Kai Wang

Zi-Qiang Zhu

Third Harmonic
Utilization in
Permanent
Magnet
Machines
Third Harmonic Utilization in Permanent Magnet
Machines
Kai Wang Zi-Qiang Zhu
•

Third Harmonic Utilization

in Permanent Magnet
Machines

123
Kai Wang Zi-Qiang Zhu
Department of Electrical Engineering Department of Electronic
Nanjing University of Aeronautics and Electrical Engineering
and Astronautics The University of Sheffield
Nanjing Sheffield
China UK

ISBN 978-981-13-0628-0 ISBN 978-981-13-0629-7 (eBook)

https://doi.org/10.1007/978-981-13-0629-7

Library of Congress Control Number: 2018943731

© Springer Nature Singapore Pte Ltd. 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Contents

1 General Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . 1
1.1 Pole Shaping Techniques . . . . . . . . . . . . . . . . . . .
. . . . . . . . . 5
1.1.1 Pulse Width Modulation . . . . . . . . . . . . . .
. . . . . . . . . 5
1.1.2 Modular Pole . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . 6
1.1.3 Halbach . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . 7
1.1.4 Pole Shaping . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . 8
1.1.5 Sinusoidal Plus 3rd Harmonic Shaped Rotor Shape . . . 10
1.2 Outline of the Book . . . . . . . . . . . . . . . . . . . ........... . . 10
2 Torque Enhancement of Three Phase Surface-Mounted
Permanent Magnet Machine Using 3rd Order Harmonic . . . . . . . . 13
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 SPM Machines with Various PM Shapes . . . . . . . . . . . . . . . . . 15
2.3 PM Shape with Optimal 3rd Harmonic and FE Validation . . . . 17
2.3.1 Sinusoidal Shaping PM . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3.2 PM Shape with Optimal Amplitude of 3rd
Harmonic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.3.3 FE Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3.4 Influence of PM Edge Thickness . . . . . . . . . . . . . . . . . 24
2.4 Finite Element Analysis of Electromagnetic Performance . . . . . 27
2.4.1 Open-Circuit Flux Density Distribution
and Back-EMFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.4.2 Torque Characteristics . . . . . . . . . . . . . . . . . . . . . . . . 27
2.5 Experimental Verification and Discussions . . . . . . . . . . . . . . . . 32
2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

v
vi Contents

3 Average Torque Improvement of Three Phase Interior

Permanent-Magnet Machine Using 3rd Harmonic in Rotor
Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2 Rotor Configurations of IPM Machines . . . . . . . . . . . . . . . . . . 41
3.3 Rotor Shaping with 3rd Harmonic . . . . . . . . . . . . . . . . . . . . . . 41
3.3.1 ICS Shaped Rotor . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3.2 Rotor Shape with Different Amplitude of 3rd
Harmonic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 45
3.4 Simplified Analytical Analysis of Average Torque
Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 47
3.5 Finite Element Analysis of Electromagnetic Performance ..... 50
3.5.1 Open-Circuit Flux Density Distribution
and Back-EMFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.5.2 Torque Characteristics . . . . . . . . . . . . . . . . . . . . . . . . 53
3.6 Experimental Verification and Discussions . . . . . . . . . . . . . . . . 57
3.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4 Third Order Harmonic Utilization in In-Wheel Machines
to Improve Output Torque . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2 In-Wheel Machines with Various PM-Shaped Rotor . . . . . . . . . 66
4.3 PM Shape with Optimal 3rd Harmonic . . . . . . . . . . . . . . . . . . 68
4.3.1 Influence of PM Edge Thickness . . . . . . . . . . . . . . . . . 68
4.4 Finite Element Analysis of Electromagnetic Performance . . . . . 70
4.4.1 Open-Circuit Flux Density Distribution
and Back-EMFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.4.2 Torque Characteristics . . . . . . . . . . . . . . . . . . . . . . . . 73
4.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5 Influence of Airgap Flux Density Waveform on Optimal
Split Ratio and Torque Density of SPM Machines . . . . . . . ...... 79
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...... 79
5.2 General Torque Density Equation . . . . . . . . . . . . . . . . ...... 81
5.3 Derivation of Optimal Split Ratio for Maximum Torque
Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.3.1 Calculation of Slot Area . . . . . . . . . . . . . . . . . . . . . . . 83
5.3.2 Optimal Split Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.3.3 TD Comparison Under Optimal Split Ratio . . . . . . . . . 86
5.4 FE Analysis of Optimal Split Ratio and Torque . . . . . . . . . . . . 91
5.5 Experimental Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Contents vii

5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6 Investigation of Stator Flux Density and Iron Loss in 3rd Order
Harmonic Shaped Surface-Mounted Permanent Magnet
Machines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 99
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 99
6.2 SPM Machines with Various PM Shapes . . . . . . . . . . . . . . . . . 101
6.3 Analytically Predicted Influence of 3rd Harmonic on Staror
Iron Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
6.3.1 Flux Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.3.2 Iron Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
6.4 Finite Element Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
6.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
7 Analysis of Cogging Torque in Surface-Mounted Permanent
Magnet Machines with Shaped Magnets . . . . . . . . . . . . . . . . . . . . . 121
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
7.2 Analytical Prediction of Cogging Torque Based on Energy
Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
7.2.1 Simplified Airgap Permanence Function . . . . . . . . . . . 124
7.2.2 Cogging Torque for Sinusoidal Airgap Flux
Density Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . 126
7.2.3 Cogging Torque for Sine+3rd Airgap Flux
Density Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . 127
7.2.4 Cogging Torque for Sine and Sine+3rd Airgap
Flux Density Distributions . . . . . . . . . . . . . . . . . . . . . 130
7.3 Three-Phase SPM Machines with Shaped PMs . . . . . . . . . . . . . 131
7.4 FE Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
7.5 Experimental Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
7.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
8 Influence of Harmonics on Radial Force and Vibration
of Surface-Mounted Permanent Magnet Machines . . . . . . . . . . . . . 143
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
8.2 12-Slot/10-Pole PM Machines with Shaped Magnets . . . . . . . . 145
8.3 Radial Force Density Distributions . . . . . . . . . . . . . . . . . . . . . . 146
8.3.1 Radial Force Density Under No Load Condition . . . . . 146
8.3.2 Radial Force Density Under Rated Load . . . . . . . . . . . 148
8.4 Mode Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
8.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
viii Contents

9 Multi-objective Optimization of Surface-Mounted Permanent

Magnet Machine with Third Harmonic Shaped Rotor . . . . . . . . . . 159
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
9.2 Topology and Parametric Model of SPM Machine . . . . . . . . . . 161
9.3 Multi-objective Optimization Procedure . . . . . . . . . . . . . . . . . . 163
9.3.1 Flowchart of Multi-objective Optimization . . . . . . . . . . 163
9.3.2 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 165
9.3.3 Multi-objective Optimization . . . . . . . . . . . . . . . . . . . . 166
9.4 Electromagnetic Performance Comparison . . . . . . . . . . . . . . . . 169
9.4.1 Open-Circuit Flux Density Distributions
and Back-EMFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
9.4.2 Torque Characteristics . . . . . . . . . . . . . . . . . . . . . . . . 173
9.4.3 Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
9.5 Experimental Validation and Discussion . . . . . . . . . . . . . . . . . . 175
9.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
10 Using Third Harmonic for Shape Optimization of Flux Density
Distribution in Slotless Linear Permanent-Magnet Machine . . . . . . 181
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
10.2 PMLMS with Various PM Shapes . . . . . . . . . . . . . . . . . . . . . . 183
10.3 PM Shape Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
10.3.1 Simplified Estimation of PM Shapes . . . . . . . . . . . . . . 184
10.3.2 Optimization of Sinusoidal PM Shapes . . . . . . . . . . . . 186
10.3.3 PM Shape with Third Harmonic . . . . . . . . . . . . . . . . . 187
10.4 Performance Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
10.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
11 Design and Analysis of Permanent Magnet Linear Synchronous
Machine with Third Harmonic Shaping Mover . . . . . . . . . . . . . . . 197
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
11.2 PMLSM with Different Shaping Movers . . . . . . . . . . . . . . . . . 199
11.3 Optimal Design of PM Edge Thickness and Pole Arc
Coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
11.4 Electromagnetic Performance Comparision . . . . . . . . . . . . . . . . 203
11.4.1 Open-Circuit Flux Density Distributions
and Back-EMFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
11.4.2 Thrust Force Characteristics . . . . . . . . . . . . . . . . . . . . 205
11.4.3 Losses and Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . 208
11.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
Abstract

This book presents various permanent magnet (PM) synchronous machines with a
third-harmonic-shaped rotor and investigates the torque (or thrust) improvement
and the influence of the third harmonic on the electromagnetic performance, such as
split ratio, cogging torque (force), stator iron loss, radial magnetic force.
The torque (thrust) density can be improved, while the torque (thrust) ripple
remains low by PM shaping utilizing third harmonic. In this book, the analytical
expression of output torque (thrust) is derived for various PM machines with third
harmonic. All the results show that the optimal third harmonic is always 1/6 of the
fundamental one and the torque (thrust) improvements for three-phase
surface-mounted PM (SPM) inner rotor and outer rotor machines, interior PM
(IPM) machine, slotless and slotted linear machines are >10%, 10%, 6%, 14.5%,
and 17.2%, respectively.
An unified analytical model for optimizing the split ratios of SPM machines
considering various airgap flux density distributions, i.e., square (also referred to as
conventional), sinusoidal (sine), and sinusoidal with third harmonic (sine+3rd),
considering both overlapping windings and nonoverlapping windings. Besides,
torque densities under the optimal split ratios are compared for the conventional,
sine, and sine+3rd. It is found that harmonics in airgap flux density will reduce the
optimal split ratio and their influence is stronger for nonoverlapping windings
compared with overlapping windings. The sine+3rd overlapping winding machines
show competence in torque density with higher value of stator slot Ns especially
under high flux density ratio, reaching about 99% of the torque of the conventional
one. For sine+3rd nonoverlapping winding machines whose slot number Ns and
pole number 2p differed by one, the torque density is larger than the conventional
one. The sine+3rd is promising in torque density for both overlapping and
nonoverlapping winding machines.
The multi-objective optimization based on multi-level optimization method is
implemented to achieve even lower torque ripple, higher power density, and effi-
ciency for the sine+3rd-shaped SPM machine. The Taguchi method is employed in

ix
x Abstract

the sensitivity analysis of design variables to divide design variables into sensitive
level and nonsensitive level. Then, the response surface (RS) method and sequential
nonlinear programming (SNP) method are combined to optimize the sensitive
design variables of the SPM machine. After optimization, the electromagnetic
performances of the initial and optimal SPM machine are contrastively analyzed
based on finite element method.
The third harmonic can be injected into the sinusoidal-shaped PMs to improve
the average torque in three-phase SPM machines without deteriorating the torque
ripple caused by back-EMF harmonics. However, the influence of the injected third
harmonic on the cogging torque is uncertain. Analytical expressions are derived for
the cogging torque in SPM machines equipped with sinusoidal shaped and sinu-
soidal shaped with third harmonic injected PMs, which are optimized to obtain
sinusoidal and sinusoidal with third harmonic airgap flux density distributions,
respectively. It is found that the cogging torque of the SPM machines with
sine-shaped PMs is negligibly small. The sine+3rd-shaped PM can increase the
amplitude of the cogging torque for the specific three-phase SPM machines with the
number of slots per pole per phase (spp) equal to 0.25, 0.5, and 1. Finite element
analyses are carried out on 6-slot/8-pole (spp = 0.25), 12-slot/8-pole (spp = 0.5),
24-slot/8-pole (spp = 1), and 12-slot/10-pole (spp = 0.4) SPM machines with
shaped PMs to verify the analyses.
The third harmonic can be utilized in the contour of the permanent magnets to
improve the output torque of SPM machines. However, the introduced third har-
monic airgap flux density will result in variation of flux density distribution in the
stator and influence the stator iron losses as well. The stator is decomposed into
three parts, i.e., tooth tips, tooth body and stator yoke, to analytically investigate the
influence of additional third harmonic airgap flux density on the stator flux density
and stator iron losses. The sine, sine+3rd, and 1.15 (sine+3rd)-shaped PM rotors are
employed for discussion, where the 1.15 (sine+3rd) is commonly utilized to
enhance average torque and the sine+3rd is proposed for comparison. It is found
that the classical eddy current loss and hysteresis loss are larger for 1.15 (sine+3rd)-
shaped PM machines compared to one with sine-shaped PM rotor. For the sine+3rd
machines, while the classical eddy current loss in stator and the hysteresis loss in
tooth body and stator yoke are larger, its value in tooth tips is dependent on the slot
and pole number combinations.
Electromagnetic vibrations of 12-slot/10-pole SPM machines with both
sine-shaped and sine+3rd-shaped PMs are investigated. Radial force density
expressions are calculated by Maxwell stress tensor method under both no-load and
on-load conditions. The magnitude of the lowest mode of the machines with
sine-shaped and sine+3rd-shaped PMs, which is the second mode for
12-slot/10-pole ones, is investigated. Since the amplitude of the fifth (fundamental
in electrical angle) airgap flux density is increased for sine+3rd-shaped PMs
compared with sine-shaped PMs, the second mode due to the interaction between
the fifth PM field and seventh harmonic (either from slotting or armature field or
Abstract xi

both) is larger. However, it is found the 15th PM field (third harmonic in electrical
angle) in sine+3rd-shaped PMs can interact with 17th field to generate a negative
amplitude of second mode to cancel part of the increase due to the fifth PM field,
though the total magnitude of second mode is about 8.8% larger compared with the
sine-shaped PMs under both no-load and rated load conditions.
Chapter 1
General Introduction

With advancement of high energy permanent magnet material, improvement of

sensor technology, advent of power electronics and modern control theories, per-
manent magnet (PM) brushless machines obtain increasing popularity over past
four decades since they offer several key features, such as high torque density, high
efficiency, and excellent dynamic performance. As a consequence, numerous PM
machine configurations emerge and are extensively investigated. One of them is the
radial field PM brushless machines, which can be categorized in terms of rotor PMs
and stator winding, as shown in Fig. 1.1.
The rotor structures are classified into four categories according to the placement
of PMs in the rotor, i.e. the SPM machine, Fig. 1.2a, the inset PM machine,
Fig. 1.2b, and the interior PM (IPM) machine, Fig. 1.2c, and the spoke PM
machine, Fig. 1.2d. The SPM machine has a large effective airgap and thus the
inductance and armature reaction are relatively low. The demagnetization with-
standing capability is relative low as the magnets directly face the armature reaction
field. In addition, rotor sleeves such as carbon fiber or Titanium are often required
for high-speed application. On the contrary, IPM machine have PMs buried in the
rotor iron and can have more robust rotor mechanical structure than SPM machine,
whilst the PMs are embedded on the rotor surface for an inset PM machine.
However, the leakage flux between magnet pole and iron pole is significant. As a
consequence, the output torque is reduced. The spoke PM machines can increase
the airgap flux density by flux focusing and thus the output torque. However, the
design and manufacture is relatively complicated. For high performance applica-
tions, such as electric power steering systems, servo motors, and direct-drive wind
power generators, it is much more challenging, since both high torque density and
low torque ripple are required. None of the rotor structure can have sinusoidal
airgap flux density distribution without special rotor design, such as rotor shaping
techniques. Therefore, the torque ripple is usually high. The torque ripples can be
reduced by either machine designs or control methods. PM shaping methods, which
are one of the simplest and most effective torque ripple reduction methods, have

© Springer Nature Singapore Pte Ltd. 2019 1

K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_1
2 1 General Introduction

Fig. 1.1 Radial field PM brushless machines

(a) SPM (b) IPM

(c) Inset (d) Spoke

Fig. 1.2 PM machines with different rotor configurations
1 General Introduction 3

been widely employed. These rotor shaping techniques will be reviewed in

Sect. 1.1.
In the stator part, the windings for a radial airgap field, inner rotor and slotted
stator machine can be classified into two categories, viz. non-overlapping and
overlapping windings, as illustrated in Fig. 1.3. The features of these two winding
configurations are compared in Table 1.1.

(a) Non-overlapping (b) Overlapping

Fig. 1.3 Stator winding

Table 1.1 Comparison of non-overlapping and overlapping windings

Non-overlapping winding Overlapping winding
Automatic manufacture Easy Difficult
End-winding Short Long
Winding factor Usually <1 ¼1 or <1
Terminal isolation Can be well isolated Not isolated
between phases (depending on design)
Cogging torque Usually low as fractional-slot Usually high as integral-slot
stator is used stator is used
Space distribution of Square-wave Can be very close to
stator MMF sine-wave
MMF harmonics High Low
Rotor losses High Low
Winding inductance High Low
Fault tolerance Yes No
Flux-weakening High Low for SPM but High for
capability IPM
Slot fill factor Usually high Usually low
Modularity Yes No
4 1 General Introduction

(a) Double-layer (b) Single-layer

Fig. 1.4 Non-overlapping winding configuration

As shown in Table 1.1, due to its relative merits compared with overlapping
winding, the non-overlapping fractional-slot winding, viz. the number of slots per
pole per phase is not an integer number and the winding pitch is unity slot pitch, is
often employed, as shown in Fig. 1.3a. Commonly, there are two winding con-
figurations for non-overlapping fractional-slot winding, viz. double- and
single-layer windings, as shown in Fig. 1.4a, b. The double-layer means that there
are two coil sides within each slot, while the single-layer winding has one coil sides
within each slot. Their relative merits have been compared in Table 1.2.
The SPM and IPM machines with non-overlapping windings will be employed
in the investigation and this book will focus on the rotor topology to have high
torque and low torque ripple.

Table 1.2 Comparison of double- and single-layer windings

Double-layer Single-layer
Winding factor Low High
Terminal isolation between phases Poor Better
MMF harmonics Lower Higher
Rotor losses Lower Higher
Mutual winding inductance High Low
Fault tolerance No Yes
1.1 Pole Shaping Techniques 5

1.1 Pole Shaping Techniques

1.1.1 Pulse Width Modulation

“Pulse width modulation technique” and “pulse amplitude modulation technique”,

which are widely employed in power electronics for achieving a sinusoidal current
or voltage, are proposed in SPM machines. As shown in Fig. 1.5, in one pole, the
PMs are segmented and their dimensions and positions are determined by the pulse
width modulation technique or pulse amplitude modulation technique similar to
PWM signals used to switch power electronic devices. In this way, the specific
harmonics in airgap flux density can be eliminated. Recently, this technique is
employed in IPM machines to segment the rotor steel similar to segmenting PMs, as
shown in Fig. 1.6. However, to reduce the harmonics of the airgap flux density, the

(b) Arrangement of PMs in conventional

(a) Conventional SPM rotor
rotor

(d) Arrangement of segmented PMs in one

(c) PWM segmented SPM rotor
pole

Fig. 1.5 PM arrangement by pulse width modulation technique

6 1 General Introduction

(a) Conventional IPM rotor (b) Arrangement of steel in conventional IPM rotor

(c) PWM segmented IPM rotor (d) Arrangement of segmented steel in one pole

Fig. 1.6 Iron pole arrangement by pulse width modulation technique

fundamental airgap flux density and electromagnetic torque are also reduced. In
addition, since the block magnets are arranged in this configuration, it is hard to
implement because putting the segments of a divided PM pole with one direction of
magnetization adjacent to each other in a PM machine with short pole pitch is very
difficult.

1.1.2 Modular Pole

The PMs with different thickness in one pole is proposed in SPM machines, which
is employed to reduce the harmonics in the back-EMF. The desired PM thickness is
obtained by simplified analytical method. This would increase the manufacturing
cost of PMs. In order to reduce such cost, modular PM pole technique, which uses
different types of PM material with the same thickness as shown in Fig. 1.7a or the
1.1 Pole Shaping Techniques 7

High remenence Br

Low remenence Br

(a) SPM rotor with different magnet material (b) Arrangement of modular pole [ISF09]

(c) SPM rotor with different PM thickness (d) Arrangement of modular pole

Fig. 1.7 Modular pole technique

same PM material with different thickness as shown in Fig. 1.7b in one pole, is
employed to reduce harmonics in the airgap flux density. The calculated results
have shown that it is an effective way of reducing the total harmonic distortion
(THD) of phase back-EMF. However, these structures will still increase the com-
plexity and the manufacturing cost. Meanwhile, to reduce the harmonics of the
airgap flux density, the fundamental airgap flux density and electromagnetic torque
are reduced as well. In addition, the low quality PM material or thin PM material at
both sides of one pole would increase the risk of PM demagnetization.

1.1.3 Halbach

The sinusoidal airgap flux density can be obtained by Halbach magnetization,

which was firstly proposed by Halbach, as shown in Fig. 1.8a. The ideal Halbach
8 1 General Introduction

(a) Conventional Halbach (b) Modular Halbach

Fig. 1.8 Halbach techniques

magnetized rotor can be achieved by bonded neodymium-iron-boron (NdFeB)

magnets. However, the low airgap flux density is limited by the bonded NdFeB.
Hence, it is achieved by segmenting magnet with different magnetization. In order
to further improve their performance, the Halbach topology combined the modular
pole topology, which has been illustrated in previous Sect. 1.1.2, is proposed to
have high PM volume utilization ratio, as shown in Fig. 1.8b. However, this
method needs many magnet pieces with different sizes and directions of magneti-
zation. Thus, it uses excessive amount of PM material, leading to a high production
cost.
In all the mentioned techniques, although these effectively reduce the harmonics
in the flux density of the PM pole as well as the machine torque ripple, all have the
similar issues which are very hard to implement because placing the segments of a
divided PM pole in one direction of magnetization vectors adjacent to each other in
a PM machine with short pole pitch is quite difficult.

1.1.4 Pole Shaping

Magnet arc shaping techniques, which have been used to produce the sinusoidal
airgap flux density, are widely employed. These techniques can be categorized as
sinusoidal shaping, eccentric shaping, and inverse cosine shaping, as shown in
Fig. 1.9. The PMs can be shaped by one sinusoidal function with rotor position or
by one eccentric arc to shape the PM surface. In order to shape the IPM rotor, the
rotor surface should be shaped by one inverse cosine function, which is different
from the SPM machines. Although this method is relatively implemented compared
with modular pole and PWM techniques, these methods still have the problems of
1.1 Pole Shaping Techniques 9

(a) Sinusoidal shaping (b) Sinusoidal shaping

(c) Eccentric shaping (d) Inverse cosine shaping

(e) Inverse cosine shaping (f) Inverse cosine shaping

Fig. 1.9 Pole shaping techniques

reducing the fundamental airgap flux density and average electromagnetic torque,
thereby reducing the power density of the machines.
10 1 General Introduction

1.1.5 Sinusoidal Plus 3rd Harmonic Shaped Rotor Shape

The 3rd harmonic current, which interacts with the 3rd harmonic EMFs, was
injected into the winding to improve the output torque capability in multi-phase
machines. However, in three-phase machines with balanced Y-connected windings,
the 3rd harmonic current cannot be injected. Alternatively, a PM shaping method
for SPM machines by utilizing the 3rd harmonic, which was able to achieve higher
fundamental EMFs than the conventional sine shaping method when the minimum
airgap length and maximum magnet thickness remain the same, as shown in
Fig. 1.10.
However, the output torque is not analytically established for the rotor shape
with 3rd harmonic or the value of 3rd harmonic into magnet/rotor shape is not an
optimal value, as will be described in Chaps. 2 and 3 and the technique is applied to
in-wheel PM machines and linear PM machines, which are carried out in Chaps. 4,
5 and 7, respectively. Since the third harmonic is introduced into the airgap flux
density, its influence on the optimal split ratio, the sensitivity analysis, cogging
torque, stator iron loss, and radial magnetic force should also be considered, which
are conducted in Chaps. 7–10, respectively. Finally, the sensitivity of design
variables is analyzed to give a reference to the manufacturing tolerance in Chap. 11.

1.2 Outline of the Book

The contents of the subsequent chapters may be summarized as follows:

This Chapter reviews the previous work, presents the main objectives and
contributions of this book.
Chapter 2 derives the relationship of output torque and 3rd harmonic into the
rotor shape in three phase SPM machine. The optimal 3rd harmonic for maximizing
the output torque is analytically obtained and verified by both finite element
analyses and experimental results.
Chapter 3 applies the 3rd harmonic shaping technique to the IPM machine and
obtains the optimal 3rd harmonic by finite element methods. The torque

Fig. 1.10 3rd harmonic shaping techniques

1.2 Outline of the Book 11

improvement is analytically derived. The difference between the SPM and IPM
machine utilizing the 3rd harmonic is highlighted.
Chapter 4 presents a permanent magnet (PM) shaping technique with optimal
3rd harmonic to enhance the output torque of in-wheel machines.
Chapter 5 develops an unified analytical model for optimizing the split ratios of
surface-mounted permanent magnet machines considering various airgap flux
density distributions, i.e., square, sinusoidal and sinusoidal with third harmonic,
considering both overlapping windings and nonoverlapping windings.
Chapter 6 investigates the influence of additional third harmonic airgap flux
density on the stator flux density and stator iron losses with consideration of slot/
pole combinations.
Chapter 7 analyzes the cogging torque of SPM machines with Sinusoidal shaped
and Sinusoidal shaped with third harmonic injected PMs considering different slot/
pole number combinations.
Chapter 8 analyzes the electromagnetic vibrations of 12-slot/10-pole SPM
machines with both Sine-shaped and Sine+3rd-shaped PMs. Radial force density
expressions are calculated by Maxwell stress tensor method under both no load and
on load conditions.
Chapter 9 presents a SPM machine with third harmonic shaped rotor for
potential applications in electrical power steering systems. The sensitivity of design
variables is analyzed to give a reference to the manufacturing tolerance and reduce
optimization design space dimension.
Chapter 10 presents a sinusoidal PM shaping technique with third harmonic to
improve the electromagnetic force in linear slotless PM machines.
Chapter 11 proposes a linear PM machines with sinusoidal plus 3rd harmonic
shaped mover to improve thrust force and efficiency without deteriorating thrust
ripple.
Chapter 2
Torque Enhancement of Three Phase
Surface-Mounted Permanent Magnet
Machine Using 3rd Order Harmonic

This chapter presents a permanent magnet (PM) shaping technique with optimal 3rd
harmonic to improve the output torque without deteriorating the torque ripple in
surface-mounted PM (SPM) machines. The optimal value of 3rd harmonic injected
into the sinusoidal PM shape for maximum torque improvement is analytically
derived and confirmed by finite element (FE) analysis. Further, the influence of
magnet edge thickness on the airgap field distribution is investigated and utilized to
compensate the inter-pole flux leakage and curvature effect. It is found that the
optimal 3rd harmonic is 1/6 of the fundamental one. For the SPM machines having
rotors without shaping, Sine shaping, Sine shaping with 3rd harmonic injected, the
electromagnetic performance including the back-EMF waveforms, cogging torque,
average torque and torque ripple are compared. It is demonstrated that the average
torque in the machine of a Sine shaping with an optimal 3rd harmonic injected can
be improved by >9% while the torque ripple remains similar to that of the one with
Sine shaping. Finally, the machines with both conventional (without shaping) and
optimal 3rd harmonic PM rotors are prototyped and measured to validate the
analyses.

2.1 Introduction

Permanent magnet (PM) machines have found wide applications from automotive
to aerospace over the past four decades since they offer several key features, such as
high torque density, high efficiency, and excellent dynamic performance [1–4].
However, for high-performance applications, such as electrical power steering
systems, servo motors, and direct-drive wind power generators, it is much more
challenging, since both high torque density and low torque ripple are required. The
torque ripples can be reduced by either machine designs or control methods [5–18].
Rotor steel or PM shaping methods, which are one of the simplest and most
effective torque ripple reduction methods, have been widely employed [8–13]. The
© Springer Nature Singapore Pte Ltd. 2019 13
K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_2
14 2 Torque Enhancement of Three Phase Surface-Mounted …

inverse cosine-shaped pole shoe shape of salient pole wound-field synchronous

machines is often employed [8–10], where the airgap permeance is a cosine
function with the rotor position. However, in surface-mounted PM
(SPM) machines, the airgap permeance is almost constant due to the relative per-
meability of PMs almost equal to unity and hence the sinusoidal airgap flux density
can be achieved by shaping PMs [11–13]. However, when compared with the
machine without shaping, the fundamental airgap flux density and thus the torque
density are notably decreased as well, as will be shown later in this chapter. This is
also the same for the other torque ripple reduction methods, i.e., to achieve low
torque ripple, the torque density is usually considerably compromised. In [18], it
was shown that for the sake of low torque ripple, it was better to reduce electric
loading and magnetic saturation during machine design, which results in lower
torque density. Different approaches have been proposed to increase the funda-
mental output from inverters to machine designs. Especially, the third harmonic has
gained wide application in pulse width modulation techniques, rectifier, sensorless
operation, and machine design [19–28]. The output voltage can be improved by
pulse width modulation with the third harmonic in inverter [19, 20]. The total
harmonic distortion of input current for rectifier can be reduced with third harmonic
current injection [21]. The third harmonic back-EMF can be employed for sen-
sorless control [22]. The flux patterns can be optimized in the five-phase induction
machine by injection of third harmonic current into the winding [23]. In [23–25],
the third harmonic current, which interacts with the third harmonic EMFs, was
injected into the winding to improve the output torque capability in multi-phase
machines. However, in three-phase machines with balanced Y-connected windings,
the third harmonic current cannot be injected. Alternatively, in [26], it mentioned a
PM shaping method for SPM machines by utilizing the third harmonic, which was
able to achieve higher fundamental EMFs than the conventional Sine shaping
method when the minimum airgap length and maximum magnet thickness remain
the same. However, the optimal third harmonic in PM shaping was not discussed in
[26]. In [27], the optimal amount of third harmonic for PM shaping, which exhibits
the maximum fundamental EMFs with the same minimum airgap length and
maximum magnet thickness, was investigated by finite-element (FE) analyses. In
[28], an inverse cosine-shaped salient pole rotor utilizing third harmonic was
developed to improve the torque density of interior PM machines. However, the
torque of the machine with third harmonic into the magnet shape is not analytically
established and hence the amount of optimal third harmonic into the magnet shape
cannot be analytically determined. The concept of PWM modulation previously
used in the inverters was employed in the PM segmentation to reduce the airgap
harmonics in machine design [14], but it is very complicated. As will be shown in
this chapter, we employ the PM shaping technique rather than the PM segmentation
technique. Further, to achieve the sinusoidal or sinusoidal with third harmonic
airgap flux density distribution, it is necessary to consider the influence of orien-
tation of PM, flux leakage, curve effect, as well as practical implementation, etc. on
the PM shaping. Hence, the PM cannot be only simply shaped by sinusoidal or
2.1 Introduction 15

sinusoidal with third harmonic function as the PWM modulation utilizing third
harmonic although the concept is similar.
In this chapter, a simplified analytical model is developed to predict the variation
of average torque with the 3rd harmonic in PM shaping. The optimal 3rd harmonic
into the PM shaping will be analytically obtained. Quantitative comparisons
between different magnet shaping methods, i.e. without shaping, Sine shaping, Sine
shaping with 3rd harmonic injected, on electromagnetic performance including the
back-EMF waveforms, cogging torque, average torque and torque ripple are carried
out. Further, the influence and utilization of magnet edge thickness which is
required in practice on the airgap field distribution will be investigated to com-
pensate the inter-pole flux leakage and curvature effect. Experimental validations
are reported as well.
This chapter will be organized as follows. In Sect. 2.2, three prototype machines
with different shaping methods, i.e. without shaping, Sine shaping, Sine shaping
with 3rd harmonic injected, are described. In Sect. 2.3, the optimal value of 3rd
harmonic injected into the sinusoidal PM shape will be analytically derived and
confirmed by FE analyses, whilst the influence of magnet edge thickness is
investigated. In Sect. 2.4, the electromagnetic performance including the back-EMF
waveforms, cogging torque, average torque and torque ripple of SPM machines
with various PM shapes are compared, and the impact of the optimal 3rd harmonic
to the PM shape on average torque improvement is analyzed. The machines with
both conventional (without shaping) and optimal 3rd harmonic PM rotors are
prototyped and tested to validate the analysis in Sect. 2.5, while Sect. 2.6 is the
conclusion.

2.2 SPM Machines with Various PM Shapes

In order to investigate the influence of PM shape with 3rd harmonic on the torque
enhancement, 12-slot/10-pole SPM machine with double-layer non-overlapping
winding is employed for illustration. Figure 2.2a shows the stator together with the
winding connections used for all the machines, whilst the rotors with different PM
shapes are shown in Fig. 2.2b–d, respectively. Figure 2.1a, b shows the resultant
normalized MMF distribution and space harmonics produced by the three phase
fractional-slot concentrated windings. As can be seen that the harmonic contents are
high but only the 5th harmonic will interact with the PM-produced magnetic field to
produce the average torque. The other harmonics may cause undesirable effects,
such as localized saturation, additional iron loss, and eddy current loss in the
magnets, etc. [29]. These rotors are designated as the conventional rotor, Sine rotor,
and Sine+3rd rotor, to capture the key PM shape features. It should be noted that the
airgap length along the stator periphery in Sine rotor and Sine+3rd rotor machines is
not uniform since the PM outer diameter is not constant. However, the effective
airgap length which includes the magnets and airgap length is almost the same due
to the PM relative permeability nearly equal to airgap permeability. For Sine rotor
16 2 Torque Enhancement of Three Phase Surface-Mounted …

(a) Waveform

(b) Harmonics

Fig. 2.1 12-slot/10-pole SPM machines with different PM shapes

machine, the minimum airgap length is along the d-axis, whilst for Sine+3rd rotor
one, the position of minimum airgap length is determined by the value of 3rd
harmonic injected into the PM shape. This will be illustrated in detail in the fol-
lowing section. Although the airgap length for all machines varies, its minimum
value and PM maximum thickness are kept to be the same, 1 mm and 3 mm,
respectively. All the stator, rotor, and PM parameters are given in Table 2.1.
2.3 PM Shape with Optimal 3rd Harmonic and FE Validation 17

Table 2.1 Main Machine design specifications and parameters

Pole number 10 Tooth width 7.4 mm
Slot number 12 Back-iron 3 mm
Rated speed 400 rpm Slot opening 2 mm
Peak current 25 A Minimum airgap length 1 mm
Turns in series per phase 48 Magnet remanence 1.2 T
Axial length 50 mm Relative permeability 1.07
Stator outer diameter 90 mm Pole arc to pole pitch ratio 1
Stator inner diameter 53 mm Maximum PM thickness 3 mm

2.3 PM Shape with Optimal 3rd Harmonic and FE

Validation

Approximate sinusoidal or sinusoidal with 3rd harmonic airgap flux density dis-
tributions can be achieved by sinusoidally shaping or sinusoidally with 3rd har-
monic shaping the PMs, respectively, as shown in Fig. 2.2c, d. Both of them are
functions of rotor position h as illustrated in Fig. 2.3a, b.
Figure 2.3 shows the variation of PM thickness Dhm ðhÞ with the rotor positions
h. The maximum thickness Dm of the arc magnet is constant. It can also be
observed that the airgap flux density distribution is a function of h. Ideally, the
variation of airgap flux density with h can be expressed as:

Br
B g ð hÞ ¼ l ðhÞ
ð2:1Þ
1 þ lr Dhgm ðhÞ

where Br and lr is the magnet remanence and the relative permeability of PM,
respectively, lg ðhÞ is the variation of airgap length with rotor position. If lr is
assumed to be unity, the airgap flux density can be further derived as:

Dhm ðhÞBr
Bg ðhÞ ¼ ð2:2Þ
Dhm ðhÞ þ lg ðhÞ

Although both Dhm ðhÞ and lg ðhÞ vary with the rotor position, the total length of
PM and airgap is constant:

Dhm ðhÞ þ lg ðhÞ ¼ l ð2:3Þ

Thus

Dhm ðhÞBr
Bg ðhÞ ¼ ð2:4Þ
l
18 2 Torque Enhancement of Three Phase Surface-Mounted …

(a) 12-slot stator (b) Conventional rotor

(c) Sine rotor (d) Sine+3rd rotor

Fig. 2.2 12-slot/10-pole SPM machines with different PM shapes

2.3.1 Sinusoidal Shaping PM

In this case, the aim of PM shaping is to obtain a sinusoidal airgap flux density
distribution which is given by:

Bg ðhÞ ¼ Bm1 sinðphÞ ð2:5Þ

2.3 PM Shape with Optimal 3rd Harmonic and FE Validation 19

P1

(a) Sine shaped

P2 P1

(b) Sine +3rd shaped

Fig. 2.3 Theoretical PM shapes for sine and sine with 3rd harmonic

where p is the number of pole-pairs and Bm1 is the desired amplitude of the flux
density. From (2.4) and (2.5), in order to obtain the sinusoidal airgap flux density,
the variation of PM thickness with h should theoretically be:

l l
Dhm ðhÞ ¼ Bg ðhÞ ¼ Bm1 sinðphÞ ð2:6Þ
Br Br

Therefore, the variation of PM thickness with h should be:

Dhm ðhÞ ¼ Dm sinðphÞ ð2:7Þ

From (2.6) and (2.7), Bm1 can be approximately given as:

Dhm ðhÞBr
Bm1 ¼ ð2:8Þ
l

Neglecting the slot-opening, the airgap flux density of Sine shaped rotor can be
approximately given by (2.5), i.e.:
20 2 Torque Enhancement of Three Phase Surface-Mounted …

DmBr
Bg ðhÞ ¼ sinðphÞ ð2:9Þ
l

The open-circuit phase flux linkage and phase back-EMF are given by:

DmBr
uphase ¼ 2Rs la Np Kdp1 sinðpxr tÞ ð2:10Þ
l
DmBr
e ¼ 2Rs la Np xr Kdp1 cosðpxr tÞ ð2:11Þ
l

where Nc, la, Rs, and xr are the number of turns per coil, the stator active axial
length, the stator bore radius, and the rotor speed (rad/s), respectively. Kdp1 is the
fundamental winding factor.
When three phase currents, ia, ib, ic, are supplied in phase with the corresponding
back-EMFs, ea, eb, ec, the instantaneous electromagnetic torque can be derived as:

ea ia þ ea ia þ ea ia
Tm ¼ ð2:12Þ
xr

From (2.11) and (2.12), the electromagnetic torque of Sine shaped rotor machine
can be derived as:

3E1 I1 3I1 DmBr

Tm ¼ ¼ 2Rs la Np Kdp1 ð2:13Þ
2xr 2 l

where E1 is the amplitude of fundamental component of phase back-EMF,

DmBr
E1 ¼ 2Rs la Np xr Kdp1 ð2:14Þ
l

and I1 is the amplitude of the fundamental component of the phase current.

2.3.2 PM Shape with Optimal Amplitude of 3rd Harmonic

With addition of third harmonic to the Sine shaped magnet, (2.7) needs to be
modified as:

Dhm ðhÞ ¼ Dma ½sinðphÞ þ a sinð3phÞ ð2:15Þ

where Dma and a are the parameters to be determined. a is the amplitude of inject
3rd harmonic, while Dma varies with the value of injected 3rd harmonic in order to
maintain the same maximum magnet thickness.
2.3 PM Shape with Optimal 3rd Harmonic and FE Validation 21

The location of maximum magnet thickness P2 of Sine+3rd rotor, as shown in

Fig. 2.3b, should be firstly determined by differentiating Dhm ðhÞ with respect to h
and equating it to zero, i.e.

dDhm ðhÞ
¼ Dma p½cosðphÞ þ 3a cosð3phÞ ð2:16Þ
dh

Therefore, the maximum PM thickness occurs at:

12
9a  1
cosðphÞ ¼ ð2:17Þ
12a

The maximum value of Δhm(h) can then be obtained as:

3
d 1 þ 3a 2
Dhm ¼ Dma 8a ð2:18Þ
12a

dm
In order to have the same maximum magnet thickness as Sine shaped PM, Dh
should be equal to Dm:

dm ¼ Dm
Dh ð2:19Þ

From (2.18) and (2.19)

 3
1 1 þ 3a 2
Dma ¼ Dm ð2:20Þ
8a 12a

From (2.4) and (2.15), the airgap flux density of Sine+3rd shaped rotor can be
approximately expressed by:

Dma Br
B g ð hÞ ¼ ½sinðphÞ þ a sinð3phÞ ð2:21Þ
l

The 3rd harmonic back-EMF will be cancelled due to the star-connected bal-
anced windings. Therefore, similar to the case of sinusoidal shaping, the torque of
Sine+3rd shaped machine can be derived as:

3E1 I1 3I1 Dma Br

Tm ¼ ¼ 2Rs la Np Kdp1 ð2:22Þ
2xr 2 l

By applying the theory of Lagrange for optimisation, in order to maximize the

torque (2.22) while maintaining the same maximum magnet thickness (2.20), the
function against the amplitude of 3rd harmonic a can be obtained as:
22 2 Torque Enhancement of Three Phase Surface-Mounted …

 3 !
3I1 Dma Br 1 1 þ 3a 2
LðDma ; aÞ ¼ 2Rs la Np Kdp1 þ k Dma  Dm ð2:23Þ
2 l 8a 12a

where k is the Lagrange factor.

Hence, the maximum torque of Sine+3rd rotor can be determined by differen-
tiating LðDma ; aÞ with respect to Dma and a, and equating them to zero under the
condition of (2.20), i.e.:

dLðDma ; aÞ
¼0 ð2:24aÞ
dDma

dLðDma ; aÞ
¼0 ð2:24bÞ
da

 3
1 1 þ 3a 2
Dma  Dm ¼ 0 ð2:24cÞ
8a 12a

From (2.24), the values of a and Dma are obtained as:

1
a¼ ð2:25aÞ
6
pffiffiffi
2 3
Dma ¼ Dm ð2:25bÞ
3

Therefore, the optimal value of a is 1/6, and the contour of PM with optimal 3rd
harmonic is:
pffiffiffi  
2 3 1
Dhm ðhÞ ¼ Dm sinðphÞ þ sinð3phÞ ð2:26Þ
3 6

and the corresponding maximum torque is:

pffiffiffi
3I1 Br 2 3
Tm ¼ 2Rs la Np Kdp1 Dm ð2:27Þ
2 l 3
pffiffi
It can be seen from (2.23) and (2.27) that the torque can be improved by 2 3 3 ¼
1.15 compared with the machine with Sine shaped PM rotor due to the increase of
P1 in the PM contour, as shown in Fig. 2.3a, b.
By the way, it is worth mentioning that the optimal 3rd harmonic is 1/6 of the
fundamental one, which is coincidentally the same as PWM technique utilizing 3rd
harmonic [19, 20].
2.3 PM Shape with Optimal 3rd Harmonic and FE Validation 23

2.3.3 FE Validation

The contour of PM with rotor angle for the Sine+3rd shaped rotor to maximize the
torque can be obtained from (2.26). In order to verify the optimal value of third
harmonic in the magnet shaping, the influences of amplitude of 3rd harmonic in the
contour of PM for the Sine+3rd shaped rotor are investigated by FE analysis.
Figure 2.4 shows the variation of the torque with the amplitude of 3rd injected
into the Sine shaped PMs by both analytical and FE methods. It can be seen that,
the analytically predicted optimal 3rd harmonic agrees well with the FE methods.
The airgap magnetic field is analytically calculated by simplified derivation,
Eq. (2.1), in which leakage flux and curve effect are neglected and then the ana-
lytical torque is obtained to optimize the 3rd harmonic in the magnet shape. In FE
model, these neglected factors are considered. Therefore, the FE predicted torque is
lower than that of analytically predicted one and the amplitude difference is mainly
due to the simplified analytical derivation, such as neglected leakage flux and curve
effect, etc. However, there is no significant effect on optimal 3rd harmonic deter-
mination. Meanwhile, the Sine shaped PM has the lowest torque and the Sine+1/6
3rd harmonic shaped PM has the highest torque among all the shaped PMs. It is
worth of noting that although from approximate analytical models the Sine+1/6 is
the optimal one for the torque improvement, the torque of Sine+1/7 and Sine+1/5
3rd harmonic are almost similar to that of Sine+1/6 one. Therefore, the value of 3rd
harmonic into the magnet is not so sensitive around the value of 1/6 which is good
for accommodating the manufacturing tolerance.
Figure 2.5 shows the airgap flux density distributions for the machines with
conventional rotor, Sine shaped rotor and Sine+3rd shaped rotor by FE analysis,
where the slot-opening effects are neglected. It can be observed that the airgap field
distribution is neither sinusoidal nor sinusoidal with 3rd harmonic distribution due
to the inter-pole flux leakage and curvature effect, etc.

Fig. 2.4 Variation of torque

with different value of 3rd
harmonic
24 2 Torque Enhancement of Three Phase Surface-Mounted …

Fig. 2.5 Airgap flux density

distribution for different
magnet shape (neglecting
effects of slot-opening)

2.3.4 Influence of PM Edge Thickness

In order to compensate inter-pole flux leakage and curvature effect, PM edge

thickness Δt as shown in Fig. 2.6a, b is introduced to ease manufacture and is
optimized for achieving nearly sinusoidal and sinusoidal with 3rd harmonic airgap

(a) Sine shaped

(b) Sine +3rd shaped

Fig. 2.6 Edge thickness compensation for Sine and Sine+3rd shaped PMs
2.3 PM Shape with Optimal 3rd Harmonic and FE Validation 25

flux density distribution. Then, the magnet thickness with such compensation is
given by:

hm ðhÞ ¼ Dt þ Dhm ðhÞ ð2:28Þ

The maximum thickness of the magnet is kept to be constant, 3 mm. Hence, Dm

will be varied with Dt. It can be seen that when the edge thickness is equal to 3 mm,
i.e. Dm = 0 mm, both the airgap length and PM thickness are uniform, i.e. con-
ventional rotor with pole arc to pole pitch ratio = 1.
Figure 2.7a, b shows the FE predicted airgap flux density distributions and their
THDs for different Δt of Sine shaped PMs. It can be seen that, the conventional PM
shape has the highest fundamental one whilst Δt of 1 mm has the lowest total
harmonic distortion of all the investigated PM shapes. Its fundamental value always
reduces with the decrease of Δt while the THD firstly reduces and then increases.
Hence, the optimal Δt of 1 mm exists for minimum THD.

(a) Waveform (for clarification, only results for ∆t=0, 1 and 2mm are shown)

(b) Fundamental and THD

Fig. 2.7 Variation of airgap flux density with PM edge thickness for Sine shaped PM
26 2 Torque Enhancement of Three Phase Surface-Mounted …

In order to further investigate the influence of PM edge thickness on the airgap

flux density of the Sine+3rd shaped PMs with 1/6 of fundamental one, Fig. 2.8a, b
shows the FE predicted airgap flux density distributions and their THDs for dif-
ferent Δt. It can be seen that, the conventional PM shape still has the highest
fundamental one while Δt from 0 mm to 1 mm has almost similar THD of all the
investigated PM shapes. The fundamental value and THD always reduce with the
decrease of Δt. Compared to Sine shaped PMs with Δt of 1 mm in Fig. 2.7b, the
THD of Sine+3rd shaped PMs is slightly high. However, when the edge thickness
of Sine+3rd shaped PMs Δt is equal to 1 mm, the torque will be enhanced and
torque ripple will not be deteriorated compared to that of Sine shaped PMs with Δt
of 1 mm. This will be illustrated in the following sections.

(a) Waveform (for clarification, only results for ∆t=0, 1 and 2mm are shown)

(b) Fundamental and THD

Fig. 2.8 Variation of airgap flux density with PM edge thickness for Sine+3rd shaped PM
2.4 Finite Element Analysis of Electromagnetic Performance 27

2.4 Finite Element Analysis of Electromagnetic

Performance

The Sine shaped PMs with Δt = 1 mm and the Sine+3rd shaped PMs with optimal
3rd harmonic of 1/6 of fundamental one and Δt = 1 mm are employed in 12-slot/
10-pole machines (outer diameter = 90 mm and active axial length = 50 mm) for
further analysis in this section.

2.4.1 Open-Circuit Flux Density Distribution

and Back-EMFs

Figure 2.9 shows the FE predicted open-circuit field distributions for the 12-slot/
10-pole SPM machines with various PM shapes. Figure 2.10a, b shows the airgap
flux density distributions and harmonic contents. As shown in Fig. 2.10b the
fundamental one in the airgap flux density of Sine shaped rotor has been reduced,
compared with that of SPM machine with the conventional rotor. However, such
decrease in the airgap flux density can be compensated by injecting the 3rd har-
monic into the Sine shaped rotor, i.e. Sine+3rd shaped rotor, as shown in
Fig. 2.10b. On the other hand, other harmonics of Sine shaped rotor SPM machine
are also reduced. For Sine+3rd shaped rotor SPM machine, other harmonics except
for 3rd harmonic do not increase compared with Sine shaped rotor.
The FE predicted back-EMFs of the SPM machines with various PM shapes are
shown Fig. 2.11. It can be seen that the conventional rotor SPM machine has the
highest amplitude but the lowest amplitude exists in Sine shaped PM machine for
both phase and line-line back-EMFs. However, injecting the 3rd harmonic to the
Sine shaped PM, both the phase and line-line back-EMFs have been increased by
10%. In order to compare its harmonics, the harmonic contents have been nor-
malized to the fundamental one, as given in Table 2.2. It can be observed that the
5th, 11th, and 13th harmonics are reduced significantly for both Sine shaped and
Sine+3rd shaped SPM machine. However, it should be noted that the Sine+3rd
shaped rotor SPM machine has the higher fundamental phase and line-line
back-EMFs compared with Sine shaped rotor SPM machine. Meanwhile, those
triplen harmonics have been eliminated in the line-line back-EMFs as shown in
Fig. 2.11 and Table 2.2.

2.4.2 Torque Characteristics

The cogging torque depends on the smallest common multiple between the slot
number and pole number [6]. Although the same stator for all the three machines is
employed, the peak cogging torque is not the same due to the influence of rotor
28 2 Torque Enhancement of Three Phase Surface-Mounted …

Fig. 2.9 Flux distributions of

12-slot/10-pole SPM
machines

(a) Conventional rotor

(b) Sine rotor

(c) Sine+3rd rotor

2.4 Finite Element Analysis of Electromagnetic Performance 29

(a) Waveforms

(b) Harmonics

Fig. 2.10 Variations of airgap flux density distributions with stator position

shape, as shown in Fig. 2.12. This is due to the fact that different rotor shape results
in different airgap flux density distribution and its harmonic contents (Fig. 2.10a,
b). However, the period of the cogging torque is the same for all machines since
they have the same smallest common multiple between the stator slot numbers and
the rotor poles. It can be also observed that the cogging torque of conventional rotor
SPM machine and Sine shaped rotor SPM machine is the highest and the lowest,
respectively, among all the machines. Compared with Sine shaped rotor SPM
machine, the cogging torque in Sine+3rd shaped rotor one is slightly increased.
Figure 2.13 shows the variation of torque with rotor position for the three
machines whilst Table 2.3 shows the torque components, where the supplied
30 2 Torque Enhancement of Three Phase Surface-Mounted …

Fig. 2.11 Back-EMFs (400 rpm)

Table 2.2 Harmonics in back-EMFs (400 rpm)

Conventional rotor (%) Sine rotor (%) Sine+3rd rotor (%)
Phase Line Phase Line Phase Line
1 100 100 100 100 100 100
(4.84 V) (8.39 V) (4.08 V) (7.07 V) (4.48 V) (7.76 V)
3 9.62 0.00 5.38 0.00 7.59 0.00
5 0.52 0.51 0.53 0.53 0.32 0.32
7 0.08 0.09 0.19 0.19 0.52 0.52
9 1.41 0.00 0.33 0.00 1.04 0.00
11 0.52 0.51 0.11 0.11 0.29 0.29
13 0.17 0.16 0.01 0.01 0.12 0.11

Fig. 2.12 Variation of cogging torque with rotor position

2.4 Finite Element Analysis of Electromagnetic Performance 31

Fig. 2.13 Variation of torque with rotor position, Ipeak = 25 A

three-phase currents are in phase with its phase back-EMFs. It shows that the
conventional rotor SPM machine has the highest average torque but Sine shaped
rotor SPM machine has the lowest average torque. The average torque for Sine+3rd
shaped rotor SPM machine has been improved by >9% compared with that of Sine
shaped rotor SPM machine. This is due to the variation of average PM thickness
and airgap length for different PM shapes, as shown in Fig. 2.3a, b. The magnet
usage and magnet usage efficiency ηpm defined by average torque to PM volume for
the machines with various PM shapes are also compared in Table 2.3.
Meanwhile, the torque ripples of Sine shaped and Sine+3rd shaped machines are
similar, but lower than that of the conventional rotor SPM machine. This is due to
the fact that the 5th, 11th, and 13th harmonics in back-EMFs are reduced

Table 2.3 Average airgap Conventional Sine Sine

length, PM height, torque +3rd
components and PM volume
6th (Nm) 0.042 0.007 0.012
12th (Nm) 0.129 0.005 0.003
Tave (Nm) 4.32 3.63 3.99
DTpeak-peak (Nm) 0.30 0.02 0.03
Tripple (%) 6.98 0.50 0.67
Tcogging (Nm) 0.114 0.001 0.011
lavg (mm) 1 1.73 1.45
lm (mm) 3 2.27 2.55
ΔTavg (%) – 100 115
(calculated)
ΔTavg (%) (FE) 120 100 110
Vpm ( 10−5 m3) 2.26 1.69 1.91
ηpm (%) 1.91 2.15 2.09
32 2 Torque Enhancement of Three Phase Surface-Mounted …

significantly (Fig. 2.11 and Table 2.2) for both Sine shaped rotor and Sine+3rd
shaped rotor SPM machine, which will interact with the sinusoidal current causing
the 6th and 12th harmonics in torque, as given in Table 2.3. Low harmonics in the
back-EMF together with low cogging torque result in low torque ripple. Most
importantly, the average torque has been improved.

2.5 Experimental Verification and Discussions

The 12-slot/10-pole SPM machines (outer diameter = 90 mm and active axial

length = 50 mm) with conventional and Sine+3rd shaped PM where the 3rd har-
monic is 1/6 of fundamental one and edge thickness Δt = 1 mm have been proto-
typed and tested for validation. Figure 2.14 shows the pictures of the same stator
together with the conventional and Sine+3rd shaped rotors, respectively.
The 2-dimensional FE calculated and measured cogging torque of the SPM
machines with conventional and Sine+3rd shaped rotor are given in Fig. 2.15. It
confirms that the cogging torque could be significantly decreased for the Sine+3rd
shaped SPM machine compared with that of the conventional machine. The vari-
ations of FEA predicted and measured phase and line-line back-EMFs with rotor
position are compared in Fig. 2.16a, b, respectively, whilst Tables 2.4 and 2.5 list
the phase back-EMF where the corresponding dominant harmonics (fundamental,
5th, 7th, 11th, and 13th) are given. It also confirms that the fundamental one is
slightly reduced as well. The triplen harmonics do not exist in the line-line
back-EMFs because of star-connected three-phase balanced windings. The varia-
tion of static output torque is measured by supplying three-phase windings with
dc-bus current Idc = Ia = −2Ib = −2Ic. Figures 2.17a, b and 2.18 show the FEA
predicted and measured torque with rotor position and peak torque with current of
SPM machine with conventional rotor and Sine+3rd shaped rotor, respectively. In
general, good agreements between all FEA and experimental results are achieved.
2.5 Experimental Verification and Discussions 33

Fig. 2.14 Prototypes of

12-slot/10-pole SPM machine

(a) Stator

(b) Conventional rotor

(c) Sine +3rd rotor

34 2 Torque Enhancement of Three Phase Surface-Mounted …

Fig. 2.15 Variation of FE predicted and measured cogging torque of conventional and Sine+3rd
rotor 12-slot/10-pole SPM machines with rotor position

(a) Phase

(b) Line-line

Fig. 2.16 Variation of FE predicted and measured phase and line-line back-EMFs of conventional
and Sine+3rd shaped rotor 12-slot/10-pole SPM machines (400 rpm)
2.6 Summary 35

Table 2.4 Harmonics in Conventional rotor (%)

back-EMFs of SPM machine
Phase Line-line
with conventional rotor,
400 rpm FE Measured FE Measured
1 100 100 100 100
(4.84 V) (4.80 V) (8.39 V) (8.27 V)
3 9.62 9.96 0.00 0
5 0.52 0.21 0.51 0.38
7 0.08 0.43 0.09 0.40
9 1.41 1.67 0.00 0
11 0.52 0.09 0.51 0.10
13 0.17 0.16 0.16 0.20

Table 2.5 Harmonics in Sine + 3rd shaped rotor (%)

back-EMFs of SPM machine
Phase Line-line
with Sine+3rd rotor, 400 rpm
FE Measured FE Measured
1 100 100 100 100
(4.48 V) (4.43 V) (7.76 V) (4.72 V)
3 7.59 7.19 0 0
5 0.32 0.16 0.32 0.20
7 0.52 0.27 0.52 0.24
9 1.04 0.87 0 0
11 0.29 0.20 0.29 0.17
13 0.12 0.19 0.11 0.15

2.6 Summary

This chapter presents a permanent magnet (PM) shaping technique with optimal 3rd
harmonic to enhance the output torque of surface-mounted PM (SPM) machines
without deteriorating the torque ripple. The influence of magnet edge thickness on
the airgap field distribution is investigated and can be utilized to compensate the
inter-pole flux leakage and curvature effect. It is found that the optimal 3rd har-
monic is 1/6 of the fundamental one by both analytical and FE analyses. It is
demonstrated that the average torque of the machine of a Sine shaping with an
optimal 3rd harmonic injected can be improved by >9%, while the torque ripple is
almost similar to that of the one with Sine shaping. Finally, the measured results on
the prototype machines with both conventional (without shaping) and optimal 3rd
harmonic PM rotors verify the analyses.
36 2 Torque Enhancement of Three Phase Surface-Mounted …

(a) Conventional rotor

(b) Sine + 3rd rotor

Fig. 2.17 Variation of FE predicted and measured torque with rotor position of conventional and
Sine+3rd shaped PM in 12-slot/10-pole SPM machines (Idc = Ia = −2Ib = −2Ic)
References 37

Fig. 2.18 FE calculated and measured peak torque with current of conventional and Sine+3rd
shaped PM in 12-slot/10-pole SPM machines

References

1. Z.Q. Zhu, D. Howe, Electrical machines and drives for electric, hybrid, and fuel cell vehicles.
Proc. IEEE 95(4), 746–765 (2007)
2. A.M. El-Refaie, Fractional-slot concentrated-windings synchronous permanent magnet
machines: opportunities and challenges. IEEE Trans. Ind. Electron. 57(1), 107–121 (2010)
3. A.M. El-Refaie, Motors/generators for traction/propulsion applications: a review. IEEE Veh.
Technol. Mag. 8(1), 90–99 (2013)
4. W.P. Cao, B.C. Mecrow, G.J. Atkinson, J.W. Bennett, D.J. Atkinson, Overview of electric
motor technologies used for more electric aircraft (MEA). IEEE Trans. Ind. Electron. 59(9),
3523–3531 (2012)
5. T.M. Jahns, W.L. Soong, Pulsating torque minimization techniques for permanent magnet AC
motor drives—a review. IEEE Trans. Ind. Electron. 43(2), 321–330 (1996)
6. Z.Q. Zhu, D. Howe, Influence of design parameters on cogging torque in permanent magnet
machines. IEEE Trans. Energy Convers. 15(4), 407–412 (2000)
7. N. Bianchi, S. Bolognani, Design techniques for reducing the cogging torque in
surface-mounted PM motors. IEEE Trans. Ind. Appl. 38(5), 1259–1265 (2002)
8. W. Schuisky, Berechnung Elektrischer Maschinen (Springer, New York, NY, USA, 1960),
pp. 95–97
9. J. Pyrhönen, T. Jokinen, V. Hrabovcova, Design of rotating electrical machines (Wiley,
Chichester, UK, 2008), p. 165
10. A. Evans, Salient pole shoe shapes of interior permanent magnet synchronous machines, in
Proceedings 19th International Conference Electrical Machine, Sept 2010, pp. 1–6
11. Y. Li, J.W. Xing, T.B. Wang, Y.P. Lu, Programmable design of magnet shape for permanent
magnet synchronous motors with sinusoidal back EMF waveforms. IEEE Trans. Magn. 44(9),
2163–2167 (2008)
12. P. Zheng, J. Zhao, J. Han, J. Wang, Z. Yao, R. Liu, Optimization of the magnetic pole shape
of a permanent-magnet synchronous motor. IEEE Trans. Magn. 43(6), 2531–2533 (2007)
13. N.R. Tavana, A. Shoulaie, Analysis and design of magnetic pole shape in linear
permanent-magnet machine. IEEE Trans. Magn. 46(4), 1000–1006 (2010)
38 2 Torque Enhancement of Three Phase Surface-Mounted …

14. S. Chaithongsuk, N. Takorabet, F. Meibody-Tabar, On the use of pulse width modulation

method for the elimination of flux density harmonics in the air-gap of surface PM motors.
IEEE Trans. Magn. 45(3), 1736–1739 (2009)
15. A.H. Isfahani, S. Vaez-Zadeh, M.A. Rahman, Performance improvement of permanent
magnet machines by modular poles. IET Electr. Power Appl. 3(4), 343–351 (2009)
16. M.Y. Kim, Y.C. Kim, G.T. Kim, Design of slotless-type PMLSM for high power density
using divided PM. IEEE Trans. Magn. 40(3), 746–749 (2004)
17. N.R. Tavana, A. Shoulaie, V. Dinavahi, Analytical modeling and design optimization of
linear synchronous motor with stair-step-shaped magnetic poles for electromagnetic launch
applications. IEEE Trans. Plasma Sci. 40(2), 519–527 (2012)
18. M.S. Islam, S. Mir, T. Sebastian, S. Underwood, Experimental verification of design
techniques of permanent-magnet synchronous motors for low-torque-ripple applications.
IEEE Trans. Ind. Appl. 47(1), 88–95 (2011)
19. G. Buja, G. Indri, Improvement of pulse width modulation techniques. Archiv Elektotech.
57(5), 281–289 (1975)
20. J.A. Houldsworth, D.A. Grant, The use of harmonic distortion to increase the output voltage
of a three-phase PWM inverter, in IEEE Transaction on Industry Application, Sept 1984, vol.
IA-20, no. 5, pp. 1224–1228
21. P. Pejovic, Z. Janda, An improved current injection network for three-phase
high-power-factor rectifiers that apply the third harmonic current injection. IEEE Trans.
Ind. Electron. 47(2), 497–499 (2000)
22. J.X. Shen, Z.Q. Zhu, D. Howe, Sensorless flux-weakening control of permanent-magnet
brushless machines using third harmonic back EMF. IEEE Trans. Ind. Appl. 40(6), 1629–
1636 (2004)
23. L.B. Zheng, J.E. Fletcher, W.B. Williams, X.N. He, Dual-plane vector control of a five-phase
induction machine for an improved flux pattern. IEEE Trans. Ind. Electron. 55(5), 1996–2005
(2008)
24. L. Parsa, H.A. Toliyat, Five-phase permanent magnet motor drives. IEEE Trans. Ind. Appl.
41(1), 30–37 (2005)
25. R.O.C. Lyra, T.A. Lipo, Torque density improvement in a sixphase induction motor with third
harmonic current injection. IEEE Trans. Ind. Appl. 38(5), 1351–1360 (2002)
26. Y. Li, J.B. Zou, Y.P. Lu, Optimum design of magnet shape in permanent-magnet synchronous
motors. IEEE Trans. Magn. 39(6), 3523–3526 (2003)
27. Z.Q. Zhu, K. Wang, G. Ombach, Optimal magnet shaping with third order harmonic for
maximum torque in SPM machines, in Proceedings of 6th IET Conference Power Electronics,
Machines Drives, PEMD, Bristol, UK, March 2012, pp. 1–6
28. K. Wang, Z.Q. Zhu, G. Ombach, W. Chlebosz, Optimal rotor shape with third harmonic for
maximizing torque and minimizing torque ripple in IPM motors, in Proceedings of 20th
ICEM, Marseille, France, Sept 2012, pp. 397–403
29. M. Freddy, L. Heinz, Parasitic effects in PM machines with concentrated windings. IEEE
Trans. Ind. Appl. 43(5), 1223–1232 (2007)
Chapter 3
Average Torque Improvement of Three
Phase Interior Permanent-Magnet
Machine Using 3rd Harmonic
in Rotor Shape

A rotor shaping technique with optimal 3rd harmonic is presented to enhance the
average torque of IPM machines in this chapter. The optimal value of 3rd harmonic
injected into the rotor outer surface shape has been derived, and further confirmed
by both finite element analyses and experiments. The impact of the optimal 3rd
harmonic to the rotor shape on the electromagnetic performance including har-
monics in the back-EMFs, cogging torque, average torque and torque ripple is
investigated. It is demonstrated that without any modification of the costly rare
earth permanent magnet employed for the inverse cosine shaped rotor, the average
torque of the machine of an inverse cosine injected with an optimal 3rd harmonic
shaped rotor can be improved by >6%. Simultaneously, the torque ripple remains
almost unchanged, and the saliency ratio is also improved, further boosting the
average torque. Finally, the machines with both conventional and optimal 3rd
harmonic rotors are prototyped and tested to validate the analysis.

3.1 Introduction

With the development of high-energy permanent magnets, high-speed computer,

and commercial finite element software, permanent magnet (PM) machines have
gained wide applications from automotive to aerospace over the past four decades
[1–3]. Interior permanent magnet (IPM) machines are particularly attractive since
they may employ rectangular block PMs buried into the rotor, good flux-weakening
capability and easy retaining of PMs [4, 5]. Meanwhile, fractional slot
non-overlapping windings are attractive in terms of low cogging torque, short
end-winding and thus low copper loss and high efficiency [6]. Therefore, IPM
machines with fractional-slot non-overlapping windings are the research subject of
the academia and industries in order to exploit their potentials for wider applications
[7–10].

© Springer Nature Singapore Pte Ltd. 2019 39

K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_3
40 3 Average Torque Improvement of Three Phase …

However, concentrated fractional-slot winding together with relatively small

airgap length in IPM machines causes the high harmonic contents in electromag-
netic motive force compared with surfaced-mounted PM machines [11].
Consequently, the torque ripples may be too high for some high performance
applications, such as electric power steering systems. Hence, it is necessary to
reduce the torque ripples for smooth operation. The torque ripples origin from the
cogging torque due to the interaction between the PMs and the stator slots, the
reluctance torque ripples arising from the variation of high order inductance har-
monics and the PM torque ripples because of interaction between non-ideal sinu-
soidal current with non-ideal sinusoidal back-EMF [12, 13]. In addition, magnetic
saturation in the stator and rotor can also affect torque ripples under load condition
[14, 15]. Thus, many techniques have been proposed for torque ripple reduction and
can be mainly divided into two categories [16]. The first one is machine-design-
based [16–19], which involves with the optimization of machine design parameters.
The second one is control-based [20–23], which needs to optimize the current
waveform. This chapter discusses the machine-design-based techniques to reduce
the torque ripple.
Several techniques based on the optimization of machine parameters can be
employed to reduce the torque ripple in IPM machines, including the ratio of
pole-arc to pole-pitch [24–26], the number of slots per pole pair [27, 28], skew/step
skew [29, 30], auxiliary slots/teeth [16], iron bridge shaping [31], rotor outer sur-
face shaping [32–34], holes in the rotor or groove on the rotor surface [19, 35], and
slot/flux-barrier number combinations [36], etc. The sinusoidal air-gap flux density
distribution can be achieved by shaping the rotor pole [37], and it has been widely
employed in electrical power steering application [12, 34, 38].
On the other hand, although rotor shaping methods can effectively minimize the
torque ripple, the fundamental air-gap flux density and, thus, the torque density
decrease as well. Different approaches have been proposed to increase the funda-
mental output from inverters to machine designs. In particular, the third harmonic
has gained wide application in power electronics, pulse width modulation tech-
niques, and machine design and sensorless operation [39–44]. The output voltage
can be improved by pulsewidth modulation with the third harmonic in inverters
[39], and the THD of input current for rectifiers can be reduced with third harmonic
current injection [40]. The third harmonic back EMF can be employed for sen-
sorless control [41]. The flux patterns can be optimized in the five phase induction
machine by third harmonic current injection into the winding [42], whereas for PM
machines, the torque density can be improved by third harmonic in the PMs’ shape
[38, 43, 44]. Shaping the IPM machine rotor outer surface with third harmonic also
changes the air-gap length along both d- and q-axes and thus their inductances [38].
However, in [38], the influence of rotor shape with third harmonic on average
torque was not analytically investigated. Hence, the impact of parameters on
average torque improvement was not revealed, and no prototype machine was built
and tested.
Therefore, a rotor shaping technique with optimal 3rd harmonic is presented to
enhance the average torque of IPM machines in this chapter. The optimal value of
3.1 Introduction 41

3rd harmonic injected into the rotor outer surface shape has been derived for
average torque improvement, and further confirmed by finite element method. The
average torque improvement using 3rd harmonic is simplified investigated, and the
impact of the optimal 3rd harmonic to the rotor shape on the harmonics in the
back-EMFs, cogging torque, average torque and torque ripple is also investigated.
Finally, the prototype IPM machines with/without 3rd harmonic are fabricated and
experiments are presented to verify the theoretical and FE analyses.

3.2 Rotor Configurations of IPM Machines

Owing to the short end-winding, the 12-slot/10-pole IPM machine with

double-layer non-overlapping winding is employed for this investigation. The
common stator together with the winding connections is shown in Fig. 3.1a. The
rotors with different outer surface profile shapes are designated as the conventional
rotor, inverse cosine shaped (ICS) rotor, and ICS+3rd shaped rotor, to capture the
key rotor shape features, as shown in Fig. 3.1b–d, respectively. In order to highlight
their difference, different rotor contours are shown in Fig. 3.1e. The rotor opti-
mization adds to the complexity and manufacturing cost of the motor compared
with the conventional one (without shaping). However, the ICS+3rd shaped method
will not increase the complexity and manufacturing cost of the motor with ICS
shaped rotor. Hence, the complexity and manufacturing cost of the machine is
similar to the ICS shaped one. The ICS shaped one has gained wide application for
the electric power steering [12, 34, 37, 38]. It should be noted that the airgap length
for the ICS shaped rotor and ICS+3rd shaped is not uniform due to the modification
of rotor shape. For ICS shaped rotor, the minimum airgap length is along the d-axis
whilst for ICS+3rd shaped rotor, the position of minimum airgap length is deter-
mined by the value of 3rd harmonic injected into the rotor shape. This will be in
detail illustrated in the following section. Although the airgap length for all
machines is varied, its minimum value is the same, 0.5 mm. All the stator, rotor,
and PM parameters including its remanence and relative permeability are given in
Table 3.1.

3.3 Rotor Shaping with 3rd Harmonic

For ICS and ICS+3rd shaped rotor in IPM machines, the ratio of shaped-arc to
pole-pitch, defined in Fig. 3.2a, cannot be unity due to existence of flux-barriers and
is an important parameter to have a nearly sinusoidal airgap flux density distribution
[37]. This ratio can be firstly determined and then the rotor shape with 3rd harmonic
can be derived based on the shaped-arc to pole-pitch ratio.
42 3 Average Torque Improvement of Three Phase …

(a) 12-slot stator (b) Conventional rotor

(c) ICS rotor (d) ICS+3rd rotor

stator Conventional-(b)

ICS - (c) ICS+3rd

- (d)

(e) Rotor contour of different rotor structures

Fig. 3.1 12-slot/10-pole IPM machines with different rotor structures

3.3 Rotor Shaping with 3rd Harmonic 43

Table 3.1 Main machine design specifications and parameters

Pole number 10 Tooth width 7.4 mm
Slot number 12 Back-iron 3 mm
Rated speed 400 rpm Slot opening 2 mm
Peak current 56 A Minimum airgap length 0.5 mm
Turns in series per phase 48 Magnet remanence 1.2 T
Axial length 50 mm Relative permeability 1.07
Stator outer diameter 90 mm PM width 12.2 mm
Stator inner diameter 53 mm PM thickness 3.4 mm

(a) Shaped-arc to pole-pitch ratio (b) Airgap length with shaped-arc θ.

Fig. 3.2 ICS shaped rotor

3.3.1 ICS Shaped Rotor

In order to achieve a sinusoidal airgap flux density distribution, the airgap length
needs to follow a specific function with the shaped-arc angle h. The variation of
airgap length with h has been derived and can be given as [37, 38]:

l
lg ðhÞ ¼
gd  ð3:1Þ
cos spp h

lgd is shown in Fig. 3.2.

From Eq. (3.1), it can be seen that when shaped-arc angle h is equal to sp/2, the
airgap length is infinite and PMs are removed from the rotor surface. Hence, the PM
volume inside the rotor of IPM machine together with the rotor shape will be
varied. Therefore, the optimal shaped-arc angle exists for achieving nearly
44 3 Average Torque Improvement of Three Phase …

sinusoidal airgap flux density distribution and lowest total harmonic distortion
(THD). To keep the identical flux leakage and saturation level in the flux-barrier for
investigation of influence of shaped-arc to pole-pitch ratio, the dimensions of the
flux-barrier is the same.
FE predicted distributions and total harmonic distortion of airgap flux density are
given in Fig. 3.3a, b. It can be seen that the waveforms of the airgap flux density
distributions are different for different shaped-arc to pole-pitch ratios. With the
increase of this ratio, the total harmonic distortion in the airgap flux density reduces
and hence the waveforms become more sinusoidal. More specifically, the con-
ventional rotor has a rectangular waveform, in which the total harmonic distortion is

(a) Waveforms

(b) Fundamental and THD.

Fig. 3.3 Variations of airgap flux density with ratio of shaped-arc to pole-pitch for ICS shaped
rotor
3.3 Rotor Shaping with 3rd Harmonic 45

highest one for all the ratios of shaped-arc to pole-pitch. Increasing the ratio of
shaped-arc to pole-pitch makes the airgap flux density sinusoidal, resulting in low
total harmonic distortion, as presented in Fig. 3.3b. However, the fundamental one
is also reduced as shown in Fig. 3.3b and this can be compensated by 3rd harmonic
in the rotor shape. All of these are due to the modification of the airgap length,
especially along the q-axis. Moreover, the shaped-arc to pole-pitch ratio of 0.833
has the lowest total harmonic distortion of the airgap flux density and this means
that the stator iron loss may be the lowest amongst all of the investigated shaped-arc
to pole-pitch ratios.

3.3.2 Rotor Shape with Different Amplitude

of 3rd Harmonic

When different amplitude of 3rd harmonic is injected into the ICS shaped rotor, the
airgap length for the ICS+3rd shaped rotor can be given:

kðaÞ lgd
lg ðhÞ ¼

 ð3:2Þ
p
cos sp h  a cos 3 spp h

Figure 3.4 shows the ICS+3rd shaped rotor with various value of 3rd harmonic.
It can be observed that with the increase of the amplitude of injected 3rd harmonic,
the airgap length in the d-axis increases whilst that in q-axis decreases. Further, the
minimum airgap length is the same for all the ICS+3rd shaped rotor but the position
of minimum airgap length varies with the value of injected 3rd harmonic into the
ICS shaped rotor. The variations of d- and q-axis airgap lengths together with the
fixed minimum airgap length mean that the optimal value of 3rd harmonic exists for
the ICS+3rd shaped rotor.
The FE predicted airgap flux density distribution of different ICS+3rd shaped
rotors is shown in Fig. 3.5, and its fundamental value is given in Fig. 3.6, where the

Fig. 3.4 Rotor shapes with various values of 3rd harmonic

46 3 Average Torque Improvement of Three Phase …

Fig. 3.5 Airgap flux density with different rotor shapes

1.2
Fundamental flux density (T)

1.0

0.8

0.6

0.4

0.2

0.0
Conventional ICS ICS+1/9 3rd ICS+1/6 3rd ICS+1/3 3rd

Fig. 3.6 Fundamental airgap flux density of different rotor shapes

slot-openings are neglected. It can be obviously seen that the airgap flux density
distribution of ICS+3rd shaped rotors is different and the 3rd harmonic in the airgap
flux density is obvious. In addition, the amplitude of airgap flux density is also
varied because of 3rd harmonic airgap flux density. In addition, the fundamental
airgap flux density for all the investigated rotors is different and this is again due to
the variation of the value of injected 3rd harmonic. Moreover, it can be found that
the ICS shaped rotor has the lowest fundamental airgap flux density and the con-
ventional rotor (without shaping) has the highest fundamental one. However, as
described previously, the total harmonic distortion of airgap flux density for the
conventional rotor is high. Although the ICS shaped rotor has nearly sinusoidal
airgap flux density (Fig. 3.6a), the fundamental airgap flux density for this rotor is
3.3 Rotor Shaping with 3rd Harmonic 47

lower than that of ICS+1/9 3rd harmonic and ICS+1/6 3rd harmonic shaped rotor.
Further, the ICS+1/6 3rd harmonic shaped rotor has the highest fundamental airgap
flux density among all the shaped rotors. This means that the ICS+1/6 3rd harmonic
shaped rotor is the optimal one for the rotor shape with 3rd harmonic. Hence, the
airgap length with shaped angle for the optimal ICS+3rd shaped rotor can be given
as:
pffiffiffi
3lgd
lg ðhÞ ¼


 ð3:3Þ
2 cos sp h  16 cos 3 spp h
p

3.4 Simplified Analytical Analysis of Average

Torque Improvement

Based on the foregoing analysis, the airgap flux density is different for three IPM
machines due to the variation of airgap length, as shown in Fig. 3.7. Therefore, the
average torque of IPM machines with different rotor shapes is analytically analyzed
in this section.
For conventional IPM rotor, Fig. 3.7a, the average airgap length can be given as:

lavg C ¼ lgd ð3:4Þ

While for 10-pole ICS shaped rotor in Fig. 3.7b, its average value over one
pole-pitch can be obtained from the integration of (3.1) as:
5sp
Z12
1 l
lavg ICS ¼ 5sp
gd  dh ¼ 1:549 lgd ð3:5Þ
12 0 cos spp h

It can be found that the average airgap length has been significantly increased
compared with that of conventional IPM rotor. However, this can be compensated
by the 3rd harmonic into the rotor shape as shown in Fig. 3.7c. The airgap length of
ICS+3rd shaped rotor can be derived from the integration of (3.3) as:

pffiffiffi Z12
5sp

3 lgd
lavgICS þ 3rd ¼ 10sp

 dh ¼ 1:175lgd ð3:6Þ
p
12 0 cos sp h  16 cos 3 spp h

Therefore, the average airgap length has been reduced compared with the ICS
shaped rotor IPM machine and thus the airgap flux density can be improved.
When the saturation and flux leakage are neglected, the average airgap flux
density can be expressed as:
48 3 Average Torque Improvement of Three Phase …

Fig. 3.7 Different rotor

shapes

(a) Conventional rotor

(b) ICS rotor

(c) ICS+3rd rotor.

Br
Bavg ¼ Ag lavg
ð3:7Þ
Am þ lr lm

The peak to peak open-circuit flux-linkage is given by:

pDL
Up ¼ Bavg ð3:8Þ
2p
3.4 Simplified Analytical Analysis of Average Torque Improvement 49

The average phase back-EMF due to the variation of flux linkage can be
expressed as:

2NUp
Eavg ¼ Kdpn 2p
¼ 2Kdpn NUp pxr ¼ Kdpn NBavg DLxr ð3:9Þ
2pxr

Hence, the average torque can be obtained as:

P Eavg I p
Tavg ¼ ¼ ¼ Kdpn D2 LBavg ð3:10Þ
xr xr 2

Therefore, compared with the IPM machine with conventional rotor, the average
torque is dependent on the variation of average airgap flux density, which is given
by:

ðkair 1Þlgd
lr lm
DBavg ¼ Ag kair lgd
 100% ð3:11Þ
Am þ l r lm

Hence, the average torque reduction is given by:

ðkair 1Þlgd
lr lm
DTavg ¼ Ag kair lgd
 100% ð3:12Þ
Am þ lr lm

The average airgap lengths of three different IPM rotors and the variation of
average torque are listed in Table 3.2. It can be observed that the average torque
reduction in ICS shaped rotor is higher than that of ICS+3rd shaped rotor machine,
i.e., the average torque of ICS+3rd shaped rotor machine has been improved
compared with that of ICS shaped rotor IPM machine. The torque improvement of
the ICS+3rd shaped rotor machine is due to the decrease of the airgap length
compared with that of ICS shaped rotor machine. Therefore, without any modifi-
cation of the rare earth permanent magnet employed for ICS shaped IPM rotor
machine, the average torque of the machine of an inverse cosine injected with an
optimal 3rd harmonic shaped rotor can be enhanced.

Table 3.2 Average airgap Conventional ICS ICS+3rd

length and torque reduction rotor rotor rotor
compared with conational
rotor lavg (mm) 0.50 0.77 0.59
ΔTavg 100 94.63 98.22
(%)
50 3 Average Torque Improvement of Three Phase …

3.5 Finite Element Analysis of Electromagnetic

Performance

The foregoing mentioned non-overlapping winding 12-slot/10-pole IPM machines

with different rotor shapes are employed to investigate the effect of rotor shape on
the electromagnetic performance.

3.5.1 Open-Circuit Flux Density Distribution

and Back-EMFs

Figure 3.8 shows the open-circuit airgap flux density distributions for different IPM
machines whilst the airgap flux density is given in Fig. 3.9a. Its harmonics in one
mechanical period are given in Fig. 3.9b and the harmonics multiplied by 5 in
mechanical period will be divided by 5 in one electrical period, for example, the 5th
harmonic will be fundamental one. Figure 3.9c can be directly obtained from
Fig. 3.9b. It can be seen that the fundamental airgap flux density of ICS shaped
rotor has been reduced due to the increase of airgap length, compared with that of
IPM machine with the conventional rotor. It means that the torque density of IPM
machine with ICS shaped rotor is lower than that of IPM machine with the con-
ventional rotor. However, such decrease in the airgap flux density can be coun-
terbalanced by injecting the 3rd harmonic into the ICS shaped rotor, i.e. ICS+3rd
shaped rotor, as shown in Fig. 3.9c. On the other hand, other harmonics of ICS
shaped rotor IPM machine are also reduced, resulting in low torque ripple and this
will be in detail analyzed in the following part. As can be seen, the 7th, 9th and 11th
harmonics in the conventional IPM machines are significantly reduced in ICS+3rd
shaped one. In addition, for ICS+3rd shaped rotor IPM machine, other harmonics
except for 3rd harmonic are not increased compared with ICS shaped rotor and this
will not deteriorate the torque ripple.
The waveforms of back-EMFs in the three IPM machines are given Fig. 3.10. It
can be seen that the conventional rotor IPM machine has the highest amplitude but
the lowest amplitude exists in ICS shaped rotor machine for both phase and
line-line back-EMFs. However, injecting the 3rd harmonic to the ICS shaped rotor,
both the phase and line-line back-EMFs have been increased. In order to compare
its harmonics, the harmonic contents have been normalized to the fundamental one,
as given in Table 3.3. It can be observed that the 5th, 11th, and 13th harmonics are
reduced significantly for both ICS shaped rotor and ICS+3rd shaped rotor IPM
machine. However, it should be noted that the ICS+3rd shaped rotor IPM machine
has the higher fundamental phase and line-line back-EMFs compared with ICS
shaped rotor IPM machine. Meanwhile, the 3rd harmonic back-EMF increases for
ICS+3rd shaped rotor IPM machine. However, it will not cause any undesirable
effect under balanced operation for 3-phase machines with star-connection
3.5 Finite Element Analysis of Electromagnetic Performance 51

Fig. 3.8 Flux distributions of

12-slot/10-pole IPM machines

(a) Conventional rotor

(b) ICS rotor.

(c) ICS+3rd rotor.

52 3 Average Torque Improvement of Three Phase …

(a) Waveforms

(b) Harmonics from (a)

(c) Harmonics from (b)

Fig. 3.9 Variations of airgap flux density distributions with rotor position
3.5 Finite Element Analysis of Electromagnetic Performance 53

Fig. 3.10 Back-EMFs (400 rpm)

Table 3.3 Harmonics in back-EMFs (400 rpm)

Conventional rotor (%) ICS rotor (%) ICS+3rd rotor (%)
Phase Line Phase Line Phase Line
1 100 100 100 100 100 100
(5.32V) (9.21 V) (4.86 V) (8.42 V) (5.05 V) (8.75 V)
3 1.21 0 2.23 0 2.72 0
5 1.05 1.04 0.71 0.70 0.66 0.67
7 0.01 0.01 0.32 0.32 0.34 0.34
9 1.47 0 0.55 0 0.48 0
11 1.87 1.86 0.35 0.35 1.59 1.57
13 1.64 1.63 0.13 0.13 0.57 0.58

windings because these harmonics can be cancelled in the line-line back-EMFs

[39], as shown in Fig. 3.10 and Table 3.3.

3.5.2 Torque Characteristics

Although the same stator for all the three machines is employed, the peak cogging
torque is not the same due to the influence of rotor shape, as shown in Fig. 3.11.
This is due to the fact that different rotor shape has different airgap flux density
distribution and its harmonic contents (Fig. 3.9a, b). However, the period of the
cogging torque is the same for all machines since they have the same smallest
common multiple between the stator slot numbers and the rotor poles. It can be also
observed that the cogging torque of conventional rotor IPM machine and ICS
shaped rotor IPM machine is the highest and the lowest, respectively, among all the
machines. Compared with ICS shaped rotor IPM machine, the undesirable effect is
54 3 Average Torque Improvement of Three Phase …

Fig. 3.11 Variation of cogging torque with rotor position

slight increase of cogging torque for ICS+3rd shaped rotor one. Although its value
of ICS+3rd shaped rotor one is higher than the ICS shaped one, it is still low (peak
value 0.036 Nm and 0.4% of rated torque) and can be negligible in some high
performance applications, such as electric power steering systems.
Figure 3.12 shows the variation of torque with rotor position for the three
machines when the supplied three-phase currents are in phase with its phase
back-EMFs. It shows that the conventional rotor IPM machine has the highest
average torque but also the highest torque ripple, and ICS shaped rotor IPM
machine has the lowest average torque and its torque ripple is also very low. When
the supplied three-phase currents are in phase with its phase back-EMFs, the
average torque Tavg for ICS+3rd shaped rotor IPM machine has been improved by
>5% compared with that of ICS shaped rotor IPM machine, as shown in Table 3.4.

Fig. 3.12 Variation of torque with rotor position (Ipeak = 56 A, current angle = 0)
3.5 Finite Element Analysis of Electromagnetic Performance 55

Table 3.4 Torque components

Conventional rotor ICS rotor ICS+3rd rotor
6th (Nm) 0.12 0.11 0.09
12th (Nm) 0.23 0.03 0.05
Tave (Nm) 9.31 8.40 8.83
DTpeak-peak (Nm) 0.61 0.23 0.23
Tripple (%) 6.53 2.69 2.65
Tcogging (Nm) 0.170 0.001 0.036
Max-Tavg/A (Nm) 9.66 8.69 9.22

Meanwhile, the torque ripples of both machines are almost the same, but lower than
that of the conventional rotor IPM machine. This is due to the fact that the 5th, 11th,
and 13th harmonics are reduced significantly (Fig. 3.10 and Table 3.3) for both ICS
shaped rotor and ICS+3rd shaped rotor IPM machine, which will interact with the
sinusoidal current causing 6th and 12th harmonics torque, as given in Table 3.4.
The cogging torque for ICS shaped rotor and ICS+3rd shaped rotor IPM machines
are also lower than that of conventional rotor IPM machine. Low harmonics in the
back-EMF together with low cogging torque result in low torque ripple. Most
importantly, the average torque has been improved but the torque ripple is almost
the same.
Saliency ratio is an important parameter for improving the output torque and
extending flux-weakening operation region. Under the condition of the same stator
and winding topologies, the rotor shape modification will cause the variation of
flux-linkage and d- and q-axis inductances. Compared with the conventional rotor
IPM machine, the ICS shaped rotor IPM machine increases the effective airgap
length along q-axis, and thus the PM flux-linkage reduces and meanwhile q-axis
inductance is reduced. In contrast, ICS+3rd shaped rotor machine has reduced the
airgap length along q-axis and slightly increased it along d-axis. Therefore, its
effects on PM flux-linkage and inductance variation are in converse way. Table 3.5
shows the PM flux-linkage, d- and q-axis inductances, and saliency ratio. As can be
seen, the PM flux-linkage and saliency of both ICS shaped rotor and ICS+3rd
shaped rotor machines have been reduced compared with that of conventional rotor
IPM machine. However, compared with ICS shaped rotor IPM machine, the PM
flux-linkage and saliency ratio have been improved from 0.022 to 0.023 Wb, and
1.156 to 1.352. Therefore, both PM flux-linkage and saliency ratio improvements
further result in >6% (maximum torque/ampere designated by Max-Tavg/A) average
torque compared with ICS shaped rotor, as given in Table 3.4. The current angle of
maximum torque per ampere is determined by the PM flux-linkage and saliency
ratio, and therefore, the variation of average torque with current angle is given in
Fig. 3.13. It can be found that its current angles are 12°, 12°, and 13°, respectively.
Figure 3.14 shows the variation of torque with rotor position using the maximum
torque per ampere control [5]. Again, the ICS shaped and ICS+3rd shaped rotor
56 3 Average Torque Improvement of Three Phase …

Table 3.5 Inductance, Conventional ICS ICS+3rd

saliency ratio, PM rotor rotor rotor
flux-linkage
Wm 0.024 0.022 0.023
(Wb)
Lq (mH) 0.481 0.419 0.487
Ld (mH) 0.339 0.362 0.360
q 1.419 1.156 1.352

Fig. 3.13 Variation of torque with current angle (Ipeak = 56 A)

Fig. 3.14 Variation of torque with rotor position at maximum torque/ampere (Ipeak = 56 A)

IPM machines still have lower torque ripple than the conventional one and the ICS
+3rd shaped IPM machine has higher average torque than ICS shaped machine.
3.6 Experimental Verification and Discussions 57

3.6 Experimental Verification and Discussions

The 12-slot/10-pole IPM machines with conventional and ICS+3rd shaped rotors
have been prototyped and tested for validation. Figure 3.15 shows the pictures of
the same stator together with the conventional and ICS+3rd shaped rotors,

Fig. 3.15 Prototypes of

12-slot/10-pole IPM machine

(a) Stator

(b) Conventional rotor

(c) ICS+3rd rotor

58 3 Average Torque Improvement of Three Phase …

respectively. The cogging torque waveform and static torque with current angle of
the prototypes are measured according to the test method reported in [45].
The FEA calculated and measured cogging torque of the IPM machines with
conventional and ICS+3rd shaped rotor are given in Fig. 3.16. It confirms that the
cogging torque could be significantly decreased for the ICS+3rd shaped IPM
machine compared with that of the conventional machine. The variations of FEA
predicted and measured phase- and line-line back-EMFs with rotor position are
compared in Fig. 3.17, whilst Tables 3.6 and 3.7 list the phase back-EMF where
the corresponding dominant harmonics (fundamental, 5th, 7th, 11th, and 13th) are
given. It clearly shows that the 5th, 11th and 13th harmonics are substantially
reduced with slight increase of 7th harmonic. It also confirms that the fundamental

Fig. 3.16 Variation of cogging torque of conventional and ICS+3rd rotor 12-slot/10-pole IPM
machines with rotor position from 2-D and experiment

Fig. 3.17 Variation of phase and line-line back-EMFs of conventional and ICS+3rd shaped rotor
12-slot/10-pole IPM machines from 2-D FEA and experiment (400 rpm)
3.6 Experimental Verification and Discussions 59

Table 3.6 Harmonics in Conventional rotor (%)

back-EMFs of IPM machine
Phase Line-line
with conventional rotor
FE Measured FE Measured
1 100 100 100 100
(5.32 V) (5.22 V) (9.21 V) (9.04 V)
3 1.21 1.25 0 0
5 1.05 0.77 1.04 0.40
7 0.01 0.27 0.01 0.29
9 1.47 1.12 0 0
11 1.87 1.59 1.86 1.45
13 1.64 1.43 1.63 1.43

Table 3.7 Harmonics in ICS+3rd shaped rotor (%)

back-EMFs of IPM machine
Phase Line-line
with ICS+3rd rotor
FE Measured FE Measured
1 100 100 100 100
(5.05 V) (4.93 V) (8.75 V) (8.58 V)
3 2.72 1.91 0 0
5 0.66 0.45 0.67 0.29
7 0.34 0.29 0.34 0.30
9 0.48 0.50 0 0
11 1.59 1.29 1.57 1.13
13 0.57 0.63 0.58 0.72

is slightly reduced as well. The triplen harmonics do not exist in the line-line
back-EMFs because of star-connected three-phase balanced windings. The varia-
tion of static output torque is measured by supplying three-phase windings with
dc-bus current Idc = Ia = −2Ib = −2Ic. Figures 3.18 and 3.19 show the FEA pre-
dicted and measured torque with rotor position and peak torque with current of IPM
machine with conventional rotor and ICS+3rd shaped rotor, respectively. In gen-
eral, good agreements between all FEA and experimental results are achieved.
60 3 Average Torque Improvement of Three Phase …

(a) Conventional rotor

(b) ICS+3rd rotor

Fig. 3.18 Variation of FEA calculated and tested torque with rotor position of conventional and
ICS+3rd shaped rotor 12-slot/10-pole IPM machines (Idc = Ia = −2Ib = −2Ic)
3.7 Summary 61

Fig. 3.19 FEA calculated and tested peak torque with current of conventional and ICS+3rd
shaped rotor 12-slot/10-pole IPM machines

3.7 Summary

This chapter presents a rotor shaping technique for IPM machines with optimal 3rd
harmonic to enhance the average torque. The impact of the optimal 3rd harmonics
to the rotor shape on the electromagnetic performance, including harmonics in the
back-EMFs, average torque, torque ripple and saliency ratio, is investigated in
detail. It is demonstrated that, without any modification of the rare earth permanent
magnet employed for the inverse cosine-shaped IPM machine rotor, the average
torque of the machine of an inverse cosine injected with an optimal 3rd
harmonics-shaped rotor can be enhanced by >6%. Simultaneously, the torque ripple
remains almost unchanged. Finally, the prototype IPM machines with/without 3rd
harmonic are fabricated and experiments are presented to verify the theoretical and
FE analyses.

References

1. Z.Q. Zhu, D. Howe, Electrical machines and drives for electric, hybrid, and fuel cell vehicles.
Proc. IEEE 95(4), 746–765 (2007)
2. G. Pellegrino, A. Vagati, P. Guglielmi, B. Boazzo, Performance comparison between
surface-mounted and interior PM motor drives for electric vehicle application. IEEE Trans.
Ind. Electron. 59(2), 803–811 (2012)
3. W.P. Cao, B.C. Mecrow, G.J. Atkinson, J.W. Bennett, D.J. Atkinson, Overview of electric
motor technologies used for more electric aircraft (MEA). IEEE Trans. Ind. Electron. 59(9),
3523–3531 (2012)
62 3 Average Torque Improvement of Three Phase …

4. T.M. Jahns, G.B. Kliman, T.W. Neumann, Interior permanentmagnet synchronous motors for
adjustable-speed drives. IEEE Trans. Ind. Appl. IA-22(4), 738–747 (1986)
5. S. Morimoto, M. Sanada, Y. Takeda, Wide-speed operation of interior permanent magnet
synchronous motors with high-performance current regulator. IEEE Trans. Ind. Appl. 30(4),
920–926 (1994)
6. S. Kwon, S. Kim, Z. Peng, J. Hong, Performance comparison of IPMSM with distributed and
concentrated windings, in 41st Conference Record IEEE IAS Annual Meeting, Tampa, FL,
USA, 8–12 Oct 2006, vol. 4, pp. 1984–1988
7. L. Chong, R. Dutta, M.F. Rahman, Application of concentrated windings in interior
permanent magnet machine, in Proceedings of Australasian University Power Engineering
Conference, 9–12 Dec 2007, pp. 1984–1988
8. A.M. EL-Refaie, Fractional-slot concentrated-windings synchronous permanent magnet
machines: opportunities and challenges. IEEE Trans. Ind. Electron. 57(1), 107–121 (2010)
9. D.G. Dorrell, M. Hsieh, M. Popescu, L. Evans, D.A. Staton, V. Grout, A review of the design
issues and techniques for radial flux brushless surface and internal rare-earth
permanent-magnet motors. IEEE Trans. Ind. Electron. 58(9), 3741–3757 (2011)
10. L. Ferraris, P. Ferraris, E. Poskovic, A. Tenconi, Theoretic and experimental approach to the
adoption of bonded magnets in fractional machines for automotive applications. IEEE Trans.
Ind. Electron. 59(5), 2309–2318 (2012)
11. E. Fornasiero, N. Bianchi, S. Bolognani, Slot harmonic impact on rotor losses in
fractional-slot permanent-magnet machines. IEEE Trans. Ind. Electron. 59(6), 2557–2564
(2012)
12. Z. Azar, Z.Q. Zhu, G. Ombach, Influence of electric loading and magnetic saturation on
cogging torque, back-EMF and torque ripple of PM machines. IEEE Trans. Magn. 48(10),
2650–2658 (2012)
13. W.Z. Fei, P.C.K. Luk, Torque ripple reduction of a direct-drive permanent-magnet
synchronous machine by material-efficient axial pole pairing. IEEE Trans. Ind. Electron.
58(9), 2601–2611 (2012)
14. M. Dai, A. Keyhani, T. Sebastian, Torque ripple analysis of a PM brushless DC motor using
finite element method. IEEE Trans. Energy Convers. 19(1), 40–45 (2004)
15. M.S. Islam, S. Mir, T. Sebastian, S. Underwood, Design considerations of sinusoidally
excited permanent-magnet machines for low-torque ripple applications. IEEE Trans. Ind.
Appl. 41(4), 955–962 (2005)
16. T.M. Jahns, W.L. Soong, Pulsating torque minimization techniques for permanent magnet AC
motor drives—a review. IEEE Trans. Ind. Electron. 43(2), 321–330 (1996)
17. L. Parsa, H. Hao, Interior permanent magnet motors with reduced torque pulsation. IEEE
Trans. Ind. Electron. 55(2), 602–609 (2008)
18. K.I. Laskaris, A.G. Kladas, Permanent-magnet shape optimization effects on synchronous
motor performance. IEEE Trans. Ind. Electron. 58(9), 3776–3783 (2011)
19. S. Chaithongsuk, B. Nahid-mobarakeh, J. Caron, Optimal design of permanent magnet
motors to improve field-weakening performances in variable speed drives. IEEE Trans. Ind.
Electron. 59(6), 2484–2494 (2012)
20. D.C. Hanselman, Minimum torque ripple, maximum efficiency excitation of brushless
permanent magnet motors. IEEE Trans. Ind. Electron. 41(3), 292–300 (1994)
21. H. Zhu, X. Xiao, Y.D. Li, Torque ripple reduction of the torque predictive control scheme for
permanent-magnet synchronous motors. IEEE Trans. Ind. Electron. 59(2), 871–877 (2012)
22. K. Gulez, A.A. Adam, H. Pastaci, Torque ripple and EMI noise minimization in PMSM using
active filter topology and field-oriented control. IEEE Trans. Ind. Electron. 55(1), 251–257
(2008)
23. I. Beerten, J. Verveckken, J. Driesen, Predictive direct torque control for flux and torque
ripple reduction. IEEE Trans. Ind. Electron. 57(1), 404–412 (2010)
References 63

24. Z.Q. Zhu, S. Ruangsinchaiwanich, N. Schofield, D. Howe, Reduction of cogging torque in

interior-magnet brushless machines. IEEE Trans. Magn. 39(5), 3238–3240 (2003)
25. S.I. Kim, J.Y. Lee, Y.K. Kim, J.P. Hong, Y. Hur, Y.H. Jung, Optimization for reduction of
torque ripple in interior permanent magnet motor by using the Taguchi method. IEEE Trans.
Magn. 41(5), 1796–1799 (2005)
26. K.C. Kim, D.K. Koo, J.P. Hong, J. Lee, A study on the characteristics due to pole-arc to
pole-pitch ratio and saliency to improve torque performance of IPMSM. IEEE Trans. Magn.
43(6), 2516–2518 (2007)
27. S.H. Han, T.M. Jahns, W.L. Soong, M.K. Guven, M.S. Illindala, Torque ripple reduction in
interior permanent magnet synchronous machines using stators with odd number of slots per
pole pair. IEEE Trans. Energy Convers. 25(1), 118–127 (2010)
28. S.H. Han, T.M. Jahns, Z.Q. Zhu, Design tradeoffs between stator core loss and torque ripple
in IPM machines. IEEE Trans. Ind. Appl. 46(1), 187–195 (2010)
29. N. Bianchi, S. Bolognani, D. Bon, M. Dai Pre, Rotor flux-barrier design for torque ripple
reduction in synchronous reluctance and PMassisted synchronous reluctance motors. IEEE
Trans. Ind. Appl. 45(3), 921–928 (2009)
30. H.S. Chen, D.G. Dorrell, M.C. Tsai, Design and operation of interior permanent-magnet
motors with two axial segments and high rotor saliency. IEEE Trans. Magn. 46(9), 3664–
3675 (2010)
31. U.J. Seo, Y.D. Chun, J.H. Choi, P.W. Han, D.H. Koo, J. Lee, A technique of torque ripple
reduction in interior permanent magnet synchronous motor. IEEE Trans. Magn. 47(10),
3240–3243 (2011)
32. D.M. Ionel, M. Popescu, M.I. McGilp, T.J.E. Miller, S.J. Dellinger, Assessment of torque
components in brushless permanent-magnet machines through numerical analysis of the
electromagnetic field. IEEE Trans. Ind. Appl. 41(5), 1149–1158 (2005)
33. D.H. Kim, I.H. Park, J.H. Lee, C.E. Kim, Optimal shape design of iron core to reduce cogging
torque of IPM motor. IEEE Trans. Magn. 39(3), 1456–1459 (2003)
34. J. Junak, G. Ombach, Performance optimisation of the brushless motor with IPM rotor for
automotive applications, in Proceedings of XIX International Conference on Electrical
Machines, 6–8 Sept 2010, pp. 1–5
35. A. Kioumarsi, M. Moallem, B. Fahimi, Mitigation of torque ripple in interior permanent
magnet motors by optimal shape design. IEEE Trans. Magn. 42(11), 3706–3711 (2006)
36. A. Vagati, M. Pastorelli, G. Franceschini, S.C. Petrache, Design of low-torque-ripple
synchronous reluctance motors. IEEE Trans. Ind. Appl. 34(4), 758–765 (1998)
37. S.A. Evans, Salient pole shoe shapes of interior permanent magnet synchronous machines, in
Proceedings of XIX International Conference on Electrical Machines, 6–8 Sept 2010, pp. 1–6
38. K. Wang, Z.Q. Zhu, G. Ombach, W. Chlebosz, Optimal rotor shape with third harmonic for
maximizing torque and minimizing torque ripple in IPM motors, in Proceedings of XX
International Conference on Electrical Machine, 2–5 Sept 2012, pp. 397–403
39. J.A. Houldsworth, The use of harmonic distortion to increase the output voltage of a
three-phase PWM inverter. IEEE Trans. Ind. Appl. IA-20(5), 1224–1228 (1984)
40. P. Pejovic, Z. Janda, An improved current injection network for three-phase
high-power-factor rectifiers that apply the third harmonic current injection. IEEE Trans.
Ind. Electron. 47(2), 497–499 (2000)
41. J.X. Shen, Z.Q. Zhu, D. Howe, Sensorless flux-weakening control of permanent-magnet
brushless machines using third harmonic back EMF. IEEE Trans. Ind. Appl. 40(6), 1629–
1636 (2004)
42. L.B. Zheng, J.E. Fletcher, W.B. Williams, X.N. He, Dualplane vector control of a five-phase
induction machine for an improved flux pattern. IEEE Trans. Ind. Electron. 55(5), 1996–2005
(2008)
43. Y. Li, J.B. Zou, Y.P. Lu, Optimum design of magnet shape in permanent-magnet synchronous
motors. IEEE Trans. Magn. 39(6), 3523–3526 (2003)
64 3 Average Torque Improvement of Three Phase …

44. Z.Q. Zhu, K. Wang, G. Ombach, Optimal magnet shaping with third order harmonic for
maximum torque in SPM machines, in Proceedings of 6th IET Conference on Power
Electronics, Machine and Drives, 27–29 March 2012, pp. 1–6 [CDROM]
45. Z.Q. Zhu, A simple method for measuring cogging torque in permanent magnet machines, in
Proceedings of IEEE Power Energy Society General Meeting, 2009, pp. 1–4
Chapter 4
Third Order Harmonic Utilization
in In-Wheel Machines to Improve
Output Torque

This chapter presents the untilziation of third harmonic in in-wheel machines to

improve the average torque. The optimal value of 3rd harmonic injected into the
sinusoidal PM shape for maximum torque improvement is analytically derived.
Further, the influence of magnet edge thickness on the airgap field distribution is
investigated and utilized to compensate the inter-pole flux leakage and curvature
effect. It is found that the optimal 3rd harmonic is 1/6 of the fundamental one. For
the SPM machines having rotors without shaping, Sine shaping, Sine shaping with
3rd harmonic injected, the electromagnetic performance including the back-EMF
waveforms, cogging torque, average torque and torque ripple are compared. It is
demonstrated that the average torque in the machine of a Sine shaping with an
optimal 3rd harmonic injected can be improved by 10% while the torque ripple is
slightly larger than that of the one with Sine shaping.

4.1 Introduction

Fractional-slot concentrated-winding (FSCW) synchronous permanent magnet

(PM) machines have been gaining interest over the last few years. This is mainly
due to the several advantages that include high-power density, high efficiency, short
end turns, and high slot fill factor particularly when coupled with segmented stator
structures, low cogging torque, flux-weakening capability, and fault tolerance.
Among various types of electric drive train for EVs [1], in-wheel electric motors
have been widely promoted as an alternative propulsion system for future drive-
trains. In particular, direct drive in-wheel motors would result many advantages
such as robustness, high efficiency, increased vehicle space and low maintenance
cost. Furthermore, vehicle manoeuvrability and control can be improved due to
independent control of each wheel. One major drawback of in-wheel drive is the
increase in unsprung mass. Thus, motors high torque density and effective for
vehicle integration are essential for the application. Outer-rotor surface PMSMs
© Springer Nature Singapore Pte Ltd. 2019 65
K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_4
66 4 Third Order Harmonic Utilization …

with fractional-slot and concentrated-winding (FSCW) configuration, which can

deliver the highest torque among the various PMSM configurations with similar
rating, lead themselves as the ideal candidate for the application [2, 3]. Compared
with conventional integral-slot configurations, the FSCW normally has compact
end windings with minimal overhang, coupled with superior electromagnetic and
thermal performance.
For surface-mounted permanent magnet machines, the torque ripple comprises
of the following two main components: (1) cogging torque arising from the PMs’
tendency to align themselves with the minimum reluctance paths formed by varying
positions between the rotor and the stator; (2) torque pulsation due to the interac-
tions between the harmonics of armature currents driven into the machine and the
harmonics of the distorted sinusoidal back EMFs. In addition, magnetic saturations
in the stator and rotor cores will further exasperate the magnitude of the output
torque of the machine [4]. Accordingly, torque ripple minimization techniques can
be categorized into machine design based and machine control based [5]. The
former concerns the optimization of cogging torque and back EMF during the
electromagnetic design stage of the machine, while the latter mainly concerns the
optimal control of armature current during the operation of the machine. Control
methods with varying degrees of sophistication have been developed to mitigate the
torque ripple while driving the PMSMs online [6–12]. On the other hand, a host of
machine design techniques has been proposed to reduce cogging torque, harmonic
contents of the back EMF, and, hence, torque pulsation of the PMSMs, including
magnet shaping [13–16], magnet pole arc width and skewing [17–20], combina-
tions of the stator slot and rotor pole number [21, 22], magnet segmentation [23–
25], magnet pole pairing [26–28]. It is noteworthy, from the techniques surveyed,
that there are invariably tradeoffs torque density and torque ripple in the imple-
mentation of these optimization schemes.
In this chapter, the PM shaping technique with optimal third harmonic is applied
to the rotor of the in-wheel machines to enhance the output torque whilst the
characteristic of low torque ripple. The optimal third harmonic into the PM shaping
will be analytically obtained. Quantitative comparisons between different magnet
shaping methods, i.e., without shaping, Sine shaping, Sine shaping with third har-
monic injected, on electromagnetic performance including the back-EMF wave-
forms, cogging torque, average torque, and torque ripple are carried out. Further, the
influence and utilization of magnet edge thickness which is required in practice on
the airgap field distribution will be investigated to compensate the inter-pole flux
leakage and curvature effect. Experimental validations are reported as well.

4.2 In-Wheel Machines with Various PM-Shaped Rotor

24-slot/28-pole machine with double-layer non-overlapping winding is employed to

investigate the influence of PM shape with 3rd harmonic on the torque enhance-
ment in inwheel machines. Figure 4.1 shows stator structures and winding
4.2 In-Wheel Machines with Various PM-Shaped Rotor 67

connections and rotor with different PM shapes on the surface, i.e., conventional,
sine-shaped and sine+3rd-shaped. It should be noted that unlike the conventional
rotor, the airgap length of machines equipped with sine rotor and sine+3rd rotor
machines is not uniform since the PM thickness is not constant. However, the
effective airgap length which includes the magnets and airgap length is almost the
same due to the PM relative permeability nearly equal to airgap permeability. For
Sine rotor machine, the minimum airgap length is along the d-axis, whilst for Sine
+3rd rotor one, the position of minimum airgap length is determined by the value of
3rd harmonic injected into the PM shape. For all the three machines, the minimum
value of the airgap length and PM maximum thickness are kept to be the same,
0.5 mm and 2.5 mm, respectively. All the stator, rotor, and PM parameters are
given in Table 4.1.

(a) In-wheel machine with conventional rotor

(b) In-wheel machine with sine-shaped rotor (c) Machine with sine+3rd-shaped rotor

Fig. 4.1 24-slot/28-pole in-wheel machines with different PM shapes

68 4 Third Order Harmonic Utilization …

Table 4.1 Main machine Parameter Value

design specifications and
parameters Stator inner diameter (mm) 63
Stator yoke thickness (mm) 2.6
Stator outer diameter (mm) 105
Rotor outer diameter (mm) 126
Maximum PM thickness (mm) 2.5
Teeht width (mm) 5.2
Airgap length (mm) 0.5
Axial length (mm) 50
Magnet remanence (T) 1.2
PM relative permeability 1.05
Stator slot number 24
Pole pair number 14
Parallel branch number 1
Turns in series per phase 160
Rated speed (rpm) 300
Current density (A/mm2) 5

4.3 PM Shape with Optimal 3rd Harmonic

Approximate sinusoidal or sinusoidal with 3rd harmonic airgap flux density dis-
tributions can be achieved by sinusoidally shaping or sinusoidally with 3rd har-
monic shaping the PMs, respectively. Both of them are functions of rotor position h
as illustrated in Fig. 4.2a, b.

4.3.1 Influence of PM Edge Thickness

In order to compensate inter-pole flux leakage and curvature effect, PM edge

thickness Δt as shown in Fig. 4.2a, b is introduced to ease manufacture and is
optimized for achieving nearly sinusoidal and sinusoidal with 3rd harmonic airgap
flux density distribution. Then, the magnet thickness with such compensation is
given by:

hm ðhÞ ¼ Dt þ Dhm ðhÞ ð4:1Þ

The maximum thickness of the magnet is kept to be constant, 2.5 mm. Hence,
Dm will be varied with Dt. It can be seen that when the edge thickness is equal to
3 mm, i.e. Dm = 0 mm, both the airgap length and PM thickness are uniform, i.e.
conventional rotor with pole arc to pole pitch ratio = 1.
4.3 PM Shape with Optimal 3rd Harmonic 69

airgap

Δm
Δt Δhm(θ)

(a) Sine shaped

airgap

Δm
Δt Δhm(θ)

(b) Sine +3rd shaped

Fig. 4.2 Edge thickness compensation for Sine and Sine+3rd shaped PMs

Figure 4.3 shows the FE predicted fundamental airgap flux density and their
THDs for different Δt of Sine shaped PMs. It can be seen that, the conventional PM
shape has the highest fundamental one whilst Δt of 0.5 mm has the lowest total
harmonic distortion of all the investigated PM shapes. Its fundamental value always
reduces with the decrease of Δt while the THD firstly reduces and then increases.
Hence, the optimal Δt of 0.5 mm exists for minimum THD.
In order to further investigate the influence of PM edge thickness on the airgap
flux density of the Sine+3rd shaped PMs with 1/6 of fundamental one, Fig. 4.4
shows the FE predicted fundamental airgap flux density and their THDs for dif-
ferent Δt. The fundamental value and THD always reduce with the decrease of Δt.
Compared to Sine shaped PMs with Δt of 0.5 mm in Fig. 4.3, the THD of Sine+3rd
shaped PMs is slightly high. However, when the edge thickness of Sine+3rd shaped
PMs Δt is equal to 0.5 mm, the torque will be enhanced and torque ripple will not
be deteriorated much compared to that of Sine shaped PMs with Δt of 0.5 mm. This
will be illustrated in the following sections.
70 4 Third Order Harmonic Utilization …

1.4 40
Flux density THD
1.2
30
1.0
Flux density (T)

THD (%)
0.8
20
0.6

0.4
10
0.2

0.0 0
0.0 0.5 1.0 1.5 2.0 Conventional

Fig. 4.3 Variation of fundamental airgap flux density and THD with PM edge thickness for Sine
shaped PM

1.4 50
Flux density THD
1.2
40
1.0
Flux density (T)

THD (%)
30
0.8

0.6
20

0.4
10
0.2

0.0 0
0.0 0.5 1.0 1.5 2.0

Fig. 4.4 Variation of airgap flux density with PM edge thickness for Sine+3rd shaped PM

4.4 Finite Element Analysis of Electromagnetic

Performance

The Sine shaped PMs with Δt = 0.5 mm and the Sine+3rd shaped PMs with
optimal 3rd harmonic of 1/6 of fundamental one and Δt = 0.5 mm are employed in
24-slot/28-pole machines (outer diameter = 126 mm and active axial length = 30
mm) for further analysis in this section.
4.4 Finite Element Analysis of Electromagnetic Performance 71

4.4.1 Open-Circuit Flux Density Distribution

and Back-EMFs

Figure 4.5 shows the FE predicted open-circuit field distributions for the 24-slot/
28-pole in-wheel machines with various PM shapes. Figure 4.6a, b shows the

1.97T
1.67T

1.60T
Flux density (T)
0.0 0.5 1.0 1.5 2.0

(a) Conventional rotor

1.97T
1.43T

1.38T
Flux density (T)
0.0 0.5 1.0 1.5 2.0

(b) Sine rotor

1.96T
1.55T

1.52T
Flux density (T)
0.0 0.5 1.0 1.5 2.0

(c) Sine+3rd rotor

Fig. 4.5 Flux distributions of 24-slot/28-pole SPM machines

72 4 Third Order Harmonic Utilization …

1.5
Sin
Sin+3rd
1.0 Conventional
Flux density (T)
0.5

0.0

-0.5

-1.0

-1.5
0 60 120 180 240 300 360
Rotor position (mech.deg)
(a) Waveforms

1.2

Sin
Sin+3rd
Air flux density (T)

0.8 Conventional

0.4

0.0
1 3 5 7 9 11 13
Harmonic order (n)
(b) Harmonics

Fig. 4.6 Variations of airgap flux density distributions with stator position

airgap flux density distributions and their harmonic contents. As shown in Fig. 4.6b
the fundamental one in the airgap flux density of Sine shaped rotor has been
reduced, compared with that of in-wheel machine with the conventional rotor.
However, injecting the 3rd harmonic into the Sine shaped rotor can compensate part
of the decrease, as shown in Fig. 4.6b. On the other hand, other harmonics of Sine
shaped rotor SPM machine are also reduced.
The FE predicted back-EMFs of the in-wheel machines with various PM shapes
are shown Fig. 4.7. It can be seen that the conventional rotor in-wheel machine has
the highest amplitude but the lowest amplitude exists in Sine shaped PM machine
for the phase back-EMFs. However, injecting the 3rd harmonic to the Sine shaped
PM, both the back-EMFs have been increased by 10.31%. In order to compare its
4.4 Finite Element Analysis of Electromagnetic Performance 73

20

Sin
10 Sin+3rd
Conventional
Back-EMF(V)

-10

-20
0 60 120 180 240 300 360
Rotor position (elec.deg)

Fig. 4.7 Back-EMFs (300 rpm)

Table 4.2 Harmonics in back-EMFs (300 rpm)

Conventional rotor (%) Sine rotor (%) Sine+3rd rotor (%)
1 16.90 (V) 14.55 (V) 16.05 (V)
3 1.50 0.38 0.76
5 0.14 0.04 0.10
7 0.04 0.00 0.05
9 0.10 0.02 0.07
11 0.02 0.01 0.02
13 0.04 0.02 0.00

harmonics, the harmonic contents are given in Table 4.2. It should be noted that the
Sine+3rd shaped rotor SPM machine has the higher fundamental phase back-EMFs
compared with Sine shaped rotor SPM machine.

4.4.2 Torque Characteristics

The cogging torque depends on the smallest common multiple between the slot
number and pole number. Although the same stator for all the three machines is
employed, the peak cogging torque is not the same due to the influence of rotor
shape, as shown in Fig. 4.8. This is due to the fact that different rotor shape results
in different airgap flux density distribution and its harmonic contents (Fig. 4.6a, b).
However, the period of the cogging torque is the same for all machines since they
have the same smallest common multiple between the stator slot numbers and the
rotor poles. It can be also observed that the cogging torque of conventional rotor
SPM machine and Sine shaped rotor SPM machine is the highest and the lowest,
74 4 Third Order Harmonic Utilization …

60

Sin
40
Cogging torque (mN)
Sin+3rd
Conventional
20

-20

-40

-60
0 5 10 15 20 25 30
Rotor position (elec.deg)

Fig. 4.8 Variation of cogging torque with rotor position

respectively, among all the machines. Compared with Sine shaped rotor SPM
machine, the cogging torque in Sine+3rd shaped rotor one is increased.
Figure 4.9 shows the variation of torque with rotor position for the three
machines whilst Table 4.3 shows the torque components, where the supplied
three-phase currents are in phase with its phase back-EMFs. It shows that the
conventional rotor SPM machine has the highest average torque but Sine shaped
rotor SPM machine has the lowest average torque. The average torque for Sine+3rd
shaped rotor SPM machine has been improved by >9% compared with that of Sine
shaped rotor SPM machine. This is due to the variation of average PM thickness
and airgap length for different PM shapes, as shown in Fig. 4.2a, b.
Meanwhile, the torque ripples of Sine shaped and Sine+3rd shaped machines are
lower than that of the conventional rotor SPM machine. This is due to the fact that

7
Sin
Sin+3rd
6 Conventional

5
Torque (Nm)

2
0 60 120 180 240 300 360
Rotor position (elec.deg)

Fig. 4.9 Variation of torque with rotor position, Ipeak = 25 A

4.4 Finite Element Analysis of Electromagnetic Performance 75

Table 4.3 Average airgap Conventional Sine Sine

length, PM height, torque +3rd
components and PM volume
6th (Nm) 0.079 0.021 0.088
12th (Nm) 0.084 0.005 0.005
Tave (Nm) 5.79 4.99 5.49
Tripple (%) 4.49 0.85 3.01
Tcogging (mN) 50.5 0.05 12.09
Volume of PMs 25.05 18.19 14.51
(cm3)

the 5th, 11th, and 13th harmonics in back-EMFs are reduced significantly (Fig. 4.7
and Table 4.2) for both Sine shaped rotor and Sine+3rd shaped rotor SPM machine,
which will interact with the sinusoidal current causing the 6th and 12th harmonics
in torque, as given in Table 4.3. Low harmonics in the back-EMF together with low
cogging torque result in low torque ripple. Most importantly, the average torque has
been improved. However, the torque ripple of the sine+3rd-shaped in-wheel is
larger than the sine+3rd-shaped inner rotor machines.

4.5 Summary

This chapter presents a permanent magnet (PM) shaping technique with optimal 3rd
harmonic to enhance the output torque of in-wheel machines. The influence of
magnet edge thickness on the airgap field distribution is investigated and can be
utilized to compensate the inter-pole flux leakage and curvature effect. It is found
that the optimal 3rd harmonic is 1/6 of the fundamental one. It is demonstrated that
the average torque of the machine of a Sine shaping with an optimal 3rd harmonic
injected can be improved by 10%, while the torque ripple is slightly higher than that
of the one with Sine shaping.

References

1. R. Wrobel, P.H. Mellor, Design considerations of a direct drive brushless machine with
concentrated windings. IEEE Trans. Energy Convers. 23(1), 1–8 (2008)
2. A.M. EL-Refaie, Fractional-slot concentrated-windings synchronous permanent magnet
machines: opportunities and challenges. IEEE Trans. Ind. Electron. 57(1), 107–121 (2010)
3. B.C. Mecrow, J.W. Bennett, A.G. Jack, D.J. Atkinson, A.J. Freeman, Drive topologies for
solar-powered aircraft. IEEE Trans. Ind. Electron. 57(1), 457–464 (2010)
4. M.S. Islam, R. Islam, T. Sebastian, Experimental verification of design techniques of
permanent-magnet synchronous motors for low-torque-ripple applications. IEEE Trans. Ind.
Appl. 47(1), 88–95 (2011)
76 4 Third Order Harmonic Utilization …

5. T.M. Jahns, W.L. Soong, Pulsating torque minimization techniques for permanent magnet ac
motor drives—a review. IEEE Trans. Ind. Electron. 43(2), 321–330 (1996)
6. S. Chen, C. Namuduri, S. Mir, Controller-induced parasitic torque ripples in a PM
synchronous motor. IEEE Trans. Ind. Appl. 38(5), 1273–1281 (2002)
7. P. Mattavelli, L. Tubiana, M. Zigliotto, Torque-ripple reduction in PM synchronous motor
drives using repetitive current control. IEEE Trans. Power Electron. 20(6), 1423–1431 (2005)
8. G.H. Lee, S.I. Kim, J.P. Hong, J.H. Bahn, Torque ripple reduction of interior permanent
magnet synchronous motor using harmonic injected current. IEEE Trans. Magn. 44(6), 1582–
1585 (2008)
9. K. Gulez, A.A. Adam, H. Pastaci, Torque ripple and EMI noise minimization in PMSM using
active filter topology and field-oriented control. IEEE Trans. Ind. Electron. 55(1), 251–257
(2008)
10. C. Ortega, A. Arias, C. Caruana, J. Balcells, G.M. Asher, Improved waveform quality in the
direct torque control of matrix-converter-fed PMSM drives. IEEE Trans. Ind. Electron. 57(6),
2101–2110 (2010)
11. T. Shi, Y. Guo, P. Song, C. Xia, A new approach of minimizing commutation torque ripple
for brushless dc motor based on dc–dc converter. IEEE Trans. Ind. Electron. 57(10), 3483–
3490 (2010)
12. Y.C. Zhang, J.G. Zhu, Direct torque control of permanent magnet synchronous motor with
reduced torque ripple and commutation frequency. IEEE Trans. Power Electron. 26(1), 235–
248 (2011)
13. M.S. Islam, S. Mir, T. Sebastian, S. Underwood, Design considerations of sinusoidally
excited permanent-magnet machines for low-torque ripple applications. IEEE Trans. Ind.
Appl. 41(4), 955–962 (2000)
14. W. Fei, P.C.K. Luk, An improved model for the back EMF and cogging torque characteristics
of a novel axial flux permanent magnet synchronous machine with a segmental laminated
stator. IEEE Trans. Magn. 45(10), 4609–4612 (2009)
15. W. Fei, P.C.K. Luk, Torque ripple reduction of axial flux permanent magnet synchronous
machine with segmented and laminated stator, in Proceedings of IEEE Energy Conversion
Congress and Exposition, 2009, pp. 132–138
16. N.N. Chen, S.L. Ho, W.N. Fu, Optimization of permanent magnet surface shapes of electric
motors for minimization of cogging torque using FEM. IEEE Trans. Magn. 46(6), 2478–2481
(2010)
17. M. Dai, A. Keyhani, T. Sebastian, Torque ripple analysis of a PM brushless dc motor using
finite element method. IEEE Trans. Energy Convers. 19(1), 40–45 (2004)
18. R. Islam, I. Husain, A. Fardoun, K. McLaughlin, Permanent-magnet synchronous motor
magnet designs with skewing for torque ripple and cogging torque reduction. IEEE Trans.
Ind. Appl. 45(1), 152–160 (2009)
19. L. Zhu, S.Z. Jiang, Z.Q. Zhu, C.C. Chan, Analytical methods for minimizing cogging torque
in permanent-magnet machines. IEEE Trans. Magn. 45(4), 2023–2031 (2009)
20. L. Parsa, L. Hao, Interior permanent magnet motors with reduced torque pulsation. IEEE
Trans. Ind. Electron. 55(2), 602–609 (2008)
21. K. Atallah, J. Wang, D. Howe, Torque-ripple minimization in modular permanent-magnet
brushless machines. IEEE Trans. Ind. Appl. 39(6), 1689–1695 (2003)
22. S.H. Han, T.M. Jahns, W.L. Soong, M.K. Guven, M.S. Illindala, Torque ripple reduction in
interior permanent magnet synchronous machines using stators with odd number of slots per
pole pair. IEEE Trans. Energy Convers. 25(1), 118–127 (2010)
23. R. Lateb, N. Takorabet, F. Meibody-Tabar, Effect of magnet segmentation on the cogging
torque in surface-mounted permanent magnet motors. IEEE Trans. Magn. 42(3), 442–445
(2006)
24. B. Boukais, H. Zeroug, Magnet segmentation for commutation torque ripple reduction in a
brushless dc motor drive. IEEE Trans. Magn. 46(11), 3909–3919 (2010)
References 77

25. M. Ashabani, Y.A.-R.I. Mohamed, Multiobjective shape optimization of segmented pole

permanent-magnet synchronous machines with improved torque characteristics. IEEE Trans.
Magn. 47(4), 795–804 (2011)
26. N. Bianchi, S. Bolognani, Design techniques for reducing the cogging torque in
surface-mounted PM motors. IEEE Trans. Ind. Appl. 38(5), 1259–1265 (2002)
27. Y. Wang, M.J. Jin, W. Fei, J.X. Shen, Cogging torque reduction in permanent-magnet
flux-switching machines by rotor teeth axial pairing. IET Elect. Power Appl. 4(7), 500–506
(2010)
28. W. Fei, P.C.K. Luk, A new technique of cogging torque suppression in direct-drive
permanent magnet brushless machines. IEEE Trans. Ind. Appl. 46(4), 1332–1340 (2010)
Chapter 5
Influence of Airgap Flux Density
Waveform on Optimal Split Ratio
and Torque Density of SPM Machines

This chapter presents an unified analytical model for optimizing the split ratios of
surface-mounted permanent magnet (SPM) machines considering various airgap
flux density distributions, i.e., square (also referred to as conventional), sinusoidal
(sine) and sinusoidal with third harmonic (sine+3rd), considering both overlapping
windings and nonoverlapping windings. Besides, torque density under the optimal
split ratios are compared for the conventional, sine and sine+3rd. It is found that
harmonics in airgap flux density will lower the optimal split ratio and their influence
is stronger for nonoverlapping windings compared with overlapping windings. The
sine+3rd overlapping windings machines shows competence in torque density with
higher value of stator slot Ns especially under high flux density ratio, reaching about
99% of the torque of the conventional one. For sine+3rd nonoverlapping windings
machines whose slot number Ns and pole number 2p differed by one, the torque
density is larger than the conventional one. The sine+3rd is promising in torque
density for both overlapping and nonoverlapping winding machines. Finite element
(FE) analysis is carried out on the 12-slot/10-pole SPM machines with conven-
tional, sine-shaped and sine+3rd-shaped PM under different flux density ratio.
Finally, a 12-slot/10-pole sine+3rd-shaped machine is prototyped and tested for
validation.

5.1 Introduction

Permanent magnet machines have gained a great deal of popularity in industry

application due to their remarkable advantages such as high torque density and high
efficiency. Split ratio, i.e. ratio of stator bore diameter to stator outer diameter (Dsi/
Dso) proves to be an important parameter for SPM machines since it has a signif-
icant influence on the torque density and efficiency [1]. Lots of literature focuses on
optimizing the split ratio to maximize the output torque either by analytical methods
or finite element (FE) analysis [2–11]. The optimal split ratio was analyzed in SPM
© Springer Nature Singapore Pte Ltd. 2019 79
K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_5
80 5 Influence of Airgap Flux Density Waveform …

machines where the airgap flux density is assumed to be independent of the split
ratio [2–4]. In [2], the author derived the mathematical expression to determine the
optimal split ratio which indicates that it is related to the air gap flux density
amplitude, the pole pairs and the maximum flux density in the steel lamination
when the copper loss is fixed, whilst the influence of its ratio on the overall cost and
thermal behavior was discussed in [3]. Based on [2, 3], in order to take the influence
of the BLDC or BLAC operations, the stator tooth-tips and the end-windings into
consideration, the optimal split ratio for maximum torque density is analytically
investigated in [4]. Different from the SPM machines, the airgap flux density in
interior permanent magnet (IPM) machines is considered to be a function of the
split ratio. Thus, the optimal split ratio together with the optimal airgap flux density
ratio were analytically derived for IPM machines with non-overlapping windings
[5]. Further, in [6–8], the analytical determination of optimal split ratio was
extended to external rotor SPM machines considering the influence of both fixed
and variable rotor back-iron thickness, curvature effect, slot number and pole
number combinations, tooth-tip, and end-winding length. Since only the copper loss
was fixed when deriving the optimal split ratio in [2–8], the authors in [9] expanded
the current knowledge about split ratio by optimizing the split ratio considering
both the copper loss and current density limitation. It should be noted that in [2–9],
the iron loss is neglected which is reasonable in low rated speed, whilst in [10], the
author investigated the influence of iron loss on the optimal split ratio for maximum
efficiency, with particular emphasis on high-speed machines.
In order to reduce the torque ripple, the airgap flux density are commonly
designed to be sinusoidal or sinusoidal with harmonics [11–14]. As pointed in [4],
the airgap flux density distribution has a significant influence on the optimal split
ratio and torque density. In [13], the comparison of torque density were made
among 12-slot/10-pole nonoverlapping winding SPM machines with three various
airgap flux density, i.e., conventional, sinusoidal and sinusoidal with third harmonic
injected whose peak values are kept to be the same whilst the same comparison was
conducted for five phase SPM machines in [14]. However, such comparison was
carried out under the same split ratio which means that the stator iron saturation
level were different for the three. In addition, it was restricted to the machines with
nonoverlapping windings. Therefore, to make a fair comparison, it is necessary to
investigate the influence of different airgap flux density waveforms on the optimal
split ratio and compare the torque density under their optimal split ratios consid-
ering both overlapping and nonoverlapping windings.
The purpose of this chapter is to investigate the influence of airgap flux density
waveforms on the optimal value of split ratio and compare their torque density for
SPM machines with both overlapping and nonoverlapping windings. General tor-
que density expression is derived for three different airgap flux density distribu-
tions, i.e., square waveform (conventional), sinusoidal waveform (sine) and
sinusoidal waveform with optimal third harmonic (sine+3rd) by introducing two
coefficients Ka and Kb. Then the optimal split ratio is obtained for both the
machines with overlapping and nonoverlapping windings based on a simplified
stator model and the torque density comparison is carried out for the three
5.1 Introduction 81

machines. Furthermore, the SPM machines with conventional, sine-shaped and sine
+3rd-shaped PM are designed to obtain the square, sinusoidal and sinusoidal with
third harmonic airgap flux density waveforms, respectively. Further, FE analysis is
carried out on the 12-slot/10-pole machines under different flux density ratios to
verify the analytical analysis. Finally, a 12-slot/10-pole sine+3rd-shaped SPM
machine with optimal split ratio is prototyped and measured to validate the
conclusion.

5.2 General Torque Density Equation

Figure 5.1 shows the waveforms of three idealized airgap flux density distributions,
i.e. square (conventional), sinusoidal (sine) and sinusoidal with optimal third har-
monic injected (sine+3rd). To give a unified torque density equation for machines
with the three airgap flux density distributions, Ka and Kb are introduced. Kc is also
introduced to facilitate to calculate stator slot area. Their definitions are given as
below.

Bg1
Ka ¼ ð5:1Þ
Bg av

Bg av
Kb ¼ ð5:2Þ
Bgm
( )
Ns
R
max pDs Bg dh
maxfBt av g slot pitch
Kc ¼ ¼ ð5:3Þ
Bgm Bgm

where Bg1 is the fundamental airgap flux density, Bg_av is the average airgap flux
density in one pole pitch, Bgm is maximum airgap flux density, and Bt_av is the
average airgap flux density in one tooth pitch.
Table 5.1 gives the values of Ka, Kb, and Kc for conventional, sine and sine+3rd
airgap flux density distributions.
Consider the machines are supplied by three phase symmetrical sinusoidal
currents, the general torque equation can be given as

3p
T ¼ pffiffiffi Kdp1 Ka Kb Lef Ds Bgm Nw Ia ð5:4Þ
2 2

where Kdp1, Lef, Ds, Nw, Ia, are the fundamental winding factor, the stator active
axial length, the stator bore diameter, the number of turns in series per phase, the
RMS phase current respectively.
82 5 Influence of Airgap Flux Density Waveform …

Fig. 5.1 Idealized airgap flux density distributions

The copper loss neglecting the influence of end windings is given by the fol-
lowing equation:
5.2 General Torque Density Equation 83

Table 5.1 Values of Ka, Kb Conventional Sine Sine+3rd

and Kc
Ka 4/p p/2 9p/19
Kb 1 2/p 1.154 * 19/(9p)
 
Kc 1 sinðNpps Þ pp
sinð3 Þ
pp 1:154 sinðNpps Þ 116 Ns
pp
Ns 3 sinð Þ
Ns
pp
Ns

Lef
Pcu ¼ 36Nw2 Ia2 q ð5:5Þ
As Ks Ns

where q, As, Ks, Ns are the resistivity of copper, the area of one slot, the slot packing
factor, the number of stator slots.
Equation (5.5) can be rewritten as Ia = f(Pcu) assuming the copper loss is an
constant, substituting it into Eq. (5.4), the torque density expression can be derived as
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
T Pcu As Ks pffiffiffiffiffiffiffiffiffiffi
¼ 2
Ka Kb cBmax k Ns As
V 2qlef Dso
sffiffiffiffiffiffiffiffiffiffiffiffi ð5:6Þ
4Ns As
¼ CKa Kb ck
D2so
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Pcu As Ks
C¼ Bmax ð5:7Þ
8qlef
sffiffiffiffiffiffiffiffiffiffiffiffi
T 4Ns As
TD ¼ ¼ Ka Kb ck ð5:8Þ
VC D2so

where Dso is the stator outer diameter, k is the split ratio (Dsi/Dso), c is flux density
ratio (Bgm/Bmax) and Bmax is the maximum flux density in the stator lamination
considering the magnetic saturation, which is usually chosen around the knee point
of the steel amination [6]. Equation (5.8) is obtained for simplicity. It is clear to see
that TD is closely related to k, c and As.

5.3 Derivation of Optimal Split Ratio for Maximum

Torque Density

5.3.1 Calculation of Slot Area

Figure 5.2 shows the simplified open circuit flux distributions for machines with
both overlapping and nonoverlapping windings. Figure 5.3 depicts the shape of the
84 5 Influence of Airgap Flux Density Waveform …

S N S

(a) overlapping winding

S N S

(b) nonoverlapping winding

Fig. 5.2 Idealized open-circuit flux distributions

Fig. 5.3 Simplified stator

teeth shape with parallel sides
Yw

Tw

Hw

simplified stator teeth with parallel sides to take the influence of tooth-tips into
consideration. It is assumed that for overlapping windings, the flux in the stator
back-iron is half of the airgap flux per pole whilst for nonoverlapping windings, the
width of stator yoke Yw is half of stator tooth wide Tw. The expressions of Yw and Tw
are given by (5.9) and (5.10). It should be noted that both Yw and Tw are functions
of k and c, so the slot area is also a function of k and c. The relationship between the
torque density and the split ratio can be established, and the optimal split ratio for
maximum torque per volume is obtained analytically for a fixed machine outer
diameter and constant copper loss.
5.3 Derivation of Optimal Split Ratio for Maximum Torque Density 85

pDso
Tw ¼ Kc ck ð5:9Þ
Ns
(
Kb ck pD so
overlapping
Yw ¼ 4p
ð5:10Þ
Kc ck pD
2Ns
so
nonoverlapping

The slot area taking the tooth-tips width Hw into consideration can be expressed
as
p h i
As ¼ ðDso  2Yw Þ2 D2si
4Ns
   
Dso  Dsi pDsi
 Tw  Yw  Hw  Tw ð5:11Þ
2 Ns
2  
pDso
¼ ak2 þ bk þ c
4Ns
8 2 2

> p K
< 4p2b þ pKpb Kc c2 þ 2Kc c  1 overlapping
a¼

>
: p p þ 2 ðKc cÞ2 þ 2ðKc cÞ  1 nonoverlapping
Ns Ns
8

>
<  pKp b þ 2Kc ð1  rÞ c  2r overlapping
b¼
ð5:12Þ
>
: 2 Np þ 1  r ðKc cÞ  2r nonoverlapping
s

c¼1
pD2o  2 
¼ ak þ bk þ c
4Ns

2Hw
r¼ ð5:13Þ
Ds

5.3.2 Optimal Split Ratio

Substituting (5.11) into (5.8), the relationship between TD and k can be obtained.
Differentiate TD with respect to k, and equate it to zero, the optimal value of split
ratio kop can be derived, which is given by:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
3b  9b2  32ac
kop ¼ ð5:14Þ
8a
86 5 Influence of Airgap Flux Density Waveform …

As can be seen from Eq. (5.14), the optimal value of split ratio is influenced by
Kb, Kc (determined by the airgap flux density waveforms), c, Ns and r. However,
the influence of r is not discussed in this chapter, the value of Hw is assumed to be
an constant, 1 mm.
Figure 5.4 shows variations of optimal split ratio with flux density c for both
overlapping and nonoverlapping windings with conventional, sine and sine+3rd
airgap flux density distributions. It can be seen that the optimal value of split ratio
always reduces with the increase of flux density ratio. The reason is that wider stator
yoke and teeth should be designed with higher c to maintain the same level of
saturation and the stator bore diameter reduces, so is the split ratio. The sine airgap
flux density has the largest optimal split whist the conventional one is the smallest
for both winding dispositions. From Fig. 5.4a–c, it can be seen that for overlapping
windings, the difference of optimal split ratio of the sine and sine+3rd is becoming
smaller with the increase of Ns, whilst the influence of Ns on the optimal split ratio
of conventional waveform is much smaller. From Fig. 5.4d–f, it is noted that for
nonoverlapping windings, the Ns has the most influence on the optimal split ratio of
the conventional. With the increase of Ns, the gap between the optimal value of split
ratio of conventional and sine+3rd tends to be smaller. Compare Fig. 5.4a–c with
Fig. 5.4d–f, it is found that the optimal split ratio difference between sine and sine
+3rd for nonoverlapping winding is larger than that for the overlapping winding,
which indicates that the third harmonic in airgap flux density has a greater influence
on the optimal split ratio for nonoverlapping winding, and the torque density
comparison of sine and sine+3rd under their optimal split ratio should be discussed.

5.3.3 TD Comparison Under Optimal Split Ratio

Variation of torque density with flux density ratio c for under optimal split ratio for
various airgap flux density waveforms with both overlapping and nonoverlapping
windings are pictured in Fig. 5.5. It can be seen from Fig. 5.5a–c that the sine+3rd
can achieve >90% of the torque density of the conventional whilst the sine witness
about 20% reduction of torque density compared with the conventional. The higher
the flux density ratio, the less difference of torque density between sine and sine
+3rd. With the number of stator slots increasing, the torque density of sine+3rd is
approaching to the conventional especially under high flux density ratio, for
example, for the 72-slot/8-pole machines, the torque density of sine+3rd is about
99% of that of the conventional when the flux density ratio is 0.8. From Fig. 5.5d–f,
it can be seen that for the 9-slot/10-pole, the torque density of sine+3rd is the
highest among the three. For the 12-slot/10-pole, the sine+3rd has the highest
torque density when flux density ratio c > 0.8 whilst for the 15-slot/10-pole, the
torque density of the sine+3rd is in the middle of the sine and conventional. With
the increase of Ns, the torque density of the conventional increase rapidly whilst that
of the sine and sine+3rd increase slightly. The sine+3rd shows more competence in
torque density for 9-slot/10-pole machines.
5.3 Derivation of Optimal Split Ratio for Maximum Torque Density 87

0.7
0.6

Optimal split ratio

0.5
0.4
0.3
Conventional
0.2 Sine
0.1 Sine+3rd
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(a) Overlapping winding (24-slot/8-pole).

0.7
0.6
Optimal split ratio

0.5
0.4
0.3
Conventional
0.2 Sine
0.1 Sine+3rd
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(b) Overlapping winding (48-slot/8-pole).

0.7
0.6
Optimal split ratio

0.5
0.4
0.3
Conventional
0.2 Sine
0.1 Sine+3rd
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(c) Overlapping winding (72-slot/8-pole).

Fig. 5.4 Variation of optimal split ratio with flux density ratio
88 5 Influence of Airgap Flux Density Waveform …

0.7
0.6

Optimal split ratio

0.5
0.4
0.3
Conventional
0.2 Sine
0.1 Sine+3rd
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(d) Nonoverlapping winding (9-slot/10-pole).

0.8
0.7
Optimal split ratio

0.6
0.5
0.4
0.3 Conventional
0.2 Sine
0.1 Sine+3rd
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(e) Nonoverlapping winding (12-slot/10-pole).

0.7
0.6
Optimal split ratio

0.5
0.4
0.3
Conventional
0.2 Sine
0.1 Sine+3rd
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(f) Nonoverlapping winding (15-slot/10-pole).

Fig. 5.4 (continued)

5.3 Derivation of Optimal Split Ratio for Maximum Torque Density 89

0.30

0.25

TD 0.20

0.15

0.10 Conventional
Sine
0.05 Sine+3rd
0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(a) Overlapping winding (24-slot/8-pole).

0.30

0.25

0.20
TD

0.15

0.10 Conventional
Sine
0.05 Sine+3rd
0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(b) Overlapping winding (48-slot/8-pole).

0.30

0.25

0.20
TD

0.15

0.10 Conventional
Sine
0.05 Sine+3rd
0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(c) Overlapping winding (72-slot/8-pole).

Fig. 5.5 Variation of TD with flux density ratio under optimal split ratio
90 5 Influence of Airgap Flux Density Waveform …

0.35
0.3
0.25
0.2
TD

0.15
Conventional
0.1 Sine
0.05 Sine+3rd
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(d) Nonoverlapping winding (9-slot/10-pole).

0.3

0.25

0.2
TD

0.15

0.1 Conventional
Sine
0.05 Sine+3rd
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(e) Nonoverlapping winding (12-slot/10-pole).

0.30

0.25

0.20
TD

0.15

0.10 Conventional
Sine
0.05 Sine+3rd
0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Flux density ratio γ
(f) Nonoverlapping winding (15-slot/10-pole).

Fig. 5.5 (continued)

5.4 FE Analysis of Optimal Split Ratio and Torque 91

5.4 FE Analysis of Optimal Split Ratio and Torque

To verify the foregoing discussions, 12-slot/10-pole SPM machines are used for
2-D FE analysis. To obtain the square, sinusoidal and sinusoidal with third har-
monic airgap flux density waveforms, the conventional (fully pitched tile type),
sine-shaped and sine+3rd-shaped PM are utilized. Sine-shaped and sine
+3rd-shaped PM are shown in Fig. 5.6. It should be noted that the maximum PM
thickness, being 3 mm, and the minimum airgap length, being 1 mm, is kept the
same for the three PM shapes. Three various airgap flux density waveforms without
considering the slotting effects and the harmonic contents are depicted in Fig. 5.7.
The detailed parameters of the machine are listed in Table 5.2. The values of
optimal split ratio for the conventional, sine and sine+3rd are obtained through FE
analysis with copper loss is fixed to be 33 W. Only two flux density ratio, i.e.,
c = 0.5 (Bgm = 0.75) and c = 0.8 (Bgm = 1.2) are chosen, which can be obtained by
designing lg and Br to be 1 mm and 1.0 T, 0.5 mm and 1.4 T respectively.
Variations of average torque with split ratio for three airgap flux density
waveforms under c = 0.5 and c = 0.8 are shown in Fig. 5.8. Compared with
Fig. 5.4e, it can be noted that the FE obtained optimal values of split ratio are
consistent with that from analytical results for both c = 0.5 and c = 0.8, as also can
be seen from Table 5.3. Under c = 0.5 and optimal split ratio, the conventional has
the highest average torque whilst the sine had the lowest. Under c = 0.8 and

(a) Sine shaped PM

(b) Sine+3rd shaped PM

Fig. 5.6 Different PM shapes

92 5 Influence of Airgap Flux Density Waveform …

1.0
Conventional

Airgap flux density (T)

Sine
0.5
Sine+3rd

0.0

-0.5

-1.0
0 60 120 180 240 300 360
Rotor position (elec.deg)
(a) Waveform

1.0
Conventional
Airgap flux density (T)

0.8
Sine
Sine+3rd
0.6

0.4

0.2

0.0
1 2 3 4 5 6 7 8 9 10 11 12 13
Harmonic order
(b) Harmonic contents

Fig. 5.7 Airgap flux density waveforms and harmonic contents (stator slotting is neglected)

optimal split ratio, the torque of the sine+3rd is about 99.5% of the conventional. It
can be concluded that the sine+3rd can be promising in torque density under higher
flux density ratio for 12-slot/10-pole machines.
Figure 5.9 shows flux density distributions for three machines with optimal k
under c = 0.5 and c = 0.8 when flux linkage of phase A is maximum. It can be seen
that the maximum flux density in stator lamination is around 1.5 T, which is below
the knee point of lamination steel B-H curve, 1.6 T. The saturation level of the
machines is low.
5.5 Experimental Verification 93

Table 5.2 Machine specifications and parameters

Parameters Value Unit
Phase number (m) 3
Slot number (Ns) 12
Pole number (2p) 10
Rated speed (r) 400 rpm
Copper loss (Pcu) 33 W
Slot fill factor (Ks) 0.4
Turns in series per phase (Nw) 48
Stator outer diameter (Dso) 90 mm
Tooth-tip width (Hw) 1 mm
Axial length (Lef) 50 mm
Minimum airgap length (lg) 1/0.5 mm
Maximum PM thickness (lm) 3 mm
Pole arc to pole pitch ratio 1
PM remanence (20 °C) (Br) 1.0/1.4 T
Relative recoil permeability (lr) 1.05
Maximum stator lamination flux density (Bmax) 1.5 T
Maximum airgap flux density (Bgm) 0.75/1.2 T

5.5 Experimental Verification

The 12-slot/10-pole SPM machine with sine+3rd-shaped rotor has been prototyped
and tested for validation. The stator and the sine+3rd-shaped rotor are pictured in
Fig. 5.10. Its stator outer diameter (Dso) and active axial length (Lef) are 90 and
50 mm respectively. It is designed with the optimal split ratio (k = 0.6,
Dsi = 27 mm) with flux density ratio is 0.5. The test platform is shown in Fig. 5.11.
Figure 5.11a shows variation of the FE predicted and measured cogging torques
with rotor position. Although the peak value of cogging torque is slightly higher
than the FE predicted one probably because of the manufacturing tolerances and
assembly deficiencies. The agreements between the FE predicted and measured of
cogging torque are good. The variation of static output torque is measured by
supplying three-phase windings with dc-bus current Idc = Ia = −2Ib = −2Ic with
the copper loss are fixed at 33 W (rated condition) and 132 W (overload condition).
Figure 5.11b shows the FEA predicted and measured torque with rotor position and
in general, good agreements between all FEA and experimental results are
achieved.
94 5 Influence of Airgap Flux Density Waveform …

4
Average Torque (Nm)
3

2 0.58 0.6 0.62

Conventional
1 Sine
Sine+3rd
0
0.3 0.4 0.5 0.6 0.7 0.8
Split ratio
(a) γ = 0.5

5
Average Torque (Nm)

2 Conventional 0.47 0.5 0.55

Sine
1
Sine+3rd
0
0.2 0.3 0.4 0.5 0.6 0.7
Split ratio
(b) γ = 0.8

Fig. 5.8 FE calculated torque with split ratio for 12-slot/10-pole machines with different flux
density ratio c

Table 5.3 Comparison of Analytical/FE

optimal split ratio
c = 0.5 Conventional 0.581/0.58
Sine 0.617/0.62
Sine+3rd 0.597/0.6
c = 0.8 Conventional 0.478/0.47
Sine 0.553/0.55
Sine+3rd 0.512/0.50
5.5 Experimental Verification 95

2.0 1.6 1.2 0.8 0.4 0

B (T)

(a) Conventional with (b) Sine with λ=0.62 (c) Sine+3rd with λ=0.6
λ=0.58
(i) γ = 0.5

(a) Conventional with (b) Sine with λ=0.55 (c) Sine+3rd with λ=0.5
λ=0.47
(ii) γ = 0.8

Fig. 5.9 Flux density distributions under different flux density ratio c with optimal k when flux
linkage of phase A is maximum
96 5 Influence of Airgap Flux Density Waveform …

Fig. 5.10 Prototypes and test

platform of 12-slot/10-pole
Sine+3rd-shaped PM machine

(a) Stator

(b) Rotor

Prototype

Balanced beam

Lathe

Digital scale

(c) Test platform

5.6 Summary 97

20
15 FE Measured

Cogging Torque (mNm) 10

5
0
-5
-10
-15
-20
0 1 2 3 4 5 6
Rotor position (mech.deg)
(a) Cogging torque

12
FE
8 Pcu=132W Measured
Torque (Nm)

4
Pcu=33W
0

-4

-8

-12
0 60 120 180 240 300 360
Rotor position (elec.deg)

(b) Static torque (Ia=-2Ib=-2Ic)

Fig. 5.11 Comparison of FE predicted and measured results

5.6 Summary

Optimal split ratio and torque density of SPM with three various airgap flux density
distributions, i.e., conventional, sine and sine+3rd, are compared in this chapter
with both overlapping and nonoverlapping winding configuration taken into con-
sideration. Harmonics in airgap flux density tends to lower the value of split ratio
especially for the nonoverlapping windings. Under the optimal split ratio, the sine
+3rd overlapping windings machines shows competence in torque density with
higher value of stator slot Ns especially under high flux density ratio, reaching about
99% of the torque of the conventional whilst for sine+3rd nonoverlapping windings
98 5 Influence of Airgap Flux Density Waveform …

machines whose slot number and pole number differed by one, the torque density is
larger than the conventional. Since the torque ripple of the sine+3rd is much smaller
than the conventional, it is promising in both overlapping and nonoverlapping
winding configurations.

References

1. V.B. Honsinger, Sizing equations for electrical machinery. IEEE Trans. Energy Convers. 2
(1), 116–121 (1987)
2. F.B. Chaaban, Determination of the optimum rotor/stator diameter ratio of permanent magnet
machines. Electr. Mach. Power Syst. 22, 521–531 (1994)
3. F.B. Chaaban, A. El-Hajj, A cost-effective design approach for permanent magnet brushless
machines. Electr. Mach. Power Syst. 28, 893–900 (2000)
4. Y. Pang, Z.Q. Zhu, D. Howe, Analytical determination of optimal split ratio for permanent
magnet brushless motors. IEE Proc. Elect. Power Appl. 153(1), 7–13 (2006)
5. L.J. Wu, Z.Q. Zhu, J.T. Chen, Z.P. Xia, G.W. Jewell, Optimal split ratio in fractional-slot
interior permanent-magnet machines with non-overlapping windings. IEEE Trans. Magn. 46
(5), 1235–1242 (2009)
6. Y. Shen, Z.Q. Zhu, L.J. Wu, Analytical determination of optimal split ratio for overlapping
and non-overlapping winding external rotor PM brushless machines, in IEEE International
Electric Machines & Drives Conference, 2011, pp. 41–46
7. Y. Shen, Z.Q. Zhu, Analytical prediction of optimal split ratio for fractional-slot external rotor
PM brushless machines. IEEE Trans. Magn. 47(10), 4187–4190 (2011)
8. W.Q. Chu, Z.Q. Zhu, Y. Shen, Analytical optimisation of external rotor permanent magnet
machines. IET Elect. Syst. Transp. 3(2), 41–49 (2013)
9. T. Reichert, T. Nussbaumer, J.W. Kolar, Split ratio optimization for high-torque pm motors
considering global and local thermal limitations. IEEE Trans. Energy Convers. 28(3), 493–
501 (2013)
10. J.D. Ede, Z.Q. Zhu, D. Howe, Optimal split ratio for high-speed permanent magnet brushless
DC motors, in Proceedings of Fifth International Conference on Electrical Machines and
Systems, 2001, vol. 2, pp. 909–912
11. K. Wang, Z.Q. Zhu, G. Ombach, W. Chlebosz, Average torque improvement of interior
permanent-magnet machine using 3rd harmonic in rotor shape. IEEE Trans. Ind. Electron 61,
5047–5057 (2014)
12. F. Scuiller, Magnet shape optimization to reduce pulsating torque for a 5-phase PM low speed
machine. IEEE Trans. Magn. 50(4), 1–9 (2014)
13. K. Wang, Z.Q. Zhu, G. Ombach, Torque enhancement of surface-mounted permanent magnet
machine using third-order harmonic. IEEE Trans. Magn. 50(3), 104–113 (2014)
14. H.L. Kang, L.B. Zhou, X. Peng, Z.X. Ling, Magnet shape optimization of five-phase
surface-mounted permanent magnet machine to improve torque performance based on
equivalent permanent magnet method. IEEJ Trans. Elec. Electron Eng 10(s1), 133–143
(2015)
Chapter 6
Investigation of Stator Flux Density
and Iron Loss in 3rd Order Harmonic
Shaped Surface-Mounted Permanent
Magnet Machines

The third harmonic can be utilized in the contour of the permanent magnets to
improve the output torque of surface-mounted permanent magnet machines (SPM).
However, third harmonic component will be introduced in the airgap flux density
which will result in variation of flux density distribution in the stator laminations
and the stator iron losses. In this chapter, to analytically investigate the influence of
additional third harmonic airgap flux density on the stator flux density and iron
losses, the stator laminations are decomposed into three parts, i.e., tooth tips, tooth
body and stator yoke. The 12-slot/10-pole machines with Sine, Sine+1/6 3rd, and
1.15(Sine+1/6 3rd) PM shaped rotor are employed for discussion. It is found that, in
tooth body and stator yoke, the classical eddy current loss will increase for both
Sine+1/6 3rd, and 1.15(Sine+1/6 3rd) PM shaped machines compared to one with
Sine-shaped PM rotor, while in tooth tips the hysteresis loss will decrease for
machines with Sine+1/6 3rd PM shaped rotor and rise for ones with 1.15(Sine+1/6
3rd) PM shaped rotor. Further, the thermal distribution in stator laminations is
carried out on three different PM shaped rotor by FE software, which verifies the
aforementioned conclusion.

6.1 Introduction

Permanent magnet (PM) machines have found wide applications from automotive
to aerospace over the past four decades since they offer several key features, such as
high torque density, high efficiency, and excellent dynamic performance [1–4].
However, for high performance applications, such as electrical power steering
systems, servo motors, and direct-drive wind power generators, it is much more
challenging, since both high torque density and low torque ripple are required. To
have high torque and low torque ripple, the magnet shape with additional 3rd
harmonic of three and five phase machines has been reported in [5–12], as shown in

© Springer Nature Singapore Pte Ltd. 2019 99

K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_6
100 6 Investigation of Stator Flux Density and Iron Loss …

Fig. 6.1, where the fundamental airgap flux density is increased by injecting of third
harmonic into the PM shape.
However, the introduction of additional 3rd harmonic in the rotor shape not only
modifies the airgap flux density distribution but also changes the flux density
distribution in the stator lamination of the machine. This will lead to variation of
iron losses distribution. This issue has not been given the deserved attention in the
published chapters. The comparison and torque increase reported in the literature
has not been related to the changes in iron losses. It must be noted that the rated
power and torque is limited by the overall losses and temperature rise in the
windings. In fact, almost all machines can produce higher output torque if the losses
and inner temperature is allowed to increase above the rated value. Thus, higher

(a) 12-slot stator (b) Conventional rotor

(c) Sine rotor (d) Sine+1/6 3rd rotor

Fig. 6.1 12-slot/10-pole SPM machines with different PM shapes

6.1 Introduction 101

power and torque have to be related to losses and temperature rise inside the
machine. Torque increase is only meaningful if losses and temperature raise
remains below the limits for a given machine.
This chapter analyses the effects on the iron losses coming from the generation
of an optimized PM shape to improve the torque. The analysis is based on finite
element field plots of a prototype machine and on analytical expressions for the
losses. The prototype data used in this chapter are described in details in [7]. The
field distribution and the peak flux density in the iron parts are compared to the case
where a sinusoidal airgap flux density exists.
This chapter will be organized as follows. In Sect. 6.2, three prototype machines
with different shaping methods, i.e. Sine shaping, Sine shaping with 3rd harmonic
injected, 1.15 times Sine shaping with 3rd harmonic injected are described. The
method for predicting the influence of 3rd harmonic on stator flux density is pro-
posed in Sect. 6.3, and the losses in iron parts are described and expression for their
determination is presented taking account of the third harmonic airgap flux density
in this part. In Sect. 6.4, a 12-slot/10-pole machine is presented to illustrate the
prediction using finite element method (FEM) analysis, and thermal analysis will
also be given. Conclusion will be given in Sect. 6.5.

6.2 SPM Machines with Various PM Shapes

In order to investigate the influence of additional 3rd harmonic on the iron loss, a
12-slot/10-pole SPM machine with double-layer non-overlapping winding is
employed for illustration. Figure 6.1a shows the stator together with the winding
connections used for all the machines, whilst the rotors with different PM shapes are
shown in Fig. 6.1b–d, respectively. These rotors are designated as the conventional
rotor, Sine rotor and Sine+1/6 3rd rotor. All the stator, rotor, and PM parameters are
given in Table 6.1.
In order to reduce the peak airgap flux density, the Sine shaped PM as shown in
Fig. 6.2a can be reshaped to Sine+1/6 3rd shaped PM as given in Fig. 6.2b. The
1.15(Sine+1/6 3rd) shaped PM in Fig. 6.2c can be employed to improve the output
torque. When the ratio of amplitude of third harmonic to fundamental one equals to
1/6, average torque can be improved by >9%. The machine with the PM shape in
Fig. 6.2a, c have the same minimum airgap length, 1 mm, while the airgap length
of the machine with the PM shape in Fig. 6.2b increased to 1.134 mm due to the
introduction of the 3rd harmonic (Table 6.1).
102 6 Investigation of Stator Flux Density and Iron Loss …

(a) Sine shaped PM

(b) Sine+1/6 3rd shaped PM (increased airgap length)

(c) 1.15(Sine +3rd) shaped PM (same airgap length as Sine shaped PM)
Fig. 6.2 PM shapes with optimal values of 3rd harmonic

6.3 Analytically Predicted Influence of 3rd Harmonic

on Staror Iron Loss

The finite element (FE) predicted airgap flux density distributions and their fun-
damental components for the machines with Sine shaped PM, Sine+1/6 3rd shaped
PM and 1.15(Sine+1/6 3rd) shaped PM are shown in Fig. 6.3 and Table 6.2,
respectively. It can be seen that, the peak airgap flux density of the machines with
6.3 Analytically Predicted Influence of 3rd Harmonic on Staror Iron Loss 103

Table 6.1 Main machine Parameters unit

design specifications and
parameters Pole number/Slot number 10/12
Rated speed 400 rpm
Peak current 20.4 A
Turns in series per phase 48
Axial length 50 mm
Stator outer diameter 90 mm
Stator inner diameter 53 mm
Minimum airgap length (Sine shaped PM) 1 mm
Minimum airgap length (Sine+1/6 3rd) 1.134 mm
shaped
Magnet remanence 1.2 T
Relative permeability 1.07
Maximum PM thickness 3 mm

Sine shaped PM and 1.15(Sine+1/6 3rd) shaped PM is almost the same. The peak
value in the machine with Sine+1/6 3rd shaped PM reduces to 0.81T compared with
the machine with Sine shaped PM and 1.15(Sine+1/6 3rd) shaped one, but its
fundamental one is maintained almost the same as Sine shaped one. The funda-
mental airgap flux density of 1.15(Sine+1/6 3rd) shaped one is improved by nearly
10% compared with that of Sine shaped and Sine+1/6 3rd shaped one.

6.3.1 Flux Density

In order to investigate the influence of additional 3rd harmonic on the stator iron
loss, the flux density in the tooth tips, tooth body and stator yoke can be simplified
analyzed. The stator laminations can be divided into five regions, as shown in
Fig. 6.4.

Fig. 6.3 FE predicted airgap 1.5

flux density with different PM Sine
shapes 1.0 Sine+1/6 3rd
1.15（Sine+1/6 3rd）
Flux density (T)

0.5

0.0

-0.5

-1.0

-1.5
0 12 24 36 48 60 72
Rotor position (elec.deg)
104 6 Investigation of Stator Flux Density and Iron Loss …

Table 6.2 Fundamental Flux density Sine Sine+1/6 3rd 1.15(Sine+1/6 3rd)
airgap flux density
components Bpeak (T) 0.92 0.81 0.91
B1 (T) 0.91 0.90 0.99

B 4
A Wt
C D
3 b0/2
1 lt 2
yttp1 yttp2
1 slot pitch

(a) Position of search coil (b) Regions of stator core

Fig. 6.4 Search coil location and region division

According to the airgap flux density expression in (6.1), the flux in tooth tips
(areas P1 and P2 in Fig. 6.4) can be analytically determined from:


2Rs la Dma Br sin pp
2Ns
uttpi ¼
pl
2
3 
   3pp
pp a sin 3pp
 4cos pxr t  5
2Ns

 cos 3pxr t  ði ¼ 1; 2Þ
2Ns 3 sin pp 2Ns
2Ns

ð6:1Þ

where Rs, la, Ns, Br, l, p and xr are the stator bore radius, the stator active axial
length, the number of slots, the remanence of PM, the maximum PM thickness, the
number of pole-pairs, and the angular velocity. Dma and a are the parameters to be
determined by the optimal 3rd harmonic to improve the output torque, which are
1.15Dm and 1/6.
Therefore, the flux density in the tooth tips can be expressed as:
uttpi
Bttpi ¼
; i ¼ 1; 2 ð6:2Þ
0:5la lt þ pR
Ns
s
 0:5b 0
6.3 Analytically Predicted Influence of 3rd Harmonic on Staror Iron Loss 105


where 0:5la lt þ pRs
Ns  0:5b0 is the effective average cross-section of tooth tips
areas P1 and P2 in Fig. 6.4.
The flux density in region 3 can be expressed as:

uttp1 þ uttp2
Bttp3r ¼ ð6:3Þ
0:5la ð2lt þ wt Þ

where 0:5la ð2lt þ wt Þ is the effective average cross-section of region 3.

The flux density in the tooth body 4 can be expressed as:
uttp1 þ uttp2
Btb4r ¼
wt la

2
 3
2Rs Dma Br sin pp 3pp ð6:4Þ
4cosðphÞ  a sin

 cosð3phÞ5
Ns Ns
¼
wt pl 3 sin pp
Ns

The flux density in the stator yoke depends primarily on the number of stator
slots and the number of pole-pairs. For the 12-slot/10-pole machine, there is only
one tooth per pole which contributes to the flux carried by the stator yoke at the
instant when it is maximum. The flux waveform can be approximated as piece-wise
linear, using two values of flux, uymax and uymin, corresponding to the instants when
the axes of a stator slot and a stator tooth coincide with the transition between
magnet poles. The flux in the stator yoke can be approximated simply as:

1
uymax ¼ utbmax ð6:5Þ
2
uymin ¼ 0 ð6:6Þ

The influence of additional 3rd harmonic on the flux density in the stator yoke 5
is similar to the flux density in the tooth body.
According to (6.1), (6.2) and (6.4), it can be seen that the influence of additional
3rd harmonic on the flux density in the stator lamination, including tooth tips, tooth
body and stator yoke, is related to slot and pole combination. It can be divided into
2 cases to discuss.
Case 1: Dma ¼ Dm
As can be seen from (6.2), the effect of additional 3rd harmonic on the peak flux
density in the tooth tips is dependent of the slot/pole number combinations. Taking
6-slot/10-pole and 6-slot/8-pole machines as example, the additional 3rd harmonic
always increase the peak flux density in the tooth tips in 6-slot/10-pole machines
while the additional 3rd harmonic has no influence on the peak flux density in the
tooth tips in 6-slot/8-pole machines, since the term sin(3pp/2Ns) is always negative
for 6-slot/10-pole machines and it is always zero for 6-slot/8-pole machines. It can
106 6 Investigation of Stator Flux Density and Iron Loss …

be concluded that when the slot/pole number combinations meet the formulation
(6.7), the additional 3rd harmonic has no influence on the peak flux density in the
tooth tips, when the slot/pole number combinations meet the formulation (6.8), the
additional 3rd harmonic always decrease the peak flux density in the tooth tips, and
when the slot/pole number combinations meet the formulation (6.9), the additional
3rd harmonic always increase the peak flux density in the tooth tips.

p 2
¼ ð6:7Þ
ns 3

p 2
0\ \ ð6:8Þ
ns 3

2 p
\ \1 ð6:9Þ
3 ns

The effect of additional 3rd harmonic on the flux density in the tooth tips under
different pole-slot combinations is summarized in Table 6.3.
As can be seen from (6.4), the effect of additional 3rd harmonic on the peak flux
density in the tooth body is also dependent of the slot/pole number combinations.
Taking 12-slot/10-pole and 12-slot/8-pole machines as example, the additional 3rd
harmonic always increase the peak flux density in the tooth body in 12-slot/10-pole
machines while the additional 3rd harmonic has no influence on the peak flux
density in the tooth body in 12-slot/8-pole machines, since the term sin(3pp/Ns) is
always negative for 12-slot/10-pole machines and it is always zero for 12-slot/
8-pole machines. It can be concluded that when the slot/pole number combinations
meet the formulation (6.10), the additional 3rd harmonic has no influence on the
peak flux density in the tooth body, when the slot/pole number combinations meet
the formulation (6.11), the additional 3rd harmonic always decrease the peak flux
density in the tooth body, and when the slot/pole number combinations meet the
formulation (6.12), the additional 3rd harmonic always increase the peak flux
density in the tooth body.

Table 6.3 The effect of 3rd harmonic on the peak flux density of tooth tips under different
pole-slot combinations
ns p2
2 4 8 10 14
3 Decrease No influence – – –
6 Decrease Decrease No influence Increase –
9 Decrease Decrease Decrease Decrease Increase
12 Decrease Decrease Decrease Decrease Decrease
24 Decrease Decrease Decrease Decrease Decrease
6.3 Analytically Predicted Influence of 3rd Harmonic on Staror Iron Loss 107

p 1 2
¼ ; ð6:10Þ
ns 3 3

2k p 2k þ 1
\ \ ; k ¼ 0; 1 ð6:11Þ
3 ns 3

1 p 2
\ \ ð6:12Þ
3 ns 3

The effect of additional 3rd harmonic on the peak flux density in the tooth body/
yoke under different pole-slot combinations is summarized in Table 6.4.
Case 2: Dma ¼ 1:15Dm
pp
If 2N s
¼ a, then 3sinsin3aa  1, the equation is set up when and only when the number
of slot tends to infinity or the number of pole tends to be zero, so 3sinsin3aa \1. When
a = 1/6, the amplitude of 1:15Dm½cos  ðphÞ  a cosð3a 3phÞ and Dm  cosðphÞ are the
same, so the amplitude of 1:15Dm cosðphÞ  a3 sin sin a cos ð3ph Þ is more than the
amplitude of Dm cosðphÞ certainly. Therefor the additional 3rd harmonic increase
the peak flux density in the tooth tips under any slot/pole number combination, also
applying to tooth body and yoke.
According to the discussion, comparison of peak flux density (per-unit value) in
the 12-slot/10-pole machines with three different PM shapes is listed in Table 6.5. It
can be seen that when Dma = Dm, the additional 3rd harmonic can reduce the peak
flux density in the tooth tips, while increase the peak flux density in the tooth body
and stator yoke. And when Dma = 1.15Dm, the additional 3rd harmonic can
increase the peak flux density in the tooth tips, tooth body and stator yoke.

Table 6.4 The effect of 3rd harmonic on the peak flux density of tooth body/yoke under different
pole-slot combinations
ns p2
2 4 8 10 14
3 No influence No influence – – –
6 Decrease No influence No influence Decrease –
9 Decrease Decrease Increase Increase Decrease
12 Decrease Decrease No influence Increase Increase
24 Decrease Decrease Decrease Decrease Increase

Table 6.5 Comparison of Sine Sine+1/6 3rd 1.15(Sine+1/6 3rd)

peak flux density (PU value)
in the 12-slot/10-pole Bttp1 peak 1 0.92 1.06
machines with different PM Btb4r peak 1 1.04 1.20
shapes By peak 1 1.04 1.20
108 6 Investigation of Stator Flux Density and Iron Loss …

6.3.2 Iron Loss

As stated previously, the variation of airgap flux density caused by the PM shape
results in flux density variation in the stator. This will lead to variation of iron losses
distribution. The iron losses including hysteresis loss Ph, classical eddy current loss
Pc and the excess eddy current loss Pe, which can be expressed as following,
respectively:

Ph ¼ Kh fBam ð6:13Þ

Z2p  2
2prd 2 f 2 dB
Pc ¼ dhe ð6:14Þ
12d dhe
0

2p 3
pffiffiffiffiffiffi 1:5 Z dB 2
Pe ¼ 2pKe f dhe ð6:15Þ
dh
e
0

where f and Bm are the frequency and amplitude of the flux density; r, d, and d are
the electrical conductivity, the mass density, and the lamination thickness,
respectively. kh, a, and ke are experimentally determined loss coefficients.
According to the expression of the flux density in the tooth tips, the hysteresis
loss and eddy current loss in the tooth tips can be given by:
2
 2
33a
2Rs Dma Br sin 2N
pp
a sin 3pp

Ph ¼ Kh f 4

s
 41

2Ns
 55 ð6:16Þ
0:5 lt þ Ns  0:5b0 pl
pRs 3 sin 2Npp
s

!2     
2p2 rd 2 f 2 2Rs Dma Br pp 2 3pp
Pc ¼ sin 2
þ a sin
2
ð6:17Þ
12d 0:5plðlt þ pR
Ns  0:5b0 Þ
s 2Ns 2Ns

The hysteresis loss and the classical eddy current loss in the tooth body can be
given by:
0 0
11a
2Rs Dma Br sinðpp sin 3pp
Ns Þ a
Ph ¼ Kh f @ @1 
 AA
Ns
ð6:18Þ
plwt 3 sin pp
Ns

      
2p2 rd 2 f 2 2Rs Dma Br 2 pp 3pp
Pc ¼ sin2 þ a2 sin2 ð6:19Þ
12d plwt Ns Ns
6.3 Analytically Predicted Influence of 3rd Harmonic on Staror Iron Loss 109

The flux density distribution in the stator yoke is similar to that in the tooth body.
Hence, the influence of the additional 3rd harmonic on the hysteresis and classical
eddy current loss in the stator yoke is similar to those in the tooth body.
Case 1: Dma ¼ Dm
As can be seen from (6.16) to (6.18), the effect of additional 3rd harmonic on the
hysteresis loss in the tooth tips and tooth body/yoke is dependent of the slot/pole
number combinations. As can be seen from (6.17) to (6.19), the additional 3rd
harmonic will increase the classical eddy current loss in any slot/pole number
combination. The influence of additional 3rd harmonic on the stator loss under
different slot/pole number combinations is summarized in Table 6.6.
Case 2: Dma ¼ 1:15Dm
Considering the effect of additional 3rd harmonic on the stator flux density, the
additional 3rd harmonic can reduce the iron loss in the stator. Therefore, the total
iron loss must be increased since the shaped PM.
According to (6.16)–(6.19), comparison of iron losses (per-unit value) in the
12-slot/10-pole machines with three different PM shapes are listed in Table 6.7. It
can be seen that when Dma = Dm, the additional 3rd harmonic can reduce the
hysteresis loss in the tooth tips due to the decrease of peak flux density, as given in
Table 6.4, but increase the classical eddy current loss in the tooth body and stator
yoke due to the additional 3rd harmonic airgap flux density. When Dma = 1.15Dm,
the additional 3rd harmonic can increase the hysteresis loss and classical eddy
current loss in the tooth tips, tooth body and stator yoke.
Table 6.6 The effect of 3rd harmonic on the iron loss of the tooth under different pole-slot
combinations
2 P
ns 2 4 8 10
Tooth Tooth Tooth Tooth Tooth Tooth Tooth Tooth
tips body/ tips body/ tips body/ tips body/
yoke yoke yoke yoke
Ph Pc Ph Pc Ph Pc Ph Pc Ph Pc Ph Pc Ph Pc Ph Pc
3 # " "# " "# " "# " – –
6 # " # " # " "# " "# " "# " " " # "
9 # " # " # " # " # " " " # " " "
12 # " # " # " # " # " "# " # " " "
24 # " # " # " # " # " # " # " # "

Table 6.7 Comparison of Sine Sine+1/6 3rd 1.15(Sine+1/6 3rd)

iron losses (PU value) in the
12-slot/10-pole machines with Ph ttp1 1 0.84 1.12
different PM shapes Pc ttp1 1 1.06 1.40
Ph tb4r 1 1.08 1.43
Pc tb4r 1 1.04 1.38
110 6 Investigation of Stator Flux Density and Iron Loss …

6.4 Finite Element Analysis

The flux density distribution in the P1, P4, and P5 of Fig. 6.4 is given in Fig. 6.5.
and summarized in Table 6.8. It can be observed that the additional 3rd harmonic
can reduce the peak flux density in the teeth tips and increase the peak flux density
in the tooth body and stator yoke when the amplitude of fundamental keeps the
same ðDma ¼ DmÞ. As can be seen from (6.2), (6.8) to (6.12), when the number of

0.8 0.9
0.6 Sine 0.8

Flux density (T)

Sine
Flux density (T)

Sine+3rd 0.7
0.4 1.15(Sine+3rd) Sine+1/6 3rd
0.6
0.2 1.15(Sine+1/6 3rd)
0.5
0.0
0.4
-0.2 0.3
-0.4 0.2
-0.6 0.1
-0.8 0
0 60 120 180 240 300 360 1 3 5 7 9 11 13
Rotor position (elec.deg) Harmonic order
(a) Tooth tips
2.0 1.6
1.5 Sine 1.4
Flux density (T)

Sine+1/6 3rd Sine

Flux density (T)

1.0 1.15(Sine+1/6 3rd) 1.2 Sine+1/6 3rd

0.5 1 1.15(Sine+1/6 3rd)
0.0 0.8
-0.5 0.6
-1.0 0.4
-1.5 0.2
-2.0 0
0 60 120 180 240 300 360 1 3 5 7 9 11 13
Rotor position (elec.deg) Harmonic order
(b) Tooth body
0.5 0.45
0.4 0.4
Sine
Flux density (T)

Sine
Flux density (T)

0.3 Sine+3rd 0.35

0.2 1.15(Sine+3rd) Sine+1/6 3rd
0.3
0.1 1.15(Sine+1/6 3rd)
0.25
0.0
0.2
-0.1
-0.2 0.15
-0.3 0.1
-0.4 0.05
-0.5 0
0 60 120 180 240 300 360 1 3 5 7 9 11 13
Rotor position (elec.deg) (c) Stator yoke Harmonic order

Fig. 6.5 Variation of flux density and harmonic contents with different rotor shapes in divided
regions
6.4 Finite Element Analysis 111

Table 6.8 Flux density components

Sine Sine+1/6 3rd 1.15(Sine+1/6 3rd)
P1 P4 P5 P1 P4 P5 P1 P4 P5
Bpeak (T) 0.69 1.28 0.35 0.63 1.31 0.38 0.69 1.45 0.51
B1 (T) 0.70 1.27 0.31 0.70 1.29 0.33 0.77 1.44 0.44
B3 (T) 0.01 0.01 0.04 0.06 0.02 0.05 0.08 0.02 0.08

slot/pole number is 12/10, the additional 3rd harmonic can decrease the peak flux
density in the tooth tips while increasing the peak flux density in the tooth body and
yoke. The additional 3rd harmonic can increase the peak flux density in the tooth
tips, tooth body and yoke when Dma ¼ 1:15Dm according to the mathematical
analysis. Due to the phenomenon in the graphs, theoretical analysis above can be
validated.
Core loss distribution of stator under no load and rated load are given in
Figs. 6.6 and 6.7. It can be seen that the loss density of the tooth and yoke in Sine
shaped PM machine is less than that in the 3rd harmonic introduced PM machine
obviously. The variation of stator loss, including hysteresis loss and classical eddy
current loss in three different machines at no load and rated load is listed in
Tables 6.9 and 6.10. As the main loss is on the tooth body and stator yoke, the
overall loss will increase. As can be seen from (6.16) to (6.19), when the number of
slot/pole number is 12/10, the additional 3rd harmonic can reduce the hysteresis
loss in the tooth tips while increasing the other losses, as the main loss is on the
tooth body and stator yoke, the overall loss will increase when Dma ¼ Dm.
Considering the effect of additional 3rd harmonic on the stator flux density, the
additional 3rd harmonic can reduce the iron loss in the stator when Dma ¼ 1:15Dm.
According to the data and graphs, the theoretical analysis on iron loss can be
validated.
The iron loss, copper loss and eddy-current loss of PM can translate into heat at
the machine operation, which cause the rise of machine temperature. It should be
noted that the temperature rise must be limited within the temperature tolerance of
material to ensure the efficiency and reliability of the machine operation. Therefore,
it is necessary to analyze the machine temperature rise. The temperature field of the
machine with sine and sine+3rd shaped PM rotor is analyzed.
The basic assumption, boundary conditions, thermal conductivity and surface
coefficient of heat dissipation are determined by the heat transfer theory and the
material properties. The three-dimension finite element model of the PM machine is
established to calculate the temperature field distributions in software ANSYS
Workbench, and the corresponding stator, rotor and winding temperature field
distributions are achieved. Figure 6.8 shows the three-dimension temperature field
distributions of the machine at no-load. It can be seen that the highest temperature
emerges at the stator teeth, the temperature is 24.23 °C, and the lowest temperature
emerges the slots isolation, the temperature is 24.14 °C. The temperature field
distributions of the various parts of the machine are shown in Fig. 6.9. The
112 6 Investigation of Stator Flux Density and Iron Loss …

(a) Scale (b) Sine shaped PM machine

(c) Sine+3rd shaped PM machine(increased (d) 1.15(Sine +3rd) shaped PM

airgap length) machine(same airgap length as Sine
shaped PM)

Fig. 6.6 Core loss density distribution of stator under no load

temperature field distributions of stator iron is shown in Fig. 6.9a. It can be seen
that the highest temperature emerges of stator iron at neutral position, it is mainly
due to the iron loss of stator teeth has the maximum proportion of iron loss, and the
heat dissipation is also the worst. The highest temperature is at the same position
with the stator iron, the temperature point can be as the judge reference for the PM
demagnetization. The rotor iron and the chassis temperature field distributions are
shown in Fig. 6.9c, d, respectively.
The highest temperature emerges at the stator windings at load, due to the
windings copper loss has the maximum proportion of the machine loss. The tem-
perature field distribution of the machine at load is shown in Fig. 6.10. The highest
temperature is 64.16 °C at windings, and the lowest temperature is 60.96 °C,
emerging at the slots isolation. The temperature field distribution of various parts at
load is shown in Fig. 6.11. It can be seen that the highest temperature of the stator is
62.23 °C, and the lowest is 61.42 °C in Fig. 6.11a. The temperature field
6.4 Finite Element Analysis 113

(a) Scale (b) Sine shaped PM machine

(c) Sine+3rd shaped PM machine (increased (d) 1.15(Sine +3rd) shaped PM machine
airgap length) (same airgap length as Sine shaped PM)

Fig. 6.7 Core loss density distribution of stator under rated load

Table 6.9 No-load stator loss

Sine Sine+3rd 1.15(Sine+3rd)
Hysteresis loss (W) 1.15 1.18 1.41
Classical eddy current loss (W) 0.09 0.10 0.11
Stator loss (W) 1.24 1.28 1.52

Table 6.10 Load stator loss

Sine Sine+3rd 1.15(Sine+3rd)
Hysteresis loss (W) 1.23 1.24 1.46
Classical eddy current loss (W) 0.09 0.10 0.11
Stator loss (W) 1.32 1.34 1.57
114 6 Investigation of Stator Flux Density and Iron Loss …

Fig. 6.8 The temperature

field distribution of whole PM
machine

(a) stator iron (b) PM

(c) rotor iron (d) chassis

Fig. 6.9 The temperature field distribution of various parts for whole PM machine

distribution of PM, rotor iron and chassis is shown in Fig. 6.11b–d. The temper-
ature of the PM is 61.89 °C, and the temperature rise is 39.89 °C, which will not
cause the PM demagnetization. The temperature field distribution stator windings
6.4 Finite Element Analysis 115

Fig. 6.10 the temperature

field distribution of whole
machine at load

and slots isolation are shown in Fig. 6.11e, f. It can be observed that the highest
emerges at the contact part between the windings and windings isolation.
The highest temperature emerges at the stator windings at load, due to the
windings copper loss has the maximum proportion of the machine loss. The tem-
perature field distribution of the machine at load is shown in Fig. 6.10. The highest
temperature is 64.16 °C at windings, and the lowest temperature is 60.96 °C,
emerging at the slots isolation. The temperature field distribution of various parts at
load is shown in Fig. 6.11. It can be seen that the highest temperature of the stator is
62.23 °C, and the lowest is 61.42 °C in Fig. 6.11a. The temperature field distri-
bution of PM, rotor iron and chassis is shown in Fig. 6.11b–d. The temperature of
the PM is 61.89 °C, and the temperature rise is 39.89 °C, which will not cause the
PM demagnetization. The temperature field distribution stator windings and slots
isolation are shown in Fig. 6.11e, f. It can be observed that the highest emerges at
the contact part between the windings and windings isolation.
In order to analyze the impact of third harmonic injection into the sine shaped
PM on the temperature field distribution. The temperature of the sine shaped
machine, sine+3rd shaped machine and 1.15 (sin+3rd) shaped machine are com-
pared. Figure 6.12 show the temperature field distribution of the three machine. It
can be seen that the sine shaped machine has the lowest temperature, and the 1.15
(sin+3rd) shaped machine has the highest temperature.
The temperature and temperature rise of the machine at the environment tem-
perature 22 °C is summarized in Table 6.11. It can be seen that the temperature is
increased by third harmonic injection under the constraint of the same fundamental
amplitude. The temperature is increased by 3.03% at no-load, and 0.07% is
increased at load for the sine and sin+3rd shaped machine. The temperature of the
1.15(Sine+3rd) shaped machine is increased by 22.72% at no-load, 0.87% at load,
respectively, compared to the sine shaped machine. The temperature of the 1.15
(Sine+3rd) shaped machine is increased by 19.11% at no-load, 0.80% at load,
respectively, compared to the sine+3rd shaped machine.
116 6 Investigation of Stator Flux Density and Iron Loss …

(a) stator iron (b) PM

(c) rotor iron (d)chassis

(e) windings (f) windings isolation

Fig. 6.11 Temperature field distribution of various parts at load
6.4 Finite Element Analysis 117

(a) Sine shaped machine at no-load (b) Sine shaped machine at load

(c) Sine +3rdshaped machine at (d) Sine +3 rd shaped machine at

no-load load

(e) 1.15(Sine+3rd)shaped machine at (f) 1.15(Sine+3rd) shaped machine at

no-load load
Fig. 6.12 temperature field distribution of three machines
118 6 Investigation of Stator Flux Density and Iron Loss …

Table 6.11 Temperature rise and temperature in the machine

Sine Sine+3rd 1.15(Sine+3rd)
No-load load No-load load No-load load
The highest temperature 23.818 63.601 23.873 63.63 24.231 63.962
(°C)
The lowest temperature 23.743 60.634 23.797 60.665 24.14 61.032
(°C)
The highest temperature 1.818 41.601 1.873 41.63 2.231 41.962
rise (°C)

6.5 Summary

In this chapter, the influence of additional 3rd harmonic in airgap on flux density
and iron losses in stator laminations are analyzed. It is found that, in tooth body and
stator yoke, the classical eddy current loss will increase for both Sine+1/6 3rd, and
1.15(Sine+1/6 3rd) PM shaped machines compared to one with Sine-shaped PM
rotor, while in tooth tips the hysteresis loss will decrease for machines with Sine+1/
6 3rd PM shaped rotor and rise for ones with 1.15(Sine+1/6 3rd) PM shaped rotor.
The total loss will increase for machines with the 3rd harmonic in the contour of the
PMs but temperature rise remains below the limits.
This chapter assesses the changes in magnetic losses in three phase induction
machines when an optimized airgap induction is produced in the airgap. In order to
simplify the analysis, the field is considered as pure alternating; this assumption is
valid for the stator tooth tips and in part for the stator core. The induction values are
obtained from finite element analysis and the losses have been estimated with
analytical expressions considering a three term model. Under the assumptions
made, the introduction of a third harmonic induction component in the airgap will
always increase the eddy current losses in the stator tooth tips, tooth body, and also
in the stator core. When Dma ¼ Dm, the hysteresis losses depends on the slot/pole
combinations. For the 12-slot/10-pole machine, the hysteresis losses decrease in the
tooth tips and increase in the stator core. This result is mainly due the change in the
peak induction. Considering that the amount of hysteresis losses in the tooth body
and yoke are higher than in the tooth tips, this change contributes to increment in
the overall losses. Finally, the anomalous losses show a light increase in the stator.
When Dma ¼ 1:15Dm, the hysteresis losses and classical eddy current losses
increase in the stator for the variation of the flux density in the stator. The chapter
shows that the inclusion of a third harmonic component in the airgap changes the
magnetic losses distribution, which can lead to higher temperature rise inside the
machine. Therefore, a torque increase has to be weighed against the losses and
temperature increase. The fundamental airgap flux density of 1.15(Sine+3rd)
shaped one has been improved by 10% compared with that of Sine shaped one.
However, the temperature rise of 1.15(Sine+3rd) shaped machine only improved by
6.5 Summary 119

0.9% compared with that of Sine shaped one under rated load. The optimal model
can obtain higher torque and lower temperature rise compared with the machine
with Sine shaped PM.

References

1. Z.Q. Zhu, D. Howe, Electrical machines and drives for electric, hybrid, and fuel cell vehicles.
IEEE Proc. 95(4), 746–765 (2007)
2. A.M. EL-Refaie, Fractional-slot concentrated-windings synchronous permanent magnet
machines: opportunities and challenges. IEEE Trans. Ind. Electron. 57(1), 107–121 (2010)
3. A.M. EL-Refaie, Motors/generators for traction/propulsion applications: a review. IEEE Veh.
Technol. Mag. 8(1), 90–99 (2013)
4. W.P. Cao, B.C. Mecrow, G.J. Atkinson, J.W. Bennett, D.J. Atkinson, Overview of electric
motor technologies used for more electric aircraft (MEA). IEEE Trans. Ind. Electron. 59(9),
3523–3531 (2012)
5. Y. Li, J.B. Zou, Y.P. Lu, Optimum design of magnet shape in permanent-magnet synchronous
motors. IEEE Trans. Magn. 39(6), 3523–3526 (2003)
6. Z.Q. Zhu, K. Wang, G. Ombach, Optimal magnet shaping with third order harmonic for
maximum torque in SPM machines, in Proceedings of 6th IET Conference on Power
Electronics, Machines and Drives, PEMD, Bristol, UK, March 2012, pp. 1–6
7. K. Wang, Z.Q. Zhu, G. Ombach, Torque enhancement of surface-mounted permanent magnet
machine using third-order harmonic. IEEE Trans. Magn. 50(3), 104–113 (2014)
8. K. Wang, Z.Q. Zhu, G. Ombach, W. Chlebosz, Optimal rotor shape with third harmonic for
maximizing torque and minimizing torque ripple in IPM motors, in XX International
Conference on Electrical Machines (ICEM 2012), Marseille, France, 2–5 Sept 2012, pp. 397–
403
9. K. Wang, Z.Q. Zhu, G. Ombach, W. Chlebosz, Average torque improvement of interior
permanent-magnet machine using 3rd harmonic in rotor shape. IEEE Trans. Ind. Electron. 61,
5047–5057 (2014)
10. K. Wang, Z.Q. Zhu, R. Nilssen, Using third harmonic for shape optimization of flux density
distribution in linear permanent-magnet machine, in International Symposium on Linear
Drives for Industry Applications (LDIA2013), Hangzhou, China, 27–29 July 2013, CD-ROM
11. P.Z. Zhao, G.J. Yang, Torque density improvement of five-phase PMSM drive for electric
vehicles applications. J. Power Electron. 11(4), 401–407 (2011)
12. K. Wang, Z.Q. Zhu, Torque improvement of five phase surface-mounted permanent-magnet
machine using 3rd order harmonic. IEEE Trans. Energy Convers. 29, 735–747 (2014)
Chapter 7
Analysis of Cogging Torque
in Surface-Mounted Permanent Magnet
Machines with Shaped Magnets

The third harmonic can be injected into the Sinusoidal shaped permanent magnets
(PM) to improve the average torque in three-phase surface-mounted PM
(SPM) machines without deteriorating the torque ripple caused by back-EMFs
harmonics. However, the influence of the injected third harmonic on the cogging
torque is uncertain. The aim of this chapter is to analytically derive the cogging
torque in SPM machines equipped with Sinusoidal shaped (Sine-shaped) and
Sinusoidal shaped with third harmonic injected (Sine+3rd-shaped) PMs, which are
optimized to obtain sinusoidal and sinusoidal with third harmonic airgap flux
density distributions, respectively. It is found that the cogging torque of the SPM
machines with Sine-shaped PMs is negligibly small. The Sine+3rd-shaped PM can
increase the amplitude of the cogging torque for the specific three-phase SPM
machines with the number of slots per pole per phase (spp) equal to 0.25, 0.5 and 1.
Finite element analyses are carried out on 6-slot/8-pole (spp = 0.25), 12-slot/8-pole
(spp = 0.5), 24-slot/8-pole (spp = 1) and 12-slot/10-pole (spp = 0.4) SPM machi-
nes with shaped PMs to verify the analyses. A SPM machine with Sine+3rd-shaped
PMs is prototyped and tested for validation.

7.1 Introduction

Cogging torque arises from the interaction between permanent magnets and stator
teeth, which exists in all types of permanent magnet (PM) machines [1, 2]. In
high-performance applications, such as electrical power steering systems, servo
motors, and direct-drive wind power generators, low cogging torque is required.
Cogging torque can be mitigated by optimizing either the stator or rotor [3, 4].
The PM or rotor steel shaping method, which is one of the simplest and most
effective approaches, have been widely employed in [5–11], where the pole-arc
optimization approach [5, 6], the PM edge shaping method [5, 7] and the PM
eccentric arc shaping method [8, 9] were proposed to effectively reduce the cogging
© Springer Nature Singapore Pte Ltd. 2019 121
K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_7
122 7 Analysis of Cogging Torque in Surface-Mounted …

torque. In addition, the PM sine shaping [10] and rotor steel inverse cosine shaping
[11] approaches have been investigated for the suppression of cogging torque and
torque ripple. As all the aforementioned PM or rotor steel shaping methods
inevitably compromise the torque density of the machines, the sine shaping with
third harmonic injected (Sine+3rd) method was firstly proposed in [12] and further
investigated in [13, 14] to increase the average torque. It should be noted that the
injected third harmonic will not cause the torque ripple associated with the elec-
tromagnet force (EMF) since the triplen harmonics are absent in the line-line
back-EMFs in the three phase machines. However, whether the injected third
harmonic will influence the cogging torque is uncertain. In [13] and [14], the
amplitude of the cogging torque witnesses a slight increase in a 12-slot/10-pole
machine with the Sine+3rd-shaped PMs or rotor steel compared to that of the
Sine-shaped counterparts. On the other hand, for a 24-slot/8-pole SPM machine, the
injected third harmonic in airgap flux density causes large cogging torque and
deteriorates the torque ripple [15]. Thus, it is necessary to investigate the influence
of the airgap flux density harmonics, especially the third harmonic, on the cogging
torque. In [4], it was firstly stated that the cogging torque can be reduced by
controlling the NLth harmonic component of the field function, i.e., the square of
the airgap flux density, where NL denotes the least common multiple of number of
poles and number of slots. In [16], it was claimed that only certain harmonic
components of the field function can cause cogging torque and the elimination of
the harmonic components of order 3n (n = 1, 2, 3) of the field function can largely
reduce the cogging torque of a 6-slot/8-pole SPM machine. In [17], the technique of
using the auxiliary grooves on the permanent magnet pole face is employed in a
6-slot/8-pole SPM machine to mitigate the cogging torque by removing the har-
monics of the field function associated with the cogging torque. In [18], the third
harmonic in airgap flux density was reduced in a 6-slot/4-pole IPM machine by the
rotor core notching design to lower the noise and vibration level. However, all these
literatures are confined to machines with a specific slot number and pole number
combination and fail to explicitly explain how the third harmonic in airgap flux
density affects the generation of the cogging torque.
The purpose of this chapter is to investigate the influence of the harmonics in
airgap flux density on the generation of the cogging torque in SPM machines.
Based on the energy method, the cogging torque are mathematically derived when
the PMs are Sine-shaped and Sine+3rd-shaped, respectively. The influence of the
different slot/pole number combinations on the cogging torque are discussed.
Furthermore, the three phase SPM machines are investigated, considering the
influence of the third harmonic in airgap flux density, and a general conclusion is
drawn. The finite element (FE) analyses are carried out on four slot/pole number
combinations, i.e., 6-slot/8-pole, 12-slot/8-pole, 24-slot/8-pole and 12-slot/10-pole
to verify the conclusion. Experimental validations are reported as well.
This chapter will be organized as follows. In Sect. 7.2, the cogging torque in the
SPM machines with sinusoidal and sinusoidal with third harmonic airgap flux density
distributions are mathematically derived based on the energy method considering slot/
pole number combinations. In Sect. 7.3, the configurations of the Sine-shaped and
7.1 Introduction 123

Sine+3rd-shaped PMs are described firstly. Then, the parameters of 6-slot/8-pole,

12-slot/8-pole, 24-slot/8-pole and 12-slot/10-pole three phase SPM machines having
Sine-shaped and Sine+3rd-shaped PMs are given. In Sect. 7.4, the cogging torque
comparison between the machines with Sine-shaped PMs and Sine+3rd-shaped PMs
for the 6-slot/8-pole, 12-slot/8-pole, 24-slot/8-pole and 12-slot/10-pole combinations
are made by using FE analysis. In Sect. 7.5, the experiments are performed on a SPM
machine with Sine+3rd-shaped PMs. Section 7.6 is the conclusion.

7.2 Analytical Prediction of Cogging Torque Based

on Energy Method

Figure 7.1 shows a part of the cross sectional area of the SPM machine. As can be
seen, the center of slot1 is defined as the zero position of the circumference. The
rotor position a is the angle between the zero position of the circumference and the
center of PM1.
Since the energy variation in PMs and iron are negligible compared to that of the air,
the cogging torque can be expressed as follows based on the energy method [4, 19].

@WðaÞairgap
Tcog ðaÞ ¼  ð7:1Þ
@ðaÞ

where W(a)airgap is the magnetic energy in the air, and can be expressed as

Z2p
1
WðaÞairgap ¼ Lef ðR2m  R2s Þ G2 ðhÞB2 ðh; aÞdh ð7:2Þ
4l0
0

stator

slot 1

airgap a
   
   

   

   

 
         
     

rotor 0 θ
PM 1

Fig. 7.1 Cross-section of simplified SPM machine

124 7 Analysis of Cogging Torque in Surface-Mounted …

where l0, Lef, Rm, Rs, h, G(h), B(h, a) are the permeability of air, the effective axial
length of the machine, the outer radius of the airgap, the inner radius of the airgap, the
angle along the rotor circumference, the airgap relative permeability function, the air
gap flux density in an equivalent slotless machine. For machines with symmetrical
stator and rotor in space, G2(h), B2(h, a) can be expanded using Fourier series.

X
1
G2 ðhÞ ¼ GkNs cosðkNs hÞ ð7:3Þ
k¼0

X
1
B2 ðh; aÞ ¼ BmNp cos½mNpðh  aÞ ð7:4Þ
m¼0

where Ns, Np, GkNs, BmNp are the number of slots, the number of poles and the
corresponding Fourier coefficients.
According to the product to sum formula of trigonometric functions, only when
equation is satisfied, the cogging torque is nonzero

kNs ¼ mNp ¼ nNL ð7:5Þ

where k, m, n are integers, NL is the least common multiple (LCM) of Ns and Np.
The cogging torque is given by (7.6)

pNL X1
TðaÞ ¼  Lef ðR22  R21 Þ nGnNL BnNL sinðnNL aÞ ð7:6Þ
4l0 n¼1

It can be seen from (7.6) that the cogging torque has a close relationship with the
value of BnNL, which is determined by the airgap flux density waveform. When the
airgap flux density waveform is sinusoidal or sinusoidal with third harmonic, the
value of BnNL may be zero for some slot and pole number combinations resulting in
no cogging torque, as will be described later.

7.2.1 Simplified Airgap Permanence Function

The simplified stator slot shape is shown in Fig. 7.2. To simplify the mathematical
derivation process, the slot depth is assumed to be infinite, which means the value
of G(h) under the slots and teeth are 0 and 1, respectively. A simplified airgap
permanence function can be obtained, as shown in Fig. 7.3.
Since the number of stator slots is Ns, the period of G2(h) is 2p/Ns. G2(h) can be
expressed as below in the interval [−p/Ns, p/Ns]:
7.2 Analytical Prediction of Cogging Torque Based on Energy Method 125

Fig. 7.2 Simplified stator 2π/Ns

slot shape

slot1

b0

-π/Ns π/Ns

Fig. 7.3 Simplified airgap G(θ)

permeance function 1

-π/Ns -b0/2 0 b0/2 π/Ns θ

(
1; ð Nps ;  b20 Þ [ ðb20 ; p
Þ
G ðhÞ ¼
2 Ns
ð7:7Þ
0; ð b20 ; b20 Þ

where b0 is the slot opening.

The Fourier series for (7.7) is:
 
b0 Ns X 1
2 kNs b0
G2 ðhÞ ¼ 1   sin cosðkNs hÞ
2p kp 2

k¼1
 ð7:8Þ
p p
h2  ;
Ns Ns

The GnNL is given in (7.9), which can be substituted into Eq. (7.6) to calculate
the cogging torque.
126 7 Analysis of Cogging Torque in Surface-Mounted …

 
2 nNL b0
GnNL ¼ sin n ¼ 1; 2; 3; . . . ð7:9Þ
np 2

7.2.2 Cogging Torque for Sinusoidal Airgap Flux

Density Distribution

Figure 7.4 shows the sinusoidal air gap flux density distribution B(h, a) produced
by one pair of rotor PM poles when the stator is slotless whilst Fig. 7.5 is the
waveform of the field function B2(h, a). The mathematical expressions of B(h, a)
and B2(h, a) are as follows:

Bðh; aÞ ¼ B1 cos½Np ðh  aÞ=2

 
p 3p ð7:10Þ
 þ a; þa
Np Np

B
B1

-π/N p+α 0 α π/N p+α 3π/N p+α

Fig. 7.4 Sinusoidal airgap flux density distribution

B2

-π/N p + α 0 α π/N p +α 3π/N p +α

Fig. 7.5 Waveform of B2(h, a)

7.2 Analytical Prediction of Cogging Torque Based on Energy Method 127

B2 B2
B2 ðh; aÞ ¼ 1 þ 1 cos½Np ðh  aÞ
2 2  ð7:11Þ
p p
 þ a; þa
Np Np

Since the airgap flux density waveform is sinusoidal, there are only two com-
ponents in the field function B2(h, a), namely the average component and the
fundamental component. The average component will not contribute to the cogging
torque. Whether the fundamental term in Eq. (7.11) can give rise to the cogging
torque is dependent on the slot/pole number combinations. Two cases can be
classified for discussion.
Case A: when Ns and Np satisfy

Np 6¼ kNs k ¼ 1; 2; 3; . . . ð7:12Þ

The cogging torque is zero. Since most commonly used slot/pole number
combinations belongs to Case A, it can be concluded that the sinusoidal airgap flux
density design can totally eliminate the cogging torque theoretically.
Case B: when Ns and Np satisfy

Np ¼ kNs k ¼ 1; 2; 3; . . . ð7:13Þ

The cogging torque is:

 
B21 Ns Np b0
TðaÞ ¼  Lef ðR2  R1 Þsin
2 2
sinðNp aÞ ð7:14Þ
4l0 2

7.2.3 Cogging Torque for Sine+3rd Airgap Flux Density

Distribution

Figure 7.6 shows the sinusoidal with third harmonic injected air gap flux density
distribution when the stator slotting effect is neglected whilst Fig. 7.7 is the
waveform of the field function B2(h, a). The mathematical expressions of B(h, a)
and B2(h, a) are given in (7.15) and (7.16) respectively assuming b = Np (h − a).

Bðh; aÞ ¼ B1 cos½Np ðh  aÞ=2

 B3 cos½3Np ðh  aÞ=2
  ð7:15Þ
p 3p
 þ a; þa
Np Np
128 7 Analysis of Cogging Torque in Surface-Mounted …

B B1

B3
-π/Np+α 0 α π/Np+α 3π/Np+α

Fig. 7.6 Sinusoidal airgap flux density distribution with injected third harmonic

B2

-π/N p + α 0 α π/N p + α 3π/N p + α

Fig. 7.7 Waveform of B2(h, a)

B21 þ B23 1
B2 ðh; aÞ ¼ þ ð B21  B1 B3 Þ cos b
2
|fflfflfflffl{zfflfflfflffl} 2
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
1 2
B2
 B1 B3 cos 2b þ 3 cos 3b ð7:16Þ
|fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl} |fflfflfflfflfflffl{zfflfflfflfflfflffl}
2
3
4
 
p p
 þ a; þa
Np Np

As shown in Eq. (7.16), the field function has four terms: the average compo-
nent, the fundamental component, the second and the third harmonic component. It
can be concluded that not every component in Eq. (7.16) will contribute to the
cogging torque [20]. It is dependent on the slot number Ns and pole number Np
combinations in terms of which component in Eq. (7.16) will cause the cogging
torque. There exist certain slot/pole number combinations, for which none of the
component in Eq. (7.16) will give rise to the cogging torque, which means that the
injected third harmonic in airgap flux density has no influence on the generation of
cogging torque, as will be discussed later.
7.2 Analytical Prediction of Cogging Torque Based on Energy Method 129

For the calculation of the cogging torque, the machines with different slot
number Ns and pole number Np combinations are classified into four cases, namely
case A, case B, case C and case D.
Case A: when Ns and Np satisfy

Np ¼ kNs k ¼ 1; 2; 3; . . . ð7:17Þ

The cogging torque is:

 
ðB21  2B1 B3 ÞNs Np b0
TðaÞ ¼  Lef ðR2  R1 Þsin
2 2
sinðNp aÞ
4l0 2
B1 B3 Ns
þ Lef ðR22  R21 Þsin(Np b0 Þ sinð2Np aÞ ð7:18Þ
2l0
 
B23 Ns 3Np b0
 Lef ðR2  R1 Þsin
2 2
sinð3Np aÞ
4l0 2

In this case, all the components in (7.18) except the average one will contribute
to the cogging torque.
Case B: when Ns and Np satisfy

2k þ 1
Np ¼ Ns k ¼ 0; 1; 2; . . . ð7:19Þ
2

The cogging torque is:

B1 B3 Ns
TðaÞ ¼ Lef ðR22  R21 Þsin(Np b0 Þ sinð2Np aÞ ð7:20Þ
2l0

In this case, only the second harmonic component in Eq. (7.16) will result in the
cogging torque.
Case C: when Ns and Np satisfy

3k þ 1 3k þ 2
Np ¼ Ns or Np ¼ Ns k ¼ 0; 1; 2; . . . ð7:21Þ
3 3

The cogging torque is:

 
B23 Ns 3Np b0
TðaÞ ¼ Lef ðR2  R1 Þsin
2 2
sinð3Np aÞ ð7:22Þ
4l0 2

In this case, only the third harmonic component in Eq. (7.16) can induce the
cogging torque, when k = 0, Np = Ns/3 or Np = 2Ns/3; when k = 1, Np = 4Ns/3 or
Np = 5Ns/3.
130 7 Analysis of Cogging Torque in Surface-Mounted …

Case D: For the rest of slot/pole number combinations, the cogging torque is
zero. In this case, no component in Eq. (7.16) has contributions to the cogging
torque, which means that the third harmonic in airgap flux density will not affect the
cogging torque.

7.2.4 Cogging Torque for Sine and Sine+3rd Airgap Flux

Density Distributions

The winding configurations which are most commonly employed in three-phase

PM machines can be classified into two categories, i.e., integral slot distributed
winding (ISDW) and fractional slot concentrated winding (FSCW) [21]. Due to the
feasibility of three phase winding, the slot number Ns and pole number Np under
such a constraint are as below:
 
Ns Np
is an integer; t ¼ GCD Ns ;
3t 2

For three-phase FSCW, to guarantee the pitch factor Ky  0.866, one constraint
on slot number Ns and pole number Np is that 0.25  spp  0.5 (spp is defined as
slot per phase per pole), which means 4 * Ns/3  Np  2 * Ns/3.
For three-phase ISDW, considering the manufacturing difficulty of too many
stator slots, the spp is confined to 1  spp  6.
Under the constraints for three phase FSCW and ISDW, when the sinusoidal
airgap flux density distribution is sinusoidal, the cogging torque is zero no matter
what the slot/pole number combinations are.
For the sinusoidal with third harmonic airgap flux density distribution, slot
number Ns and pole number Np combinations can be classified into two cases for
simplification.
Case one: When Np = 4 * Ns/3 (spp = 0.25) or Np = 2 * Ns/3 (spp = 0.5) or
Np = Ns/3 (spp = 1).
Cogging torque can be expressed as Eq. (7.22). In this case, the third harmonic
in airgap flux density can give rise to the cogging torque. Slot/pole number com-
binations commonly used belong to this case are 6-slot/8-pole (spp = 0.25), 12-slot/
8-pole (spp = 0.5) and 24-slot/8-pole (spp = 1), as will be investigated in next
sections.
Case two: For the rest of slot/pole number combinations, i.e., spp 6¼ 0.25, 0.5,
and 1, the cogging torque is zero, which indicates that third harmonic in air gap flux
density has no influence on cogging torque.
7.3 Three-Phase SPM Machines with Shaped PMs 131

7.3 Three-Phase SPM Machines with Shaped PMs

Figure 7.8 shows the 8-pole rotors equipped with Sine-shaped and Sine+3rd-shaped
PMs on their surfaces, the magnetization direction of the PMs are parallel.
Figure 7.8a and b depict one piece of Sine-shaped PM and Sine+3rd-shaped PM
respectively. It is demonstrated that the optimal third harmonic is 1/6 of the fun-
damental one for maximizing output torque [13]. The PM edge thickness Δt, as
shown in Fig. 7.9a, is optimized for achieving nearly sinusoidal and sinusoidal with
third harmonic airgap flux density distributions.
Figure 7.10a and b shows the airgap flux density distributions and their harmonic
contents respectively by FE analysis in SPM machines with Sine-shaped and Sine
+3rd-shaped PMs whilst the stator slotting effect is neglected. The PM edge thickness
Δt is set to be 1 mm. As can be seen in Fig. 7.10b, the open-circuit airgap flux density
is purely sinusoidal for the Sine-shaped PM design whilst only fundamental and third
harmonic components survive in the open-circuit airgap flux density owing to the
Sine+3rd-shaped PM design. Although the ratio of the amplitude of third harmonic to
the fundamental one is not 1/6, which is optimal for maximizing the torque, it will not
prohibit us from investigating the influence of the third harmonic airgap flux density
on the cogging torque, which will be carried out later.

(a) Sine rotor (b) Sine+3rd rotor

Fig. 7.8 8-pole SPM rotors with different PM shapes

132 7 Analysis of Cogging Torque in Surface-Mounted …

(a) Sine shaped PM

(a) Sine+3rd shaped PM

Fig. 7.9 Different PM shapes

Based on the analyses in Sect. 7.2, for three phase SPM machines whose spp
equals to 0.25, 0.5, and 1, injecting third harmonic into sinusoidal airgap flux
density will cause an increase of amplitude of the cogging torque. Figure 7.11
shows four SPM machines with different slot/pole number combinations and
Sine-shaped and Sine+3rd-shaped PMs, namely, 6-slot/8-pole (spp = 0.25), 12-slot/
8-pole (spp = 0.5), 24-slot/8-pole (spp = 1), 12-slot/10-pole (spp = 0.4). For all the
four machines to be analyzed, PM edge thickness is keep to be the same, i.e., 1 mm.
Tooth body width and yoke height are optimized to have a good stator core uti-
lization. All the stator, rotor, and PM parameters of the four machines are listed in
Table 7.1.

7.4 FE Analysis

Figure 7.12 shows the variation of cogging torque with rotor position for 6-slot/
8-pole machines equipped with Sine-shaped PMs and Sine+3rd-shaped PMs, whilst
Figs. 7.13 and 7.14 for 12-slot/8-pole and 24-slot/8-pole machines, respectively.
The period of the cogging torque is the same for all the three machines since they
have the same smallest common multiple between the stator slot numbers and the
7.4 FE Analysis 133

1.2
Sine Sine+3rd
0.8
Airgap flux density (T)
0.4

0.0

-0.4

-0.8

-1.2
0 60 120 180 240 300 360
Rotor position (elec.deg)
(a) Waveforms
1.2
Sine Sine+3rd
1.0
Airgap flux density (T)

0.8

0.6

0.4

0.2

0.0
1 3 5 7 9 11 13
Harmonic order
(b) Harmonics

Fig. 7.10 Open-circuit airgap flux density distributions in slotless SPM machines with
Sine-shaped and Sine+3rd-shaped PMs

rotor poles [3]. It can be observed that the peak cogging torque values of the three
machines with Sine-shaped PMs is very small, as listed in Table 7.2. It should be
noted that the FE calculated cogging torque differs from the cogging torque
obtained by the analytical derivation in Eq. (7.22) due to the simplified analytical
model in this chapter. However, what is important is that for all the three SPM
machines, the injected third harmonic airgap flux density will greatly increase the
peak cogging torque.
134 7 Analysis of Cogging Torque in Surface-Mounted …

(a) 6-slot/8-pole Sine-shaped (b) 6-slot/8-pole Sine+3rd-shaped

(c) 12-slot/8-pole Sine-shaped (d) 12-slot/8-pole Sine+3rd-shaped

Fig. 7.11 Sine-shaped and Sine+3rd-shaped SPM machines with different slot and pole number
combinations
7.4 FE Analysis 135

(e) 24-slot/8-pole Sine-shaped (f) 24-slot/8-pole Sine+3rd-shaped

(g) 12-slot/10-pole Sine-shaped (h) 12-slot/10-pole Sine+3rd-shaped

Fig. 7.11 (continued)

Table 7.1 Main machine Phase number 3

design specifications and
Stator outer diameter 90 mm
parameters
Stator inner diameter 53 mm
Axial length 50 mm
Minimum airgap length 1 mm
Maximum PM thickness 3 mm
Pole arc to pole pitch ratio 1
PM remanence (20 °C) 1.2 T
Relative recoil permeability 1.05
136 7 Analysis of Cogging Torque in Surface-Mounted …

60
sine sine+3rd

Cogging Torque (mNm)

40

20

-20

-40

-60
0 3 6 9 12 15
Rotor position (mech.deg)

Fig. 7.12 Cogging torque waveforms of 6-slot/8-pole SPM machines with Sine-shaped and Sine
+3rd-shaped PMs

120
sine sine+3rd
Cogging Torque (mNm)

80

40

-40

-80

-120
0 3 6 9 12 15
Rotor position (mech.deg)

Fig. 7.13 Cogging torque waveforms of 12-slot/8-pole SPM machines with Sine-shaped and Sine
+3rd-shaped PMs

Figure 7.15 depicts the variation of cogging torque with rotor position in a
12-slot/10-pole machine with Sine-shaped Sine+3rd-shaped PMs. After injecting
third harmonic into the sinusoidal airgap flux density, the amplitude of cogging is
about 5 mN m, which is much smaller than that of the aforementioned 6-slot/
8-pole, 12-slot/8-pole and 24-slot/8-pole SPM machine. The influence of the
injected third harmonic in aigap flux density on the cogging torque of a 12-slot/
10-pole machine is negligible, which agrees with the analysis.
7.4 FE Analysis 137

80
sine sine+3rd
60

Cogging Torque (mNm) 40

20
0
-20
-40
-60
-80
0 3 6 9 12 15
Rotor position (mech.deg)

Fig. 7.14 Cogging torque waveforms of 24-slot/8-pole SPM machines with Sine-shaped and
Sine+3rd-shaped PMs

Table 7.2 Peak cogging torque of 6-slot/8-pole, 12-slot/8-pole and 24-slot/8-pole machines with
Sine-shaped and Sine+3rd-shaped PMs
Peak cogging torque (mN m)
6-slot/8-pole Sine-shaped 3.74
Sine+3rd-shaped 47.23
12-slot/8-pole Sine-shaped 4.52
Sine+3rd-shaped 96.38
24-slot/8-pole Sine-shaped 10.13
Sine+3rd-shaped 66.30
12-slot/10-pole Sine-shaped 2.05
Sine+3rd-shaped 5.19

6
Sine Sine+3rd
Cogging Torque (mNm)

-2

-4

-6
0 1 2 3 4 5 6
Rotor position (mech.deg)

Fig. 7.15 Cogging torque waveforms of 12-slot/10-pole SPM machines with Sine-shaped and
Sine+3rd-shaped PMs
138 7 Analysis of Cogging Torque in Surface-Mounted …

7.5 Experimental Verification

The 12-slot/10-pole SPM machine (outer diameter = 90 mm and active axial

length = 50 mm) with Sine+3rd-shaped PMs whose edge thickness Δt = 1 mm has
been prototyped and tested for validation. Figure 7.16 shows the pictures of the
stator together with the rotor equipped with Sine+3rd-shaped PMs.
A platform which is frequently used for the measurement of cogging torque and
static torque in electric machines is established [22], as shown in Fig. 7.17. The
stator of the 12-slot/10-pole SPM machine is fixed to the lathe while the rotor

(a) Stator

(b) Sine+3rd rotor

Fig. 7.16 Prototype 12-slot/10-pole Sine+3rd-shaped SPM machine
7.5 Experimental Verification 139

Prototype

Balanced beam

Lathe

Digital scale

Fig. 7.17 Platform of cogging torque and static torque measurement

rotates. The shaft is connected to one side of a balanced beam, while the other side
to a digital scale. Through the weight scaled on the digital scale, the value of
cogging torque can be obtained accordingly.
The 2-D FE calculated and measured cogging torques of the SPM machines with
Sine+3rd-shaped PMs are given in Fig. 7.18. The measured cogging torque is
slightly higher than that obtained by FE analysis owing to the existence of some

20
15
Cogging Torque (mNm)

10
5
0
-5
-10
-15
-20
0 2 4 6 8 10 12
Rotor position (mech.deg)

Fig. 7.18 Variation of FE predicted and measured cogging torques of Sine+3rd-shaped 12-slot/
10-pole SPM machines with rotor position
140 7 Analysis of Cogging Torque in Surface-Mounted …

manufacturing tolerance. However, since the measured peak cogging torque value
is less than 10 mN m, one can still conclude that the cogging torque of a 12-slot/
10-pole SPM machine with Sine+3rd-shaped PMs can be negligible under ideal
condition.

7.6 Summary

This chapter analyzes the cogging torque of SPM machines with shaped magnet
considering different slot/pole number combinations. Expressions of cogging torque
are derived when the SPM machines are equipped with Sine-shaped and Sine
+3rd-shaped PMs on the rotor. It is found that the Sine+3rd-shaped PMs will
increase the amplitude of cogging torque for three phase machines whose slot
number and pole number satisfying spp = 0.25 or spp = 0.5 and spp = 1 compared
to the PM Sine-shaped design, whilst for other slot/pole number combinations, its
influence on cogging torque is negligible. FE analysis carried out on four three
phase SPM machines, i.e., 6-slot/8-pole (spp = 0.25), 12-slot/8-pole (spp = 0.5),
24-slot/8-pole (spp = 1) and 12-slot/10-pole (spp = 0.4) verifies the conclusion.
Finally, a SPM machine with Sine+3rd-shaped PMs is prototyped and the measured
results verify the analysis.

References

1. Z.Q. Zhu, D. Howe, Analytical prediction of the cogging torque in radial-field permanent
magnet brushless motors. IEEE Trans. Magn. 28(2), 1371–1374 (1992)
2. T. Ishikawa, G.R. Slemon, A method of reducing ripple torque in permanent-magnet motors
without skewing. IEEE Trans. Magn. 29(2), 2028–2031 (1993)
3. Z.Q. Zhu, D. Howe, Influence of design parameters on cogging torque in permanent magnet
machines. IEEE Trans. Energy Convers. 15(4), 407–412 (2000)
4. S.M. Hwang, J.B. Eom, Y.H. Jung, D.W. Lee, B.S. Kang, Various design techniques to
reduce cogging torque by controlling energy variation in permanent magnet motors. IEEE
Trans. Magn. 37(4), 2806–2809 (2001)
5. N. Bianchi, S. Bolognani, Design techniques for reducing the cogging torque in
surface-mounted PM motors. IEEE Trans. Ind. Appl. 38(5), 1259–1265 (2002)
6. Y. Yang, X. Wang, R. Zhang, T. Ding, R. Tang, The optimization of pole arc coefficient to
reduce cogging torque in surface-mounted permanent magnet motors. IEEE Trans. Magn.
42(4), 1135–1138 (2006)
7. Y. Pang, Z.Q. Zhu, Z.J. Feng, Cogging torque in cost-effective surface-mounted
permanent-magnet machines. IEEE Trans. Magn. 47(9), 2269–2276 (2011)
8. P. Zheng, J. Zhao, J. Han, J. Wang, Z. Yao, R. Liu, Optimization of the magnetic pole shape
of a permanent-magnet synchronous motor. IEEE Trans. Magn. 43(6), 2531–2533 (2007)
9. Y. Li, J.W. Xing, T.B. Wang, Y.P. Lu, Programmable design of magnet shape for permanent
magnet synchronous motors with sinusoidal back EMF waveforms. IEEE Trans. Magn. 44(9),
2163–2167 (2008)
References 141

10. S. Ruangsinchaiwanich, Z.Q. Zhu, D. Howe, Influence of magnet shape on cogging torque
and back-emf waveform in permanent magnet machines, in Proceedings of International
Conference on Electrical Machines and Systems, , vol. 1 (2005), pp. 284–289
11. A. Evans, Salient pole shoe shapes of interior permanent magnet synchronous machines, in
Proceedings of 19th International Conference on Electrical Machines, Sept 2010, pp. 1–6
12. Y. Li, J.B. Zou, Y.P. Lu, Optimum design of magnet shape in permanent-magnet synchronous
motors. IEEE Trans. Magn. 39(6), 3523–3526 (2003)
13. K. Wang, Z.Q. Zhu, G. Ombach, Torque enhancement of surface-mounted permanent magnet
machine using third-order harmonic. IEEE Trans. Magn. 50(3), 104–113 (2014)
14. K. Wang, Z.Q. Zhu, G. Ombach, W. Chlebosz, Average torque improvement of interior
permanent-magnet machine using 3rd harmonic in rotor shape. IEEE Trans. Ind. Electron. 61,
5047–5057 (2014)
15. H.D. Lin, D. Wang, D.Z. Liu, J.Q. Chen, Influence of magnet shape on torque behavior in
surface-mounted permanent magnet motors, in 17th International Conference on Electrical
Machines and Systems, Hangzhou, China (2014), pp. 44–47
16. C.S. Koh, J.S. Seol, New cogging-torque reduction method for brushless permanent-magnet
motors. IEEE Trans. Magn. 39, 3503–3506 (2003)
17. L. Wu, W. Jin, J. Ni, J. Ying, A cogging torque reduction method for surface mounted
permanent magnet motor, in Proceedings of International Conference on Electrical Machines
and Systems, Seoul, Korea (2007), pp. 769–773
18. H. Jin, J.W. Reu, B.W. Kim, G.H. Kang, Vibration reduction of IPM-type BLDC motor using
negative third harmonic elimination method of air-gap flux density. IEEE Trans. Ind. Appl.
47(3), 1300–1309 (2010)
19. L. Zhu, S.Z. Jiang, Z.Q. Zhu, C.C. Chan, Analytical methods for minimizing cogging torque
in permanent-magnet machines. IEEE Trans. Magn. 45(4), 2023–2031 (2009)
20. Z.Q. Zhu, S. Ruangsinchaiwanich, D. Howe, Synthesis of cogging-torque waveform from
analysis of a single stator slot. IEEE Trans. Ind. Appl. 42(3), 650–657 (2006)
21. A.M. EL-Refaie, Fractional-slot concentrated-windings synchronous permanent magnet
machines: Opportunities and challenges. IEEE Trans. Ind. Electron. 57(1), 107–121 (2010)
22. Z.Q. Zhu, A simple method for measuring cogging torque in permanent magnet machines, in
Proceedings of IEEE Power Energy Society General Meeting (2009), pp. 1–4
Chapter 8
Influence of Harmonics on Radial Force
and Vibration of Surface-Mounted
Permanent Magnet Machines

This chapter analyzes the electromagnetic vibrations of 12-slot/10-pole SPM

machines with both Sine-shaped and Sine+3rd-shaped PMs. Radial force density
expressions are calculated by Maxwell stress tensor method under both no load and
on load conditions. The magnitude of the lowest mode of the machines with
Sine-shaped and Sine+3rd-shaped PMs, which is 2nd mode for 12-slot/10-pole
ones, are investigated. Since the amplitude of the 5th (fundamental in electrical
angle) airgap flux density is increased for Sine+3rd-shaped PMs compared with
Sine-shaped PMs, the 2nd mode due to the interaction between the 5th PM field and
7th harmonic (either from slotting or armature field or both) is larger. However, it is
found the 15th PM field (third harmonic in electrical angle) in Sine+3rd-shaped
PMs can interact with 17th field to generate an negative amplitude of 2nd mode to
cancel part of the increase due to 5th PM field, though the total magnitude of 2nd
mode is about 8.8% larger compared with the Sine-shaped PMs under both no load
and rated load conditions. Furthermore mode analysis and vibration behaviors are
conducted through ANSYS. Finally, experiments are carried out on the prototyped
12-slot/10-pole SPM machines.

8.1 Introduction

Permanent magnet machines equipped with fractional slot concentrated windings

are proved to be competitive in torque density, fault tolerance and flux weakening
capability and efficiency. However, the electromagnetic vibrations of fractional slot
PM machines associated with the radial magnetic force is much more serious
compared with their integral slot counterparts owing to their rich sub-harmonic and
super-harmonic contents in stator magnetomotive force [1, 2].

© Springer Nature Singapore Pte Ltd. 2019 143

K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_8
144 8 Influence of Harmonics on Radial Force …

Various analytical models [3–6] and finite element analysis (FEA) models [7, 8]
have been proposed to evaluate the air-gap magnetic field flux density components
and resultant radial magnetic force density harmonics in surface-mounted perma-
nent magnet (SPM) machines. Maxwell’s stress tensor method is generally used to
calculate the radial magnetic force density between the PM magnetic field and the
stator MMF. The influence of slot and pole number combination on radial magnetic
force and vibration has been fully investigated [3, 4, 9, 10]. It has been demon-
strated that low modes of vibration are excited by the radial magnetic force dis-
tribution of FSCW machines resulting much higher vibration level than that of
traditional integral slot PM machine.
Plenty of efforts have been made to reduce vibrations induced by radial magnetic
force both in SPM and IPM machines. In [11], by enlarging the air gap partly where
the magnetic force density is large, the vibration of the machine with concentrated
windings can be reduced. In [12], by using field weakening, the current waveforms
of permanent-magnet motors are adjusted to minimize radial magnetic force ripple
at the expense of increased power consumption. The techniques of optimizing
shape of flux barrier in the rotor core and shape of stator tooth to reduce vibration
and noise of the motor were proposed in [13], whilst in [14] the relationship
between stepped skewed rotor and vibration are established. The multi-objective
optimization method was utilized in [15] to contribute to reduction of torque and
radial force ripple in external-rotor SPM machines. Rotor pole shaping and rotor
notching in IPM machines were studied in [16] to reduce the dominant order of the
radial force, which will improve the acoustic behavior. In [17, 18], the minimization
of vibration and noise is realized by forming optimal notches corresponding to the
shape of the magnetic field on the rotor pole face.
PM shaping technique is one of the effective methods to reduce cogging torque
and torque ripple and to lower noise and vibration level in SPM machines [19–24].
The Sine+3rd-shaping technique is proved to be competitive in both high torque
density and low torque ripple. It is proved that machines with Sine+3rd-shaped PM
can achieve about 9% higher torque density than ones with the sine-shaped PMs.
However, the third harmonic airgap flux density is introduced in the Sine
+3rd-shaped PM and its influence on the characteristic of radial magnetic force is
unclear.
This chapter investigated the difference of radial magnetic force between
machines with Sine-shaped or Sine+3rd-shaped PMs in 12-slot/10-pole machines.
Airgap flux density for both shaped PMs under no load and rated load conditions
are obtained from finite element (FE) analysis. Then, the radial magnetic force
density is calculated by using Maxwell’s stress tensor method. With only the lowest
mode of radial magnetic force considered, main components of 2nd are discussed.
Particular attention is paid to the contribution of the third harmonic in PM field to
the 2nd mode radial magnetic force.
8.2 12-Slot/10-Pole PM Machines with Shaped Magnets 145

8.2 12-Slot/10-Pole PM Machines with Shaped Magnets

Figure 8.1 shows the 10-pole rotors equipped with Sine-shaped and Sine
+3rd-shaped PMs on their surfaces, the magnetization direction of the PMs are
parallel. Figure 8.2a and b depict one piece of Sine-shaped PM and Sine
+3rd-shaped PM respectively. It is demonstrated that the optimal third harmonic is
1/6 of the fundamental one for maximizing output torque. The PM edge thickness
Δt, as shown in Fig. 8.2a, is optimized for achieving nearly sinusoidal and sinu-
soidal with third harmonic airgap flux density distributions.

(a) Sine rotor

(b) Sine+3rd rotor

Fig. 8.1 8-pole SPM rotors with different PM shapes

146 8 Influence of Harmonics on Radial Force …

(a) Sine shaped PM

(b) Sine+3rd shaped PM

Fig. 8.2 Different PM shapes

8.3 Radial Force Density Distributions

8.3.1 Radial Force Density Under No Load Condition

The radial component of flux density which is generated by the rotor Sine-shaped or
Sine+3rd-shaped magnets without considering the stator slotting effects can be
expressed as:

Bm slotless ðhs ; tÞ ¼ B1 cos½pðhs  xr tÞ

ð8:1Þ
or ¼ B01 cos½pðhs  xr tÞ  B3 cos½3pðhs  xr tÞ

where, hs, t, p, xr, B1, B01 , B3 are angular position along the stator referred to the
phase A winding axis, the time, the number of pole pairs, the mechanical speed of
rotor, the amplitude of the fundamental airgap flux density with Sine-shaped PMs,
and the amplitude of fundamental and third harmonic airgap flux density with Sine
+3rd-shaped PMs respectively. The simplified relative airgap permeance function is
given by (8.2):

kðhs Þ ¼ k0 þ k1 cosðNs hs Þ ð8:2Þ

According to Eqs. (8.1) and (8.2), open circuit airgap flux density produced by
PMs can be expressed by
8.3 Radial Force Density Distributions 147

Bg PM ðhs ; tÞ ¼ Bm slotless ðhs ; tÞkðhs Þ ð8:3Þ

The harmonic order and angular frequency of the open-circuit air gap flux
density can be summarized as:
For machines with Sine-shaped PMs

½p; pxr ; ½Ns  p; pxr  ð8:4Þ

For machines with Sine+3rd-shaped PMs

½p; pxr ; ½Ns  p; pxr 

ð8:5Þ
½3p; 3pxr ; ½Ns  3p; 3pxr 

Based on the Maxwell stress tensor method, the open-circuit radial force density
in the airgap can be given by

B2g PM ðhs ; tÞ
fr PM ðhs ; tÞ ¼
2l0
1 X
X 1 ð8:6Þ
¼ fr max cosðmhs  kxr tÞ
m¼0 k¼0

where m is the mode number and k time denotes harmonic order whilst fr_max is the
amplitude of each component of radial force density. Since the deformation
amplitude is inversely proportional to m4 and the frequency of lower mode of the
stator can be more likely in the range of machines’ speed. Besides the zeroth mode
and mode with whose frequency is zero will not cause any vibration of the stator.
Only the lowest nonzero mode whose frequency is not zero will be considered in
this chapter.
Tables 8.1 and 8.2 shows the lowest harmonic order of radial force density under
no load condition for Sine-shaped and Sine+3rd-shaped PMs respectively. It can be
seen that both the lowest order of radial force density for Sine-shaped and Sine
+3rd-shaped PMs are |2p − Ns| whose frequency is 2pxr. However, the |2p − Ns|
order of mode for machines with Sine+3rd-shaped PMs comes from not only the
interaction between the fundamental component and fundamental one modulated
the stator, but also the fundamental component and the third harmonic modulated
the stator, the third harmonic and fundamental one modulated the stator.

Table 8.1 Lowest harmonic order of radial force density under no load condition for sine-shaped
PMs
Harmonic order of magnetic field Mode number m Time harmonic k
p and (Ns − p) |2p − Ns| 2p
148 8 Influence of Harmonics on Radial Force …

Table 8.2 Lowest harmonic order of radial force density under no load condition for sine
+3rd-shaped PMs
Harmonic order of magnetic field Mode number m Time harmonic k
p and (Ns − p) |2p − Ns| 2p
p and (Ns − 3p) |2p − Ns| 2p
3p and (Ns+ p) |2p − Ns| 2p

8.3.2 Radial Force Density Under Rated Load

For three phase PM machines with symmetry 60° phase belt winding supplied by
balanced three-phase, sinusoidally time-varying currents, the airgap flux density pro-
duced by armature currents neglecting the stator slotting effects can be expressed as:
X
1
Ba slotless ðhs ; tÞ ¼ Bn sinðnhs  pxr tÞ
n¼6k1
k¼1;2;...
ð8:7Þ
P1
þ Bn sinðnhs þ pxr tÞ
n¼6k þ 1
k¼0;1;2;...

where Bn is the amplitude of nth harmonic. Here one of the harmonic in 6k − 1 is

chosen to be the working one to generate the average torque.
The armature airgap flux density considering the stator slotting effects is given by:

Bg a ðhs ; tÞ ¼ Ba slotless ðhs ; tÞkðhs Þ ð8:8Þ

Thus, ignoring the saturation effects of the machine, the airgap flux density
under both PM and stator currents excitations can be expressed as:

Bg ðhs ; tÞ ¼ BPM ðhs ; tÞ þ Bg a ðhs ; tÞ ð8:9Þ

The radial force density under rated load condition can be given by:

B2g ðhs ; tÞ
fr ðhs ; tÞ ¼ ð8:10Þ
2l0

Tables 8.3 and 8.4 shows the lowest harmonic order of radial force density under
rated load for Sine-shaped and Sine+3rd-shaped PMs respectively. Since the
amplitude of the armature airgap flux density is relatively low compared to the field
due to PMs, radial force density due to the armature itself is ignored. Compared
with Sine-shaped situation when only the pth PM field interacts with the armature
magnetic field to the lowest mode of radial force density, the 3pth PM field can also
induce the lowest mode by interacting with some of the harmonics in the armature
field.
8.3 Radial Force Density Distributions 149

Table 8.3 Main harmonic orders of radial force density under rated load for sine-shaped PM
Harmonic order of magnetic field Mode number m Time harmonic k
p and (Ns − p) |2p − Ns| 2p
p and 6k + 1 |6k + 1 − p| 2p

Table 8.4 Main harmonic orders of radial force density under rated load for sine+3rd-shaped PM
Harmonic order of magnetic field Mode number m Time harmonic k
p and (Ns − p) |2p − Ns| 2p
p and (Ns − 3p) |2p − Ns| 2p
p and 6k + 1 |6k + 1 − p| 2p
3p and (Ns+ p) |2p − Ns| 2p
3p and 6k + 1 |6k − 1 − 3p| 2p
3p and 6k − 1 |6k + 1 − 3p| 4p

For 12-slot/10-pole machines, the main harmonic which generates the average
torque is the 5th component. MMF harmonics produced by stator currents are 1st,
5th, 7th, 11th, 13th, 17th, etc., among which only the 5th component will interact
with the PM MMF to generate useful torque.
Figure 8.3 shows the radial airgap flux density distributions for machines with
either Sine-shaped or Sine+3rd-shaped PM under both no load and rated load
conditions. The 15th harmonic arises due to the Sine+3rd-shaped design while the
amplitude of the fundamental one increases about 8%. It can also be seen that the
7th, 17th and 19th appear mainly owing to the slotting effects whilst the 1st, 11th
and 13th coming from the armature field whose amplitude is relatively small.
Figure 8.4 depicts variations of airgap flux density distributions with time at
hs = 0 under both no load and rated load in one electrical period. As can been seen,
only the fundamental airgap flux density survives due to the Sine-shaped PM design
whilst for Sine+3rd-shaped PMs, the third harmonic is injected into the airgap flux
density with small portion of the 5th harmonic.
Figure 8.5 is the radial force density distributions and their harmonic contents
under both no load and rated load condition for the two PM designs. Under either
no load or load condition, the Sine+3rd-shaped design sees an increase about 9%
for the second mode amplitude compared with the Sine-shaped. The 4th and 6th
harmonic arise under load condition for both Sine-shaped and Sine+3rd-shaped
design whilst the Sine+3rd-shaped design witnesses a slight reduction of 10th
harmonic in radial force density compared to the Sine-shaped one. However, this is
out of the scope of this chapter.
150 8 Influence of Harmonics on Radial Force …

Radial Airgap Flux Density (T)

Sine_no load Sine_on load
1.2
Sine+3rd_no load Sine+3rd_on load
0.8

0.4

0.0

-0.4

-0.8

-1.2
0 60 120 180 240 300 360
Rotor position (mech.deg)
(a) Waveforms.
1.0
Radial Airgap Flux Density (T)

Sine_no load
0.8 Sine_on load
Sine+3rd_no load
0.6 Sine+3rd_on load

0.4

0.2

0.0
1 3 5 7 9 11 13 15 17 19
-0.2
Harmonic order
(b) Harmonics.

Fig. 8.3 Variations of airgap flux density distributions with angular position along the stator
under both no load and rated load

The 2nd mode of radial force density mainly origins from two sources, i.e., the
interaction between 5th PM field and 7th airgap flux density due to either slotting
effects or armature current or both, and the interaction between 15th PM field and
17th airgap field. Figure 8.6 shows the main components of 2nd mode of radial
force density. It is important to note that for the Sine+3rd-shaped design the
magnitude of 2nd mode resulting from the interaction between the injected 15th
harmonic of PM field and the 17th harmonic field (either from slotting effects or
armature current) is negative. Since the amplitude of fundamental PM field for the
Sine+3rd-shaped design is increased compared with the Sine-shaped one, the total
magnitude of 2nd mode is larger.
8.3 Radial Force Density Distributions 151

Radial Airgap Flux Density (T)

1.2 Sine_no load Sine_on load
Sine+3rd_no load Sine+3rd_on load
0.8

0.4

0.0

-0.4

-0.8

-1.2
0 4 8 12 16 20
Time (ms)
(a) Waveforms.
1.2
Radial Airgap Flux Density (T)

Sine_no load
1.0 Sine_on load
Sine+3rd_no load
0.8 Sine+3rd_on load

0.6

0.4

0.2

0.0
1 3 5 7 9 11 13 15 17 19
Harmonic order
(b) Harmonics.

Fig. 8.4 Variations of airgap flux density distributions with time at hs= 0 under both no load and
rated load

Figure 8.7 shows the variations of radial force density distributions with time at
h = 0 under both no load and rated load. It can be seen that the 4th harmonic of
electrical frequency arises for the Sine+3rd-shaped design under both no load and
rated load conditions.
152 8 Influence of Harmonics on Radial Force …

500
Sine_no load

Radial Force Density (kN/m2 )

Sine_on load
400 Sine+3rd_no load
Sine+3rd_on load
300

200

100

0
0 60 120 180 240 300 360
Rotor position (mech.deg)
(a) Waveforms.
160
Radial Force Density (kN/m2 )

Sine_no load
140 Sine_on load
120 Sine+3rd_no load
Sine+3rd_on load
100
80
60
40
20
0
1 2 3 4 5 6 7 8 9 10
Harmonic order
(b) Harmonics.

Fig. 8.5 Variations of radial force density distributions with rotor position under both no load and
rated load
8.4 Mode Analysis 153

30

of Radial Force Density (kN/m2)

Main Components of 2nd Mode
Sine_no load
25 Sine_on load
Sine+3rd_no load
20
Sine+3rd_on load
15
10
5
0
-5 5th and 7th 15th and 17th total

-10
Harmonic order

Fig. 8.6 Main harmonic components of 2nd mode of radial force density

8.4 Mode Analysis

The mechanical equation for the mechanical system with free boundary condition
can be represented as

½Mf€xg þ ½Kfx_ g ¼ 0 ð8:11Þ

where x is the node displacement, M and K are the global mass and stiffness
matrices. In order to get the mode shapes and corresponding natural frequencies,
3-D structural FEA analysis is carried out. The first two mode shapes and their
natural frequencies are obtained and illustrated in Fig. 8.8.

8.5 Summary

The electromagnetic vibrations of 12-slot/10-pole SPM machines with both

Sine-shaped and Sine+3rd-shaped PMs is analyzed. it is found the 15th PM field
(third harmonic in electrical angle) in Sine+3rd-shaped PMs can interact with 17th
field to generate an negative amplitude of 2nd mode to cancel part of the increase
due to 5th PM field, though the total magnitude of 2nd mode is about 8.8% larger
for the Sine-shaped PMs under both no load and rated load conditions.
Furthermore, mode analysis and vibration behaviors are conducted through
ANSYS.
154 8 Influence of Harmonics on Radial Force …

400
Sine_no load Sine_on load

Radial Force Density (kN/m2 )

Sine+3rd_no load Sine+3rd_on load
300

200

100

0
0 4 8 12 16 20
Time (ms)
(a) Waveforms.
200
Radial Force Density (kN/m2 )

Sine_no load
Sine_on load
150 Sine+3rd_no load
Sine+3rd_on load
100

50

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

Fig. 8.7 Variations of radial force density distributions with time at h = 0 under both no load and
rated load
8.5 Summary 155

(a) 860 Hz (b) 2234.7Hz

(c) 3846.5Hz (d) 5308.8Hz

(e) 5968.3Hz

Fig. 8.8 The modal results by 3D FEM simulation

156 8 Influence of Harmonics on Radial Force …

References

1. A. El-Refaie, Fractional-slot concentrated-windings synchronous permanent magnet machi-

nes: opportunities and challenges. IEEE Trans. Ind. Electron. 57(1), 107–121 (2010)
2. F. Magnussen, H. Lendenmann, Parasitic effects in PM machines with concentrated windings.
IEEE Trans. Magn. 42(10), 3395–3397 (2006)
3. J. Wang, Z.P. Xia, S.A. Long, D. Howe, Radial force density and vibration characteristics of
modular permanent magnet brushless ac machine. IEE Proc. Inst. Elect. Eng. Elect. Power
Appl. 153(6), 793–801 (2006)
4. Z.Q. Zhu, Z.P. Xia, L.J. Wu, G.W. Jewell, Analytical modeling and finite-element
computation of radial vibration force in fractional-slot permanent-magnet brushless machines.
IEEE Trans. Ind. Appl. 46(5), 1908–1918 (2010)
5. R. Islam, I. Husain, Analytical model for predicting noise and vibration in permanent-magnet
synchronous motors. IEEE Trans. Ind. Appl. 46(6), 2346–2354 (2010)
6. G. Dajaku, D. Gerling, The influence of permeance effect on the magnetic radial forces of
permanent magnet synchronous machines. IEEE Trans. Magn. 49(6), 2953–2966 (2013)
7. Y.S. Chen, Z.Q. Zhu, D. Howe, Vibration of PM brushless machines having a fractional
number of slots per pole. IEEE Trans. Magn. 42(10), 3395–3397 (2006)
8. G. Dajaku, D. Gerling, Magnetic radial force density of the PM machine with 12-teeth/
10-poles winding topology, in Proceedings of IEEE International Electric Machines and
Drives Conference, 3–6 May 2009, pp. 157–164
9. T. Sun, J.-M. Kim, G.-H. Lee, J.-P. Hong, M.-R. Choi, Effect of pole and slot combination on
noise and vibration in permanent magnet synchronous motor. IEEE Trans. Magn. 47(5),
1038–1041 (2011)
10. M. Valavi, A. Nysveen, R. Nilssen, R.D. Lorenz, T. Rølvåg, Influence of pole and slot
combinations on magnetic forces and vibration in low-speed PM wind generators. IEEE
Trans. Magn. 50(5), Art. no. 8700111 (2014)
11. Y. Asano, Y. Honda, H. Murakami, Y. Takeda, S. Morimoto, Novel noise improvement
techniques for a PMSM with concentrated winding, in Proceedings of IEEE Power
Conversion Conference, vol. 2 (2002), pp. 460–465
12. G. Jiao, C. Rahn, Field weakening for radial force reduction in brushless permanent-magnet
dc motors. IEEE Trans. Magn. 40(5), 3286–3292 (2004)
13. J.-W. Jung, S.-H. Lee, G.-H. Lee, J.-P. Hong, D.-H. Lee, K.-N. Kim, Reduction design of
vibration and noise in IPMSM type integrated starter and generator for HEV. IEEE Trans.
Magn. 46(6), 2454–2457 (2010)
14. J.-W. Jung, D.-J. Kim, J.-P. Hong, G.-H. Lee, S.-M. Jeon, Experimental verification and
effects of step skewed rotor type IPMSM on vibration and noise. IEEE Trans. Magn. 47(10),
3661–3664 (2011)
15. J. Krotsch, T. Ley, B. Piepenbreier, Reduction of torque and radial force fluctuation in
permanent magnet synchronous motors by means of multi-objective optimization, in
Proceedings of 1st International Electric Drives Production Conference (EDPC), Sept
2011, pp. 40–48
16. A.K. Putri, S. Rick, D. Franck, K. Hameyer, Application of sinusoidal field pole in a
permanent-magnet synchronous machine to improve the NVH behavior considering the
MTPA and MTPV operation area. IEEE Trans. Ind. Appl. 52(3), 2280–2288 (2016)
17. G.-D. Lee, M.-C. Kang, G.-T. Kim, The equilibrium design of radial magnetic force for
reduction of vibration in IPM type BLDC, in Proceedings of IEEE International Electric
Machines and Drives Conference, 10–13 May 2015, pp. 555–561
18. J. Hur, J.W. Reu, B.W. Kim, G.H. Kang, Vibration reduction of IPM-type BLDC motor using
negative third harmonic elimination method of air-gap flux density. IEEE Trans. Ind. Appl.
47(3), 1300–1309 (2011)
References 157

19. Y. Yang, X. Wang, R. Zhang, T. Ding, R. Tang, The optimization of pole arc coefficient to
reduce cogging torque in surface-mounted permanent magnet motors. IEEE Trans. Magn. 42
(4), 1135–1138 (2006)
20. Y. Pang, Z.Q. Zhu, Z.J. Feng, Cogging torque in cost-effective surface-mounted
permanent-magnet machines. IEEE Trans. Magn. 47(9), 2269–2276 (2011)
21. P. Zheng, J. Zhao, J. Han, J. Wang, Z. Yao, R. Liu, Optimization of the magnetic pole shape
of a permanent-magnet synchronous motor. IEEE Trans. Magn. 43(6), 2531–2533 (2007)
22. S. Ruangsinchaiwanich, Z.Q. Zhu, D. Howe, Influence of magnet shape on cogging torque
and back-emf waveform in permanent magnet machines, in Proceedings of International
Conference Electrical Machines and Systems, vol. 1 (2005), pp. 284–289
23. Y. Li, J.B. Zou, Y.P. Lu, Optimum design of magnet shape in permanent-magnet synchronous
motors. IEEE Trans. Magn. 39(6), 3523–3526 (2003)
24. K. Wang, Z.Q. Zhu, G. Ombach, Torque enhancement of surface-mounted permanent magnet
machine using third-order harmonic. IEEE Trans. Magn. 50(3), 104–113 (2014)
Chapter 9
Multi-objective Optimization of
Surface-Mounted Permanent Magnet
Machine with Third Harmonic
Shaped Rotor

This chapter presents a surface-mounted permanent magnet (SPM) machine with

third harmonic shaped rotor. However, it is desired that the SPM machine offers the
advantages of low torque ripple, high power density and efficiency for potential
application in electrical power steering systems. To realize the design objectives,
the multi-objective optimization based on multi-level optimization method is
implemented. The Taguchi method is employed in the sensitivity analysis of design
variables to divide design variables into sensitive level and nonsensitive level.
Then, the response surface (RS) method and sequential non-linear programing
(SNP) method are combined to optimize the sensitive design variables of the SPM
machine. After optimization, the electromagnetic performances of the initial and
optimal SPM machine are contrastively analyzed based on finite element method.
Finally, the optimal prototype machine is fabricated to verify the related analyses
and operation performances of the SPM machine.

9.1 Introduction

Permanent magnet (PM) machines have obtained wide applications due to the
advantages of simple structure, low torque ripple and excellent dynamic perfor-
mance [1, 2]. However, it is a challenging issue to have low torque ripple, high
torque density and efficiency simultaneously, which is especially crucial for its
application in high-performance areas, such as electrical power steering systems,
numerically-controlled machine tool, robot, etc.
The torque ripple can be reduced by making flux density distribution close to
sinusoidal wave. At present, numerous methods have been proposed in some lit-
eratures to achieve this goal, such as the PM sine shaping technique, the pulse width
modulation method of PM, the Halbach PM array, the sinusoidal and symmetrical
PM shaping technique in the axial direction, the modular poles technique and the
eccentric structure of PM [3–8]. However, the fundamental air-gap flux density of
© Springer Nature Singapore Pte Ltd. 2019 159
K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_9
160 9 Multi-objective Optimization of Surface-Mounted Permanent …

these magnetic pole structures are decreased inevitably which result in lower torque
density, compared with the machine without shaping. In order to compensate the
output torque of the machine with PM shaping, in [9–11], the PM shaping tech-
niques by utilizing third harmonic injection are proposed for the surface-mounted
permanent magnet (SPM) machines. The higher average torque can be achieved
compared to the conventional PM sine shaping technique. This is mainly due to
more fundamental air-gap flux density is obtained when the third harmonic is
injected into the sine shaped PM under the constraint of the same PM height.
Further, the relationship of average torque and the amplitude of third harmonic
injected into the SPM is analytically established and confirmed by finite element
method [12]. However, there still have some practical issues which should be
considered for the SPM machine with third harmonic shaped rotor. For example,
the machine with shaped rotor will increase the complexity of the manufacturing,
and the parametric variation of the PM edge thickness, the third harmonic amplitude
and some stator design variables have impact on the machine performances.
Meanwhile, there have conflict among the average torque, efficiency and torque
ripple. Therefore, the sensitivity of design variables should be analyzed, and then
the multi-objectives optimization should be performed to search out the optimal
comprehensive performances among the conflicting optimization objectives.
In recent researches, some optimization algorithms, such as the genetic algo-
rithm, the particle swarm algorithm and other optimization algorithms are adopted
in the multi-objectives optimization design of machines [13–15]. Nevertheless,
these methods are complicated to establish the objective functions. In [16, 17], the
response surface (RS) method and some algorithms are combined into multi-level
optimization method for parametric modeling and machine optimization design. In
addition, multi-level optimization approaches are also adopted to optimize the
machine and controller simultaneously, in which the optimization framework are
divided into the machine level and controller level [18, 19]. However, these
multi-level optimization algorithms are consisted of some relative complicated
methods. How to simplify the multi-level optimization method is also worthy of
attention in the multi-objective optimization of the machine. The aim of the chapter
is to analyze the sensitivity of the design variables by Taguchi method, which
divides the design variables into sensitive level and non-sensitive level to give the
guidelines for the machine manufacture tolerance and simplify the design variables
number of multi-objectives optimization. Then, the RS method and sequential
non-linear programing method (SNP) are combined into a relatively simple
multi-level optimization method to seek the optimal objectives among the average
torque, efficiency and torque ripple, which avoids the establishing the complicated
objectives functions. This paper is organized as follows. In Sect. 9.2, the SPM
machine structure with third harmonic shaped rotor and parametric model are
described. In Sect. 9.3, the Taguchi method is adopted for sensitivity analysis of
design variables, and the RS method and SNP method are combined to realize
multi-objectives optimization of the SPM machine. In Sect. 9.4, the quantitative
comparisons of the SPM machines with initial and optimal design variables on
electromagnetic performances are carried out by finite element (FE) method. The
9.1 Introduction 161

prototype of the SPM machine with optimal dimensions is manufactured and tested
to verify related analyses in Sect. 9.5, while Sect. 9.6 is the conclusion.

9.2 Topology and Parametric Model of SPM Machine

Figure 9.1 shows the structure of the SPM machine with third harmonic shaped
rotor. The SPM machine possesses 12-slot stator with the double-layer
non-overlapping winding, which is conducive to the short end winding. The third
harmonic shaped rotor with third harmonic injecting into sine-shaped SPM poles is
employed to enhance average torque.
The stator and SPM design variables of parametric models are shown in Fig. 9.2,
which are used for the later multi-objective optimization. x1 is the amplitude of third
harmonic injected into sine-shaped PM, which is related with the fundamental one
[4]. Meanwhile, there is a constraint that the minimum air-gap length and maximum
magnet thickness maintain constant, when the x1 is varied. x2 is the PM edge
thickness. x3 is the width of stator teeth. x4 is the depth of the slot opening. x5 is the
thickness of stator yoke. The main design specifications and parameters of the SPM
machine with third harmonic shaped rotor are shown in Table 9.1.

Fig. 9.1 SPM machine with third harmonic shaped rotor

162 9 Multi-objective Optimization of Surface-Mounted Permanent …

(a) SPM machine

(b) Sine-shaped PM with third harmonic injection

Fig. 9.2 Parametric model of SPM machine with third harmonic shaped rotor

Table 9.1 Main machine Parameters Value Unit

design specifications and
parameters Pole number 10
Slot number 12
Rated speed 400 rpm
Peak current 25 A
Turns in series per phase 48
Axial length 50 mm
Stator outer diameter 90 mm
Stator inner diameter 53 mm
Air-gap length 1 mm
Magnet remanence 1.2 T
Relative permeability 1.07
Pole arc to pole pitch ratio 1
Maximum thickness of PM 3 mm
9.3 Multi-objective Optimization Procedure 163

9.3 Multi-objective Optimization Procedure

Considering the potential applications in the electrical power steering systems, high
average torque is required to realize quick response of acceleration and decelera-
tion, and high efficiency is also demanded to reduce energy consumption. In
addition, to improve servo performance and reduce operating noise of the SPM
machine with third harmonic shaped rotor, torque ripple also needs to been con-
sidered. Therefore, the average torque, torque ripple and efficiency are preferred to
be selected as optimization objectives.
The average torque of the SPM machine with third harmonic shaped rotor can be
expressed as follow [12].

3I1 Br
Tavg ¼ 2Rs la Np Kdq1 Dma ð9:1Þ
2 l

 3
1 1 þ 3a 2
Dma ¼ Dm ð9:2Þ
8a 12a

where I1, Rs, la, Np, Br, l, Kdq1, a and Δm are the fundamental component of phase
current, the stator bore radius, the stator active axial length, the number of turns per
phase, the effective air-gap length, the fundamental winding factor, amplitude of
third harmonic injected into the sine-shaped PM and the maximum thickness of
PM, respectively.
The torque ripple is defined as:

Tmax  Tmin
Tr ¼  100% ð9:3Þ
Tavg

where Tmax, Tmin are the maximum torque, and minimum torque, respectively.
The efficiency ignoring mechanical loss is expressed as following.

Tavg x
g¼ ð9:4Þ
Tavg x þ Pc þ PFe þ Pe

where x is the mechanical angular velocity, Pc is the copper loss, PFe is the stator
and rotor iron loss, Pe is the PM eddy current loss.

9.3.1 Flowchart of Multi-objective Optimization

To illustrate the multi-objective optimization procedure more clearly, the flowchart

of the overall multi-objective optimization procedure is depicted in Fig. 9.3.
164 9 Multi-objective Optimization of Surface-Mounted Permanent …

Fig. 9.3 Flowchart of multi-objective optimization

9.3 Multi-objective Optimization Procedure 165

9.3.2 Sensitivity Analysis

The five design variables and three optimization objectives are selected in the
aforementioned multi-objectives optimization model. In order to reduce the opti-
mization space dimension and give a reference to manufacturing tolerance, it is
necessary to implement sensitivity analysis of design variables, which denotes the
weights of the each design variable on optimization objectives of machine. In this
chapter, the Taguchi method, as one of the design of experiment (DOE) method
[20, 21], is employed in the sensitivity analysis.
In the Taguchi method, an orthogonal array of experiments depending on the
number of design variables and levels is used to design experiments. The orthog-
onal array of experiments can be expressed uniformly as Ln(ji), in which the n is the
number of experiments, j is the levels of design variables and i is the number of
design variables. There are five design variables x1, x2, x3, x4 and x5, and each
design variable has four levels which are chosen in vicinity of corresponding initial
value. Then the orthogonal array of experiments based on L16(45) is established,
the satisfied results are gained by only 16 experiments (compared to 45 = 1024 for
full factorial approach). The experiments are performed by FE method to calculate
the response of optimization objectives, and the sensitivity index is evaluated based
on the results. The sensitivity index can be expressed as:
P4  2
j¼1 Mxi ðjÞ ðSÞ  MðSÞ
Hðxi Þ ¼ P5 P4  2 ð9:5Þ
i¼1 j¼1 Mxi ðjÞ ðSÞ  MðSÞ

where S is the objectives, Mxi ðjÞ ðSÞ is the average effect of design variables xi at
level j on S, and M(S) is the overall means of objectives.
Figure 9.4 shows the sensitivity indexes of design variables on optimization
objectives, which are calculated by formula (2.5). Generally, the sensitivity indexes
present positive correlation to the impact on optimization objectives. It can be
observed from Fig. 9.4 that the design variables x1 and x3 have significant influence
on the average torque, torque ripple and efficiency, whose sensitivity indexes H(xi)
are greater than 0.2. The effects of other design variables x2, x4 and x5 are almost
negligible due to the less sensitivity index H(xi). Therefore, the design variables can
be divided into sensitive level and non-sensitive level depending on the sensitivity
indexes.
The design variables of non-sensitive level can bear certain manufacturing tol-
erance during manufacture of prototype machine, due to little effect on performance
of the SPM machine. Simultaneously, these design variables also need not to be
further optimized, and the initial values are adopted. The sensitive level contains
two design variables x1 and x3, which need to be optimized. Therefore, the number
of design variables which need to be optimized could be reduced from five to two in
the following multi-objective optimization procedure.
166 9 Multi-objective Optimization of Surface-Mounted Permanent …

Fig. 9.4 Sensitivity indexes of design variables to optimization objectives

9.3.3 Multi-objective Optimization

Based on the previous sensitivity analysis, only two design variables x1 and x3 are
selected to be optimized. Hence, the complicated multi-objective optimization issue
of the SPM machine is converted to investigate the relationship among the two
design variables and the three optimization objectives. The RS method is widely
used to reveal the relationship between design variables and multiple optimization
objectives. Thus the method is employed to quickly obtain the corresponding
variation relationship. The responses of optimization objectives under sensitive
design variables are achieved, as shown in Fig. 9.5.
It can be observed from Fig. 9.5 that the average torque and efficiency appear the
same trend with the variation of design variables x1 and x3. Therefore, the optimal
results of the two optimization objectives can be achieved simultaneously.
Unfortunately, the conflicts between average torque and torque ripple exists. For
instance, when the design variable x1 is fixed at 0.3, the average torque decreases
gradually with the decrease of design variable x3. However, the torque ripple
decreases firstly and then increases with the decrease of design variable x3. The
same conflict issue exists between efficiency and torque ripple. Consequently, it is
necessary to adopt the compromise design in average torque, efficiency and torque
ripple in the multi-objectives optimization procedure.
In order to obtain the optimal tradeoff results of average torque, torque ripple and
efficiency, the SNP method is employed by establishing multi-objective optimiza-
tion model f(xi)max, which is expressed as a weighted sum of individual objectives.
The optimization model is shown as follow equation.

Tavg ðxi Þ Tr0 gðxi Þ

f ðxi Þmax ¼ x1 0
þ x2 þ x3 0 ð9:6Þ
Tavg Tr ðxi Þ g
9.3 Multi-objective Optimization Procedure 167

Fig. 9.5 Response of

average torque, torque ripple
and efficiency under sensitive
design variables

(a) Average torque distribution between x1 and x3

(b) Torque ripple distribution between x1 and x3

(c) Efficiency distribution between x1 and x3

168 9 Multi-objective Optimization of Surface-Mounted Permanent …

0
where xi is the ith design variable, Tavg ; Tr0 and g0 are initial values of the average
torque, the torque ripple and the efficiency, and x1, x2, and x3 are their weight
coefficients, which meet the equation x1+ x2+ x3= 1. Because the average torque
and efficiency can be achieved simultaneously, and the high power density is
required, the weight coefficient x1, x2, and x3 are selected to be 0.6, 0.4, and 0,
respectively.
In addition, in order to make optimization objectives reach desired targets, the
other performance of the SPM machine are the design requirements. The opti-
mization function needs to obey some boundary constraints, which are described as
follows:
8
>
> Tr  0:6%
>
>
< g  92%
Pout  160 W ð9:7Þ
>
>
>
> T  3:9 Nm
: avg
40%  kfill  55%

where Pout is the output power, and kfill is the winding fill factor.
Eventually, the equation of optimization model is solved based on the response
results. The variation of optimization model f(xi)max with the optimization objec-
tives is shown in Fig. 9.6.
From Fig. 9.6, it can be seen that the optimization model f(xi)max are variable
with the conflicting optimization objectives for average torque and torque ripple,
and the maximum value of f(xi)max is equal to 1.24, at which a red cube point is
marked. In the case, the corresponding design variables are optimized as Table 9.2.
The optimal comprehensive performance of the SPM machine is achieved.

Fig. 9.6 Variation of optimization functions with optimization objectives

9.4 Electromagnetic Performance Comparison 169

Table 9.2 Optimal design Items Initial values Optimal values

variables
x1 0 1/7
x2 1 mm 1 mm
x3 6.5 mm 7 mm
x4 0.5 mm 0.5 mm
x5 3.75 mm 3.75 mm

9.4 Electromagnetic Performance Comparison

In order to verify the multi-objectives optimization, the electromagnetic perfor-

mance of the SPM machine with initial and optimal design variables is contrastively
analyzed based on FE method in this section.

9.4.1 Open-Circuit Flux Density Distributions

and Back-EMFs

The FE predicted open-circuit field distributions and flux density maps of the SPM
machine with initial and optimal design variables are shown in Fig. 9.7. It can be
observed that the inter-pole leakage flux is reduced and the flux density in stator
teeth and stator yoke is increased for the optimal SPM machine. Figure 9.8 shows
the air-gap flux density distributions and the harmonic spectrum analysis. It can be
seen that the optimal SPM machine has higher fundamental air-gap flux density
compared with the initial SPM machine, which is due to the fact that the optimal
SPM machine dimensions are more reasonable through multi-objective optimiza-
tion design. The 7th, 11th and 13th harmonic in initial SPM machine and optimal
SPM machine are almost same, while the 3rd, 5th and 9th harmonics in optimal
SPM machine are slightly higher than that in initial SPM machine. Surprisingly, the
corresponding torque ripple will not be deteriorated, as will be explained later.
Figure 9.9 shows the FE predicted no-load back EMFs waveforms and harmonic
spectrum analysis of the SPM machine with initial and optimal design variables. It
can be seen that the fundamental no-load back EMF is improved remarkably by
multi-objective optimization, and 12.3% has been increased, which will contribute
to improve the average torque. By comparing their harmonics, it can be observed
that the optimal SPM machine has lower difference between 5th and 7th harmonic
and difference between 11th and 13th harmonic, despite of the higher no-load back
EMFs harmonic contents caused by higher air-gap flux density harmonic contents.
Meanwhile, those triplen harmonics will not result in torque ripple, because they
have been eliminated in line-line no-load back EMF for windings in balanced Y
connection.
170 9 Multi-objective Optimization of Surface-Mounted Permanent …

Fig. 9.7 Field distributions

and flux density maps of
initial and optimal SPM
machine

(a) Initial SPM machine

(b) Optimal SPM machine

9.4 Electromagnetic Performance Comparison 171

(a)Waveforms

(b)Spectrums
Fig. 9.8 Comparison of air-gap flux density of initial and optimal SPM machine
172 9 Multi-objective Optimization of Surface-Mounted Permanent …

(a) Waveforms

(b) Spectrums
Fig. 9.9 Phase back-EMFs of initial and optimal SPM machines (400 rpm)
9.4 Electromagnetic Performance Comparison 173

9.4.2 Torque Characteristics

The variation of output torque with rotor position for the SPM machines with initial
and optimal design variables under id = 0 control is given in Fig. 9.10a.
Figure 9.10b shows the cogging torque of the initial and optimal SPM machines.
The variation of the average torques with current angle for the two machines is
illustrated in Fig. 9.10c. It can be seen that the current angles of maximum torque
per ampere for the two machines are all 0 electrical degree, but the torque per
ampere for the optimal SPM machines is higher than that of the initial SPM
machine. The torque components of the two SPM machines for the average torque,
torque ripple, cogging torque and torque harmonic contents are shown in Table 9.3.
It can be observed that the average torque of the optimal SPM machine is effec-
tively improved, which has been increased by 10.87%, when compared with that of
the intimal SPM machine. The main reason is the increase of fundamental no-load
back-EMF. Unfortunately, the peak value of cogging torque is increased from 3 to
11 mN m, but the harmonic contents of output torque is decreased from 16 to
12 mN m, which result in low torque ripple. This is due to the fact that difference
between 5th and 7th harmonic and difference between 11th and 13th harmonic are
reduced significantly for the optimal SPM machine.

9.4.3 Efficiency

The loss components and efficiency of the initial and optimal SPM machine is
shown in Table 9.4. It can be seen that the undesired copper loss is decreased about
19.15%, but the corresponding iron loss and eddy loss of PM is slightly increased
under the constraints of same output power. Meanwhile, the iron loss and PM eddy
loss have a certain increase, and the output power is increased by 10.87% of initial
SPM machine under the constraints of same copper loss. The reasons for the
phenomenon are due to the increase of fundamental and partial harmonic air-gap
density by multi-objectives optimization. Most importantly, the efficiency is
increased by 0.708 and 1.494% of initial values under the constraints of same
copper loss and the constraints of same output power, respectively.
According to the above analysis, the average torque, torque ripple and efficiency
can satisfy the design requirements simultaneously. Therefore, the FE analytical
results show that the multi-level optimization method can effectively deal with the
conflict among optimization objectives.
174 9 Multi-objective Optimization of Surface-Mounted Permanent …

(a) Variation of torque with rotor position at id=0 control

(b) Variation of cogging torque with rotor position

(b) Average torque versus current angle

Fig. 9.10 Torque

9.5 Experimental Validation and Discussion 175

Table 9.3 Torque Items Initial Optimal

components of SPM machine
Tr (%) 0.69 0.49
Tavg (Nm) 3.68 4.08
Cogging torque (Nm) 0.003 0.011
Torque harmonic contents 0.016 0.012

Table 9.4 Loss components Items Initial Initial Optimal

and efficiency of SPM
machine Output power (W) 154.15 170.90 170.90
Copper loss (W) 12.41 15.35 12.41
Iron loss (W) 1.39 1.41 1.58
Eddy loss of PM (mW) 23.32 27.74 29.36
Efficiency (%) 91.77 91.06 92.42

9.5 Experimental Validation and Discussion

In order to further verify the validity of the related multi-objective optimization

method and the operation performances of the optimal SPM machine, the SPM
machine with optimal size has been manufactured and the test platform is estab-
lished. Figure 9.11a–c show the pictures of 12-slot stator, 10-pole rotor with third
harmonic injecting into sine-shaped PM and test platform, respectively.
The experiments on the optimal SPM machine are carried on. Figure 9.12 shows
the measured phase no-load back-EMFs waveforms and spectrums for FE predic-
tion and experiment. The third harmonic component is little higher than the FE
predicted results, mainly due to manufacturing error such as inaccurate SPM shape.
The measured steady state output torque under id = 0 control and current waveform
are shown in Fig. 9.13, it can be seen that the amplitude of the current is 25 A, and
the corresponding output average torque is 4.09 Nm, In general, good agreements
between the FE predicted and experimental results are gained.

9.6 Summary

This chapter presents a SPM machine with third harmonic shaped rotor for potential
applications in electrical power steering systems. The sensitivity of design variables
is analyzed to give a reference to the manufacturing tolerance and reduce opti-
mization design space dimension. To handle the conflict among multiple opti-
mization objectives, the multi-objective optimization based on the multi-level
optimization method is implemented. Then, the electromagnetic performances of
176 9 Multi-objective Optimization of Surface-Mounted Permanent …

Fig. 9.11 Prototype machine

and test platform

(a) Stator

(b) Third harmonic shaped rotor

(c) Test platform

9.6 Summary 177

(a) Measured back-EMFs (5V/div)

(b) Spectrums of FE predicted and experimental back-EMFs

Fig. 9.12 Back-EMFs under no-load

the optimal SPM machine and the initial SPM machine are compared. The opti-
mization results demonstrate that the optimal SPM machine can obtain 10.87%
higher average torque than the initial SPM machine, the efficiency is increased by
0.708 and 1.494% of the initial SPM machine under the constraints of same copper
loss and the constraints of same output power, respectively. Simultaneously, the
torque ripple can be reduced by 28.99%. Finally, the measured results on the
optimal prototype machine verify the related analyses.
178 9 Multi-objective Optimization of Surface-Mounted Permanent …

Fig. 9.13 Measured output torque and current

References

1. A.M. El-Refaie, Fractional-slot concentrated-windings synchronous permanent magnet

machines: opportunities and challenges. IEEE Trans. Ind. Electron. 57(1), 107–121 (2010)
2. Z.Q. Zhu, D. Howe, Electrical machines and drives for electric, hybrid, and fuel cell vehicles.
Proc. IEEE 95(4), 746–765 (2007)
3. N.R. Tavana, A. Shoulaie, Analysis and design of magnetic pole shape in linear permanent
magnet machine. IEEE Trans. Magn. 46(4), 1000–1006 (2010)
4. S. Chaithongsuk, N. Takorabet, F. Meibody-Tabar, On the use of pulse width modulation
method for the elimination of flux density harmonics in the air-gap of surface PM motors.
IEEE Trans. Magn. 45(3), 1736–1739 (2009)
5. Z.Q. Zhu, D. Howe, Halbach permanent magnet machines and applications: a review. IEE
Proc. Electr. Power Appl. 148(4), 299–308 (2001)
6. A.H. Isfahani, S.V. Zadeh, M.A. Rahman, Using modular poles for shape optimization of flux
density distribution in permanent magnet. IEEE Trans. Magn. 44(8), 2009–2015 (2008)
7. W.L. Zhao, T.A. Lipo, B. Kwon, Torque pulsation minimization in spoke-type interior
permanent magnet motors with skewing and sinusoidal permanent magnet configurations.
IEEE Trans. Ind. Electron. 61(10), 5779–5787 (2014)
8. Y. Zhou, H.S. Li, G.W. Meng, S. Zhou, Q. Cao, Analytical calculation of magnetic field and
cogging torque in surface-mounted permanent magnet machines accounting for any eccentric
rotor shape. IEEE Trans. Ind. Electron. 62(6), 3438–3447 (2015)
9. Y. Li, J.B. Zou, Y.P. Lu, Optimum design of magnet shape in permanent-magnet synchronous
motors. IEEE Trans. Magn. 39(6), 3523–3526 (2003)
10. K. Wang, Z.Q. Zhu, Y. Ren, G. Ombach, Torque improvement of dual three-phase
permanent-magnet machine with third-harmonic current injection. IEEE Trans. Ind. Electron.
62(11), 6833–6844 (2015)
11. K. Wang, Z.Q. Zhu, G. Ombach, Torque Improvement of five-phase surface-mounted
permanent magnet machine using third-order harmonic. IEEE Trans. Energy Covers. 29(3),
735–747 (2014)
12. K. Wang, Z.Q. Zhu, G. Ombach, Torque enhancement of surface-mounted permanent magnet
machine using third-order harmonic. IEEE Trans. Magn. 50(3), 104–113 (2014)
References 179

13. A. Mahmoudi, S. Kahourzade, N.A. Rahim, W.P. Hew, Design, analysis, and prototyping of
an axial-flux permanent magnet motor based on genetic algorithm and finite-element analysis.
IEEE Trans. Magn. 49(4), 1479–1492 (2013)
14. J.N. Yin, X.Y. Zhu, L. Quan, C. Zhang, Z.X. Xiang, Comprehensive multi-objective
scalarisation optimization of a permanent magnet machine with correlation parameters
stratified method. IET Electr. Power Appl. 11(1), 72–79 (2017)
15. C. Ma, L.Y. Qu, Multi-objective optimization of switched reluctance motors based on design
of experiments and particle swarm optimization. IEEE Trans. Energy Covers. 30(3), 1144–
1153 (2015)
16. P. Asef, R.B. Perpina, M.R. Barzegaran, A. Lapthorn, D. Mewes, Multi-objective design
optimization using dual-level response surface methodology and booths algorithm for
permanent magnet synchronous generators. IEEE Trans. Energy Covers. https://doi.org/10.
1109/TEC.2017.2777397 (2017)
17. X.Y. Zhu, Z.M. Shu, L. Quan, Z.X. Xiang, X.Q. Pan, Multi-objective optimization of an
outer-rotor V-shaped permanent magnet flux switching motor based on multi-level design
method. IEEE Trans. Magn. 52(10), (2016)
18. G. Lei, T. Wang, J.G. Zhu, Y.G. Guo, S.H. Wang, System-level design optimization methods
for electrical drive systems: robust appraoch. IEEE Trans. Ind. Electron. 62(8), 4702–4713
(2015)
19. X.Y. Zhu, Z.X. Xiang, L. Quan, W.Y. Wu, Y. Du, Multi-mode optimization design
methodology for a flux-controllable stator permanent magnet memory motor considering
driving cycles. IEEE Trans. Ind. Electron. 65(7), 5353–5366 (2017)
20. M. Si, X.Y. Yang, S.W. Zhao, S. Gong, Design and analysis of a novel spoke-type permanent
magnet synchronous motor. IET Electr. Power Appl. 10(3), 571–580 (2016)
21. S.J. Lee, K. Kim, S. Cho, J.Y. Jang, T. Lee, J. Hong, Optimal design of interior permanent
magnet synchronous motor considering the manufacturing tolerances using Taguchi robust
design. IET Electr. Power Appl. 8(1), 23–28 (2014)
Chapter 10
Using Third Harmonic for Shape
Optimization of Flux Density
Distribution in Slotless Linear
Permanent-Magnet Machine

This chapter presents a sinusoidal permanent magnet (PM) shaping technique with
third harmonic to improve the electromagnetic thrust force in linear slotless PM
machines without sacrificing the thrust force ripple. Slotless PM linear machine
possesses relatively low thrust force ripple due to the absence of cogging force,
compared with slotted topology. However, thrust force ripple of the machine with
rectangular PM shape still exists due to nonsinusoidal airgap flux density distri-
bution produced by PMs. Sinusoidal shaping techniques can be used to reduce the
thrust force ripple but the average thrust force is reduced as well. Therefore, a
simple PM shaping technique with optimal 3rd harmonic is presented to improve
the output thrust force but not to increase the thrust force ripple. The sinusoidal plus
3rd harmonic shaping technique is analytically demonstrated together conventional
sinusoidal shaping method and verified with finite element method. The results
show that the electromagnetic performance can be significantly improved.

10.1 Introduction

Compared with the rotary machine for applications where linear and reciprocating
motions are required, a linear machine offers several key advantages, such as high
efficiency, excellent dynamic performance, and high reliability [1–3]. Linear slotless
permanent-magnet machine as shown in Fig. 10.1 exhibits low thrust ripple without
cogging force due to the slotting and finite length of moving part and has gained wide
application in semiconductor industry and electronics board assembly, etc.
However, the thrust force density will be limited with additional effective airgap
by the slotless winding. Hence, various approaches have been proposed to improve
the thrust force density. The ratio of coil width to pole pitch has been optimized to
improve the thrust force density [4]. The analytical method and genetic algorithm
optimization has been proposed to improve the thrust force by optimizing the
dimensions of magnets [5]. The thrust force density can be increased by optimizing
© Springer Nature Singapore Pte Ltd. 2019 181
K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_10
182 10 Using Third Harmonic for Shape Optimization …

(a) Conventional

(b) Sinusoidal shaping

(c) Sinusoidal + 3rd harmonic shaping

Fig. 10.1 Slotless linear PM machine with various PM shapes

the number of segmentations and magnetized orientation of permanent magnets in

Hallbach array structure [6].
On the other hand, although the cogging thrust is absent in slotless linear per-
manent magnet machines, the thrust force ripple of slotless linear machine still
exists due to nonsinusoidal airgap flux density distribution generated by rectangular
PMs. In order to reduce the harmonics in airgap flux density and, thus, thrust force
ripple, numerous techniques have been proposed. A modular pole consists of three
pieces of magnet attached together. A so-called “modular permanent magnet pole”
is proposed and used to shape air-gap flux density [2]. In this model, the pole
consists of three or more magnet pieces with different types and specifications.
Unfortunately, this technique increases the complexity and cost of manufacturing of
PM machines. The direct modeling of thrust ripple has been presented for an
air-core permanent magnet linear synchronous (PMLMS) machine with modular
pole and has been used for optimal design of motor parameters [7]. A segmented
pole consists of two or more PM pieces located beside each other with a certain gap
between each piece [8]. A segmented PM pole has been proposed for cogging force
10.1 Introduction 183

reduction in a single-sided PMLSM [9]. A three-piece segmented type PM pole has

been used for cogging torque reduction of rotary PMS motor [10]. A pulse width
modulation (PWM) technique has been applied to the segmented pole to produce
pure sinusoidal air-gap flux density [11]. In this configuration, the block magnets
are arranged similarly to PWM signals used to switch the power electronic devices.
This technique reduces the flux density produced by the PM pole as well as the
machine thrust or torque. Moreover, it is very hard to implement because placing
the segments of a divided PM pole in one direction of magnetization vectors
adjacent to each other in a PM motor with short pole pitch is quite difficult. One
method is the magnet-arc-shaping technique [12–14]. In this model, the outer
surface of the poles in a PM motor is considered as a part of a cylinder. However,
this method increases the complexity and cost of manufacturing. Halbach magne-
tized topology has been employed to have sinusoidal airgap flux density [15–17].
However, this method needs many magnet pieces with different sizes and directions
of magnetization. Thus, it uses excessive amount of PM material, leading to a high
production cost. The stair-step-shaped pole is proposed in the PMLSM to shape
their air-gap flux density distribution [18]. The stair-step-shaped pole can develop
high thrust or torque and reduce flux density as well as EMF harmonics substan-
tially. Furthermore, the design complexity of this method is considerably less than
those of the aforementioned methods. However, this technique still increases the
cost of manufacturing.
One of the commonly employed techniques, PM shaping, Fig. 10.1b, can
directly make the airgap flux density distribution sinusoidal and effectively decrease
the thrust force ripple. Nevertheless, although the harmonics in the airgap flux
density are suppressed, it suffers from low average thrust force due to the reduction
in fundamental airgap flux density.
Therefore, this chapter presents a sinusoidal shaping technique with third har-
monic to improve the electromagnetic force in linear slotless permanent magnet
machines. A simple PM shaping technique with optimal 3rd harmonic is presented
to improve the output thrust force but not to increase thrust force ripple. The
sinusoidal plus 3rd harmonic shaping technique is analytically demonstrated
together conventional sinusoidal shaping method and verified with finite element
method. The results show that the electromagnetic performance can be significantly
improved.

10.2 PMLMS with Various PM Shapes

In order to investigate the effect of various PM shapes on the electromagnetic

performance, including average thrust force and thrust force ripple, the double-side
slotless permanent magnet linear machines are employed for illustration, as shown
in Fig. 10.1. All the machines share the identical airgap length of 1 mm and the
same current density of 5 A/mm2. The parallel magnetized PMs with pole arc to
pole pitch ratio equal to 1 are used and its remanence and relative permeability are
184 10 Using Third Harmonic for Shape Optimization …

Table 10.1 Main design Unit

parameters and specifications
Number of coil/phase/pole 1 –
Current density, J 5 A/mm
Winding height, hw 6 mm
Winding length/phase, lw 14 mm
Pole pitch, sp 42 mm
Air-gap length, lg 1 mm
PM width, wpm 21 mm
Maximum PM height, hpm 3 mm
PM remanence, Br 1.2 T
PM relative permeability, lr 1.05

shown in Table 10.1. The only difference among them is the PM shape and
therefore uniform airgap length for the three linear PM machines. The main
dimensions and specifications are also given in Table 10.1.

10.3 PM Shape Optimization

Obviously, it is desirable for the airgap flux density distribution to be as close to

sinusoidal as possible in order to have low thrust force ripple. This can be achieved
by shaping the PMs and, consequently, both the PMs and airgap length will be
varied with the position of PMs. However, the fundamental airgap flux density is
also reduced due to the decrease of the PMs volume. However, for the
star-connected three phase windings, the triplen harmonic in the phase back-EMF
can be eliminated in the line-line back-EMF and thus its impact on electromagnetic
performance can be negligible. Therefore, the optimal third harmonic can be used to
improve the fundamental airgap flux density and hence the output force capability,
as described in [13, 19].

10.3.1 Simplified Estimation of PM Shapes

The equivalent magnetic circuit model of one pole pair is shown in Fig. 10.2a.
Figure 10.3 shows the variation of airgap length permanent magnet shape with PM
positions x, and thus the variable reluctance magnetic circuit model in which the
reluctance will vary with x. Its simplified model can be constructed as shown in
Fig. 10.2a and b, respectively. In Fig. 10.2a, Rs are the reluctances of the stator
back iron, Rg is the reluctance of the air gap, /r is the flux source of a magnet, Rm is
the reluctance of a magnet, /m is the magnet flux, and Rml are the flux leakage
between the magnets and the reluctance modeling that flux leakage, and is the air
10.3 PM Shape Optimization 185

(a) Variable magnetic circuit model (b) Simplified model

Fig. 10.2 Reluctance magnetic circuit of linear slotless machine

Fig. 10.3 Variation of PM shape and airgap length with x

gap flux flowing through one-half of the air gap cross-sectional area. In Fig. 10.2b,
Pg, Pm, and Pml represent the permeances for the air gap, the magnets and the flux
leakage respectively. It can be noted that the flux density and permeance in the air
gap are the variables as functions of x. From the Kirchhoff’s law, the airgap flux
density can be obtained as:

Pg ð xÞBr Am
Bg ð xÞ ¼ ð10:1Þ
Pg ð xÞ þ Pm ð xÞ þ Pml Ag

If the flux leakage is neglected, the airgap flux density distribution can be
expressed as:

Br
Bg ð xÞ ¼ Ag ð xÞ l ð xÞ
ð10:2Þ
A m ð xÞ þ lr lmg ðxÞ
186 10 Using Third Harmonic for Shape Optimization …

Then, it can be assumed that the areas of PM and airgap is the same due to the
pole arc to pole pitch ratio equal to 1. The airgap flux density distribution can be
further simplified as:

Br
B g ð xÞ ¼ l ð xÞ
ð10:3Þ
1 þ lr lmg ðxÞ

Therefore, it can be seen that the desirable sinusoidal airgap flux density or
sinusoidal with third harmonic can be obtained by optimizing the PM height and
airgap length. If lm(x) = sin(x) or = sin(x) + asin(3x), the airgap flux density dis-
tribution will be sinusoidal or sinusoidal with third harmonic. The PM shape can be
simply estimated by the variable magnetic circuit without consideration
flux-leakage and saturation. Therefore, the PM shape needs to be optimized to
achieve a sinusoidal or sinusoidal with third harmonic distribution.

10.3.2 Optimization of Sinusoidal PM Shapes

The method of sinusoidally shaped magnets, as shown in Fig. 10.4, has been
proposed to reduce the cogging torque and make back-EMF waveform sinusoidal
distribution [14]. The maximum thickness of the magnet is 3 mm. The edges of the
magnet are varied from 3 to 0 mm. The shape of the magnet can be expressed as:


px
lm ð xÞ ¼ Dh sin þ lm  Dh ð10:4Þ
sp

where x is the mechanical displacement position, lm is the maximum magnet

thickness and Dh is the difference between the maximum magnet thickness and the
edge thickness, as shown in Fig. 10.4. The airgap flux density is further optimized
with the investigation of the edge thickness. The edge thickness varies from 0 to
3 mm with the interval of 0.5 mm, and the corresponding airgap flux density
distribution, variations of fundamental flux density, and total harmonic distortion
(THD) with the edge thickness of the magnet are shown in Fig. 10.5a–c, respec-
tively. As can be seen, with the increase of edge thickness, the fundamental airgap
flux density decreases, whereas the THD of airgap flux density firstly decreases and
then increases. Therefore, it can be deduced that the better performance will be
obtained at the point of fixed edge thickness of 0.5 mm.
10.3 PM Shape Optimization 187

Fig. 10.4 Sinusoidal PM shape

10.3.3 PM Shape with Third Harmonic

The method of sinusoidally shaped magnets with third harmonic, as shown in

Fig. 10.3, has been proposed to reduce harmonic in the airgap flux density, but
improve the output force capability. The magnet edges also need to be optimized to
have minimum airgap flux density THD, as described in previous section. The
edges of the magnet are also varied from 3 to 0 mm. When the optimal third
harmonic is injected into the magnet shape, the shape of the magnet can be
expressed as:
pffiffiffi 


2 3 px 1 3px
lm ð xÞ ¼ Dh sin þ sin þ lm  Dh ð10:5Þ
3 sp 6 sp

The airgap flux density is also optimized with the investigation of the edge
thickness. The corresponding airgap flux density distribution, variations of funda-
mental flux density and total harmonic distortion (THD) with the edge thickness of
the magnet are shown in Fig. 10.6a–c, respectively. As can be seen, with the
increase of edge thickness, the fundamental airgap flux density decreases, whereas
the THD of airgap flux density firstly decreases and then increases. Therefore, it can
be deduced that the better performance will be obtained at the point of fixed edge
thickness, and the edge thickness equal to 1 mm has the minimum value of THD.

10.4 Performance Evaluation

The performance of the PMLMS machine having sinusoidally shaped magnets with
third harmonic, with the specification listed in Table 10.1, is investigated and
compared to the machine with the conventional (rectangular shape) and sinusoidal
shaped PM poles. The flux lines and flux density distribution are depicted in
Figs. 10.7 and 10.8, respectively. The flux density distribution under one pole pair
of all the machines is shown in Fig. 10.9a. It can be seen that the sinusoidal shaped
PM produces much closer to a sinusoidal shape than that of the conventional PM
poles and sinusoidal with third harmonic shaped PM poles. However, the sinusoidal
188 10 Using Third Harmonic for Shape Optimization …

0.8
0.6
0.4
Flux density (T)
0.2
0
-0.2
-0.4 Δh=0mm Δh=0.5mm
Δh=1.5mm Δh=2.5mm
-0.6 Δh=3mm Pure sine

-0.8
0 5 10 15 20 25 30 35 40
Displacement (mm)
(a) Waveforms
0.8
Fundamental flux density (T)

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3
Δ h (mm)
(b) Fundamental
50

40
THD (%)

30

20

10

0
0 0.5 1 1.5 2 2.5 3
Δ h (mm)
(c) THD

Fig. 10.5 Airgap flux density distribution with PM edge thickness in sinusoidal PM shapes
10.4 Performance Evaluation 189

0.8
0.6
0.4
Flux density (T)
0.2
0
-0.2
-0.4 Δh=0mm Δh=0.5mm
Δh=1.5mm Δh=2.5mm
-0.6 Δh=3mm Pure sine

-0.8
0 5 10 15 20 25 30 35 40
Displacement (mm)
(a) Waveforms
0.8
Fundamental flux density (T)

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3
Δ h (mm)
(b) Fundamental
30

25

20
THD (%)

15

10

0
0 0.5 1 1.5 2 2.5 3
Δ h (mm)
(c) THD

Fig. 10.6 Airgap flux density distribution with PM edge thickness in sinusoidal PM shape with
third harmonic
190 10 Using Third Harmonic for Shape Optimization …

(a) Conventional

(b) Sinusoidal

(c) Sinusoidal + 3rd harmonic

Fig. 10.7 Magnetic flux lines in the LPMS machine

10.4 Performance Evaluation 191

(a) Conventional

(b) Sinusoidal

(c) Sinusoidal + 3rd harmonic

Fig. 10.8 Distribution of normal component of magnetic flux density in LPMS machine
192 10 Using Third Harmonic for Shape Optimization …

0.8
0.6
0.4
Flux density (T)

0.2
0
-0.2
Conventional
-0.4 Sinusoidal
Sinusoidal+3rd
-0.6
-0.8
0 10 20 30 40
Displacement (mm)
(a) Waveforms
0.8
Conventional
Sinusoidal
0.6
Flux density (T)

Sinusoidal+3rd

0.4

0.2

0
1 3 5 7 9 11 13
Harmonic order (n)
(b) Harmonics

Fig. 10.9 Comparison of Airgap flux density distribution among conventional, sinusoidal and
sinusoidal +3rd PM shapes

shaped magnets with optimal third harmonic have significantly improved the fun-
damental airgap flux density, as shown in Fig. 10.9b.
The FEM predicted back-EMFs are compared in Fig. 10.10 and its harmonic
contents normalized to the fundamental one are listed in Table 10.2. It can be seen
that the use of the sinusoidally shaped PM poles is also effective in reducing the
harmonic content of EMF. Less harmonics in back-EMFs and sinusoidal currents
also lead to lower ripple in machine thrust force. However, meanwhile, the fun-
damental back-EMF is also reduced. This can be compensated by sinusoidally
shaped PM poles with third harmonic. It can be seen that the fundamental
10.4 Performance Evaluation 193

1.5
Conventional
1.0 Sinusoidal
Back-EMF (V/turn) Sinusoidal +3rd
0.5

0.0

-0.5

-1.0

-1.5
0 7 14 21 28 35 42
Displacement (mm)

Fig. 10.10 Back-EMFs, 1 m/s

Table 10.2 Phase and line-line back-EMF, 1 m/s

Conventional (%) Sine (%) Sine+3rd (%)
Phase Line Phase Line Phase Line
1 100 (0.57 V) 100 (0.98 V) 100 (0.46 V) 100 (0.80 V) 100 (0.53 V) 100 (0.92 V)
3 11.89 0 0.03 0 6.73 0
5 1.05 1.04 0.15 0.13 0.04 0.04
7 0.30 0.30 0.04 0.01 0.02 0.01
9 0.21 0 0.05 0 0.02 0
11 0.06 0.06 0.05 0 0.01 0.02
13 0.03 0.03 0.05 0 0.02 0.01

back-EMF in sinusoidally shaped PM poles with third harmonic has been improved
significantly compared with that of sinusoidally shaped ones, being 15.22%.
Clearly, the third harmonic back-EMF is also increased but it can be eliminated in
the line-line back-EMF. Therefore, its influence on the electromagnetic perfor-
mance for three-phase balanced star-connected windings can be negligible.
Also, all the designs provide a lower thrust ripple as seen in both Fig. 10.11 and
Table 10.3. The average thrust force in the machine having sinusoidally shaped PM
poles has been reduced by 19% compared with the conventional one, while the
average thrust force can be improved by 14.5% compared with sinusoidal shaped
PM poles. The thrust force ripple for both machines is almost the same, being lower
than the conventional one.
194 10 Using Third Harmonic for Shape Optimization …

30
Conventional
28 Sinusoidal
Thrust force (N) Sinusoidal +3rd
26

24

22

20
0 7 14 21 28 35 42
Displacement (mm)

Fig. 10.11 Thrust force with displacement, J = 5 A/mm2

Table 10.3 Thrust force Conventional Sine Sine+3rd

components
Average thrust (N) 25.31 20.60 23.60
Thrust ripple (%) 2.75 0.32 0.12
Maximum thrust (N) 25.67 20.63 23.62
Minimum thrust (N) 24.97 20.56 23.59
Peak-peak thrust (N) 0.70 0.07 0.03

10.5 Summary

This chapter presents a sinusoidal PM shaping technique with third harmonic to

improve the electromagnetic force in linear slotless PM machines. Slotless PM
linear machine possesses relatively low thrust force ripple due to the absence of
cogging force, compared with slotted topology. However, thrust force ripple of the
machine with rectangular PM shape still exists due to nonsinusoidal airgap flux
density distribution produced by PMs. Sinusoidal shaping techniques can be used to
reduce the force ripple but the average force is reduced as well. Therefore, a simple
PM shaping technique with optimal 3rd harmonic is presented to improve the
output thrust force but not to increase the thrust force ripple. The sinusoidal plus 3rd
harmonic shaping technique is analytically demonstrated together conventional
sinusoidal shaping method and verified with finite element method. The results
show that both the sinusoidally shaped PM poles and sinusoidal with third har-
monic shaped PM poles machine has less back-EMF harmonics and thrust force
ripple compared with the conventional one. The average thrust force of sinusoidal
with third harmonic shaped PM machines has been improved by 14.5% compared
10.5 Summary 195

with sinusoidal shaped PM poles one. The exploitation of full potential of

permanent-magnet materials used in the pole is a salient feature of the proposed
shaped method which contributes to the cost reduction of permanent-magnet
machines using this kind of pole shaping. The technique is also applicable to slotted
linear PM machines.

References

1. S. Vaez-Zadeh, A.H. Isfahani, Multiobjective design optimization of air-core linear

permanent-magnet synchronous motors for improved thrust and low magnet consumption.
IEEE Trans. Magn. 42(3), 446–452 (2006)
2. A.H. Isfahani, S. Vaez-Zadeh, M.A. Rahman, Using modular poles for shape optimization of
flux density distribution in permanent-magnet machines. IEEE Trans. Magn. 44(8), 2009–
2015 (2008)
3. N.R. Tavana, A. Shoulaie, Analysis and design of magnetic pole shape in linear
permanent-magnet machine. IEEE Trans. Magn. 46(4), 1000–1006 (2010)
4. J.K. Kim, S.C. Hahn, J.P. Hong, D.H. Kang, D.H. Koo, Static characteristics of linear BLDC
motor using equivalent magnetic circuit and finite element method. IEEE Trans. Magn. 40(2),
742–745 (2004)
5. A.H. Isfahani, S. Vaez-Zadeh, Design optimization of a linear permanent magnet synchronous
motor for extra low force pulsations. IEEE Trans. Magn. 48, 443–449 (2007)
6. L.Y. Li, M.N. Ma, B.Q. Kou, Q.Q. Chen, Analysis and optimization of slotless
electromagnetic linear launcher for space use. IEEE Trans. Plasma Sci. 39(1), 127–132 (2011)
7. A.H. Isfahani, S. Vaez-Zadeh, M.A. Rahman, Performance improvement of permanent
magnet machines by modular poles. IET Electr. Power Appl. 3(4), 343–351 (2009)
8. A.H. Isfahani, Analytical framework of thrust enhancement in permanent-magnet (PM) linear
synchronous motors with segmented PM poles. IEEE Trans. Magn. 46(4), 1116–1122 (2010)
9. M.Y. Kim, Y.C. Kim, G.T. Kim, Design of slotless-type PMLSM for high power density
using divided PM. IEEE Trans. Magn. 40(3), 746–749 (2004)
10. R. Lateb, N. Takorabet, F. Meibody-Tabar, Effect of magnet segmentation on the cogging
torque in surface-mounted permanent-magnet motors. IEEE Trans. Magn. 42(3), 442–445
(2006)
11. S. Chaithongsuk, N. Takorabet, F. Meibody-Tabar, On the use of pulse width modulation
method for the elimination of flux density harmonics in the air-gap of surface PM motors.
IEEE Trans. Magn. 45(3), 1736–1739 (2009)
12. Y. Li, J. Xing, T. Wang, Y. Lu, Programmable design of magnet shape for permanent magnet
synchronous motors with sinusoidal back EMF waveforms. IEEE Trans. Magn. 44(9), 2163–
2167 (2008)
13. Y. Li, J. Zou, Y. Lu, Optimum design of magnet shape in permanent magnet synchronous
motors. IEEE Trans. Magn. 39(6), 3523–3526 (2003)
14. P. Zheng, J. Zhao, J. Han, J. Wang, Z. Yao, R. Liu, Optimization of the magnetic pole shape
of a permanent-magnet synchronous motor. IEEE Trans. Magn. 43(6), 2531–2533 (2007)
15. K.J. Meessen, B.L.J. Gysen, J.J.H. Paulides, E.A. Lomonova, Halbach permanent magnet
shape selection for slotless tubular actuators. IEEE Trans. Magn. 44(11), 4305–4308 (2008)
16. J.S. Choi, J. Yoo, Design of a Halbach magnet array based on optimization techniques. IEEE
Trans. Magn. 44(10), 2361–2366 (2008)
196 10 Using Third Harmonic for Shape Optimization …

17. M. Markovic, Y. Perriard, Optimization design of a segmented Halbach permanent-magnet

motor using an analytical model. IEEE Trans. Magn. 45(7), 2955–2960 (2009)
18. N.R. Tavana, A. Shoulaie, V. Dinavahi, Analytical modeling and design optimization of
linear synchronous motor with stair-step-shaped magnetic poles for electromagnetic launch
applications. IEEE Trans. Plasma Sci. 40(2), 519–527 (2012)
19. Z.Q. Zhu, K. Wang, G. Ombach, Optimal magnet shaping with third order harmonic for
maximum torque in SPM machines, in 6th IET Conference on Power Electronics, Machines
and Drives, (PEMD2012), Bristol, UK, 27–29 Mar 2012, CD-ROM
Chapter 11
Design and Analysis of Permanent
Magnet Linear Synchronous Machine
with Third Harmonic Shaping Mover

This chapter proposes a novel sine with optimal third harmonic injected shaping
mover to improve the thrust force and efficiency without deteriorating the thrust
ripple in permanent magnet linear synchronous machine (PMLSM). The optimal
value of third harmonic injected into the sinusoidal permanent magnet (PM) shape
for maximizing thrust force is analytically derived and confirmed by finite element
(FE) analysis. Further, to maximize thrust force and efficiency, and minimize thrust
triple, the PM edge thickness and pole arc coefficient are optimized by non-linear
programing method. It is found that the optimal third harmonic is 1/6 fundamental
one, and the optimal PM edge thickness and pole arc coefficient are 0 mm and 0.85,
respectively. For the PMLSM with sine and third harmonic (SIN+3rd) shaping
mover, sine (SIN) shaping mover and conventional (without shaping) shaping
mover, the electromagnetic performances are compared. The results demonstrate
that the average thrust force and efficiency of the PMLSM with SIN+3rd shaping
mover are improved significantly, while the thrust ripple is not deteriorated, com-
pared with that of the PMSLM with SIN shaping mover.

11.1 Introduction

Permanent magnet linear synchronous machines (PMLSM) have obtained wide

applications due to the advantages of no transmission, large thrust force, high
precision and excellent dynamic performance [1–3]. However, it is a challenging
issue to have low thrust ripple, high power density and efficiency simultaneously,
which is crucial for its applications in high-performance areas, such as unmanned
aerial vehicle electromagnetic launch systems, automatic sorting systems, vertical
hoist systems, etc.
For the PMLSM, there are two type additional forces, which can cause thrust
ripple, except the electromagnetic forces. The one is detent force, which is arisen
from the interaction between the PM and the armature iron teeth. The other is due to
© Springer Nature Singapore Pte Ltd. 2019 197
K. Wang and Z.-Q. Zhu, Third Harmonic Utilization in Permanent Magnet Machines,
https://doi.org/10.1007/978-981-13-0629-7_11
198 11 Design and Analysis of Permanent Magnet …

the interaction of harmonic components of the current and back-EMF. The two
forces can be minimized by machines design methods, for example, optimizing
primary iron core end teeth shape and primary iron core length, adopting auxiliary
poles technique, modular primary iron core technique, PM shaping technique and
providing sinusoidal current simultaneously [4–19]. The detent force could be
decreased by adjusting end teeth length of primary iron core at outlet edge [4]. The
auxiliary poles technique was often used to suppress the end effect, which decrease
detent force by counteracting the end force between primary iron core and auxiliary
poles [5, 6]. While the end force is sensitive to the position and size of the auxiliary
poles, the inappropriate design will deteriorate the thrust ripple. An approach to a
suitable primary iron core length, which is one of the simplest and most effective
detent force weakening methods was widely adopted. The key to establishing the
optimal primary iron core length is the determination of the position of the peak of
the single-end detent force [7, 8]. The modular primary iron cores, which consist of
two or more sections, were also employed to reduce thrust ripple, the force ripple
from different sections can be canceled by phase differences among sections [9, 10].
However, the primary iron core optimization techniques are not suitable to reduce
thrust ripple of long primary PMLSM or slotless structures.
The PM shaping methods have been proposed to reduce the thrust ripple by
making flux density distribution close to sinusoidal wave. Halbach magnetized PM
can offer sinusoidal air-gap flied distributions, which make for low detent force and
an essentially sinusoidal back-EMF waveform [11–13]. However, ideal Halbach
magnetization technique was full of challenges in fabrication, and some Halbach
magnetic poles structures, which were combined by tangential PM array and normal
PM array, were unsatisfactory to reduce thrust ripple. In [14], the PM was divided
into two pieces to reduce detent force and eliminate high-order space harmonics
affecting the thrust ripple. To improve average thrust force of PMLSM with PM
segments, the number and dimensions of PM segments were optimized whereas the
consumed PM volume was kept constant [15]. Based on PM segments shaping
technique, the modular poles shaping technique was proposed to reduce the air-gap
flux density as well as back-EMF harmonics contents produced by the poles, which
make for low thrust ripple [16]. In [17], a trapezoidal shaping PM structure was
analyzed to suppress thrust ripple. However, there are still some harmonics in
air-gap flux density and back-EMF result in thrust ripple, whether PM segments,
modular poles or trapezoidal shaping PM. In order to obtain more ideal sinusoidal
air-gap flux density, the near-sinusoidal shaping technique was adopted. Literature
[18] presented a stair-step-shaped magnetic pole structure and [19] proposed a PM
arc shaping technique. However, the fundamental air-gap flux density of these
magnetic poles of PM shaping structures is decreased inevitably which results in
lower power density, compared to the PMLSM without shaping. In order to
compensate average thrust force of the PMLSM with PM shaping without deteri-
orating the thrust ripple, this chapter proposes a novel PM shaping technique which
injects third harmonic into sinusoidal shaping PM poles for a long primary
PMLSM.
11.1 Introduction 199

In this chapter, a novel PM shaping technique is proposed to improve average

thrust force and efficiency of the long primary PMLSM without deteriorating thrust
ripple for potential application in small unmanned aerial vehicle electromagnetic
launch systems. In Sect. 11.2, the PMLSM with sine and third harmonic (SIN+3rd)
shaping mover, sine (SIN) shaping mover and conventional (without shaping)
shaping mover are described. In Sect. 11.3, the optimal third harmonic is analyti-
cally derived and conformed by finite element (FE) method, and optimal PM edge
thickness and pole arc coefficient are determined by non-linear programing method.
In Sect. 11.4, the quantitative comparisons of the PMLSM with different shaping
movers on electromagnetic performances including the air-gap flux density,
back-EMF, average thrust force, thrust ripple, losses and efficiency are carried out.
Section 11.5 is conclusion.

11.2 PMLSM with Different Shaping Movers

In order to compare the electromagnetic performances of the PMLSM with different

shaping movers, the 9-slot/8-pole long primary PMSLM with double-layer
non-overlapping windings are employed in this chapter. Figure 11.1a shows the
proposed PMLSM with SIN+3rd shaping mover, Fig. 11.1b, c show the PMLSM
with SIN shaping mover and the conventional shaping mover. For the SIN+3rd
shaping mover, the position of maximum PM thickness is variable with the

(a) Proposed SIN+3rd shaping mover

(b) SIN shaping mover

(c) Conventional shaping mover

Fig. 11.1 9-slot/8-pole PMLSM with different shaping movers

200 11 Design and Analysis of Permanent Magnet …

Table 11.1 Main design Item Value/unit

specifications and parameters
Pole-pitch 27.0 mm
Maximum PM thickness 4.0 mm
Minimum air-gap length 1.5 mm
Primary tooth width 8.0 mm
Primary height 34.5 mm
Primary yoke thickness 10.0 mm
Mover length 216.0 mm
Mover back iron thickness 10.0 mm
Axis length 70.0 mm
Number of turns/coil 60
Peak current 10.0 A
Rated speed 1.08 m/s

amplitude of third harmonic injected into SIN+3rd shaping PM, while for SIN
shaping mover, the position of maximum PM thickness is along d-axis, for con-
ventional shaping PMs, the PM thickness is constant. Although the PM contour is
various, the minimum air-gap length and maximum PM thickness maintain the
same for all PMLSM topologies, and the parameters of primary and mover back
iron are also uniform. All main specification parameters are given in Table 11.1.

11.3 Optimal Design of PM Edge Thickness

and Pole Arc Coefficient

The optimal amplitude of third harmonic injected into SIN+3rd shaping mover are
determined to 1/6 by aforementioned analytical and FE analysis. Meanwhile, in
account of significant influences of the PM edge thickness he and pole arc coeffi-
cient a on the thrust force, thrust ripple and efficiency, it is necessary to optimize
these parameters simultaneously. Therefore, the non-linear programing method is
employed to realize the multi-objectives optimization design in PMLSM with SIN
+3rd shaping mover. To better illustrate the PM edge thickness, the structure is
shown in Fig. 11.2.
The thrust ripple can be defined as

Fmax  Fmin
Fr ¼  100% ð11:1Þ
Favg

where the Fmax, Fmin and Favg are the maximum thrust force, minimum thrust force
and average thrust force, respectively.
11.3 Optimal Design of PM Edge Thickness and Pole Arc Coefficient 201

Fig. 11.2 Contour of SIN+3rd shaping PM with edge thickness

The efficiency ignoring mechanical losses can be expressed by

Favg v
g¼ ð11:2Þ
Favg v þ Pc þ PFe þ Pe

where the Pc, PFe and Pe are the copper losses, the iron losses, and the PM eddy
current losses.
The non-linear programing equation f(a,he) can be given, which is expressed as a
weighted sum of individual objectives.

Favg ða; he Þ Fir gða; he Þ

f ða; he Þ ¼ x1 þ x2 þ x3 ð11:3Þ
Fiavg Fr ða; he Þ gi

where the Favg(a,he), Fr(a,he) and η(a,he) are the variable values of average thrust
force, thrust ripple and efficiency with design variables he and a, the Fiavg, Fir and ηi
are the initial values, x1, x2 and x3 are the weight coefficients of average thrust
force, thrust ripple and efficiency, which satisfy the equation x1+ x2+ x3= 1.
The responses of thrust force Favg(a,he), thrust ripple Fr(a,he) and efficiency η(a,he)
with design variables he and a are predicted by FE method, and the response rela-
tionship between design variables and optimization objectives is shown in Fig. 11.3.
According to Fig. 11.3a, b, it can be observed that the thrust force and thrust
ripple appear the same trend, which increases with the increase of design variables
he and a. However, it is desired that the proposed machine have large thrust force
and low thrust ripple simultaneously. Therefore, there are some conflicts during
determination of design variables he and a. Although the efficiency also presents the
gradually increased trend, there are some randomness in variation of efficiency with
the increase of design variables he and a. For instance, when the design variables he
is constant and the design variables a is greater than 0.95, the efficiency does not
increase any more. Consequently, it is necessary to adopt the tradeoff design in
thrust force, thrust ripple and efficiency to obtain optimal comprehensive perfor-
mances by solving equation f(a,he).
202 11 Design and Analysis of Permanent Magnet …

(a) Thrust force distribution between he and α

(b) Thrust ripple distribution between he and α (Unit: %)

(c) Efficiency distribution between he and α (Unit: %)

Fig. 11.3 Response of machine between a and he

11.3 Optimal Design of PM Edge Thickness and Pole Arc Coefficient 203

Fig. 11.4 Variation of f(a,he) with design variables he and a

The non-linear programing equation f(a,he) is solved based on the aforemen-

tioned response results, which are shown in Fig. 11.3. The variation of f(a,he) with
design variables he and a is shown in Fig. 11.4.
From Fig. 11.4, it can be observed that the results of the equation f(a,he) emerges
the random variation with design variables he and a. There is a maximum value of f
(a,he)max, which is equal to 1.435, at which a red cube point is marked. In the case,
the corresponding optimal values of design variables a and he are determined to be
0.85 and 0 mm, respectively.

11.4 Electromagnetic Performance Comparision

In previous sections, the optimal third harmonic injected into SIN+3rd shaping PMs
to maximize the thrust force in the 9-slot/8-pole PMLSM with the SIN+3rd shaping
mover is analytically derived and verified by FE methods, and the optimal PM edge
thickness and pole arc coefficient are determined by non-linear programing method.
The performances of PMLSM with SIN+3rd shaping mover are evaluated and
compared with PMLSM with SIN shaping mover and conventional shaping mover.
For the sake of contrastive analysis, the specification parameters,materials and
electrical ratings of the primary for the three machines should be uniform, and the
minimum air-gap length, maximum PM thickness, PM edge thickness and pole arc
coefficient are also same except the PM contour.
204 11 Design and Analysis of Permanent Magnet …

11.4.1 Open-Circuit Flux Density Distributions

and Back-EMFs

The FE predicted open-circuit field distributions of the 9-slot/8-pole PMLSM with

different shaping movers are shown in Fig. 11.5. Figure 11.6 shows the harmonic
contents of air-gap flux density. It can be observed that the PMLSM with con-
ventional shaping mover has the highest air-gap flux density fundamental ampli-
tude, but the lowest fundamental amplitude exists in PMLSM with SIN shaping
mover. However, the fundamental air-gap flux density of PMLSM with SIN+3rd
shaping mover is improved remarkably by injecting third harmonic into SIN+3rd
shaping PM compared to PMLSM with SIN shaping mover, which will lead to the
back-EMFs improvement. Obviously, there are mainly odd order harmonics in
air-gap flux density of PMLSM with different shaping mover. The third harmonic
contents of PMLSM with SIN+3rd and conventional shaping mover are much
higher than that of one with SIN shaping mover, but this will not deteriorate the
thrust ripple. As will be shown later.
Figure 11.7 shows the FE predicted back-EMF waveforms and harmonic
spectrums of the 9-slot/8-pole PMLSM machine with different shaping mover. It
can be seen that the amplitude of back-EMF has the same variation trend with that
of air-gap flux density amplitude, and 12.3% has been increased for PMLSM with
SIN+3rd shaping mover compared to PMLSM with SIN shaping mover, which will

(a) SIN+3rd shaping mover

(b) SIN shaping mover

(c) Conventional shaping mover

Fig. 11.5 Field distributions of 9-slot/8-pole PMLSM

11.4 Electromagnetic Performance Comparision 205

Fig. 11.6 Harmonic spectrums of the air-gap flux density

be contributed to improving the average thrust force of the PMLSM. From the
harmonic spectrums of back-EMF, as shown in Fig. 11.7b, it can also be observed
that the third harmonic contents of the PMLSM with SIN+3rd shaping mover and
conventional shaping mover are much higher than that of the one with SIN shaping
mover. However, these triplen harmonics caused by triplen harmonics air-gap flux
density will not result in thrust ripple, due to the triplen harmonics have been
eliminated in line-line back-EMF in the balanced Y connection windings.

11.4.2 Thrust Force Characteristics

For the PMLSM with different shaping mover, the variation of detent force and
normal force with mover position under no-load condition are obtained by 2D FE
prediction, as shown in Fig. 11.8a, b. The peak-to-peak value of detent force Fdent
and average normal force Fn are listed in Table 11.2. It can be seen that the Fdent of
PMLSM with SIN+3rd shaping mover is decreased by 77.07% of the one with
conventional shaping mover, and only increased 2.07 N compared to PMLSM with
SIN shaping mover. The normal force of PMLSM with SIN+3rd shaping mover is
reduced by 27.74% of one with conventional shaping mover.
Figure 11.8c shows the variation of thrust force with mover position under id= 0
control, the average thrust force Fav and thrust ripple Fr are listed in Table 11.2. It
can be observed that the PMLSM with conventional shaping mover has the highest
average thrust force, but it also has the most large thrust ripple. The average thrust
206 11 Design and Analysis of Permanent Magnet …

(a)Waveforms

(b) Spectrums

Fig. 11.7 Phase back-EMFs of SPMLSM with different shaping mover (1.08 m/s)
11.4 Electromagnetic Performance Comparision 207

(a)Variation of detent force with mover position

(b)Variation of normal force with mover position

(c)Variation of thrust force with mover position under id=0 control

Fig. 11.8 Thrust force and detent force

208 11 Design and Analysis of Permanent Magnet …

Table 11.2 Force Items Conventional SIN SIN+3rd

components of PMLSM with
different shaping mover Fav 269.25 N 200.83 N 235.36 N
Fpk2pk 25.40 N 4.24 N 7.47 N
Fr 9.43% 2.11% 3.17%
Fdent 22.55 N 3.1 N 5.17 N
Fn 1827.69 N 963.35 N 1320.68 N

force of PMLSM with SIN+3rd mover is improved by 17.19% compared with that
of the PMLSM with SIN shaping mover. Meanwhile, the thrust ripple is similar for
the PMLSM with SIN+3rd and SIN shaping mover, which is much lower than that
of the one with conventional shaping mover. This is due to lower detent force and
more sinusoidal back-EMFs, which jointly result in low thrust ripple.

11.4.3 Losses and Efficiency

The PM eddy-current losses density distributions of PMLSM with different shaping

mover at rated operation condition are predicted by FE method, as shown in
Fig. 11.9. For the PMLSM with SIN+3rd shaping mover, the PM eddy-current
losses density is much smaller than that of PMLSM with the conventional shaping
mover, due to the effect of the q-axis armature reaction flux is away from the edge
of PM with shaping. It is noted that PM eddy-current losses density is not increased

(a) SIN+3rd (b) SIN (c) Conventional

Fig. 11.9 PM eddy-current losses density distributions

11.4 Electromagnetic Performance Comparision 209

Fig. 11.10 Iron losses density distributions. a SIN+3rd. b SIN. c Conventional

compared to one with the SIN shaping mover. However, by injecting third har-
monic into SIN shaping mover, the volume of PM for the PMLSM with SIN+3rd
shaping mover is a little higher than that of one with SIN shaping mover, which will
cause corresponding increase of PM eddy-current losses, as shown in Fig. 11.11.
Figure 11.10 shows the iron losses density distributions of PMLSM with dif-
ferent shaping mover, the total losses and efficiency are shown in Fig. 11.11 when
all machines operate under rated condition. The total losses consist of copper losses,
PMs eddy-current losses and iron losses. It can be seen that the iron losses for the
PMLSM with SIN+3rd shaping mover is much higher than that of the PMLSM with
SIN shaping mover, due to the higher fundamental and third harmonic air-gap flux
density by injecting third harmonic into SIN shaping mover. However, the iron
losses are still smaller than that of PMLSM with conventional shaping mover.
Importantly, the efficiency of PMLSM with SIN+3rd shaping mover is increased by
1.12%, under the constraints of same copper losses, compared with that of one with
SIN shaping mover.
210 11 Design and Analysis of Permanent Magnet …

Fig. 11.11 Total losses and efficiency

11.5 Summary

This chapter proposes a PMLSM with SIN+3rd shaping mover to improve thrust
force and efficiency without deteriorating thrust ripple. The optimal third harmonic
is obtained by analytically deriving and confirmed by FE method. Meanwhile, the
PM edge thickness and pole arc coefficient are optimized. It is demonstrated that the
optimal third harmonic is 1/6 fundamental one, and the optimal PM edge thickness
and pole arc coefficient are 0 mm and 0.85, respectively. The electromagnetic
performances of the PMLSM with different shaping mover are compared. The
results show that the PMLSM with SIN+3rd shaping mover can obtain 17.19%
higher average thrust force than that of one with SIN shaping mover, the efficiency
is increased by 1.12% under the constraints of same copper losses, while the thrust
ripple is not deteriorated.

References

1. T.D. Barkand, W.J. Helfrich, Application of dynamic braking to mine hoisting systems. IEEE
Trans. Ind. Appl. 24(5), 884–896 (1998)
2. D. Patterson, A. Monti, C.W. Brice, R.A. Dougal, R.O. Pettus, S. Dhulipala, D.C. Kovuri, T.
Bertoncelli, Design and simulation of a permanent magnet electromagnetic aircraft launcher.
IEEE Trans. Ind. Appl. 41(2), 566–575 (2005)
3. I.S. Jung, S.B. Yoon, J.H. Shim, D.S. Hyun, Analysis of forces in a short primary type and a
short secondary type permanent magnet linear synchronous motor. IEEE Trans. Energy
Covers. 14(4), 1265–1270 (1999)
4. Y.J. Kim, M. Watada, H, Dohmeki, Reduction of the cogging force at the outlet edge of a
stationary discontinuous primary linear synchronous motor. IEEE Trans. Magn. 43(1), 40–45
(2007)
References 211

5. Y.W. Zhu, S.G. Lee, K.S. Chung, Y.H. Cho, Investigation of auxiliary poles design criteria on
reduction of end effect of detent force for PMLSM. IEEE Trans. Magn. 45(6), 2863–2866
(2009)
6. K.H. Shin, K.H. Kim, K. Hong, J.Y. Choi, Detent force minimization of permanent magnet
linear synchronous machines using subdomain analytical method considering auxiliary teeth
configuration. IEEE Trans. Magn. 53(6), 8104504 (2017)
7. M. Inoue, K. Sato, An approach to a suitable stator length for minimizing the detent force of
permanent magnet linear synchronous motors. IEEE Trans. Magn. 36(4), 1890–1893 (2000)
8. Z.Q. Zhu, Z.P. Xia, D. Howe, P.H. Mellor, Reduction of cogging force in slotless linear
permanent magnet motors. IEE Proc. Electr. Power Appl. 144(4), 277–282 (1997)
9. S.U. Chung, H.J. Lee, B.C. Woo, J.W. Kim, J.Y. Lee, S.R. Moon, S.M. Hwang, A feasibility
study on a new doubly salient permanent magnet linear synchronous machine. IEEE Trans.
Magn. 46(6), 1572–1575 (2010)
10. Y.X. Cheng, J.J. Yang, G.X. Liu, A novel low thrust ripple 3-phase 6-slot permanent magnet
linear synchronous, in Proceedings of IEEE International Conference on CECNet, pp. 49–53
(2011)
11. Z.X. Jiao, T.Y. Wang, L. Yan, Design of a tubular linear oscillating motor with a novel
compound Halbach magnet array. IEEE Trans. Mech. 22(1), 498–508 (2017)
12. Z.Q. Zhu, D. Howe, Halbach permanent magnet machines and applications: a review. IEE
Proc. Electr. Power Appl. 148(4), 299–308 (2001)
13. X.Z. Huang, Q. Tan, Q.L. Wang, J. Li, Optimization for the pole structure of slot-less tubular
permanent magnet synchronous linear motor and segmented detent force compensation. IEEE
Trans. Appl. Supercond. 26(7), 0611405 (2016)
14. M.Y. Kim, Y.C. Kim, G.T. Kim, Design of slotless-type PMLSM for high power density
using divided PM. IEEE Trans. Magn. 40(2), 746–749 (2004)
15. A.H. Isfahani, Analytical framework for thrust enhancement in permanent-magnet (PM) linear
synchronous motors with segments PM poles. IEEE Trans. Magn. 46(4), 1116–1122 (2010)
16. A.H. Isfahani, S.V. Zadeh, M.A. Rahman, Using modular poles for shape optimization of flux
density distribution in permanent magnet. IEEE Trans. Magn. 44(8), 2009–2015 (2008)
17. Y.Q. Zhang, M.H. Yu, X. Liu, Y.Y. Yue, Field and thrust analysis of linear servo motor with
permanent magnet of different shapes, in Proceedings of 13th International Conference on
Electrical Machines and Systems (ICEMS), pp. 1516–1519 (2010)
18. N.R. Tavana, A. Shoulaie, V. Dinavahi, Analytical modeling and design optimization of
linear synchronous motor with stair-step-shaped magnetic poles for electromagnetic launch
applications. IEEE Trans. Plasma Sci. 40(2), 519–527 (2012)
19. N.R. Tavana, A. Shoulaie, Analysis and design of magnetic pole shape in linear permanent
magnet machine. IEEE Trans. Magn. 46(4), 1000–1006 (2010)