance comparison

of IPMSM with distributed and concentrated windings,” Conference Record -
IAS Annual Meeting (IEEE Industry Applications Society), vol. 4, no. c, pp.
1984–1988, 2006.

[66] J. Goss, D. Staton, R. Wrobel, and P. Mellor, “Brushless AC interior-permanent

magnet motor design: Comparison of slot/pole combinations and distributed vs.

130
 concentrated windings,” 2013 IEEE Energy Conversion Congress and Exposi-
tion, ECCE 2013, pp. 1213–1219, 2013.
[67] L. Chong, R. Dutta, and M. F. Rahman, “Performance comparison between con-
centrated and distributed wound IPM machines used for field weakening applica-
tions,” International Aegean Conference on Electrical Machines and Power Elec-
tronics, ACEMP 2011 and Electromotion 2011 Joint Conference, no. September,
pp. 616–619, 2011.
[68] M. Galea, C. Gerada, T. Raminosoa, and P. Wheeler, “A thermal improvement
technique for the phase windings of electrical machines,” IEEE Transactions on
Industry Applications, vol. 48, no. 1, pp. 79–87, 2012.
[69] D. C. Hanselman, Brushless Motor Design, 1994.
[70] N. Bianchi, H. Mahmoud, and S. Bolognani, “Fast synthesis of permanent mag-
net assisted synchronous reluctance motors,” vol. 10, pp. 312–318, 2016.
[71] C. López-Torres, A. Garcia-Espinosa, and J.-R. Riba, “Reliable Design of
PMaSynRM,” New Trends in Electrical Vehicle Powertrains, 2019.
[72] M. G. M. G. Say, The performance and design of alternating current machines :
transformers, three-phase induction motors and synchronous machines, 3rd ed.
London: Pitman Paperbacks, 1968.
[73] J. H. J. H. Walker, Large synchronous machines : design, manufacture, and
operation, ser. Monographs in electrical and electronic engineering ; [14]. Oxford:
Clarendon Press, 1981.
[74] R. A. Inte, F. N. Jurca, and C. Martis, “Design And Analysis Of Outer Rotor Per-
manent Magnet Assisted Synchronous Reluctance Machine With Concentrated
Winding For Small Electric Propulsion,” pp. 14–19, 2019.
[75] P.-m. Motors, Y. Kano, and N. Matsui, “A Design Approach for Direct-Drive,”
vol. 44, no. 2, pp. 543–554, 2008.
[76] F. Libert and J. Soulard, “Investigation on pole-slot combinations
for permanent-magnet machines with concentrated windings,” Interna-
tional Conference on Electrical Machines ( . . . , no. June, pp. 5–8, 2004.
[Online]. Available: http://www.ee.kth.se/php/modules/publications/reports/
2004/IR-EE-EME 2004 005.pdf
[77] N. Bianchi and M. Dai Pré, “Use of the star of slots in designing fractional-slot
single-layer synchronous motors,” IEE Proceedings: Electric Power Applications,
vol. 153, no. 3, pp. 459–466, 2006.
[78] K. F. Rasmussen, J. H. Davies, T. J. Miller, M. L. McGilp, and M. Olaru, “Ana-
lytical and numerical computation of air-gap magnetic fields in brushless motors
with surface permanent magnets,” IEEE Transactions on Industry Applications,
vol. 36, no. 6, pp. 1547–1554, 11 2000.

131