Celik 11

Segmental Rotor Switched
Reluctance Drives
Tuncay Celik
A thesis submitted for the degree of
Doctor of Philosophy
August, 2011
University of Newcastle upon Tyne

School of Electrical, Electronic and Computer Engineering
To my wife, Azime
and son, Tymur
ABSTRACT
Abstract
One of the well-known drawbacks of switched reluctance machines is the relatively high
output torque ripple. Techniques aiming to reduce machine torque ripple either
compromise the machine performance or the simplicity of the inverter and the controller.
The work presented in this thesis shows that low torque ripple over a wide speed range
can be achieved without severe penalties in terms of the machine performance and the
size, cost and complexity of the power electronics and the controller. This is achieved by
designing a 6-phase machine and driving it from a three-phase full bridge circuit.
Switched reluctance motors with segmented rotors are a relatively recent advancement in
the electromagnetic design of doubly-salient reluctance motors, having only been
introduced in 2002. By replacing the conventional toothed rotor with individual segments,
it has been proven that higher torque density than conventional switched reluctance
machines could be achieved. Early work by Mecrow and El-Kharashi has demonstrated
the operation of prototype machines with short-pitched and fully-pitched windings. The
machine design work presented here builds on this early work by examining aspects of
the machine design and its operation. Two six-phase machines one with a segmented
rotor and the other with a toothed rotor - have been designed. Performance comparisons
have been made between the two six-phase machines and a three phase segmented rotor
machine that was previously designed at Newcastle University.
Additionally, a three phase single tooth winding and a two phase segmented rotor
switched reluctance machine have been studied in simulation and experimentally.
Detailed comparison of inverter ratings and machine efficiencies are made under equal
conditions for a 2-phase machine driven from h-bridge and asymmetric half-bridge
inverters. This is achieved with results from a test rig and the use of accurate dynamic
simulation. Simulation models for 3-phase and 6-phase machines have also been
generated. Detailed comparison of inverter ratings and machine efficiencies are made
under equal conditions for the 3-phase and 6-phase drives in the dynamic simulation.
Comparisons between simulated and measured results are shown to be very good for all
of the drives.
ii
TABLE OF CONTENTS
Table of Contents
Abstract...ii
Table of Contents......iii
List of Figures.............x
List of Tablesxxvii
Acknowledgements....xxxi
Symbols.................xxxii
CHAPTER 1
INTRODUCTION ........................................................... 1
1.1
Introduction .......................................................................................................... 1
1.2
Switched Reluctance Machines In Industry......................................................... 2
1.2.1
1.2.2
1.3
Target Industries for SR Motor Drive Technology...................................... 3

Examples of SR Drives in Industry ............................................................. 5
Switched Reluctance Machines With Segmental Rotors ..................................... 9
1.3.1
1.3.2
Multi-tooth Winding Segmental Rotor SRM............................................. 10

Single-tooth Winding Segmental Rotor SRM ........................................... 10
1.4
Converters And Controllers For Switched Reluctance Machines ..................... 11
1.5
Switched Reluctance Machines vs. Permanent Magnet Brushless DC Machines

............................................................................................................................ 14
1.6
Objectives And Contribution To Knowledge .................................................... 16
1.7
Thesis Overview ................................................................................................ 17
CHAPTER-2
SRM TOPOLOGIES .................................................... 19
2.1
Introduction ........................................................................................................ 19
2.2
SRM Torque Ripple ........................................................................................... 20
2.2.1
2.3
Torque Ripple Minimisation through Machine Design ............................. 21
SRM Topologies ................................................................................................ 22

iii
TABLE OF CONTENTS
2.3.1
2.3.2
Conventional Toothed Rotor SRMs........................................................... 22

Segmented Rotor Switched Reluctance Machines .................................... 26
2.3.2.1 Segmented Rotor Switched Reluctance Motors with Fully-Pitched

Windings ................................................................................................................ 29
2.3.2.2 Segmented Rotor Switched Reluctance Motors with Short-Pitched
Windings ................................................................................................................ 33
2.3.2.3 Other Segmented Rotor Switched Reluctance Motor Designs .............. 39
2.4
Summary ............................................................................................................ 41
CHAPTER-3
SR DRIVE TOPOLOGIES .......................................... 43
3.1
Introduction ........................................................................................................ 43
3.2
Torque Ripple Minimisation through Drive Control ......................................... 45
3.2.1
3.3
Measuring Torque Ripple .......................................................................... 45
Drive Topologies ............................................................................................... 46
3.3.1
3.3.2
Review of Drives ....................................................................................... 47

Review of Control Methods ....................................................................... 48
3.3.2.1
3.3.2.2
3.3.2.3
3.3.3
Hysteresis Current Control .................................................................... 49

PWM Control ......................................................................................... 49
Current Profiling .................................................................................... 50
Review of Converter Topologies ............................................................... 50
3.3.3.1
3.3.3.2
3.3.3.3
3.3.3.4
3.3.3.5
3.3.3.6
3.3.3.7
3.4
Asymmetric Half-Bridge Converter ...................................................... 51

Shared Switch Converter (Miller Circuit).............................................. 53
C-Dump Converter................................................................................. 54
Split DC-Link Converter........................................................................ 55
Bifilar Winding Converter ..................................................................... 56
H-Bridge Converter ............................................................................... 57
Three-Phase Full Bridge Converter ....................................................... 58
Summary ............................................................................................................ 63
CHAPTER-4
MACHINE DESIGN AND TESTING ........................ 64
4.1
Introduction ........................................................................................................ 64
4.2
Design of the 6-Phase Segmented Rotor SRM .................................................. 65
4.2.1
4.2.2
4.3
Determining the Machine Dimensions ...................................................... 71

Design Space Analysis ............................................................................... 73
Design of the 6-Phase Toothed Rotor SRM ...................................................... 83
iv
TABLE OF CONTENTS
4.4
Summary ............................................................................................................ 89
CHAPTER-5
MACHINE CONSTRUCTION TECHNIQUES AND
TEST RIG DESCRIPTION ....................................................................... 90
5.1
Machine with Separate Stator Cores .................................................................. 90
5.2
Construction of the Prototype Machine ............................................................. 94
5.2.1
5.2.2
5.2.3
5.3
Stator Construction .................................................................................... 95

Windings .................................................................................................... 98
Rotor Construction ................................................................................... 102
Test Rig Description ........................................................................................ 103
5.3.1
5.3.2
5.3.3
5.3.4
5.3.5
Load Arrangement ................................................................................... 104

High Voltage Power Converter ................................................................ 105
Control Electronics .................................................................................. 109
Drive Software Explanation ..................................................................... 111
Prototype Testing ..................................................................................... 113
5.3.5.1
5.3.5.2
5.3.5.3
5.4
Flux-Linkage Measurements ............................................................... 113

Static Torque Measurements................................................................ 116
Thermal Characterisation of the Prototype Machine ........................... 118
Summary .......................................................................................................... 119
CHAPTER-6
SIMULATION MODELS .......................................... 120
6.1
Introduction ...................................................................................................... 120
6.2
SRM Modelling and Simulation ...................................................................... 122
6.2.1
6.2.2
6.2.2.1
6.2.2.2
6.3
Background .............................................................................................. 122

Review of SRM Simulation Methods ...................................................... 123
Early Work ........................................................................................... 123
Recent Work ........................................................................................ 126
Simulation Database ........................................................................................ 128
6.3.1
6.3.2
6.3.2.1
6.3.2.2
6.3.2.3
6.3.2.4
6.3.2.5
6.3.2.6
Machine Model ........................................................................................ 129

Control Modelling.................................................................................... 134
Angle and Speed Calculations ............................................................. 134
Current Sampling ................................................................................. 137
PID Controller ...................................................................................... 137
PWM Controller................................................................................... 139
Converter Modelling ............................................................................ 143
Phase Diode Modelling ........................................................................ 147
v
TABLE OF CONTENTS
6.3.2.7
6.3.3
6.3.3.1
6.3.3.2
6.3.3.3
Electronics Loss Modelling ................................................................. 147

Inverter Modelling ................................................................................... 149
Asymmetric Half-Bridge Inverter Drive .............................................. 149
Six-Phase Delta Connected Drive ........................................................ 149
Six-Phase Star Connected Drive .......................................................... 150
6.4
Modelling Of SRMs with Mutually Coupled Phase Windings ....................... 153
6.5
Performance Calculations ................................................................................ 162
6.6
Summary .......................................................................................................... 163
CHAPTER-7
MACHINE OPERATION and COMPARISON to
SIMULATION - TWO PHASE DRIVES ............................................... 165
7.1
Introduction ...................................................................................................... 165
7.2
Measurements and Simulation Results ............................................................ 166
7.2.1
7.2.2
7.3
Inverter VA-Rating Comparison...................................................................... 187
7.3.1
7.3.2
7.4
Two-Phase Machine Driven by H-Bridge Inverter .................................. 168

Two-Phase Machine Driven by Asymmetric Half-Bridge Inverter ......... 179
Drive under Voltage Control ................................................................... 189

Drive under Current Control .................................................................... 193
Summary .......................................................................................................... 196
CHAPTER-8
MACHINE OPERATION and COMPARISON to
SIMULATION - THREE PHASE DRIVES ........................................... 199
8.1
Introduction ...................................................................................................... 199
8.2
Measurements and Simulation Results ............................................................ 200
8.2.1
8.2.1.1
8.2.1.2
8.2.1.3
8.2.1.4
8.2.2
8.3
Comparison of Measurements against Simulations ................................. 200

Test-1 (Voltage Control) ...................................................................... 202
Test-2 (Current Control) ...................................................................... 205
Test-3 (Current Control) ...................................................................... 207
Test-4 (Voltage Control) ...................................................................... 209
Dynamic Torque vs. Speed Characteristics ............................................. 214
Summary .......................................................................................................... 217
vi
TABLE OF CONTENTS
CHAPTER-9
MACHINE OPERATION and SIMULATION - SIX
PHASE DRIVES........................................................................................ 218
9.1
Introduction ...................................................................................................... 218
9.2
Dynamic Simulation Results............................................................................ 220
9.2.1
Drive
9.2.1.1
9.2.1.2
9.2.2
Drive
9.2.2.1
9.2.2.2
9.3
Mutually De-Coupled Star Connected Line Current Controlled Six-Phase

220
Drive under Current Control ................................................................ 221
Drive under Voltage Control ............................................................... 232
Mutually De-Coupled Delta Connected Line Current Controlled Six-Phase
240
Drive under Current Control ................................................................ 241
Drive under Voltage Control ............................................................... 248
Dynamic Measurements Of The Prototype Six-Phase Segmental Rotor SRM .....

.......................................................................................................................... 254
9.3.1
9.3.2
Dynamic Torque against Speed Characteristics ...................................... 254

Dynamic Measurements........................................................................... 257
9.3.2.1 Low Speed Measurements of the Prototype Six-Phase Star-Connected

Segmental Rotor Machine Driven by the Three-Phase Bridge Circuit................ 263
9.3.2.1.1
9.3.2.1.2
9.3.2.1.3
Low Speed Measurement-1 ........................................................... 263

9.3.2.2 High Speed Measurements of the Prototype Six-Phase Star-Connected

Segmental Rotor Machine Driven by the Three-Phase Bridge Circuit................ 272
9.3.2.2.1
9.3.2.2.2
9.3.2.2.3
9.3.2.2.4
9.4
High Speed Measurement-1 ........................................................... 273

Summary .......................................................................................................... 287
CHAPTER-10
COMPARISON OF MACHINES and DRIVES .... 289
10.1 Electromagnetic Performance Comparison ..................................................... 289

10.1.1 Number of turns per slot, slot fill factor and conductor diameter as
designed .................................................................................................................. 291
10.1.2
Number of turns per slot, slot fill factor kept the same ........................... 292
vii
TABLE OF CONTENTS
10.2 Torque Ripple Comparison of Three and Six Phase Drives ............................ 296
10.3 Comparison of Inverter VA-Ratings ................................................................ 308
10.3.1
10.3.2
10.3.3
Drive under Voltage Control ................................................................... 309

Drive under Current Control .................................................................... 311
Discussion on Converter VA-Rating Comparison ................................... 313
10.4 Discussion on Torque Ripple Comparison ...................................................... 314

10.5 Summary .......................................................................................................... 317
CHAPTER-11
CONCLUSIONS and FUTURE WORK ................ 319
11.1 Conclusions ...................................................................................................... 319

11.1.1
11.1.2
11.1.3
11.1.4
11.1.5
Machine Performance .............................................................................. 320

Machine Construction Techniques .......................................................... 321
Drive Modelling ....................................................................................... 321
Inverter VA-Ratings and Drive Testing ................................................... 322
Torque Ripple .......................................................................................... 325
11.2 Future Work ..................................................................................................... 326
APPENDIX-A
FLUX-LINKAGE MEASUREMENT and POSTPROCESSING of DATA .......................................................................... 329
A.1
A.2
FLUX-LINKAGE MEASUREMENTS .......................................................... 329

FLUX-LINKAGE VIEWER GUI FOR DATA POST-PROCESSING .......... 330
APPENDIX-B
LAMINATION DRAWINGS of the SIX-PHASE
SEGMENTAL AND CONVENTIONAL ROTOR MACHINES......... 333
B.1 STATOR SEGMENT LAMINATION DRAWING OF THE SEGMENTAL
ROTOR SIX-PHASE MACHINE ............................................................................. 334
B.2 ROTOR SEGMENT LAMINATION DRAWING OF THE SEGMENTAL
B.3 STATOR SEGMENT LAMINATION DRAWING OF THE TOOTHED
B.4 ROTOR LAMINATION DRAWING OF THE TOOTHED ROTOR SIXPHASE MACHINE ................................................................................................... 337
APPENDIX-C
SIMULATION PACKAGE and GRAPHICAL USER
INTERFACE.............................................................................................. 338
C.1
C.2
INTRODUCTION ........................................................................................... 338

SIMULATION DATABASE .......................................................................... 340
APPENDIX-D ESTIMATION of MACHINE SLOT AREA and PHASE

RESISTANCE ........................................................................................... 348
viii
TABLE OF CONTENTS
D.1
D.2
CALCULATION OF MACHINE SLOT AREA ............................................ 348

ESTIMATION OF MACHINE PHASE RESISTANCE ................................ 355
LIST OF REFERENCES ......................................................................... 360
ix
LIST OF FIGURES
List of Figures
Chapter-1
Figure 1 Rectilinear representation of a 3-phase multi-tooth winding segmental rotor
SRM in the aligned position. ............................................................................................. 10
Figure 2 Rectilinear representation of a 3-phase single-tooth winding segmental rotor
SRM in the aligned position. ............................................................................................. 11
Figure 3 Star-connected 3-phase fully-pitched winding SRM driven from 3-phase full
bridge inverter. ................................................................................................................... 13
Figure 4 Delta-connected 3-phase fully-pitched winding SRM driven from 3-phase full
bridge inverter. ................................................................................................................... 13
Figure 5 Flux-Linkage loci for the 3-phase segmental rotor 12-10 SRM (in dark blue), 3phase segmental rotor 12-8 SRM (in red), 3-phase conventional toothed rotor 12-8 SRM
(in green) and a brushless d.c. PM machine (in light blue), operating under current control
............................................................................................................................................ 15
Chapter-2
Figure 6 Schematic of a three-phase conventional switched reluctance machine showing
various motor sections ....................................................................................................... 23
Figure 7 Variation of inductance and positive and negative torque production regions with
ideal pulse unidirectional current. Lmax and Lmin are the maximum and minimum phase
inductances, respectively. .................................................................................................. 24
Figure 8 Description of several machine dimensions for a rectilinear representation of a
conventional toothed rotor SRM........................................................................................ 25
Figure 9 Three-phase segmental rotor prototype machine schematic layout and magnetic
flux plots in aligned and unaligned rotor positions ............................................................ 27
Figure 10 Cross-sectional view of Xu and Lipo's axially laminated single salient variable
reluctance motor [95] ......................................................................................................... 28
Figure 11 Cross-sectional view of Horst's two-phase segmental rotor SRM [96]............. 28
Figure 12 Rectilinear representation of a single phase segmental rotor SRM with 2D flux
plots in aligned and unaligned positions [23] .................................................................... 30
x
LIST OF FIGURES
Figure 13 Comparison of aligned and unaligned magnetisation curves of toothed rotor
(both red traces) and segmental rotor (both blue traces) when tooth width to pole pitch
ratio is 0.3........................................................................................................................... 31
Figure 14 Comparison of aligned and unaligned magnetisation curves of toothed rotor
(both red traces) and segmental rotor (both blue traces) when tooth width to pole pitch
ratio is 0.7........................................................................................................................... 31
Figure 15 Rectilinear representation of three-phase segmented rotor SRMs [23] ............ 35
Figure 16 Three-phase 12/10 segmental rotor SRM with single-tooth windings - shown in
aligned position .................................................................................................................. 36
Figure 17 Magnetic vector potential of a phase coil as a function of coil MMF for both
segmental designs and a conventional SRM ...................................................................... 36
Figure 18 Measured static torque characteristics of 12/10 single tooth design with a single
phase excited. Each curve is for a constant current, rising in 2.0A steps to 20.0A [97] ... 37
Figure 19 Possible orientation of magnetisation for single tooth winding segmental rotor
SRM (all shown in aligned position) ................................................................................. 38
Figure 20 Cross-sectional view of Oyama et als segmental rotor SRM with segments
embedded in an aluminium block [98] .............................................................................. 40
Figure 21 Segmental rotor SRM with circular stator teeth work of Vattikuti et al [99] . 40
Figure 22 Outer rotating segmental rotor SRM with 18 stator teeth and 15 rotor segments
for use in aero-engine embedded starte / generator application - R.Hall et al [8] ............. 41
Chapter-3
Figure 23 Simulation of 3-phase and 6-phase 12-10 segmented rotor SRMs under perfect
current control .................................................................................................................... 48
Figure 24 Asymmetric half-bridge converter driving a three-phase machine ................... 52
Figure 25 Shared switch converter driving a three-phase machine ................................... 54
Figure 26 C-dump converter driving a three-phase machine............................................. 55
Figure 27 Split dc-link converter driving a two-phase machine ........................................ 56
Figure 28 Three-phase bifilar winding converter .............................................................. 56
Figure 29 H-bridge converter driving a four-phase machine............................................. 57
Figure 30 H-bridge converter driving a back-to-back connected two-phase machine ...... 58
xi
LIST OF FIGURES
Figure 31 Three-phase full bridge circuit driving a delta connected machine .................. 59
Figure 32 Star-connected six-phase machine driven by a 3-phase full bridge circuit ....... 60
Figure 33 Delta-connected six-phase SRM operated from a conventional 3-phase bridge
circuit ................................................................................................................................. 61
Figure 34 Ideal line currents for the delta (a) and star (b) connected six-phase drives ..... 62
Figure 35 Drawing of the 6-phase 12-10 short-pitched segmented rotor SRM (to establish
design rules and equations) in the unaligned position with respect to conduction of Phases
A and F ............................................................................................................................... 66
Chapter-4
Figure 36 Drawing of the 6-phase 12-10 short-pitched segmented rotor SRM (to establish
design rules and equations) in the aligned position with respect to conduction of Phases A
and F. Phase configuration is the same as shown in Figure 35.......................................... 67
Figure 37 Flux plot of the 6-phase segmental rotor 12-10 SRM in the unaligned position.
The energised coils are shown with the conventional dot / cross representation. Only one
half of the machine is shown. ............................................................................................ 67
Figure 38 Flux plot of the 6-phase segmental rotor 12-10 SRM in the aligned position.
The energised coils are shown with the conventional dot / cross representation. Only one
half of the machine is shown. ............................................................................................ 68
Figure 39 2D rectilinear representation of six-phase SRM showing the direction of slot
MMFs and teeth fluxes ...................................................................................................... 68
Figure 40 Aligned and unaligned average vector potential vs. MMF c urves for the
prototype machine .............................................................................................................. 71
Figure 41 Mean torque vs. MMF curve for the prototype machine................................... 72
Figure 42 Magnetic flux plots for the 6-phase 12-10 short-pitched segmented rotor SRM
with two phases excited simultaneously in the aligned position (a) and in the unaligned
position (b) ......................................................................................................................... 72
Figure 43 Aligned and unaligned magnetisation curves for each design option where the
slot fill factor and the number of turns per coil are fixed .................................................. 78
Figure 44 Flux density plot of Mk3 design in the aligned position with 20A of coil current
(adjacent coils are excited simultaneously) ....................................................................... 82
Figure 45 Rectilinear representation of the toothed rotor machine design variables ........ 84
Figure 46 Magnetisation curves in the aligned and unaligned positions for the
conventional toothed rotor 6-phase SRM with 11.00mm of rotor core-back depth ......... 86
xii
LIST OF FIGURES
Figure 47 A Single stator tooth of the 6-phase conventional toothed rotor SRM showing
the stator tooth tip tang ...................................................................................................... 87
Figure 48 Magnetic flux plots of the 6-phase 12-10 short-pitched conventional toothed
rotor SRM with two phases excited simultaneously in the aligned position (a) and in the
unaligned position (b) ........................................................................................................ 88
Figure 49 Magnetisation curves of the 6-phase 12-10 SRM with conventional toothed
rotor. Magnetisation curves from unaligned to aligned position with 2o stepping shown . 88
Chapter-5
Figure 50 Graphical representation of typical winding operation with a nozzle threaded
through the stator slot opening. .......................................................................................... 91
Figure 51 Ulrike Hoefer's high speed permanent magnet vacuum pump motor made of
SMC core with three separate teeth, two back-iron rings and preformed windings [157] 92
Figure 52 Winding process with joint lapped core and picture of joint lapped core [158] 94
Figure 53 Picture of single stator core lamination for the 6-phase 12-10 segmented rotor
prototype machine .............................................................................................................. 95
Figure 54Jig used to produce the 50mm long stator tooth pieces ...................................... 96
Figure 55 50mm stator piece inside the jig and a close-up on the glued and cured 50mm
piece ................................................................................................................................... 96
Figure 56 Pictures of the jig used to produce the 150mm long stator tooth pieces ........... 97
Figure 57 Pictures of the complete stator tooth pieces ...................................................... 97
Figure 58 Curving and bulging of conductors during winding operation ......................... 99
Figure 59 Stator tooth piece covered with slot liner against winding to lamination shorts
............................................................................................................................................ 99
Figure 60 Manual winding of the stator pieces ................................................................ 100
Figure 61 Jig used to press the bulging windings and make them ready for bonding ..... 100
Figure 62 Bonding of wires through applying a current surge ........................................ 101
Figure 63 Stator tooth with bonded windings .................................................................. 101
Figure 64 Frame and stator assembly - each coil connection brought out to have full
flexibility in connection ................................................................................................... 102
xiii
LIST OF FIGURES
Figure 65 Construction of the segmented rotor of the prototype 6-phase 12-10 SR
machine ............................................................................................................................ 102
Figure 66 Finished and machined segmented rotor assembly of the prototype 6-phase 1210 SR machine ................................................................................................................. 103
Figure 67 Schematic representation of the test rig........................................................... 104
Figure 68 Schematic description of the load arrangement dc machine with a resistive
load bank connected in parallel........................................................................................ 105
Figure 69 Electric power converter.................................................................................. 106
Figure 70 Terminal box connections - phase diodes are mounted on an aluminium sheet in
this box and all coil connections are brought inside this box .......................................... 107
Figure 71 Phase connections see Figure 70 in relation to some of the labelling .......... 108
Figure 72 Control electronics........................................................................................... 110
Figure 73 Implementation of current control with the drive software ............................. 112
Figure 74 Measured flux-linkage curves for the prototype 6-phase 12-10 segmented rotor
SRM (Mk3 design) with two phases excited at the same time. Each curve is at a constant
position, running from unaligned to aligned in 2.0 degree steps ..................................... 114
Figure 75 Comparison of 2D FE predicted aligned and unaligned magnetisation curves
with those of measurements ............................................................................................. 115
Figure 76 Measured vs. predicted static torque characteristics of the prototype 6-phase
segmented rotor SRM at various excitation levels .......................................................... 117
Figure 77 Measured static torque characteristics of the prototype 6-phase segmented rotor
SRM up to 20A of excitation ........................................................................................... 117
Figure 78 Measured and predicted temperature measurements of the windings and the
machine frame .................................................................................................................. 118
Chapter-6
Figure 79 Schematic representation of modelling one phase of a conventional SRM .... 131
Figure 80 Fl ux-linkage - Current - Position characteristics ............................................ 131
Figure 81 Current - Position - Flux-Linkage characteristics created from Flux-Linkage Current - Position characteristics ..................................................................................... 132
Figure 82 3D Flux-linkage characteristics ....................................................................... 133
xiv
LIST OF FIGURES
Figure 83 3D Static torque characteristics ....................................................................... 133
Figure 84 (a) 2D torque-current-position look-up table in Simulink............................ 134
Figure 85 Magnetic flux plots for a 12-10 single tooth winding segmental rotor machine
with a single phase excited illustrating the zero position implementation in simulation 135
Figure 86 Rotor position signal with respect to Phase-A of the machine in simulation .. 136
Figure 87 Operation of PID controller in simulation of a 2-phase machine driven by an hbridge converter shown for a few electrical cycles. ......................................................... 138
Figure 88 Zoomed version of measured and demanded current signals (along with the
error signal in green) showing the digitised nature of the measured current ................... 139
Figure 89 Back-to-back connected two-phase SRM driven by an H-Bridge inverter ..... 140
Figure 90 PWM control of H-bridge inverter. +Vref and Vref are the voltage control
signals generated by the PID controller. T1, T2, T3, T4 are the transistor gate signals, Va
is the voltage subsequently imposed on the winding ....................................................... 141
Figure 91 Screenshot of the PWM controller model in Simulink ................................ 141
Figure 92 PWM controller operation in simulation - Simulation inputs are rotor speed =
300rpm, advance angle = 0 degrees, conduction angle = 120 degrees, current demand =
10A................................................................................................................................... 142
Figure 93 Traces in Figure 92 shown for a few PWM cycles ......................................... 143
Figure 94 Schematic of controller for the two-phase drive given in Figure 89 ............... 144
Figure 95 Screenshot of simulation subsystem calculating phase voltages and power
device switching states ..................................................................................................... 146
Figure 96 Phase Diode Modelling in Simulink (Inputs are phase current and phase
energisation signal, output is HIGH when Phase Diode operates, i.e. blocks the phase
current reverse direction) ................................................................................................. 147
Figure 97 IGBT conduction loss characterisation Red trace is A representation of the
loss curve from the datasheet and green trace is the linear approximation within the
operating current range of the application ....................................................................... 148
circuit ............................................................................................................................... 150
Figure 99 Star-connected six-phase machine driven by a 3-phase full bridge circuit ..... 151
Figure 100 Raw and filtered star point voltage waveforms from simulation of 6-phase
SRM driven by a 3-phase full-bridge inverter Blue trace: Raw star-point voltage; Red
trace: Filtered star-point voltage ...................................................................................... 152
xv
LIST OF FIGURES
Figure 101 Raw and filtered star point current waveforms from simulation of 6-phase
SRM driven by a 3-phase full-bridge inverter (Star-point voltage in Figure 100 Blue
trace: Raw star-point current; Red trace: Filtered star-point current ............................... 153
Figure 102 Schematic explanation of simulation of six-phase SRM with mutually coupled
phase windings ................................................................................................................. 157
Figure 103 2D rectilinear representation of six-phase SRM showing the direction of slot
MMFs and teeth fluxes .................................................................................................... 157
Figure 104 Data Exporting for Offline Statistical Calculations Once the Simulation Has
Completed (Output Ports Circled in Red)........................................................................ 162
Figure 105 Configuration parameters pane used for data importing / exporting in
Simulink ........................................................................................................................ 163
Chapter-7
Figure 106 Six-phase segmental rotor machine configured to run as a two-phase machine
.......................................................................................................................................... 167
Figure 107 H-bridge converter driving a back-to-back connected two-phase machine .. 169
Figure 108 Simulation results of the two-phase machine driven from an H-bridge inverter
under current control. Conduction angle is 120 degrees (electrical), no advancing, current
demand is 30A and the dc-link voltage is 175V. Top trace = phase voltages red for L1,
blue for L2, black for filtered L1 voltage; middle trace = phase currents red for L1, blue
for L2; bottom trace = line current. Speed of operation is 350RPM. .............................. 170
Figure 109 Simulation results of the two-phase machine driven from an H-bridge inverter
under voltage control. Conduction angle is 120 degrees (electrical), no advancing, current
demand is 30A and the dc-link voltage is 175V. Top trace = phase voltages (red for L1,
blue for L2); middle trace = phase currents (red for L1, blue for L2); bottom trace = line
current. Speed of operation is 1000RPM. ........................................................................ 172
Figure 110 Phase current and machine total output torque comparison of simulated and
measured results for h-bridge inverter with Advance Angle = 10Degrees; Conduction
Angle = 115.4Degrees; Speed = 1013.5RPM; Current Demand = 30A; Vdclink = 175V
.......................................................................................................................................... 173
Angle = 127.7Degrees; Speed = 366.2RPM; Current Demand = 30A; Vdclink = 175V 174
Angle = 175.6Degrees; Speed = 1181.7RPM; Current Demand = 30A; Vdclink = 175V
.......................................................................................................................................... 175
xvi
LIST OF FIGURES
Angle = 176.8Degrees; Speed = 387.5RPM; Current Demand = 30A; Vdclink = 175V 176
Figure 114 Asymmetric half-bridge converter driving a two-phase machine ................. 179
measured results for asymmetric half-bridge inverter with Advance Angle = 12Degrees;
Conduction Angle = 175.75Degrees; Speed = 945RPM; Current Demand = 30A; Vdclink
= 175V ............................................................................................................................. 181
Conduction Angle = 114.8Degrees; Speed = 805.4RPM; Current Demand = 30A; Vdclink
= 175V ............................................................................................................................. 182
= 175V ............................................................................................................................. 183
measured results for asymmetric half-bridge inverter with Advance Angle = -20Degrees;
= 175V ............................................................................................................................. 184
Figure 119 Estimated flux-linkage locus for h-bridge inverter driven system with
Advance Angle = 40Degrees; Conduction Angle = 176.8Degrees; Speed = 387.5RPM;
Current Demand = 30A; Vdclink = 175V ....................................................................... 186
Figure 120 Phase and IGBT currents for the h-bridge and asymmetric half-bridge inverter
two-phase drives at 1200 rpm with 175V dc-link and 50A of current demand. .............. 192
Figure 121 Phase and IGBT currents for the h-bridge and asymmetric half-bridge inverter
two-phase drives at 100 rpm with 175V dc-link. ............................................................. 195
Figure 122 Zoomed-in IGBT currents for the h-bridge and asymmetric half-bridge
inverter two-phase drives at 100 rpm with 175V dc-link. ............................................... 195
Chapter-8
Figure 123 Phase voltage waveform with Advance Angle = 12o; Conduction Angle =
119.2o; Speed = 978.8RPM; Current Demand = 10A; Vdclink = 562V ......................... 202
Figure 124 Comparison of simulated and measured phase current and torque with
Advance Angle = 12o; Conduction Angle = 119.2o; Speed = 978.8RPM; Current Demand
= 10A; Vdclink = 562V ................................................................................................... 202
xvii
LIST OF FIGURES
Figure 125 IGBT and recovery diode currents with Advance Angle = 12o; Conduction
Angle = 119.2o; Speed = 978.8RPM; Current Demand = 10A; Vdclink = 562V ........... 203
Figure 126 Phase voltage waveform with Advance Angle = -8.8o; Conduction Angle =
164.6o; Speed = 1047RPM; Current Demand = 10A; Vdclink = 567V .......................... 205
Advance Angle = -8.8o; Conduction Angle = 164.6o; Speed = 1047RPM; Current Demand
= 10A; Vdclink = 567V ................................................................................................... 205
Figure 128 Phase voltage waveform with Advance Angle = 0o; Conduction Angle =
118.6o; Speed = 588.2RPM; Current Demand = 10A; Vdclink = 575V ......................... 207
Advance Angle = 0o; Conduction Angle = 118.6o; Speed = 588.2RPM; Current Demand =
10A; Vdclink = 575V ...................................................................................................... 207
Figure 130 Phase voltage waveform with Advance Angle = -10o; Conduction Angle =
154.4o; Speed = 1109.1RPM; Current Demand = 10A; Vdclink = 567V ....................... 209
Advance Angle = -10o; Conduction Angle = 154.4o; Speed = 1109.1RPM; Current
Demand = 10A; Vdclink = 567V ..................................................................................... 209
Figure 132 IGBT and recovery diode currents with Advance Angle = -10o; Conduction
Angle = 154.4o; Speed = 1109.1RPM; Current Demand = 10A; Vdclink = 567V ......... 210
Figure 133 Simulated phase currents of Test-1, Test-2, Test-3 and Test-4 ..................... 212
Figure 134 Simulated device currents of Test-1, Test-2, Test-3 and Test-4.................... 212
Figure 135 Simulated diode currents of Test-1, Test-2, Test-3 and Test-4 ..................... 213
Figure 136 Measured torque-speed curves of 3-phase single-tooth segmental rotor SRM.
Conduction angle varied from 120 degrees to 180 degrees. Phase current demand was
fixed at 10A and dc-link voltage was 580V..................................................................... 216
Chapter-9
Figure 137 Star-connected six-phase machine driven by a 3-phase full bridge circuit ... 219
circuit ............................................................................................................................... 219
Figure 139 Simulated phase voltage (in black), phase-A current (x5 in blue), phase-D
current (x5 in magenta), line-X current (x5 in green) and filtered phase voltage (in red)
for the 3-phase bridge circuit driven star connected 6-phase conventional rotor SRM with
mutually de-coupled phase windings running at 250rpm. ............................................... 222
xviii
LIST OF FIGURES
Figure 140 Six phase machine is star configuration ........................................................ 223
Figure 141 Simulated line potential (in black) and current (x5 in green) for the 3-phase
bridge circuit driven star connected 6-phase conventional rotor SRM with mutually decoupled phase windings running at 250rpm. ................................................................... 224
Figure 142 Simulated phase-A current (in red), phase-D current (in blue), line-X current
(in green) for the 3-phase bridge circuit driven star connected 6-phase conventional rotor
SRM with mutually de-coupled phase windings running at 250rpm. ............................. 225
Figure 143 Simulated phase-A current (x5 in red), phase-D current (x5 in blue), line-X
current (x5 in green) and star-point potential (in black) for the 3-phase bridge circuit
driven star connected 6-phase conventional rotor SRM with mutually de-coupled phase
windings running at 250rpm. ........................................................................................... 226
Figure 144 Schematic explanation of inverter states for the 3-phase bridge circuit driven
star connected 6-phase conventional rotor SRM: a-) PhA conducting and line current in
positive direction, b-) PhD conducting and line current in negative direction. Conductive
paths in red ....................................................................................................................... 227
Figure 145 Simulated star-point potential (top trace) and star-point current (bottom trace)
mutually de-coupled phase windings running at 250rpm. ............................................... 228
Figure 146 Simulated raw phase, line and star point voltages (top trace) and filtered
phase, line and star point voltages (bottom) for the 3-phase bridge circuit driven star
connected 6-phase conventional rotor SRM with mutually de-coupled phase windings
running at 250rpm. ........................................................................................................... 229
Figure 147 Simulated phase currents (top trace) and line currents (bottom trace) for the 3phase bridge circuit driven star connected 6-phase conventional rotor SRM with mutually
de-coupled phase windings running at 250rpm. .............................................................. 230
Figure 148 Simulated phase torques and resulting total output torque for the 3-phase
bridge circuit driven star connected 6-phase conventional rotor SRM with mutually decoupled phase windings running at 250rpm. ................................................................... 231
mutually de-coupled phase windings running at 1200rpm. ............................................. 232
bridge circuit driven star connected 6-phase conventional rotor SRM with mutually decoupled phase windings running at 1200rpm. ................................................................. 233
(in green) for the 3-phase bridge circuit driven star connected 6-phase conventional rotor
SRM with mutually de-coupled phase windings running at 1200rpm. ........................... 234
xix
LIST OF FIGURES
Figure 152 Simulated raw phase, line and star point voltages (top trace) and filtered
phase, line and star point voltages (bottom) for the 3-phase bridge circuit driven star
connected 6-phase conventional rotor SRM with mutually de-coupled phase windings
running at 1200rpm. ......................................................................................................... 235
Figure 153 Simulated phase-A current (x5 in red), phase-D current (x5 in blue), line-X
current (x5 in green) and star-point potential (in black) for the 3-phase bridge circuit
windings running at 1200rpm. ......................................................................................... 236
Figure 154 Simulated star-point potential (top trace) and star-point current (bottom trace)
mutually de-coupled phase windings running at 1200rpm. ............................................. 237
Figure 155 Simulated phase currents (top trace) and line currents (bottom trace) for the 3phase bridge circuit driven star connected 6-phase conventional rotor SRM with mutually
de-coupled phase windings running at 1200rpm. ............................................................ 238
bridge circuit driven star connected 6-phase conventional rotor SRM with mutually decoupled phase windings running at 1200rpm. ................................................................. 239
Figure 157 Graphical illustration of the line, link and phase currents of the 6-phase delta
connected machine with back-to-back connected phase windings driven from 3-phase full
bridge circuit. ................................................................................................................... 241
for the 3-phase bridge circuit driven delta connected 6-phase conventional rotor SRM
with mutually de-coupled phase windings running at 250rpm. ....................................... 242
bridge circuit driven delta connected 6-phase conventional rotor SRM with mutually decoupled phase windings running at 250rpm. ................................................................... 243
Figure 160 Delta-connected six-phase machine driven from 3-phase full bridge inverter
.......................................................................................................................................... 244
(in green), link-1 current (in magenta) and link-3 current (in black) for the 3-phase bridge
circuit driven delta connected 6-phase conventional rotor SRM with mutually de-coupled
phase windings running at 250rpm .................................................................................. 244
Figure 162 Simulated link currents (top trace) and line currents (bottom trace) for the 3phase bridge circuit driven delta connected 6-phase conventional rotor SRM with
mutually de-coupled phase windings running at 250rpm ................................................ 245
Figure 163 Simulated phase currents (top trace) and line currents (bottom trace) for the 3phase bridge circuit driven delta connected 6-phase conventional rotor SRM with
mutually de-coupled phase windings running at 250rpm ................................................ 246
xx
LIST OF FIGURES
bridge circuit driven delta connected 6-phase conventional rotor SRM with mutually decoupled phase windings running at 250rpm. ................................................................... 247
for the 3-phase bridge circuit driven delta connected 6-phase conventional rotor SRM
with mutually de-coupled phase windings running at 1500rpm. ..................................... 249
bridge circuit driven delta connected 6-phase conventional rotor SRM with mutually decoupled phase windings running at 1500rpm. ................................................................. 250
(in green), link-1 current (in magenta) and link-3 current (in black) for the 3-phase bridge
circuit driven delta connected 6-phase conventional rotor SRM running at 1500rpm. ... 251
Figure 168 Simulated link currents and line currents for the 3-phase bridge circuit driven
delta connected 6-phase conventional rotor SRM running at 1500rpm. ......................... 251
Figure 169 Simulated phase currents (top trace) and line currents (bottom trace) for the 3phase bridge circuit driven delta connected 6-phase conventional rotor SRM running at
1500rpm. .......................................................................................................................... 252
bridge circuit driven delta connected 6-phase conventional rotor SRM running at
1500rpm. .......................................................................................................................... 252
Figure 171 Torque-Speed curves of the prototype segmental rotor star-connected 6-phase
SRM driven from a 3-phase bridge circuit. Top trace is when the line current demand is
15A and bottom trace is when the line current demand is 30A ....................................... 255
Figure 172 Total machine copper loss vs. Total mean output torque curves of the
prototype segmental rotor star-connected 6-phase SRM driven from a 3-phase bridge
circuit. Top trace = 15A current demand and bottom trace = 30A current demand ........ 256
Figure 173 Mechanical losses vs. speed of the prototype 6-phase segmental rotor machine
.......................................................................................................................................... 257
Figure 174 Flux density plot of the 6-phase segmental rotor SRM in the aligned position
.......................................................................................................................................... 259
Figure 175 Slot (Slot1 in Figure 18) Flux vs. MMF characteristics from unaligned
position to aligned position with 1 degree steps up to 3520 ampere-turns. ..................... 260
Figure 176 Torque-Position-MMF characteristics per slot generated from the Flux-MMFPosition characteristics per slot ........................................................................................ 261
xxi
LIST OF FIGURES
Figure 177 Phase currents (top trace) and line currents (bottom trace) with Advance
Angle = -30o; Line Current Demand = 30A; Speed = 335.57RPM; DC-Link Voltage =
175V. ................................................................................................................................ 263
Figure 178 Instantaneous slot torques and total machine output torque (in red) with
Advance Angle = -30o; Line Current Demand = 30A; Speed = 335.57RPM; DC-Link
Voltage = 175V. ............................................................................................................... 264
Figure 179 Line current (top trace), line-to-line voltage (middle trace) and line-to-star
point voltage (bottom trace) with Advance Angle = -30o; Line Current Demand = 30A;
Speed = 335.57RPM; DC-Link Voltage = 175V. ............................................................ 264
175V. ................................................................................................................................ 267
Voltage = 175V. ............................................................................................................... 267
Figure 182 Line current(top trace), line-2-line voltage (middle trace) and line-2-star point
voltage (bottom trace) with Advance Angle = -50o; Line Current Demand = 30A; Speed =
341.67RPM; DC-Link Voltage = 175V. .......................................................................... 268
175V. ................................................................................................................................ 270
Voltage = 175V ................................................................................................................ 270
voltage (bottom trace) with Advance Angle = -40o; Line Current Demand = 15A; Speed =
255.32RPM; DC-Link Voltage = 175V ........................................................................... 271
Figure 186 Flux vs. MMF loci for the low speed measurements .................................... 272
Angle = 0o; Line Current Demand = 30A; Speed = 1271.19RPM; DC-Link Voltage =
175V................................................................................................................................. 273
Advance Angle = 0o; Line Current Demand = 30A; Speed = 1271.19RPM; DC-Link
Voltage = 175V ................................................................................................................ 274
Figure 189 Line current (top trace), line-2-line voltage (middle trace) and line-2-star point
voltage (bottom trace) with Advance Angle = 0o; Line Current Demand = 30A; Speed
=1271.19RPM; DC-Link Voltage = 175V ...................................................................... 274
xxii
LIST OF FIGURES
Figure 190 IGBT currents (top trace) and reverse recovery diode currents (bottom trace)
based on the measured line currents presented in Figure 187 . ....................................... 275
175V. ................................................................................................................................ 276
Voltage = 175V. ............................................................................................................... 277
=1342.28RPM; DC-Link Voltage = 175V. ..................................................................... 277
based on the measured line currents presented in Figure 191. ........................................ 278
175V. ................................................................................................................................ 279
Voltage = 175V. ............................................................................................................... 280
175V. ................................................................................................................................ 282
Voltage = 175V. ............................................................................................................... 283
Figure 201 Line current (top trace), line-2-line voltage (middle trace) and line-2-star point
Figure 203 Flux vs. MMF loci for the high speed measurements ................................... 286
xxiii
LIST OF FIGURES
Chapter-10
Figure 204 6-phase 12-10 SRM phase current waveforms under perfect current control
.......................................................................................................................................... 290
.......................................................................................................................................... 290
Figure 206 MMF vs. flux plots of the 3-phase segmental (in black), 6-phase segmental (in
blue) and 6-phase conventional (in red) rotor machines in the aligned and unaligned
(dotted curves) positions .................................................................................................. 294
Figure 207 2D flux plots of a-) 3-phase segmental rotor b-) 6-phase segmental rotor c-) 6phase conventional rotor machines in the aligned position ............................................. 295
Figure 208 Assumed phase current pattern for the torque ripple comparison study ....... 297
Figure 209 Graphical explanation of the total conduction period for the slot MMF....... 298
Figure 210 Static torque characteristics of the 6-phase 12-10 segmental rotor machine
from 0At to 3520At .......................................................................................................... 299
Figure 211 Static torque characteristics of the 6-phase 12-10 conventional rotor machine
from 0A to 45A ................................................................................................................ 300
Figure 212 Static torque characteristics of the 3-phase 12-10 single-tooth winding
segmental rotor machine from 0A to 28A ....................................................................... 300
Figure 213 Simple flow-chart explanation of the Matlab script analysing the batch-run
results of the perfect current control simulation. ............................................................. 301
Figure 214 Torque ripple vs. average output torque characteristics for the 6-phase
conventional toothed rotor drive, 6-phase segmental rotor drive and 3-phase segmental
rotor drive (with three different conduction angles). ....................................................... 302
Figure 215 Current demand vs. average output torque characteristics for the 6-phase
Figure 216 Advance angle vs. average output torque characteristics for the 6-phase
Figure 217 a-) Instantaneous slot torque (top trace) and slot MMF (bottom trace)
waveforms for the 6-phase segmental rotor drive with ~47% torque ripple. .................. 305
Figure 218 Torque ripple vs. average output torque characteristics for the 6-phase
conventional toothed rotor drive. Blue trace when conduction angle = 120o and red trace
when conduction angle = 150o. ........................................................................................ 306
xxiv
LIST OF FIGURES
Figure 219 Current demand vs. average output torque characteristics for the 6-phase
Figure 220 Advance angle vs. average output torque characteristics for the 6-phase
Figure 221 Phase current, phase torque and rotor position waveforms for drives tabulated
in Table 42 ....................................................................................................................... 315
Figure 222 Phase torques and total torque waveforms for drives tabulated in Table 42 . 315
Figure 223 Phase current, phase torque and rotor position waveforms for drives tabulated
in Table 43 ....................................................................................................................... 316
Figure 224 Phase torques and total torque waveforms for drives tabulated in Table 43 . 316
Appendix A
Figure A. 1 Screenshot of the GUI used to post-process the voltage and current
measurements and generate flux-linkage characteristics of the prototype machine ........ 331
Figure A. 2 Effects of undesirable voltage and current waveforms on the calculated fluxlinkage characteristics ...................................................................................................... 332
Appendix B
Figure B. 1 Stator segment lamination of the segmental rotor six-phase switched
reluctance machine........................................................................................................... 334
Figure B. 2 Rotor segment lamination of the segmental rotor six-phase switched
reluctance machine........................................................................................................... 335
Figure B. 3 Stator segment lamination of the toothed rotor six-phase switched reluctance
machine ............................................................................................................................ 336
Figure B. 4 Rotor lamination of the toothed rotor six-phase switched reluctance machine
.......................................................................................................................................... 337
xxv
LIST OF FIGURES
Appendix C
Figure C. 1 A screenshot of GUIDE in Matlab environment for creating GUIs .......... 339
Figure C. 2 SRMSim Database Main GUI Window Initial State .................................... 341
Figure C. 3 SRMSim Database Main GUI Window File Menu List ........................... 342
Figure C. 4 SRMSim Database Main GUI Window Once Motor Data Loaded.............. 343
Figure C. 5 SRMSim Database Main GUI Window Once Simulink Model Loaded ... 344
Figure C. 6 SIM RESULTS GUI Window (Left-Hand Side = Statistics Calculation
Results; First Figure = Filtered Phase Voltage; Second Figure = Phase Current; Third
Figure = Total Machine Torque)...................................................................................... 345
Figure C. 7 FlxLkgLocus GUI Window depicting Phase Flux-Linkage vs. Phase Current
.......................................................................................................................................... 346
Figure C. 8 STAR POINT VOLTAGE GUI Window..................................................... 347
xxvi
ACKNOWLEDGEMENTS
List of Tables
Chapter-3
Table 1 Summary of the useful switching states and potentials at points U, V, and W for
each operating condition for the 6-phase star and delta connected drives. ........................ 62
Chapter-4
Table 2 List of dimensions of 6-phase segmented design that are kept the same as those of
the 3-phase single-tooth segmented design of Mecrow et al ............................................. 65
Table 3 Full list of design parameters of the prototype machine ....................................... 73
Table 4 Stator tooth tip and rotor segment arc angles for chosen design possibilities for
the 6-phase segmented rotor SRM for further analysis ..................................................... 76
Table 5 Stator tooth width and overlap and gap angles for each design option ................ 77
Table 6 Summary of the performance comparison of the six design options (results
normalised with respect to the prototype design) .............................................................. 80
Table 7 Comparison of OptDesign parameters for each design (results are normalised
with respect to Design Option-3) ....................................................................................... 82
Table 8 Critical design parameters for the conventional toothed rotor 6-phase SRM....... 84
Table 9 Machine dimensions for the 6-phase SRM with conventional toothed rotor ....... 85
Chapter-6
Table 10 All possible switching states of the transistors in the H-Bridge inverter of the 2phase SRM model and operating conditions depending on the states of the power switches
.......................................................................................................................................... 145
Chapter-7
Table 11 Comparison of results obtained from direct measurements, predictions from the
Matlab script and the simulation for the h-bridge driven two-phase machine ............. 177
xxvii
ACKNOWLEDGEMENTS
Table 12 Script calculated performance parameters for the 2-phase 12-10 segmental rotor
machine driven from the asymmetric half-bridge inverter .............................................. 178
Table 13 Comparison of results obtained from direct measurements, predictions from the
Matlab script and the simulation for the asymmetric half bridge driven two-phase
machine ............................................................................................................................ 180
Table 14 Script calculated performance parameters for the 2-phase 12-10 segmental rotor
machine driven from the h-bridge inverter ...................................................................... 185
Table 15 Summary of comparison of the performances of h-bridge and asymmetric halfbridge driven two-phase machine drive systems under voltage control .......................... 191
Table 16 Explanation of abbreviations used in Table 15 and Table 17 ........................... 192
Table 17 Summary of comparison of the performances of h-bridge and asymmetric halfbridge driven two-phase machine drive systems under current control .......................... 194
Chapter-8
Table 18 Tabulated summary of test conditions for the measurements presented in the
proceeding sections .......................................................................................................... 200
Table 19 Summary of measured, script and simulation determined results for Test-1
where Advance Angle = 12o; Conduction Angle = 119.2o; Speed = 978.8RPM; Current
where Advance Angle = -8.8o; Conduction Angle = 164.6o; Speed = 1047RPM; Current
where Advance Angle = 0o; Conduction Angle = 118.6o; Speed = 588.2RPM; Current
where Advance Angle = -10o; Conduction Angle = 154.4o; Speed = 1109.1RPM; Current
Chapter-9
Table 23 Summary of simulated performance parameters for the 3-phase bridge circuit
xxviii
ACKNOWLEDGEMENTS
driven delta connected 6-phase conventional rotor SRM with mutually de-coupled phase
driven delta connected 6-phase conventional rotor SRM with mutually de-coupled phase
Table 27 Summary of measured and predicted performance parameters for the starconnected 6-phase segmental rotor SRM when driven from 3-phase bridge circuit with
Voltage = 175V. ............................................................................................................... 266
Voltage = 175V. ............................................................................................................... 269
Voltage = 175V. ............................................................................................................... 271
Voltage = 175V. ............................................................................................................... 275
Table 31 Summary of estimated power electronic performance parameters based on
measured line currents with Advance Angle = 0o; Line Current Demand = 30A; Speed =
1271.19RPM; DC-Link Voltage = 175V. ........................................................................ 276
Voltage = 175V. ............................................................................................................... 278
Voltage = 175V. ............................................................................................................... 281
xxix
ACKNOWLEDGEMENTS
Table 36 Summary of measured and predicted performance parameters for the starconnected 6- phase segmental rotor SRM when driven from 3-phase bridge circuit with
Voltage = 175V. ............................................................................................................... 284
measured line currents with Advance Angle = 0o; Line Current Demand = 15A; Speed
Chapter-10
Table 38 Summary of torque per copper loss as per machine design for the 6-phase and 3phase segmental and 6-phase conventional machines ..................................................... 291
Table 39 Comparison of torque per copper loss of 6-phase and 3-phase segmental and 6phase conventional machines for fixed number of turns per slot and slot fill factor ....... 292
Table 40 Number of turns, number of active coils and stator tooth widths of all three
machines .......................................................................................................................... 292
Table 41 Table of the areas covered by the magnetisation curves of all three machines 292
Table 42 Summary of dynamic simulation results with the delta and star connected sixphase conventional rotor SRM driven from the 3-phase full bridge inverter and the threephase single-tooth winding segmental rotor SRM driven from the asymmetric half-bridge
inverter at 1200 rpm generating approximately 20Nm average output torque ................ 310
Table 43 Summary of dynamic simulation results with the delta and star connected sixphase conventional rotor SRM driven from the 3-phase full bridge inverter and the threephase single-tooth winding segmental rotor SRM driven from the asymmetric half-bridge
inverter at 100 rpm generating approximately 20Nm average output torque .................. 312
Appendix D
Table D. 1 Table of machine design variables ................................................................. 350
xxx
ACKNOWLEDGEMENTS
Acknowledgements
Just to be writing the Acknowledgements of this thesis means that it is the end of a very
long era of my life which is almost surreal. At times, it has been very difficult. Without
the help, support and encouragement of many people, it could not have happened.
First of all, my biggest thanks go to my supervisor Professor Barrie Mecrow and my dear
wife Azime. I feel very lucky to have been the student of Barrie to be able to have direct
access to his expertise and experience in all aspects of electric machines and drives. His
support has always been more than technical throughout the years (and I mean, a lot of
years!). Thank you - Barrie, for believing in me when things looked gloomy. I felt the
luckiest man when I met Azime and I still do more than ever. She has sacrificed many
things for this PhD. It has been tough on her within the last few months. I cannot thank
her enough; she is the most understanding, supportive and loving person in my life. Dont
worry dear; the balance will revert back to 90-10 again.
Another short but big thanks go to my one year old boy Tymur for giving me the biggest
incentive to finish writing-up my PhD after a very long time. Of course, many thanks go
to my family for their support throughout my student life and always accepting what I do
and never doubting me in my student life.
This PhD also involved a lot of building and stripping of test rigs which would not be
possible without the help of the technicians at the mechanical workshop. Thanks to all the
academic staff one way or another involved in my PhD during my time in Newcastle.
I am especially grateful to my proof-reader, Steve Greetham for taking the time to read this
thesis despite his own busy life and for providing invaluable feedback. Special thanks to
Mark Leaver and Yu Chen for continued talks, support and encouragement they gave me
at Dyson Ltd.
Finally, big thanks to all the nice people of Geordie-land for making me feel like home
and making my time in Newcastle so enjoyable.
xxxi
LIST OF SYMBOLS
List of Symbols
Symbols
1, 2, 3
Subscripts referring to slot quantities
A, B, C
Subscripts referring to tooth quantities
Air-gap
Idem
Demand current
Resistance
Torque
Tesla unit of flux density
Voltage
Ampere
Vstar or Vstar
Star point potential
VDC or Vdclink
DC Link Voltage
12/10 or 12-10
Denotes a machine with 12 stator and 10 rotor poles
12/8 or 12-8
4/2 or 4-2
6/4 or 6-4
8/6 or 8-6
Istar or Istar
Current flowing into the star point
t/
Tooth-width to pole pitch ratio
Flux-linkage
Stator pole pitch
Rotor pole pitch
Nm
Newton-meter
MW
Mega Watt
kW
Kilo Watt
Watt
RPM / rpm
Revolutions per Minute
kg
Kilogram
xxxii
LIST OF SYMBOLS
max
Maximum
min
Minimum
Resistivity
Pshaft / Psh
Shaft power
Efficiency
Permeability of air
Magnetic flux density
Tmax
Maximum torque
Tmin
Minimum torque
Tave or Taverage
Average torque
Tripple
Torque ripple in [%]
Lmax
Maximum inductance (aligned position)
Lmin
Minimum inductance (unaligned position)
Tooth-width
Lg
Air-gap length
Vref
The voltage reference value defining the required duty

cycle to be applied to the phase windings
MagNet
The name of the finite element analysis software package

from Infolytica Corporation
Nr
Number of rotor poles
Ns
Number of stator poles
Nph
Number of machine phases
Stator tooth width
Mk
Mark
PFe or Pfe-specific
Total specific iron loss
Pfe-machine
Total machine iron loss
Frequency
Bmax
Maximum flux density
Steinmetz constant
Kh
Hysteresis loss coefficient
Ke
Eddy loss coefficient
Mmachine
Total machine mass
OptDesign
Number that is used to quantify how good a machine design

is
xxxiii
LIST OF SYMBOLS
ToutRatio
Percentage comparison of the design options in terms of

average output torque
ToutPcuRatio

average output torque per machine copper loss
ToutMassRatio

average output torque per total machine mass
IndRatio

unaligned linear inductance
Magnetic flux
Number of turns
v(t )
Instantaneous terminal voltage
i (t )
Instantaneous phase current
(t )
Instantaneous flux-linkage
Initial flux-linkage
uF
Micro-farads
Ohms
CBDS
Stator Core-Back Depth
TWS
Stator Tooth Width
Stator Pole Pitch
Stator Bore Diameter
R1
Rotor Outside Diameter
R2
Rotor Inside Diameter
shaft
Shaft Diameter
CBDR
Rotor Core-Back Depth
TWR
Rotor Tooth Width
Lg
Air-gap Length
OUT
Machine Outside Diameter
hS
Stator Tooth Height
hR
Rotor Tooth Height
Rphase
Phase resistance
Mean of the total length of a coil winding
Bare conductor diameter with no insulation layer
Conductivity of copper
xxxiv
LIST OF SYMBOLS
W
Co-energy
Posrotor
Rotor position with respect to a phase
Vtri
PWM carrier signal
I_Line
Line current
IPhA or IphaseA or Ia
Current flowing in phase A
IPhB or IphaseB or Ib
Current flowing in phase B
IPhC or IphaseC or Id
Current flowing in phase C
IPhD or IphaseD or Id
Current flowing in phase D
IPhE or IphaseE or Ie
Current flowing in phase E
IPhF or IphaseF or If
Current flowing in phase F
Eon
Energy per switch at turn-on
Eoff
Energy per switch at turn-off
Iu
Current flowing in Line U
Iv
Current flowing in Line V
Iw
Current flowing in Line W
I_LineX
Current flowing in Line X
I_LineY
Current flowing in Line Y
I_LineZ
Current flowing in Line Z
di/dt
Rate of change of current
Dcond calculated
Calculated total conductor diameter
sff
Slot Fill Factor
Aslot
Total available winding area
N turns
Number of turns per slot
Pcopper or Pcu
Copper losses
xxxv
LIST OF SYMBOLS
Abbreviations
2D
Two-dimensional
3D
Three-dimensional
PI
Proportional-Integral
PID
Proportional-Integral-Derivative
DC / dc
Direct current
AC / ac
Alternating Current
BDCM
Brushless Direct Current Machine
CSA
Cross-Sectional Area
PM
Permanent Magnet
BLDC
Brushless Direct Current
DSP
Digital Signal Processor
BEMF
Back Electromotive Force
PCB
Printed Circuit Board
EMF
Electro Motive Force
FE
Finite Element
FEA
Finite Element Analysis
FPGA
Field Programmable Gate Array
IGBT
Insulated Gate Bipolar Transistor
MMF
Magneto Motive Force
MPL
Mean Path Length
MOSFET
Metal Oxide Field Effect Transistor
SCR
Silicon Controlled Rectifier
PID
Proportional,
Integral,
Derivative
parameters)
PWM
Pulse Width Modulations
SP
Short Pitched
FP
Fully Pitched
SR
Switched Reluctance
SRM
Switched Reluctance Machine
VA
Volt-Ampere
kVA
Kilo Volt-Ampere
xxxvi
(current
controller
LIST OF SYMBOLS
TW
Tooth-Width
GUI
Graphical User Interface
EU
European Union
USA
United States of America
US
United States
UK
United Kingdom
MagFine
Name derived from the combination of 'MAGnet' and the

'FINE' structure of d-HDDR magnet powder
VAT
Value Added Tax
NdFeB
Neodymium-Iron-Boron
InD
Rotor outer diameter
OutD
Stator outer diameter
Hz
Hertz
MHz
Mega Hertz
kHz
Kilo Hertz
CBD
Core-Back Depth
ROD
Rotor Outer Diameter
SOD
Stator Outer Diameter
PhA
Phase A
PhB
Phase B
PhC
Phase C
PhD
Phase D
PhE
Phase E
PhF
Phase F
FRFS
Flux-Reversal-Free-Stator
RMS or rms
Root Mean Square
A2D
Analogue to Digital
EMI
Electromagnetic Interference
PEDM
Power Electronics Drives and Machines
mm
Milimeter
Meter
SD
Segment Depth
VB
Visual Basic
PC
Personal Computer
xxxvii
LIST OF SYMBOLS
SMC
Soft Magnetic Composite
DAC
Digital to Analog Conversion
LED
Light Emitting Diode
ADC
Analogue to Digital Conversion
MEC
Magnetic Equivalent Circuit
CPU
Centra Processing Unit
At
Ampere-turn
xxxviii
Chapter 1 INTRODUCTION
CHAPTER 1
1 - Introduction
1.1 INTRODUCTION
With
advancements
in
power
electronic
devices,
microcontroller
technology,
manufacturing processes and materials science, the use of electric motors in every sector
from the aerospace industry to domestic appliances has been rapidly increasing. Electric
motors now account for more than 60% of total energy consumption in industry within
the EU and the USA. Consequently the design requirements of electric motors have
become more demanding than ever. The main driving factors for the design requirements
are cost, efficiency, size and torque density. There are more application specific
requirements such as torque ripple, wide operating speed range, thermal performance,
fault tolerance and acoustic noise that also need consideration and continuous
development.
At the start of this PhD, previous academic study of segmental rotor SRMs was conducted
in a PhD at Newcastle University and resulted in two 3-phase machines, one with multitooth phase windings, spanning more than one stator tooth, and the other with single-tooth
phase windings. Fundamental operating principles were investigated on a single phase
rectilinear model and design rules were established. As a result two 3-phase prototype
machines were built and tested. Both of these machines were very successful and resulted
in approximately 40% more output torque compared to conventional SRMs.
The author was involved in the dynamic measurements taken from these prototype
machines at the beginning of this PhD project.
1.2 SWITCHED RELUCTANCE MACHINES IN

INDUSTRY
Switched reluctance drives are no novelty and have been around since the 1800s.
However, until the 1960s the optimum operation of an SR drive was somewhat
impossible as control of SR motors was heavily dependant on sophisticated switching
control. Although it is yet to have a significant share in the marketplace, with
advancements in power electronics and materials technology the switched reluctance
machine has entered into the electric machines and motor drives market. Compared to
other conventional drive solutions, such as ac and permanent magnet based drives, SR
drives have a small share in domestic and industrial applications. Well known issues of
high torque ripple, high acoustic noise and vibration, cost and (potential) complexity of
the power converter prohibit the SRMs greater market penetration. In simple terms,
SRMs are still not perceived as good value for money. On the other hand, advantages
such as simple and low cost machine construction due to the absence of rotor windings,
rotor cage or permanent magnets, low rotor inertia, high torque to inertia ratio and high
starting torque without excessive inrush current make the SRM an attractive alternative
for several applications.
Academic research continues to drive SRM technology forwards (see for example the
works of Prof. R. Krishnan in Virginia, Prof. B. C. Mecrow in Newcastle, Prof. T. J. E.
Miller in Glasgow, Prof. D. A. Torrey in Rensselaer, Prof. M. Ehsani in Texas, Prof. I.
Husain in Akron, Prof. C.Pollock in Leicester). However, this is not reflected in the
industry. In fact, the use of SRMs in industry has been in decline in recent years. This is
partly because of the fact that the same technological advancements making SRMs
attractive also allow other machine technologies to develop further and maintain or even
increase their market dominance. For example, recent years have seen an exponential
increase in the development of permanent magnets that are both mechanically and
magnetically stronger (e.g. MagFine compression and injection molded anisotropic
bonded magnets from Aichi Steel Corporation) than their earlier counterparts. These new
permanent magnets also tolerate operating temperatures as high as 150oC without
significant flux loss (typically 4-5%). Moreover, as the number of magnet manufacturers
increases, the cost of permanent magnets has reduced. Because of their high torque
density, permanent magnet machines have started to dominate some sectors of the
2
industry. They are the primary candidate for high efficiency / small size domestic
appliances (see for example the new generation Dyson Digital Motor used in handheld
products DC30 and DC31) and have a significant share in the wind turbine generators
market (V112-3.0 MW of Vestas Wind Systems, GEs 2.5MW wind turbines, G128-4.5
MW of Gamesa, etc.). However, a recent rise in rare earth material cost, along with
concerns over security of supply has led to debate over whether permanent magnet
machines will continue to dominate.
The AC induction machine is still the preferred option for many industrial applications.
However, it is well known to have problems with heat extraction from the rotor, as the
rotor also carries windings and current. There is also concern over the induction
machines efficiency over a wide speed and power range.
Until the performance issues of SRMs are resolved in an economical fashion this motor
drive technology will struggle to have a deeper penetration and a larger market share of
either industrial or domestic applications. Significant improvements are required in the
field of converter topologies, control algorithms, torque ripple, acoustic noise reduction
and sensorless operation.
1.2.1 Target Industries for SR Motor Drive Technology

The author believes that there are four target sectors in order to achieve greater
penetration for the SR motor drive technology. These can be summarised as follows:
Energy:
The level of interest in the use of SR machine technology in the energy sector has risen
over the years but activity has mainly remained at research level in academia [1-4]. Wind
power is growing at an annual rate of 30%, with a worldwide installed capacity of
157,900 MW in 2009, and is widely used in Europe, Asia, and the U.S.A. During the fiveyears from the end of 2004 through to 2009, worldwide renewable energy capacity grew
at rates of 10 to 60 percent annually for many technologies. For wind power and many
other renewable technologies, growth accelerated in 2009 relative to the previous four
years; more wind power capacity was added during 2009 than any other renewable
technology. Nowadays, 19% of the electricity used in Denmark is supplied by wind
3
power generators. This is 9% in Spain and Portugal, and 6% in Germany and the Republic
of Ireland [5]. The United States is an important growth area and installed U.S. wind
power capacity reached 25,170 MW at the end of 2008 [6]. Unfortunately when it comes
to machine drive technologies employed in the wind power industry, the top 10 wind
turbine generator manufacturers opt for either permanent magnet synchronous machine or
squirrel cage / wound rotor induction machine technologies. However, the potential of the
SRM for this type of application is high [1]: the SR generator has been considered for
wind power applications for a long time [2].
Much of the research on SR generators has so far focussed on aircraft [7-11]. Greater
penetration of SR machine technology into the energy sector can be achieved with further
research aiming to improve the behaviour of SRMs in generator mode, specifically in
wind power applications. However, it must be said that unless significant improvements
are to be made, penetration into the wind turbine generator market is quite difficult. There
are two mains reasons to this: first of all, new generator concepts of MW scale are very
expensive to prototype and secondly, new production lines can require enormous
investment, prohibiting companies investing in SRM technology.
Domestic Appliances:
Domestic energy consumption has been increasing with continuously growing rates since
1990s.
As the number of households, population and household disposable income
increase, the demand for energy by households also amplifies. Furthermore, as the
number of appliances owned per household increases and more functionality is demanded
from new products, the energy consumption by the domestic appliances and its impact on
our environment has become very important. There are new EU legislations imposing
strict energy limits to some of the household appliances and there will be a greater need
for more efficient and compact electric motor drives in the domestic appliances sector.
From the companies point of view, these new products will also need to be financially
feasible.
SRMs offer much greater efficiency than brushed motors: if magnet costs continue to rise
they may start to displace permanent magnet brushless products.
Automotive:
There are now over one billion cars on our planet. There is an undeniable reasoning to
develop vehicles that are clean, efficient and sustainable. This is driven by environmental,
as well as economical, issues. For an era when reducing carbon emissions and controlling
global warming is one of the biggest driving factors, the exhaust emissions of
conventional internal combustion engine vehicles are deemed to be unacceptable.
Moreover, as the oil reserves of the planet diminish, alternative ways to run these vehicles
are paramount for the automotive industry. High efficiency electric motor drive
technologies provide the means for a clean, efficient and environmentally friendly urban
transportation system.
There are incentives given by governments to promote the use of electric vehicles such
as: reduced VAT and other taxes, insurance facilities, parking and battery charging
facilities (including free recharging on street or in the parking areas), toll free travel on
motorways and exemption from congestion charging.
SRMs are much more amenable to field weakening than permanent magnet motors: they
offer good competition to induction motors and may be superior to permanent magnet
motors in the extended constant power range.
Aerospace:
Once again the sector and the motor drive technologys link to the sector are both driven
by the global aim of minimising carbon emissions. SRMs have a big advantage because
they are inherently fault tolerant, i.e. the drive will continue to work even when one or
more of the phases of the machine fail. The aircraft industry is moving towards an
increased use of electrical auxiliaries, particularly to replace hydraulic and mechanically
powered systems. As well as very high performance, the industry requires extremely
reliable and fault tolerant systems. Albeit continued interest, it is still early days for the
more electric aircraft and continuous commitment from the industry and research
institutes is paramount to the increased use of electrically driven systems in the aerospace
industry.
1.2.2 Examples of SR Drives in Industry

There are only a few companies specialising in SR motor and drive system development.
The foremost example is SR Drives in Harrogate, UK. SR Drives was a start-up company
founded by the pioneering works of Prof. Lawrenson and his colleagues at Leeds
University. Many of the specific industrial examples of SR machines and drives given
within the next few pages are developed by SR Drives.
5
Rocky Mountain Technologies Inc. in Montana, USA was incorporated in 2000 with a
strong focus on SR motors and drives. The company produces series of three and four
phase SR motor drives ranging from 0.4kW to 25kW for motion and alternative energy
applications.
VS Technology Corporation (VST) in Texas, U.S.A develops switched reluctance drives
and drive-trains for small to large applications including vehicles. VST has been working
in the automotive field since 1990 and in 2007 acquired all of the advanced motor drives
technology from Honeywell Corporation, expanding its reach. VST has a range of vehicle
traction motors from 20kW to 200kW and wind turbine generators from 50kW to
Megawatts. They also develop high speed motors up to 100,000 rpm.
Technelec Ltd. was established in 1992 by Professor Charles Pollock, aiming to exploit
his work in Warwick and Leicester Universities. The company specialises in design and
development of electronically controlled switched reluctance machines, flux switching
machines and hybrid stepping motors.
Specific examples of SR machines and drives in the market place are summarised below.
1. DOMESTIC & COMMERCIAL APPLIANCES:
Dyson DigitalMotor:
Used in Japanese Vacuum Cleaners (DC12) and Airblade (AB01 and AB03). It is a high
efficiency (more than 80%) 1600W motor, capable of spinning at 100000 rpm with a
lifetime of 1000 hours in constant use. The drive has built-in electronics resulting in
compact size.
AMETEK Blower Motors:
Blower motors from the Infin-A-Tek series of AMETEK Lamb Electric operating from
24V, 26V and 48V DC, and 120V and 230V AC use SR drive technology. The use of SR
drive in this application is claimed to have increased the service life three to five times
longer than designs commonly utilising series universal motors.
Beckman Instruments Inc:
SR drive technology is used to power Beckmans AVANTI J & JE series of high-speed
centrifuge systems. The system is claimed to be a major improvement on conventional
high-speed centrifuges, providing faster acceleration, faster deceleration and higher gforces.
2. TRANSPORTATION AND AUTOMOTIVE APPLICATIONS:

Green Propulsion:
In Green Propulsions (a Belgian company specialising in developing cleaner prototype
vehicles) Series and Parallel hybrid transmission scheme, two switched reluctance
motor-generators were utilised for a hybrid power-train; a 50kW and a 160kW SR drive
both working in conjunction with a 2.5 litre diesel engine.
Renault:
Renaults TIVA system (Transmission Infiniment Variable) uses two switched reluctance
drives linked via a two-stage epicyclic gearbox. This is a hybrid vehicle with the main
power source being a diesel combustion engine.
Caterpillar:
The US Department of Energy has provided funding to Caterpillar to investigate the
benefits of increased electrical auxiliaries for highway trucks. Electrical auxiliaries
include the oil pump, the water pump, the air compressor for the vehicle brake system and
an electric HVAC (Heating Ventilation & Air Conditioning) system. The heart of the
system is an SRM based starter/generator unit (340V, 1200Nm, 30kW) which is mounted
on the crankshaft between the engine and gearbox. The intention of the starter/generator
is both to provide a direct-drive functionality to start the engine and also to generate the
required electrical power for the various auxiliaries on board.
Deutsche Bahn AG:

German railway company Deutsche Bahn AG uses switched reluctance motor technology
in the heart of the air conditioning system for its ICE 3 high speed passenger trains. At
maximum cooling, the motor power is 40kW operating at a speed of 23,000rpm. The unit
weighs less than a conventional air conditioning unit, which allows it to be entirely roof
mounted.
3. INDUSTRIAL APPLICATIONS:
CompAir Ltd:
CompAir Ltd. specialises in compressed air equipment. Designed by SR Drives Ltd., SR
motor drives are used in its LSR and D-HSR series of variable-speed rotary-screw
compressors and are said to achieve high efficiency.
Drax Power Ltd:

Drax Power Ltd. operates the UKs largest coal-fired power station. As part of a 1
million investment in plant modifications, Drax Power Ltd has installed a fuel blending
plant which allows a variety of alternative fuels to be accurately blended with the main
coal feed. This was achieved with five SR Drive systems, ranging from 40kW to 160kW.
Two of these were new installations and three replaced fixed speed induction motors.
Picanol:
Based in Ieper, Belgium, Picanol is recognized as the worlds leading company in the
design and production of weaving machines. Picanol has been using switched reluctance
drive technology (named the Sumo motor) since 2000 with its OMNIplus range.
LeTourneau Inc:
Based in Longview, Texas, LeTourneau Inc. employs SR Drive technology in its latest
50 Series digital loaders. The L-1350 electric-wheel loader is the first machine of its
type to be fitted with an SR drive system that provides independent traction for each
wheel. Four 300kW SR motors were used. Relative to the original brushed DC motors, a
major increase in service and inspection life has been achieved (from around 500 hours to
20,000 hours) due to the absence of brushes and a commutator.
1.3 SWITCHED RELUCTANCE MACHINES WITH

SEGMENTAL ROTORS
Conventionally both the stator and the rotor of an SRM are made up of toothed steel
laminations. However, the toothed rotor structure is known to suffer from poor utilisation
of the machines magnetic circuit. A measure of the magnetic utilisation of the machine is
often indicated by the tooth width to pole pitch ratio (t/). t/ gives an idea of how much
of an excited rotor pole carries magnetic flux. With conventional toothed rotor structures
this ratio is optimally around 0.42. Once the ratio is over 0.5 there is always an overlap
between the rotor and the stator teeth which results in increased unaligned inductance
[12].
Mecrow et al in 2002 [12] suggested that the segmental rotor arrangement was very well
suited to the concentrated stator MMFs of SRMs. Unlike the conventional toothed
structure, the rotor of a segmental SRM is constructed from a series of discrete segments.
A segmental design does not suffer like the toothed design when t/ is increased beyond
0.5. In fact t/ values of beyond 0.7 are easily achievable without compromising the
unaligned inductance. With the segmental rotor design magnetic utilisation of the
machine is increased, i.e. the proportion of the air carrying flux is greatly increased.
Although segmental rotor structures are similar to that of segmental rotor synchronous
reluctance motors there are three major differences:
1. The phase windings of the segmental rotor SRM are concentrated.

2. There is negligible magnetic coupling between the phases of the segmental rotor
SRM.
3. The segmental rotor SRM is excited with direct current pulses, rather than sinusoidal
AC.
With the conventional toothed rotor SRM the stator teeth permeances are regulated as the
rotor rotates cyclically. In the case of the segmental rotor SRM, torque is generated as the
rotor segments modulate slot permeances. Slot permeance is minimised when the
segments are centred under a stator tooth (unaligned position) and maximised when the
segments bridge a slot opening (aligned position).
The rotor structure of a segmental SRM only permits neighbouring teeth to be
magnetically linked, which requires the magnetic flux to enclose a single stator slot. This
was achieved with two different winding arrangements:
1.3.1 Multi-tooth Winding Segmental Rotor SRM

Each coil spans more than one stator tooth (as many teeth as the number of phases).
Rectilinear representation of a 3-phase multi-tooth winding segmental rotor SRM in the
aligned position is depicted in Figure 1. Use of coils spanning more than one stator teeth
result in substantial end-winding lengths. This reduces the electric loading, which makes
the design impractical for applications that combine short lamination stack length with a
large pole pitch. Nevertheless, 40% more torque within a given frame size was achieved
compared to the conventional toothed rotor SRM [12].
Figure 1 Rectilinear representation of a 3-phase multi-tooth winding segmental rotor SRM in the
aligned position
1.3.2 Single-tooth Winding Segmental Rotor SRM

Slots S2 and S3 and tooth T3 in Figure 1 were removed to have a segmental rotor design
where any one slot is still occupied by windings of a single phase. Each coil of the stator
is now wound around a single tooth which achieves short and non-overlapping endwindings. With the single-tooth winding arrangement only every other stator tooth
accommodates phase windings. The wound teeth span a complete rotor pole. The
rectilinear representation of a 3-phase single-tooth winding segmental rotor SRM in the
aligned position is illustrated in Figure 2. The width of the stator teeth carrying phase
windings is doubled compared to the stator teeth with no phase windings. The stator teeth
with no windings act as the return path for the flux generated by the slot MMFs. The
10
single-tooth winding arrangement has a poorer magnetic utilisation compared to the
multi-tooth winding arrangement. However the advantage of the single-tooth winding
design is the shortened end-windings. Overall, the single-tooth winding design has similar
performance to the multi-tooth winding design but requires less copper windings.
Figure 2 Rectilinear representation of a 3-phase single-tooth winding segmental rotor SRM in the
aligned position
As the rotor depth is shallow, another advantage of the segmental design is that it can be
suitable for applications where machines with large shafts are required.
1.4 CONVERTERS AND CONTROLLERS FOR

SWITCHED RELUCTANCE MACHINES
The performance and cost of an SR drive is highly affected by the chosen converter
topology and the control method. It is well known that the direction of current has no
effect on torque production in a switched reluctance machine. This means that an SR
drive can operate with unipolar phase currents which, along with phase independence
opens up the possibility of using single switch per phase converters. As a result many
different converter topologies are currently available to drive switched reluctance
machines. The converter topologies usually differ in the way they handle the stored
magnetic energy at the end of each stroke. Some of the topologies return the stored
magnetic energy back to the supply, some use this energy to charge a capacitor to aid with
the turn-on and turn-off of phase windings, some use it to energise a closely coupled
second winding and some simply dissipate this energy across resistors. Each converter
topology has its own advantages and disadvantages. It is best to match the converter to
11
the specific application requirements. There is always a trade-off between the
performance, cost and the complexity of the converter circuit and the controller.
For example an asymmetric half-bridge converter has the maximum control flexibility (it
is able to control phases separately and provide positive, zero and negative voltage loops)
but requires two devices per phase, increasing the size and cost of the drive. A bifilar
winding converter requires only a single switch per phase but the ratings of the devices
are high; the C-dump converter offers fast demagnetisation of phases during commutation
but requires complicated control. Basic requirements of a converter topology can be
summarised as follows:
Provide positive voltage to increase the flux in a phase.

Have the ability to control the flux when the required level of current is reached.
Provide negative voltage to reduce the phase current to zero.
The Volt-Ampere (VA) rating of the converter is usually used as a measure to indicate
how effectively the ratings of the power switches have been utilised and, hence, the
relative cost of the converter circuit. This is clearly a very crude model: for example it
incorrectly suggests that a converter with fewer switches does not necessarily reduce the
cost of the power electronics if the switches require higher current, voltage and / or power
dissipation ratings.
Use of the 3-phase bridge circuit with SR drives has already been demonstrated by
Clothier et al [13] with a 3-phase fully-pitched winding conventional toothed rotor SRM.
When driven from a 3-phase full-bridge circuit the machine can be connected in star
(Figure 3) or in delta (Figure 4) configuration. There are two phases conducting at any
given time with a 3-phase full-bridge circuit driven machine. There are distinct
advantages to using the 3-phase full-bridge circuit, which can be summarised as:
The 3-phase full-bridge circuit is one of the most commonly known and readily
available topologies. The volume of production of 3-phase full bridge converters
dwarfs any other topology available for SR drives. They usually come with the
current transducers and the controller electronics.
With the 3-phase full bridge circuit, devices with internally packaged diodes can be
used.
12
The line currents, rather than the phase currents, can be controlled. As a result, the
number of current transducers can be reduced to two for a three phase machine (based
on the fact that the line currents should sum to zero).
The number of connections between the motor and the converter are reduced to three.
The controller can be separated and placed far away from the motor.
Figure 3 Star-connected 3-phase fully-pitched winding SRM driven from 3-phase full bridge inverter
Figure 4 Delta-connected 3-phase fully-pitched winding SRM driven from 3-phase full bridge
inverter
The disadvantages of using the 3-phase full bridge circuit for driving SRMs are:
There is no inherent shoot-through protection.

The line currents are restricted in that they must sum to zero.
There are additional diode losses.
13
1.5 SWITCHED RELUCTANCE MACHINES vs.

PERMANENT MAGNET BRUSHLESS DC
MACHINES
With recent advances made in the magnet material and manufacturing technologies,
permanent magnet brushless dc machines (PM BLDC) have become the first choice for
many applications. The competition is a tough one, especially in the aerospace,
automotive and renewable energy sectors (wind turbine generators specifically).
It is possible to compare the performance of machines using the flux-linkage versus
current characteristics. Per phase flux-linkage versus current locus for an SRM is only in
a single quadrant. On the other hand, for a PM machine the flux-linkage versus current
locus occupies four quadrants. This is illustrated in Figure 5 for the 3-phase 12-8 and 1210 segmental rotor designs [12], 3-phase 12-8 conventional toothed rotor design and a 3phase PM BLDC design (that has 12 slots and 8 poles with Neodymium-Iron-Boron
(NdFeB) magnets with each coil wound around a single tooth). In Figure 5 all machines
are running under current control, generating similar amounts of copper losses and, when
loaded, the peak machine flux density levels in all the machines are identical. The locus
for the PM machine is shown for one half cycle of conduction, with the machine
connected in a star configuration. The areas of flux/MMF loci are directly related to
torque production (it should be stated that the 12-10 design will be traversing its loop
25% more times than the others and hence produces 25% more torque per unit area of
locus).
Each SRM phase conducts for 1/3 of a cycle whereas each of the phases of a PM machine
conducts for 2/3 of an electrical cycle. PM machines have greater power to weight ratios
and are generally more efficient than SR machines. Furthermore, the ripple content of the
total output torque in a PM machine is significantly lower than that can be achieved with
an SRM, hence making the PM machine the preferred option in torque ripple sensitive
applications. SRMs have inherent fault tolerance, as the phases can be physically (by
winding each phase coil around a single stator tooth), electrically (by driving them from
separate half-bridges) and magnetically (by having little or no mutual coupling) isolated.
Moreover, as the SR machines are singly excited, fault current will not continue to be
driven into a faulted phase, once excitation to that phase is removed.
14
0.006
0.005
0.004
Flux per turn (Wb)
0.003
0.002
0.001
0
-0.001
500
1000
1500
-0.002
2000
3-phase
3-phase
3-phase
3-phase
-0.003
2500
3000
3500
PM brushless d.c.
Segmental 12-8 SRM
Conventional 12-8 SRM
Segmental 12-10 SRM
-0.004
Phase MM F (A)
Figure 5 Flux-Linkage loci for the 3-phase segmental rotor 12-10 SRM (in dark blue), 3-phase
segmental rotor 12-8 SRM (in red), 3-phase conventional toothed rotor 12-8 SRM (in green) and a
brushless d.c. PM machine (in light blue), operating under current control
With careful design PM machines can achieve a similar degree of fault tolerance with
higher torque densities (although the potential torque density is compromised to achieve
fault tolerance) which is a very significant merit especially in the aerospace industry
where total mass is crucial [14].
High performance permanent magnets are from the group of elements on the periodic
table known as rare earth materials. The supply risk and the price volatility (in 2008 there
was a shortage of Neodymium and Praseodymium for use in permanent magnets and the
price rose by a factor of five to a peak of $50/kg for Neodymium [15]) of these materials
must be highlighted here. As the Chinese leader Deng Xiaoping famously stated in 1992:
The Middle East has oil, China has rare earths, currently China dominates the world
supply of rare earth magnet raw materials. Added to the sheer size of available rare earth
resources, highly developed separation and processing power resulted in China
representing over 95% of the worlds rare earth materials supply. Rare earth materials are
not only used in the electric motors industry. Published global production and
consumption estimates show that as early as 2015 Asia alone could consume almost all of
the world production for many of the rare earth materials if no new production facilities
are established.
15
The Chinese rare earth industry is government controlled and can easily control the
pricing of rare earth materials to eliminate competition in mining, separation and
processing of rare earth materials. Recently China increased taxes and imposed export
restrictions and VAT penalties on all Chinese exporters of rare earth materials.
There will inevitably be a growth in demand for rare earth materials for permanent
magnets in the near future with growing numbers of hybrid and electric vehicles and high
performance wind turbine generators. Where permanent magnet motors fall short due to
price volatility, or purely because of insufficient materials resources, the SR machines
will be the prime candidate to substitute them.
1.6 OBJECTIVES AND CONTRIBUTION TO

KNOWLEDGE
This PhD thesis aims to reduce the torque ripple of SRMs by simply increasing the
number of phases. The number of power devices, current transducers and the number of
connections between the machine and the converter are kept to a minimum by connecting
phase windings of the machine back-to-back and driving them from a conventional 3phase full bridge circuit. Instead of controlling the phase currents, the controller deals
with the line currents, reducing the number of current transducers needed to two as the
third line current can be derived from the other two line currents.
The research conducted during this PhD study contributes to knowledge in the following
areas:
A comprehensive assessment of segmental rotor SRMs and comparison against other

machine types.
Approaches which enable SRMs to be driven from conventional three phase bridge
inverters.
A concept which enables six phase SRMs to be employed, whilst only having three
wires connecting the SRM to the power inverter.
Machine construction methods which allow greater slot fill factors.
A concept which allows reduction of torque ripple in a switched reluctance machine.
A method which enables driving a two-phase SRM with a single current transducer.
16
Simulation methods which enable accurate modelling of various machine and inverter
topologies.
A set of matrix transformations which enables modelling of SRMs even when there is
strong mutual coupling between phases.
1.7 THESIS OVERVIEW

This thesis is divided into eight chapters and appendices.
Various SR machine design topologies centred on the conventional toothed stator and
rotor structures are reviewed in Chapter 2 SR Machine Topologies. The same chapter
also reviews the concept of employing a segmented rotor design in a doubly salient
reluctance motor. Fundamental design principles are given and two demonstrator
machines are compared with conventional and axially laminated SRMs.
A review of drive topologies and control methods for switched reluctance machines is
discussed in Chapter 3 SR Drives. Converter topologies that are commonly employed
in SR drives are presented and basic operating principles are explained.
Chapter 4 Machine Design and Testing describes the detailed design processes of 6phase 12-10 segmental and conventional toothed rotor SRMs. Measurements of the fluxlinkage characteristics, static torque and thermal characteristics are all presented.
In Chapter 5 Machine Construction Techniques the method of constructing stator for

prototype machines is explained. The method of constructing stators from separate cores
is decribed and discussed.
Chapter 6 Simulation Models describes the simulation models developed to model

various parts of the drive system. The prototype segmental rotor machine has strong
mutual coupling between phases. A set of matrix transformations is presented to decouple the phase quantities into slot quantities.
17
In Chapter 7 Machine Operation and Comparison to Simulation Two Phase
Drives a two phase configuration is used to compare the performance of asymmetric
half-bridge and h-bridge inverters. Dynamic measurement results under various operating
conditions are presented and compared to the dynamic simulation results.
In Chapter 8 Machine Operation and Comparison to Simulation Three Phase

Drives dynamic measurement results under various operating conditions for the three
phase 12-10 single tooth segmental rotor SRM are presented. These measurements are
also used to validate the dynamic simulation and deduce inverter rating requirements.
In Chapter 9 Machine Operation and Comparison to Simulation Six Phase

Drives the six phase 12-10 segmental rotor SRM is connected in star and driven from a
3-phase full bridge circuit. Dynamic measurement results are presented and analysed.
In Chapter 10 Comparison of Machine Performances the torque ripple and inverter

VA requirements of the six phase 12-10 segmental and toothed rotor SRMs are compared
against those of the 3-phase 12-10 single tooth segmental rotor SRM.
With the final chapter, Chapter 11 Conclusion and Future Work the overall aims are
revisited and major achievements and shortcomings of the research are highlighted.
Suggested future work arising from the research carried out in this PhD thesis is provided
and an overall assessment of the obtained results is presented in this chapter.
18
Chapter 2 SRM TOPOLOGIES
CHAPTER 2
2 - SRM Topologies
2.1 INTRODUCTION
With the development of new materials, continuing advancements and cost reduction in
power semiconductors, more effective use of mathematical optimisation techniques and
increased capabilities of 2D and 3D FE analysis, interest in both commercial development
and research in switched reluctance machine based drive systems has continued to
produce new concepts in recent years. Early development of switched reluctance
machines was driven by the activities of Bedford [16, 17] at General Electric Company in
Schenectady, of Lawrenson and Stephenson [18] at Leeds University, of Byrne and his
colleagues [19] at University College in Dublin and of Ray and Davis [20] at Nottingham
University. In [18] Lawrenson et al outline the basic operating and design principles of
doubly salient reluctance motors. Pioneering work of Lawrenson, Stephenson, Ray and
Davis from Leeds and Nottingham Universities has grown into a commercial
establishment in the name of SR Drives Ltd founded in 1980. Also the work of the
SPEED Laboratory led by Prof. Miller at the University of Glasgow, the Power
Electronics Drives and Machines group of the University of Newcastle Upon Tyne, and
Prof. Charles Pollock in Leicester and Warwick Universities have led the way in SR
machine and drive research for the last three decades.
This chapter starts with a brief explanation of the torque ripple phenomenon in SRMs and
gives an overview of torque ripple minimisation techniques in SR drives through the
electromagnetic design of the machine. Recent papers by Mecrow et al [12, 23] have
introduced the concept of employing segmented rotor design to SRMs. Instead of the
19

usual toothed arrangement, the rotor is constructed from a series of discrete segments,
each of which is magnetically isolated from each other. Section 2.3.2 of this chapter
reviews Mecrows 3-phase designs. This chapter will also review various switched
reluctance motor designs giving examples of some of the relevant work aiming to
improve the performance of SRMs by altering the magnetic and / or electric design of the
machine. Section 2.3.1 is dedicated to explaining several different design variations of the
conventional toothed rotor structure. Section 2.3.3 covers recently published work on
segmental rotor switched reluctance machines.
2.2 SRM TORQUE RIPPLE

In an ideal switched reluctance machine, reluctance varies linearly with tooth overlap and
for a constant MMF, constant torque is produced. SRMs are designed to operate deep into
the magnetic saturation zone of the lamination material to maximise the output torque
density. Consequently in a real machine saturation and field fringing flux mean that the
torque produced for a constant phase current is a non-linear function of position.
Saturation and variation of magnetic reluctance in the SRM, due to its doubly salient
structure, generate highly non-linear flux and torque characteristics. As a result generated
torque can contain significant torque ripple. Torque ripple is not only a highly undesirable
characteristic for many applications (i.e. direct drive electric power steering and robotics)
on its own but is also believed to contribute significantly to another drawback of SR
drives, acoustic noise [24]. Torque ripple in a drive system can be quantified by Equation
1.
Tripple (%) =
T max T min
100
Tave
Equation 1
where Tave is the average machine output torque.
Torque ripple minimisation can be realised by machine design and / or employing torque
smoothing control strategies. The following subsection summarises the methods aiming to
reduce the torque ripple in SR drives through machine electromagnetic design. A
summary of methods relating to the drive control will be presented in the next chapter.
20
2.2.1 Torque Ripple Minimisation through Machine Design

As one phase gets closer to the end of its energisation cycle and the energisation of the
next phase is about to start a torque dip appears. As a result, the majority of torque ripple
is said to happen in this phase overlap region. The width of the overlapping region
between torque outputs of adjacent phases depends on the pole overlap angle between the
stator and rotor, material properties, number of poles and number of phases. Optimising
the individual phase torque-angle characteristics by salient pole shape profiling [25], or
skewing of the rotor axially [26] can help to minimise the inherent torque ripple.
Increasing the number of phases appears to be a better approach in reducing torque ripple,
since the higher the number of phases the higher the overlap will be as the phases get near
the commutation zone. The obvious downside of having a machine with increased
number of phases is the potentially increased cost and complexity of the drive system due
to increased number of power devices required to drive the phases of the machine.
The vibration modes and torque ripple are associated with the level of air-gap flux
density; it is believed that the higher the air-gap flux density level the worse the torque
ripple, vibration (due to normal forces) and the acoustic noise levels. It now appears that
to reduce the torque ripple, vibrations and consequently the acoustic noise, the air-gap
length can be increased. However increasing the air-gap will compromise the power
density and the efficiency of operation [27]. Another design method to reduce torque
ripple is to employ an asymmetric rotor pole geometry [28]. However skewed rotor poles
compromise the total output torque. In [29] adjacent rotor poles have unequal widths
resulting in a two-stage flux pattern, which was believed to operate with a low torque
ripple (more on this machine later in this chapter).
In [30] the authors analyse the effect of simple variations of the rotor pole face structure
on the torque production of a 60kW - 6000rpm experimental SRM. A notable design
procedure, focusing on torque ripple minimisation, is given in [31]. The design presented
in [31] is heavily centred around the feasible design triangle given in [18]. In [32] authors
use FE to analyse the effects of shaping the rotor and stator pole tips on the torque ripple
in short-pitched and fully-pitched winding SRMs. In [33] stator and rotor pole arcs are
optimised through a genetic algorithm based routine in order to reduce torque ripple,
without reducing the average torque output of the machine. The paper deals with a
specific design and appears to be (more of) an application of genetic algorithms rather
21

than a generalised approach for reducing torque ripple in switched reluctance machines.
Other examples of torque ripple reduction through pole arc shape optimisation are given
in [34, 35].
2.3 SRM TOPOLOGIES

Theory and general design considerations of switched reluctance machines are well
documented in a number of papers [30, 36-58].
The performance of a drive comprising a switched reluctance motor heavily depends not
only on the motor itself but also on how well the power electronic controller is matched to
the particular motor design and application requirements. Therefore successful switched
reluctance motor designs should take the converter and the motor as a whole, fulfilling the
requirements of the specific application. A few examples of application specific switched
reluctance machine design work are given in [59-67]. When a design is required for a
specific application, often the design space and performance requirements are known
parameters. What is left for the design engineer to do is to produce the optimal machine
and controller design to fulfil the requirements. This obviously is not a trivial task.
Optimal machine design can be regarded as a multi-dimensional optimisation problem
and the solution can be found by using genetic algorithms [68-70], artificial neural
networks [71] or non-linear optimisation methods [72].
2.3.1 Conventional Toothed Rotor SRMs

In simple terms, a switched reluctance machine has salient poles on both its stator and
rotor, but has windings on its stator only (see Figure 6).
22
Figure 6 Schematic of a three-phase conventional switched reluctance machine showing various

motor sections
Conventionally, the phase windings are short-pitched, with each coil wound around a
single stator pole. Each phase has series and / or parallel connected coils forming stator
poles. Use of short-pitched windings results in negligible magnetic coupling between
phases, i.e. flux generated by a single phase current only links the windings of that phase.
Of course, in reality there is some mutual coupling through core-back saturation and
cross-slot leakage but this is generally considered to be negligible [49]. Similar to its
stator structure, the rotor of a conventional SRM is formed from a toothed lamination
structure. Motoring torque is generated when current flows into a phase when the rate of
change of phase inductance is positive. Similarly, generating torque is achieved by
injecting current into a phase when the rate of change of phase inductance is negative (see
Figure 7).
23
Lmax
Phase
Inductance
Lmin
Motoring
Torque
Region
Generating
Torque
Region
Figure 7 Variation of inductance and positive and negative torque production regions with ideal pulse
unidirectional current. Lmax and Lmin are the maximum and minimum phase inductances,
respectively
For the conventional toothed rotor structure in Figure 8 there are four main parameters
that are considered to be of importance in its design. These are:
Tooth width to pole pitch ratio ( t / )
Air-gap length( L g )
Ratio of stator and rotor outside diameters i.e. Split ratio (
Ratio of core-back width to half of tooth width (
SOD
)
ROD
CBD
)
t/2
The dimensional abbreviations above are shown graphically in Figure 8.
24
Figure 8 Description of several machine dimensions for a rectilinear representation of a conventional

toothed rotor SRM
Tooth width to pole pitch ratio ( t / - Figure 8) is often used as a measure of magnetic
utilisation in switched reluctance machines because it gives a good measure of how much
an excited rotor pole carries magnetic flux. The limitation with a toothed design is that
t / cannot be increased beyond 0.5 as the stator and rotor poles start overlapping even at
the unaligned position, increasing unaligned permeance and consequently reducing torque
capability. Typical values of t / range from 0.3 to 0.45. An increased unaligned curve
also increases the VA requirements of the drive as it becomes more difficult to drive the
phase current when the minimum inductance of the machine is high. Higher converter VA
requirements will compromise the size and the cost of the overall drive system.
Variations on the basic topology have emerged. Significant contributions include:
Use of short flux paths [73-75]

Flux reversal machines [76-78]
Common pole E-cores [79]
Staggered rotor teeth [29, 78]
Flux switching motors [51, 80-82, 92]
Shark tooth air-gaps [83-85]
Multi-teeth per pole [86-88]
Overlapping windings to exploit mutual coupling between phases [13, 51, 89-92]
25
2.3.2 Segmented Rotor Switched Reluctance Machines

The SRM design variations listed in the preceding section have all been attempts to
alleviate various weaknesses and to improve the performance of switched reluctance
machines i.e. output torque, efficiency, torque ripple, etc. However, the magnetic
geometry of all these designs has been effectively fixed, employing a series of stator teeth
with short-pitched or fully-pitched windings and a series of rotor teeth that are
magnetically coupled through the stator and rotor back iron. The fundamental shortcoming of the magnetic circuit of a conventional toothed rotor SRM is the limitation
imposed on tooth pitch to pole pitch ratio (t/) by the unaligned inductance of the
machine. If the ratio of (t/) was to be increased above 0.5 then there would be an overlap
between stator and rotor teeth, even when the rotor was in the unaligned position. This
would result in excessive unaligned inductance and reduce the aligned to unaligned
inductance ratio, which in turn would reduce the output torque of the machine.
A patent in 2002 by Mecrow [93] and subsequent conference [94] and journal papers [12]
by Mecrow et al have introduced the concept of using a segmental rotor in SR machines.
The idea stems from the developments seen in synchronous reluctance motor magnetic
design and aims at increasing the magnetic utilisation of switched reluctance motors by
having more than half of the magnetic iron structure carry machine flux with short flux
paths at any time during machine operation. Instead of the conventional toothed structure,
the rotor is constructed from a series of discrete segments. Each rotor segment is
magnetically isolated from its neighbour (Figure 9). As the stator phase windings are
energised the rotor segments tend to move to a minimum reluctance position, modulating
the stator slot permeances in the process. This is similar to a toothed structure where the
rotor tooth moves to the minimum reluctance position, but differs in the fact that the slot
permeances are modulated by rotor rotation, rather than teeth permeances in a
conventional toothed structure.
Prior to Mecrows segmental rotor structure there have been a couple of notable
exceptions to the conventional toothed SRM design. Xu and Lipos work [95], where they
have examined the axially laminated rotor structures for variable reluctance motors and
26
Figure 9 Three-phase segmental rotor prototype machine schematic layout and magnetic flux plots in
aligned and unaligned rotor positions
Horsts patent [96], concerning the use of a segmented rotor design for unidirectional
operation of two-phase SRMs, should be stated here as examples of designs towards
changing the fundamental magnetic structure of conventional toothed SRMs. In [95] Xu
and Lipo describe a new type of reluctance motor with concentrated windings and an
axially laminated rotor driven by unidirectional currents. A cross-sectional view of their
design is given in Figure 10. Although Xu and Lipo have called their machine a
synchronous reluctance motor it had similarities with doubly salient switched reluctance
machines in that it had concentrated windings and operated in a switched fashion with
unipolar currents. Xu and Lipo were able to show that with axially laminated rotor
27

structure synchronous reluctance motors were able to generate higher torque densities.
The machine in Figure 10 has its rotor laminations bent to produce paths of minimum
reluctance in the direction of laminations and maximum reluctance in the path normal to
the laminations.
Figure 10 Cross-sectional view of Xu and Lipo's axially laminated single salient variable reluctance
motor [95]
A cross-sectional view of Horsts two-phase segmental SRM is shown in Figure 11. Horst
was aiming to increase the power density of the machine by having shorter flux paths and
using all salient poles simultaneously. As can be seen from Figure 11 the air-gap was
distorted to achieve a preferred direction of rotation. It is not clear from the patent, which
is limited to two phase machines, whether a prototype was ever built.
Figure 11 Cross-sectional view of Horst's two-phase segmental rotor SRM [96]
28

In [12] Mecrow et al demonstrated that SRMs with segmental rotors outperform
reluctance machines with axially laminated rotors. In [12] Mecrow states: It can be
concluded that axially laminated rotors work well with the distributed MMF pattern of a
synchronous reluctance motor, but do not offer the best solution for an SRM. With an
SRM the MMF is concentrated, and it is better to have a rotor magnetic circuit which has
an equally concentrated reluctance, as displayed by the segmental design. Mecrows
work produced two successful prototype machines with segmental rotors. The first
prototype was called a multi-tooth winding segmental rotor SRM and the second one was
called a single-tooth winding segmental rotor SRM. The following two sections will
discuss these two prototypes in more detail.
2.3.2.1 Segmented Rotor Switched Reluctance Motors with FullyPitched Windings

The first prototype designed and built by Mecrow et al was a 3-phase machine with 12
stator teeth and 8 rotor segments [12]. Initial design work was carried out on a single
phase rectilinear machine (Figure 12) to determine the advantages of employing
segmental rotors in switched reluctance motors. To reveal the principles of the new
design, the geometry is simplified as much as possible. For this reason a single phase,
simple rectilinear model with a segmented rotor has been considered initially. This is
illustrated in Figure 12 along with the magnetic flux plots for aligned and unaligned
positions.
In Figure 12 the magnetic flux flows down one tooth, through a rotor segment and returns
via the adjacent stator tooth. All the conductors in each slot only couple with flux driven
by their own magneto-motive force, with very little mutual coupling between one slot and
another. Torque production can therefore be considered on a per slot basis in which the
slot permeance is modulated by the rotor segments [12]. In the aligned position the rotor
segments magnetically short the stator slots, producing a high permeance, whilst in the
unaligned position the segments are located over the tooth tips so that the unaligned
permeance is limited by the slot opening [12].
29
UNALIGNED POSITION
ALIGNED POSITION
MAGNETIC FLUX PLOT OF A SINGLE PHASE SEGMENTED ROTOR SRM
ALIGNED POSITION
UNALIGNED POSITION
Figure 12 Rectilinear representation of a single phase segmental rotor SRM with 2D flux plots in
aligned and unaligned positions [23]
The definition of tooth width (t) for this machine is the width in the main body of the
tooth, not at the tooth tip as is the case for conventional toothed structures. As in the
conventional toothed SRMs is the pole pitch. Contrary to conventional toothed rotor
designs the segmental design has been shown to have the ability to maintain a low
unaligned permeance, even when the effective t/ is greater than 0.7, thereby resulting in
a much greater torque capability. Magnetisation characteristics at the maximum
permeance (i.e. aligned) position for both conventional toothed rotor SRMs and
segmented rotor SRMs correspond closely with each other. The real difference is
observed in the minimum permeance (i.e. unaligned) magnetisation characteristics. When
t/ is less than 0.5 (Figure 13) the segmental design exhibits a higher unaligned fluxlinkage, mainly due to the semi-closed nature of the slot, which increases the cross-slot
flux. However, once t/ is bigger than 0.5 (Figure 14) the opposite occurs and the
segmental design exhibits a much lower unaligned flux-linkage characteristic.
30

T ooth Width to Pole Pitch Ratio is 0.3
0.008
ConvUnaligned
ConvAligned
SegUnaligned
SegAligned
Magnetic Vector Potentail [Wb/m]
0.007
0.006
0.005
0.004
0.003
0.002
0.001
0
0
500
1000
1500
2000
2500
M M F [Ampere T urns]
Figure 13 Comparison of aligned and unaligned magnetisation curves of toothed rotor (both red
traces) and segmental rotor (both blue traces) when tooth width to pole pitch ratio is 0.3
T ooth Width to Pole Pitch Ratio is 0.7

0.018
ConvUnaligned
ConvAligned
SegUnaligned
SegAligned
Magnetic Vector Potentail [Wb/m]
0.016
0.014
0.012
0.01
0.008
0.006
0.004
0.002
0
0
500
1000
1500
M M F [Ampere T urns]
2000
2500
Figure 14 Comparison of aligned and unaligned magnetisation curves of toothed rotor (both red
traces) and segmental rotor (both blue traces) when tooth width to pole pitch ratio is 0.7
31

The single phase rectilinear design comparison is useful to show that with a segmental
rotor structure it is possible to substantially increase the torque capability of switched
reluctance machines. The advantage of the segmental design becomes more apparent with
increased tooth width to pole pitch ratio.
Mecrow et al demonstrated that although having a machine design with t/ > 0.5 would
be impractical (as this would leave insufficient room for the windings) in a rectilinear
single phase machine example, this would not apply to a multi-phase machine. Figure 9
shows the three-phase 12-8 segmental rotor SRM in aligned and unaligned positions.
Mecrow et al showed that the tooth width of the single phase design could be halved with
the multi-phase design, which would allow enough room for phase windings whilst
retaining a high value of tooth width to pole pitch ratio. This is due to the fact that in a
three-phase design, with one phase excited, each stator tooth carries only the flux of one
rotor segment [12]. Whereas in a single phase design, with one phase excited, each stator
tooth has to carry the flux of two rotor segments (see Figure 12). The multi-tooth
winding, three-phase design had 12 stator teeth and 8 rotor segments.
A tooth width to pole pitch ratio of 0.67 was achieved. This meant that the segmental
rotor design could carry much more flux than the conventional toothed rotor design with
similar machine dimensions having tooth width to pole pitch ratio of 0.33. Although the
windings of the three-phase segmental design in [12] were fully-pitched, torque was
produced as a result of changing self-inductance, not as a result of mutual inductance
experienced with toothed rotor SRMs with fully-pitched windings [12]. Compared to a
conventional toothed rotor SRMs with short-pitched windings, segmental rotor SRMs
with fully-pitched windings have considerably longer end-windings, resulting in
additional copper mass and winding loss. The importance of this additional loss depends
on the machine axial length and pole number [12]. For a short stack length machine the
ratio of end-winding length to overall winding length will be substantial.
Static torque results and comparison of the mean torque capability at thermal limit (i.e
mean torque vs. winding copper loss) between fully-pitched segmental rotor, fullypitched toothed rotor and short-pitched toothed rotor designs are given in [12] and are
explained in more detail in [97].
32
2.3.2.2 Segmented Rotor Switched Reluctance Motors with ShortPitched Windings

As described in the previous section, for SRMs with segmental rotors the flux-linkage per
phase is doubled and hence torque as a function of electric loading is doubled (flux of a
single phase is carried by two-thirds of stator teeth rather than one-third used in a
conventional toothed rotor SRM). However, as the coils spanned as many stator teeth as
the number of phases the end-winding length was increased. For a prototype machine this
additional end-winding length resulted in 20% reduction in electric loading at thermal
limit [12, 23]. Overall a net increase of 40% in the measured torque per unit volume at
thermal limit was achieved compared to that achieved by conventional toothed rotor
SRMs within a given frame size. However, the torque per unit winding mass of the
multitooth winding prototype was raised only modestly compared to conventional toothed
rotor SRM.
To develop a three-phase design which is suitable for short stack length machines it is
essential that the windings span a single tooth to keep the end-windings short, whilst
maintaining the condition that only the conductors of a single-phase occupy any one slot.
The idea is realised by Mecrow et al [23]. The design is depicted in its rectilinear form in
Figure 15, along with the rectilinear form of the fully pitched design. The new shortpitched segmented rotor SRM (or single-tooth segmented rotor SRM) is achieved by
removing the shaded regions from the fully pitched design (see Figure 15) as these
regions carried no flux and their removal would not significantly influence the level of
flux-linkage. In the short-pitched segmented design the coil spans a single rotor segment
and have a shorter end-winding, wrapped around a single-tooth [23]. The unwound teeth
of the short-pitched design are the same width as the width of the teeth of the fully
pitched design, whilst the wound teeth are double this width. The reason is that the wound
teeth now carry the flux of two adjacent slots.
Two-dimensional flux plots of the prototype single tooth winding segmental rotor SRM in
the aligned and unaligned positions are given in Figure 16.
Figure 17 shows two-dimensional finite element predictions of magnetization curves for
both segmental rotor designs and an equivalent conventional toothed rotor design. Once
again it is quite obvious from this figure that with a conventional toothed rotor design,
only half of the flux-linkage of segmental designs can be achieved. Inspection of Figure
17 also reveals that the flux-linkage characteristics of multitooth and single tooth winding
33

segmental rotor designs are virtually identical. In the aligned position the difference is
about 3%, whereas in the unaligned position the difference between the magnetisation
curves of single tooth and multitooth segmental designs is as low as 1%. This suggests
that the force produced by one phase is identical with a multitooth and a single tooth
segmental design. However, the force exerted per unit area of airgap is reduced by 20%
compared to a multitooth winding design when the MMF per phase is fixed. This is due to
the fact that the single tooth winding design requires more space for the wound teeth,
which require being twice as wide. A set of design rules for single tooth winding
segmental design is given in [23]. The assembly method for the single tooth designs is
similar to that employed for multitooth designs. For both multitooth and single tooth
winding designs static torque curves are somewhat different to those usually experienced
with conventional toothed rotor SRMs.
34
Figure 15 Rectilinear representation of three-phase segmented rotor SRMs [23]
35
ALIGNED FLUX PLOT
UNALIGNED FLUX PLOT
Figure 16 Three-phase 12/10 segmental rotor SRM with single-tooth windings - shown in aligned
position
0.014
Magnetic Vector Potential [Wb/m]
0.012
0.01
0.008
0.006
0.004
0.002
0
0
100
200
300
400
500
600
700
800
900
1000
MMF [At]
Conventional SRM - unaligned
M ultitooth Winding Segmental SRM - unaligned
Singletooth Winding Segmental SRM - unaligned
Conventional SRM - aligned

Multitooth Winding Segmental SRM - aligned
Singletooth Winding Segmental SRM - aligned
Figure 17 Magnetic vector potential of a phase coil as a function of coil MMF for both segmental
designs and a conventional SRM
36

This is explained with the aid of Figure 18. At high excitation levels the form of the static
torque curves are similar to those of a conventional toothed rotor machine when the
machine is saturated. However, the shape differs significantly when the machine is in the
unsaturated region (i.e. low excitation levels). This difference is due to the permeance
variation experienced with the segmental rotor structure as the rotor moves from the
unaligned to the aligned position. With a segmental design the area of overlap between
stator teeth and rotor segments increases on one side of the magnetic circuit but it
decreases on the other. This results in a static torque which rises to reach a peak at
approximately 12.0 degrees from the aligned position. Running waveforms of a single
tooth winding segmental rotor have been presented in [23], where the machine phases
were energised with an asymmetric half-bridge converter.
80
75
70
Current
65
60
Torque [Nm]
55
50
45
40
35
30
25
20
15
10
5
0
0
10
11
12
13
14
15
16
17
18
Position [Degrees]
Figure 18 Measured static torque characteristics of 12/10 single tooth design with a single phase
excited. Each curve is for a constant current, rising in 2.0A steps to 20.0A [97]
In [23] the effect of orientation of magnetisation is briefly explained. Apparent from the
magnetic flux plots of Figure 16 there is minimal coupling between individual phase coils
of a phase. Therefore, it can be said that the magnetic characteristics of any one phase are
independent of the polarity of the two coils. The orientation of magnetisation can be any
of the forms depicted in Figure 19 without having a measurable effect on phase
37

parameters. Mutual coupling between the phases of the single-tooth winding segmental
rotor SRM has been shown to be minimal [23].
Figure 19 Possible orientation of magnetisation for single tooth winding segmental rotor SRM (all
shown in aligned position)
Direction of energisation for each phase coil becomes an important factor when two
phases are excited simultaneously, which is most likely to occur when the machine is
operating under voltage control and a positive voltage is applied for more than one third
of a cycle (for a three-phase machine). There are two possible scenarios:
The unwound teeth of the machine are dimensioned to take only the flux of a single
phase. If the two phases are excited simultaneously and the direction of energisation
in adjacent phase coils both act in the same direction, i.e. both inwards or both
outwards, the unwound teeth will have to take the sum of the fluxes from two phases.
38

This will result in saturation of the unwound teeth (cross-saturation) resulting in
reduced torque.
If the direction of energisation in adjacent phase coils acts in opposite directions, i.e.
inwards, outwards, inwards, outwards, etc. the unwound teeth will carry the difference
between the fluxes of adjacent coils. Therefore simultaneous excitation of two phases
will actually reduce the saturation conditions. However, this time the unwound teeth
will experience bi-directional flux.
With both of the above scenarios the direction of the stator coil MMFs will influence the
iron loss in the machine.
2.3.2.3 Other Segmented Rotor Switched Reluctance Motor Designs

After the introduction of new segmental rotor SRMs by Mecrow, several researchers used
the idea of segmental rotors in their designs. In [98] Oyama et al present a 3-phase design
with 6 stator teeth and 4 rotor segments with fully-pitched phase windings. This work is
heavily based on the principles outlined by Mecrow et al. The only difference appears to
be the way the segments are held together and the shape of the segments. An aluminium
block is used to hold the segments in place but there is no discussion of mechanical
strength. A cross-sectional view of Oyama et als machine is given in Figure 20.
39
Figure 20 Cross-sectional view of Oyama et als segmental rotor SRM with segments embedded in an
aluminium block [98]
Another piece of interesting work utilising segmental rotors in SRMs is that of Vattikuti
et al [99]. A cross-sectional view of Vattikutis segmental rotor SRM is shown in Figure
21. The machine illustrated in Figure 21 has circular stator slots with fully-pitched
windings. In this design the flux is forced to flow in a circular path. Due to long endwindings of the fully-pitched winding distribution, the design is unlikely to be suitable for
designs with short stack lengths. Furthermore, all the analysis was based on FE analysis
and none of this theoretical work was proven with measurements from a prototype.
Figure 21 Segmental rotor SRM with circular stator teeth work of Vattikuti et al [99]
40

In [8] Hall et al employ a segmented rotor structure for an embedded aero-engine starter /
generator application. The design was driven by the requirements / limitations imposed by
the application, such as the available space, operating speed, operating temperature and
power. An outer rotating segmental rotor SRM with 18 stator teeth and 15 rotor segments
was the chosen machine topology which was able to satisfy fault tolerance, power,
operating temperature and speed requirements of the application. A cross-sectional view
of the design is given in Figure 22.
Figure 22 Outer rotating segmental rotor SRM with 18 stator teeth and 15 rotor segments for use in
aero-engine embedded starte / generator application - R.Hall et al [8]
2.4 SUMMARY
Fundamentals of the magnetic structure of switched reluctance motors have been
effectively unchanged for decades. Almost without exception both the stator and the rotor
of an SRM are of the toothed form (Figure 6). Design principles of these conventional
toothed rotor structures are well established in literature. The effects of variations in
various design parameters of these conventional toothed rotor machines on machine
performance (such as torque output, torque ripple, acoustic noise) are all well
documented. Some of the important design work on the conventional toothed rotor
structure is summarised in this chapter.
41

Mecrow et al introduced the concept of segmental rotors for use in switched reluctance
motors by applying some of the developments of synchronous reluctance motor magnetic
design (with modification). This segmental rotor design better utilises the magnetic circuit
and increases the torque output of the machine. The multitooth winding version has the
drawback of increased end-winding length. As a result, only a small gain is achieved in
torque per copper volume and loss compared to the conventional toothed rotor SRMs of
the same frame size. Consequently, for a short stack length application a multitooth
winding segmental rotor SRM would not be the preferred option. Segmental rotor SRMs
with single tooth windings are aimed to overcome the long end-winding penalty of the
multitooth winding option by having every other tooth wound with concentrated
windings. Consequently, a single tooth winding segmental design has much shorter endwindings compared to a multitooth winding segmental design. Mecrow et al showed that
a single tooth winding segmental design could be designed with 29% less copper
compared to a multitooth winding segmental design of the same frame size [23].
Moreover, single tooth winding solutions would be more suitable for fault tolerant
applications as the windings are magnetically, thermally and mechanically isolated from
each other to a greater degree than they are in a conventional toothed rotor and multitooth
winding segmental rotor SRMs. Recently a few designs using the segmental rotor
principle have been reported, aiming to ease the manufacturing process as well as
increasing the performance of the machine compared to conventional toothed rotor
options.
42
Chapter 3 SR DRIVE TOPOLOGIES
CHAPTER 3
3 SR Drive Topologies
3.1 INTRODUCTION
The performance and the cost of any motor drive system are highly affected by the
performance of the selected converter topology. Probably more so than the other drive
systems, the relatively low cost of manufacturing the core SR motor for an SR drive
system is counter-balanced by the potentially high power electronics costs. One often
comes across the phrase matching the converter to the particular application for an SR
drive. Indeed, the type of converter required for a particular SR motor drive is intimately
related to motor construction and number of phases. In other words, a designer is faced
with a decision to find the most appropriate combination of motor and power electronic
converter for a particular application. For this reason no single combination of motor and
power electronic converter has become the standard for SR drives.
For an SR machine the direction of the generated torque does not depend on the signs or
the values of the flux-linkage and current. As a matter of fact the torque generated by an
SR motor only depends on the sign of the rate of change of phase inductance with rotor
position. This feature of SR machines along with phase independence has been
extensively used to the advantage of SR drives and resulted in many different converter
topologies over the years. As there are many topologies to choose from for an SR drive,
there is always a trade-off between gaining some of the advantages and losing some with
each topology. Whatever the chosen converter topology is, the power electronic converter
for an SR drive is required to build the flux in a phase winding from zero at each stroke
by providing a positive voltage loop; it must then be able to reduce the applied voltage if
43

the desired current level is reached. Finally it must be able to provide a negative voltage
loop at turn-off to quickly return the phase current to zero before the sign of rate of
change of inductance changes. For an SR drive, at the end of each stroke the power
electronic converter must provide a path to recover the stored energy in the magnetic
field. SR converters differ in terms of how they achieve this. Some of the converters
dissipate the energy, whereas others use magnetic or capacitive recovery techniques. It is
also beneficial to have a converter topology which is capable of generating a zero-voltage
loop (i.e. freewheeling period) before phase commutation instances. In freewheeling
mode the phase current is circulated through the power electronic devices and phase
windings instead of being returned back to the supply with a negative voltage loop. This
mode of operation has the potential to reduce the current ripple. It also reduces the
circulating energy from the dc bus and the machine phases, increases efficiency and
reduces frequent energy reversals in the machine phases. Energy reversals are believed to
be a significant acoustic noise source. Elimination of these energy reversals during the
phase energisation cycle by introducing a freewheeling period just before the
commutation point potentially eliminates noise associated with them altogether [101].
In Chapter 2, it was mentioned that there were two distinct ways to tackle the torque
ripple problem in SR drives and the methods relating to the electromagnetic design of the
machine were summarised. In this chapter, the torque ripple minimisation techniques
relating to the control strategy employed in an SR drive system will be presented.
Consequently, section 2 of this chapter summarises the work focusing on current profiling
techniques.
Section 3 of this chapter summarises:
The SR drives based on the number of machine phases
The drive control methods e.g. high speed PWM control, advanced flux-linkage
contol and sensorless control, etc.
Some of the most common converter topologies used to drive SR machines, e.g.
asymmetric half-bridge converter, C-dump converter, etc.
Finally in section 4 a summary of the chapter is presented.
44
3.2 TORQUE RIPPLE MINIMISATION THROUGH

DRIVE CONTROL
As torque is a nonlinear function of phase current and rotor position, the flat topped phase
current is believed to be far from ideal in terms of generating ripple free output torque in
an SR drive. Furthermore, due to the non-linear relationship between rotor position,
torque and current, no reference frame transformation can be applied to SR machine
control. This necessitates instantaneous torque control in order to reduce torque ripple if
electronic control is used [102]. Several notable papers have been published by
researchers investigating ways of defining and generating ideal phase current waveforms
which result in minimal output torque ripple. Lack of space precludes a detailed
description of all methods adopted, but interested readers are referred to references [103127].
3.2.1 Measuring Torque Ripple

Torque ripple can be measured using in-line torque transducers, strain gauge bridges,
observer algorithms or accelerometers measuring vibrations of the drive housing.
Methods of measuring torque ripple are given in [24]. In [128] off-line static torque
measurements are used to estimate the running torque waveforms from measured currents
and rotor position. Absolute accuracy of the method in [128] is questionable, as it
depends on the bandwidth of the current sensor, A2D sampling rate, encoder resolution,
accuracy of the initial static torque measurements, accuracy of the encoder and the
accuracy of current transducers. However, it is believed to be a simple and effective way
of measuring the dynamic torque ripple of the drive system for comparative analysis. In
[129] Corda et al outline how to determine machine torque during operation. Torque
ripple measurement can be done by storing static torque characteristics and measuring
phase current and rotor position to look-up an instantaneous torque value.
An interesting solution to determination of torque ripple in mass-produced switched
reluctance drives is presented in [130]. In [130] Koenig and Pekarek present a specially
designed piezoelectric washer that is attached to an anchor bolt (used to mount the
machine to the mechanical structure) and senses harmonics resulting from machine
45

torsional vibrations (i.e. torque ripple). The method seems to be more suitable for large
machines, where the relative cost of the sensor is low.
3.3 DRIVE TOPOLOGIES

For the majority of SR machines the phases are said to be magnetically de-coupled,
permitting independent control of phase windings. At very high speeds, this is especially
beneficial as considerable overlapping between adjacent phase currents is present.
However, lack of mutual coupling means that the stored magnetic field energy during the
energisation of the phase windings needs to be provided with a path during commutation
of a phase. This stored energy can be handled in a number of different ways [131-134].
As torque is independent of the direction of the current for an SR machine, the fluxlinkage and the current, as well as the topology of the inverter circuit, may be unipolar.
This mode of operation is preferred because it permits a simpler form of controller.
Ability to operate with unipolar currents and the need to have a path for the stored
magnetic energy at the end of each stroke resulted in considerable amount of unique
converter topologies for an SR drive system.
It is highly desirable to have a simple form of torque control in electric drives. In DC
drives this is achieved naturally by simply controlling the product of flux and current in
the machine. The simplicity of a DC drive torque control comes from the fact that the flux
and current are independently controllable. In most AC machines the principles of DC
machine torque control is achieved with the use of field-oriented control, which is based
on some form of mathematical transformation. Generally speaking in classical DC and
AC machines, control of the flux is maintained constant while the current is varied in
response to the torque demand. The SRM is a highly non-linear machine in its nature with
torque a nonlinear function of not only the phase current but also of the rotor position.
Moreover, for a machine with mutually coupled phases - such as a fully-pitched winding
machine - torque is dependant on the rotor position and current in all of the machine
phase windings. Therefore there is no simple torque constant for an SRM.
46
3.3.1 Review of Drives

This section is a brief summary of SR drives based on the machine number of phases.
Single phase SR drives are the simplest, with fewest connections between the machine
and electronics. However, they cannot produce positive torque at all angular positions.
Starting problems can be overcome by the introduction of a permanent magnet in the
stator [135] or by using a graded air-gap (i.e. graded reluctance) to extend the angle over
which positive torque is available [65].
With two phase drives problems of starting are reduced, but still pose challenges.
Methods used include stepping the air-gap, or providing asymmetry in the rotor poles
[136]. This machine may be of interest where the cost of winding connections is
important, but again high torque ripple may be detrimental, with the torque falling close
to zero at some rotor positions. Compared to a three-phase drive, a two-phase drive
requires two thirds the number of power devices and the switching frequency is reduced
for a given rotor speed, lowering the losses in the two phase drive [82]. If cost is the
primary concern and the performance is not measured in terms of torque ripple and
starting ability, single and two phase drives can be of interest.
Three and four phase drives offer the simplest solution to starting without resorting to
high numbers of phases. Significant torque can be developed at all rotor positions, but
torque ripple is still much higher than in a conventional 3 phase ac machine.
Compared to three and four phase drives, the torque ripple is greatly reduced with a 6
phase drive. The output torque is made up of six pulses per electrical cycle, as in a 3phase PM motor drive. However, as the number of phases is higher, the number of power
devices and the number of connections between the inverter and the motor are also
higher. An example of the torque ripple reduction with a 6-phase drive is given in Figure
23. The top trace shows the torque output from a 3-phase 12-10 segmented rotor SRM
whereas the bottom trace shows the total output torque from a 6-phase 12-10 segmented
rotor SRM. In both cases perfect current control is assumed. Both machines are running
@ 200rpm under current control. Control angles are adjusted for both drives in order to
achieve similar average output torque values.
It is concluded that an SRM will only approach the low torque ripple of a conventional
PM when there are six pulses of torque per electrical cycle. This requires a 6 phase SRM
47

but, with conventional converter topologies, this will not be cost effective: there will be
12 controlled power switches and 12 cables connecting the inverter with the machine,
compared to 6 switches and three cables in a conventional ac machine. The challenge
posed is Can a 6 phase SRM be run with only 6 power switches and 3 cables between
motor and inverter?
Figure 23 Simulation of 3-phase and 6-phase 12-10 segmented rotor SRMs under perfect current
control
3.3.2 Review of Control Methods

In general, torque is directly related to the currents flowing in phase windings. Therefore
in most electric drives phase current is the controlled variable and it is often designed to
track a pre-defined reference value to achieve the required output torque. However the
48

overall shape of the current waveform is not only dependant on the average torque
required but also on the torque ripple requirements of the drive system.
The next step is to decide on the control method. With the right choice, the drive
performance can be significantly improved. With good controller design:
Motor performance can be improved.

Overall efficiency of the drive system can be increased.
A quieter drive operation can be achieved.
The drive system can operate more reliably.
The size and the number of components required in the system can be reduced.
Some of the most common control methods used for SR machines are summarised below.
3.3.2.1 Hysteresis Current Control

At low speeds the phase current has to be constrained to protect the electronics because of
the high available Volt-seconds. This is typically achieved by hysteresis current chopping
where the power transistors are switched off and on according to whether the current is
greater or less than a reference current. As the upper and lower current thresholds are
control parameters, this chopping current strategy allows a very precise current control.
However, as the switching frequency is not fixed, acoustic and electromagnetic noise is
difficult to filter [49].
3.3.2.2 PWM Control

Current control can also be achieved by an advanced PID controller combined with a
PWM voltage control strategy. Current chopping can be achieved either by varying the
applied voltage from negative to positive or from negative to zero and then to positive.
The former method is called hard-chopping and the latter one is called soft chopping [24].
In terms of current control, SR machines differ from commonly used ac and dc machines.
As the effective phase inductance varies significantly with the rotor position the controller
will only be tuned optimally at one point and for all other points the system response will
be over or under damped, making stable operation difficult [89, 137].
49

Above base speed the phase current potentially cannot reach the desired level and the
phase needs to be energised before the phase inductance starts rising (i.e. advancing of
commutation point). Eventually a speed is reached where the phase is energised for the
entire commutation interval. This form of operation is often called Single Pulse mode
as there is no phase voltage chopping, but a continuous energisation of the phase
windings for the entire conduction period.
3.3.2.3 Current Profiling

Due to the highly non-linear characteristics of the SR machine, a current waveform
tracking a constant current reference will result in high torque ripple. By profiling the
current a constant shaft torque with minimal torque ripple can be achieved. Since torque
cannot be controlled directly, due to the lack of adequate torque sensors, this can only be
accomplished using a priori information about the torque-current-angle characteristics of
the motor. Additionally these characteristics must be known fairly accurately in order to
achieve satisfactory results. The torque control strategy is based on following a contour
for each of the phases of the SR motor such that the sum of torque produced by each
phase is constant and equals the desired torque.
The principle of current profiling can be extended and used with a flux-linkage controller.
Barrass [110] successfully applied the use of flux-based control to 3-phase short-pitched
and fully-pitched SR machines. In this method 2D look-up tables are used to estimate
machine flux-linkage from current and rotor position in real time. The flux-linkage
controller is de-coupled from the dc-link voltage variations and phase winding
temperature variations. Flux-linkage controllers are said to have a fast and well-damped
response [110] and are well suited for constant torque operation.
3.3.3 Review of Converter Topologies

Although SRMs are conventionally easy to manufacture and relatively cheap to make, the
largest cost of an SR drive system is its power electronic converter. Therefore,
performance of the converter plays an important role in cost and performance of the
overall drive system.
50

The cost and size of power electronic drives vary significantly for an SRM. The selection
of a converter, as mentioned earlier, depends upon the application. The elements of cost
associated with the power electronics side of an SR drive system can be summarised as
follows:
Number and rating of power semiconductor switches and diodes.

Cooling system related to the power electronic losses. The higher the losses the
more costly the cooling system will be.
The amount of passive components needed i.e. dc-link capacitors, filter inductors.
This can be a significant factor for applications with strong EMI-restrictions.
Gate drive circuits. If the number of floating point gate-drive circuits is high, the
cost of the drive will be higher than a converter with fewer floating point and more
grounded switches.
Control logic (including sensors).
Following the above statements it is important to match the converter to the particular
application for an SR drive application. A large number of converter topologies and their
advanced versions are available for an SR drive. This is due to:
The phase independence.

The unipolar current requirement.
Different methods of dealing with the stored magnetic field energy at the end of each
stroke of the machine.
Of particular interest are the simplicity, flexibility and the cost of the converter for a
given application. A few of the most common topologies are reviewed below.
3.3.3.1 Asymmetric Half-Bridge Converter

Converters with a single switch per phase can be used but by far the most common
converter topology for controlling SRMs is the asymmetric half-bridge converter. A
simple diagram of an asymmetric half-bridge converter is given in Figure 24. In an
asymmetric half-bridge converter, the power devices are connected to either side of the dc
51

link rail and in series with a phase winding (e.g. T1 and T2 in Figure 24) and two diodes
(e.g. D1 and D2 in Figure 24) form the return path for the phase current. The diodes
cannot be anti-parallel (i.e. already packaged within the power switches). This adds to the
cost of the drive. Requirement of separate freewheeling diodes is not so much of an issue
at high power levels, as the inverter switches do not usually come with anti-parallel
diodes as a single package [142].
On the plus side an asymmetric half-bridge configuration is inherently immune to shootthough faults, as there is always a phase winding in series with power devices
Figure 24 Asymmetric half-bridge converter driving a three-phase machine
Asymmetric half-bridge converters provide maximum control flexibility and are able to
apply all of the available dc-link voltage (minus the power electronic device voltage
drops) across the phase windings [49]. This is a significant advantage in controlling the
current in the phase windings. With an asymmetric half-bridge converter each phase can
be controlled independently and power devices are rated only at the dc link voltage.
On the negative side, the total component count is high as each phase requires two
switches and two diodes and two connections between inverter and motor. This increases
the size and the cost of the system.
The asymmetric half-bridge converter is able to supply all three operating modes. These
are:
1. Positive Voltage Loop: where a positive voltage is applied across the phase
windings to increase the current. Both top and bottom devices are turned on and
conducting current in this loop.
52

2. Zero Voltage Loop: where the phase current is freewheeled through a diode and a
power device rather than returned back to the supply. This mode of operation is
advantageous as it is believed to generate less torque ripple, and hence less
machine noise. It also minimises the ripple content of the phase current and
potentially reduces the dc-link capacitor RMS ripple current rating, making it
smaller.
3. Negative Voltage Loop: where at the end of a phase energisation cycle a negative
supply voltage is applied across the phase winding to reduce the phase current
down to zero before the rate of change of inductance reverses direction. Both
power switches are turned off and negative supply voltage appears across phase
windings (anti-parallel diodes are conducting the current) quickly reducing the
current down to zero.
In terms of cost, size and the number of connections between the motor and the converter,
as there are two switches per phase the asymmetric half-bridge circuit is more suitable for
machines with fewer phase windings. Moreover, as there are always two series connected
devices conducting current; the asymmetric half-bridge topology is unsuitable for low
voltage applications.
3.3.3.2 Shared Switch Converter (Miller Circuit)

A schematic of this converter topology is given in Figure 25. With this circuit converter
switches and diodes are shared between windings of more than one phase, reducing the
number of switches used per phase. First reported by Miller [144] with this converter
topology, one power device is common to all phases of the machine [Switch Q in Figure
25]. It is this device which performs the chopping of the phase current. The main
drawback of this topology is that phases cannot be controlled independently. Once a
motor phase is turned off the on-coming phase becomes the priority. As the phase current
starts falling in the off-going phase and rising in the on-going phase, the current controller
will try controlling the current in the newly energised phase. When the current controller
starts chopping the on-coming phase this will affect the reverse voltage available to
completely de-flux the off-going phase. This inevitably slows down the de-energisation
53

process and results in a tail current in the off-going phase. This becomes more of a
problem at higher speeds, where the amount of time available to energise and de-energise
a phase is significantly reduced. The tail current will last into the region where the phase
inductance reduces. This will generate negative torque and reduce the output power and
the efficiency of the drive system.
Figure 25 Shared switch converter driving a three-phase machine
3.3.3.3 C-Dump Converter

For an SRM, flux in a phase is built up from zero at the beginning of a cycle and returned
to zero before the end. Stored energy at the end of each stroke needs to be provided with a
path to travel. In the case of a C-dump converter the stored energy is dumped into a
capacitor [Cx in Figure 26] and returned back to the supply through the chopper circuit
consisting Tc, Dc and Lc in Figure 26. Like the Miller circuit, this topology also requires
an additional switch [Tc in Figure 26] on top of a single switch per phase. A chopper, the
operation of which is based on the buck principle, is used to discharge the additional
capacitor [131]. With the dump capacitor a faster demagnetisation of phases is possible.
With this topology, the number of passive components is high and the control complexity
(requires the monitoring of the dump capacitor voltage and control of the chopper switch)
makes the converter expensive and rather complicated. As more components are added to
the system and the control method complexity is increased, there is the compromise of
adding more tolerances to the overall system. This means that the system performance
becomes more susceptible to component and/or parameter value variations. This is not a
desired feature for a high volume application.
54

Moreover the voltage ratings of the dc-link capacitor and switching devices are high (the
average voltage across the dump capacitor is approximately twice the voltage across the
dc-link capacitor in order to have equal magnetizing and demagnetizing voltages), which
potentially increases the cost and the overall size of the drive system. The system has
potential reliability issues as any control failure will lead to over-charging of the dump
capacitor and over-stressing the power devices, leading to their failure. Modified versions
of this topology have been proposed [145, 146].
Figure 26 C-dump converter driving a three-phase machine
3.3.3.4 Split DC-Link Converter

As the name suggests, the single dc-link capacitor is split into two capacitors in series.
This circuit topology is only suitable for machines with an even number of phases, as in
Figure 38. Although the power devices are rated at the supply voltage, they are only
capable of supplying half of the supply voltage.
This converter uses one extra capacitive component in the supply rail, which will
potentially increase the cost and the size of the drive and potentially reduce the overall
drive efficiency. It is also intolerant to phase unbalance. Under such a case there is
dangerous voltage build-up in the split dc-link capacitors [145]. Therefore the control
circuit must balance the phase currents to keep the centre point voltage at the correct
level.
55
Figure 27 Split dc-link converter driving a two-phase machine
3.3.3.5 Bifilar Winding Converter

As with the split dc link converter, at first glance this converter seems very attractive as it
only requires a single switch per phase (Figure 28). An extra winding is used in each
phase to recover the stored energy during commutation. Although the converter has one
power switch per phase, the switch needs to be rated at least twice the supply voltage
level, because when the secondary windings conduct negative dc-link voltage is induced
across the primary windings. This converter also suffers from reduced efficiency, as only
one of the bifilar windings carries current at any time, resulting in poor copper utilisation
for the machine [145]. Three to four electrical connections per phase are needed
compared to the usual two connections per phase for other drive circuits. Furthermore,
snubber circuitry is required to protect the power devices against large di/dt effects in
case of leakage. Overall this converter is a good example to show that a single switch per
phase drive is not necessarily a cheap and an efficient solution.
Figure 28 Three-phase bifilar winding converter
56
3.3.3.6 H-Bridge Converter

An H-bridge converter for SR drives is generally used in the format given in Figure 29. In
this format the h-bridge circuit meets the minimum switch per phase requirement for SR
drives and is suitable for 4-phase machines. Like the split dc-link circuit it can only utilise
half of the dc-link voltage. Obvious, from Figure 29, two phases are always on
simultaneously. However, only one of these phases may be contributing to torque
production at any instant. Therefore torque per copper loss is degraded [145].
Figure 29 H-bridge converter driving a four-phase machine
An alternative H-bridge converter is implemented in this PhD project. Figure 30

illustrates an h-bridge circuit for a two phase SRM where phase windings are connected
back-to-back through two diodes (named as phase diodes hereafter). The phase diodes are
used to achieve operation with unipolar phase currents, prohibiting the flow of phase
current in the reverse direction when the direction of applied phase voltage is reversed.
Therefore, when a phase is de-energised, its current decays to zero and stays at zero until
the next positive voltage cycle. The control can be achieved either with a single current
transducer, measuring the difference between adjoining phase currents (line current), or
with two current transducers measuring each phase current separately. With the
arrangement in Figure 30 the loss in power transistors is reduced as each converter phase
57

leg only conducts the difference between adjoining phase currents. The remainder of the
current simply circulates within the phases. Unfortunately the use of a h-bridge converter
in this manner is only suitable for machines with even number of phases. Moreover the
diodes in series with the windings of each phase increase the total inverter losses.
Simulation results and lab measurements of the h-bridge circuit driving a two-phase SRM
where phases are connected back to back are given in Chapter 7.
Figure 30 H-bridge converter driving a back-to-back connected two-phase machine
3.3.3.7 Three-Phase Full Bridge Converter

A three-phase full bridge converter generates AC currents and is commonly used for AC
drives. It cannot directly supply the unipolar currents desired for almost all SR machines.
With a delta connected drive, unipolar phase currents can be achieved with alternating
line currents so long as a circulating current can be maintained by the circuit. The basic
delta connection prevents a net voltage being applied around the delta, so the circulating
current cannot be maintained.
It was shown [89] that the circulating current in the delta configuration could be
maintained by placing diodes in series with each of the phase windings (Figure 31).
58
Figure 31 Three-phase full bridge circuit driving a delta connected machine
With the use of the three-phase full-bridge circuit, the number of connections between the
motor and the inverter is reduced to only three. Providing that the phase diodes are
connected at the motor end, the converter and the motor can be far away from each other.
As opposed to the asymmetric half-bridge inverter, which directly controls the phase
voltages, the standard three-phase inverter can only control the line quantities.
In [148] another use of three-phase bridge inverter for driving a switched reluctance
motor was given. In this paper, the three phase bridge was used to drive a short-pitched
winding machine. The machine was connected in star. Unlike the basic star-connection
the star point was connected at the mid point of two dc-link capacitors. Two phases were
energised in order to keep the voltage balance between the dc-link capacitors. In this
configuration, one of the energised phases did not contribute to torque production (in fact
producing negative torque and hence reducing the total output torque of the machine) and
generated extra copper losses.
Earlier in this chapter, the challenge was identified as running a 6-phase SRM with only 6
power devices and 3 cables between motor and inverter. Within this research the concept
of using a three phase bridge is extended to six phase drives in an attempt to reduce
torque ripple, without resorting to large numbers of power devices or motor connections.
Two drive concepts have been introduced, which are briefly explained in the next two
sub-sections.
59
A. Six-Phase Star Connected Drive:

The phases of the machine have been paired and connected back-to-back through series
connected diodes. The diodes in series with the phase windings have been called phase
diodes and were used to prohibit the flow of bipolar phase currents. The paired phases
form a module. The three modules were then connected in star and driven from the
three-phase bridge inverter. The star point has been left unconnected, forcing the sum of
the three line currents to be zero. The star-connected six phase drive is illustrated in
Figure 32.
Figure 32 Star-connected six-phase machine driven by a 3-phase full bridge circuit
B. Six-Phase Delta Connected Drive:

In the case of the delta connected six-phase drive, the modules were connected in delta
configuration. This drive circuitry is depicted in Figure 33. The voltage applied to any
one module directly affects the voltage available to be supplied to the other two modules.
This affects the ability of the converter to supply voltage to phases on different modules.
60
Figure 33 Delta-connected six-phase SRM operated from a conventional 3-phase bridge circuit
For both the star (Figure 32) and delta (Figure 33) connected drives the current controller
can be based on phase currents or line currents. Within this research, line current based
control has been implemented as it required two current transducers, rather than six (third
line current can be derived from the other two). Another advantage of measuring line
currents instead of phase currents is that the current sensors do not need to be at the
machine end and can be close to the inverter with short cabling and better noise
immunity. A controller based on line currents requires a separate PID and a PWM
controller for each line. The controller controls the potentials at points U, V, and W in
Figure 34. The line demand current is compared with the measured current (Iu, Iv or Iw in
Figure 34) and the error is fed to the PID controller. Depending on the error current and
the PID controller gains, the PID controller of each line only produces a positive
reference voltage (Vref), which is then compared to the triangular wave (Vtri) in the
PWM controller. The top transistors of each bridge leg (T1-T2, T3-T4 or T5-T6) are
switched on if Vref of the corresponding leg is greater than Vtri. If not, the bottom
transistors are switched on. Table 1 summarises the useful switching states and potentials
at points U, V, and W for each condition.
61
Figure 34 Ideal line currents for the delta (a) and star (b) connected six-phase drives
Table 1 Summary of the useful switching states and potentials at points U, V, and W for each
operating condition for the 6-phase star and delta connected drives.
SWITCH STATES
PHASES
DESIRED ON
LEG 1
T1
T2
LEG 2
T3
T4
POTENTIALS AT
LEG 3
T5
DIRECTION OF
T6
F&A
ON OFF OFF OFF OFF ON
A&B
ON OFF OFF ON OFF OFF
B&C
OFF OFF OFF ON
C&D
OFF ON OFF OFF ON OFF
D&E
OFF ON
E&F
OFF OFF ON OFF OFF ON
ON OFF
ON OFF OFF OFF
62
POINTS U, V, W
LINE CURRENT
Vu
Vv
Vw
Iv > 0 D4 conducts
V(+)
V(-)
V(-)
Iv < 0 D3 conducts
V(+)
V(+)
V(-)
Iw > 0 D6 conducts
V(+)
V(-)
V(-)
Iw < 0 D5 conducts
V(+)
V(-)
V(+)
Iu > 0 D2 conducts
V(-)
V(-)
V(+)
Iu > 0 D1 conducts
V(+)
V(-)
V(+)
Iv > 0 D4 conducts
V(-)
V(-)
V(+)
Iv < 0 D3 conducts
V(-)
V(+)
V(+)
Iw > 0 D6 conducts
V(-)
V(+)
V(-)
Iw < 0 D5 conducts
V(-)
V(+)
V(+)
Iu > 0 D2 conducts
V(-)
V(+)
V(-)
Iu > 0 D1 conducts
V(+)
V(+)
V(-)

Dynamic simulation of the star and delta connected six-phase machines driven from
three-phase full bridge circuit will be presented in Chapter 6. Simulated and measured
results of the 6-phase star and delta connected drives will be presented in Chapter 9.
3.4 SUMMARY
AC electrical machines produce two pulses of torque per phase per electrical cycle. Hence
a 3-phase AC machine has 6 pulses of torque per cycle. The pulses overlap strongly and
consequently the level of torque ripple produced is generally low (of the order 1%-10%).
SRMs, on the other hand, produce one pulse of torque per phase per electrical cycle.
Hence a 3-phase SRM has only 3 pulses of torque per electrical cycle. Moreover, in the
periods between peaks of pulses the output torque dips considerably, producing torque
ripple in the order of 30-100% of the mean torque value. For many applications the high
level of torque ripple is unacceptable and can only be easily rectified by increasing the
number of phases of the machine to six or more; thereby making it comparable to a 3phase AC machine. However, a negative consequence of a high phase number is both the
large number of power electronic components and electrical connections between the
motor and the converter, i.e. increased drive cost.
Alternative converter arrangements have been reviewed: these all reduce the number of
power devices, but also reduce the level of controllability of phase voltage and current.
Earlier work which used a conventional three phase bridge for a 3 phase SRM is reviewed
and it is postulated that this concept may be suitable for a 6 phase drive. Consequently,
two drive topologies have been introduced; one with the phases of the machine connected
in star and the other in delta.
63
Chapter 4 MACHINE DESIGN & TESTING
CHAPTER 4
4 Machine Design &
Testing
4.1 INTRODUCTION
It is the aim of this work to produce a high phase number SRM generating low torque
ripple, without resorting to complex power electronic supplies or control strategies. The
machine designed and constructed for this PhD project is a 6-phase SRM with 12 separate
stator cores and 10 rotor segments. The details of the machine construction are given in
this chapter.
Prototype designs were evaluated with the aid of 2D finite element analysis to generate
the flux linkage characteristics. Both a commercially available FE package (MagNet of
Infolytica Corp.) and the FEA package of the Newcastle Universitys Power Electronics,
Drives and Machines (PEDM) group were used to model both segmental and toothed
rotor 6-phase designs. Early FE analysis was carried out in the FE package of PEDM and
the design space analysis was carried out in Infolyticas MagNet. Results of magnetic and
thermal testing of prototype machines are also presented in this chapter.
64
4.2 DESIGN OF THE 6-PHASE SEGMENTED

ROTOR SRM
For direct comparison, the 6-phase segmented design has its outer diameter, inner
diameter, air-gap length and lamination stack length equal to those of the three phase
segmented rotor machines designed by Mecrow et al. These are summarised in Table 2.
For 3-phase multi-tooth and single-tooth winding segmented rotor SRMs presented in
[12] and [23], respectively, each stator coil enclosed a single-tooth, which resulted in very
little coupling between one slot and another.
For a 6-phase design where each slot contains only the windings of a single phase, the
machine would require 24 stator teeth and 22 rotor segments. A 12 stator teeth and 11
rotor segments arrangement could also be possible. However, this would produce
unbalanced magnetic pull, resulting in increased acoustic noise and potentially reduced
bearing life. To limit the switching frequency and to avoid excessive core losses, 12 stator
teeth and 10 rotor segments were chosen for the 6-phase design.
Table 2 List of dimensions of 6-phase segmented design that are kept the same as those of the 3-phase
single-tooth segmented design of Mecrow et al
Stator Outside Diameter [mm]
150.00
Stator Bore Diameter [mm]
91.40
Rotor Outside Diameter [mm]
90.80
Air-gap Length [mm]
0.30
Lamination Stack Length [mm]
150.00
Choosing a 12-10 arrangement for a 6-phase design requires each stator slot to contain
windings of two phases, with each coil of a phase wound around a single tooth, resulting
in short-pitched windings. As will be explained in the following chapters, at any one time
during the machine operation, the 6-phase design will have two phases conducting current
simultaneously. Simultaneous conduction of two separate phases has an effect on the
machine design, especially in the definition of the aligned and unaligned positions for this
machine. In Figure 35 the 6-phase segmented rotor machine is shown in the unaligned
position. Note that this unaligned position is shown with respect to phases A and F
65

conducting simultaneously. As two phases will be conducting at any one time, in the
unaligned position the gap z in Figure 35 should be as large as possible to minimise the
machine unaligned inductance. This ensures that neither the rotor nor the stator
excessively contribute to the unaligned flux-linkage. Y1 and Y2 in Figure 35 are the
stator tooth tip arc angle and the rotor segment arc angle, respectively. The direction of
current flow in each coil is highlighted by the conventional dot / cross representation in
Figure 35.
Phase A
eF
as
Ph
eC
as
Ph
Ph
as
eE
Ph
as
eB
Phase D
Phase D
Figure 35 Drawing of the 6-phase 12-10 short-pitched segmented rotor SRM (to establish design rules
and equations) in the unaligned position with respect to conduction of Phases A and F
Similarly in Figure 36 the machine is shown in the aligned position. Flux plots for the
unaligned and aligned positions in Figure 35 and Figure 36 are depicted in Figure 37 and
Figure 38, respectively.
66
Figure 36 Drawing of the 6-phase 12-10 short-pitched segmented rotor SRM (to establish design rules
and equations) in the aligned position with respect to conduction of Phases A and F. Phase
configuration is the same as shown in Figure 35
Figure 37 Flux plot of the 6-phase segmental rotor 12-10 SRM in the unaligned position. The
energised coils are shown with the conventional dot / cross representation. Only one half of the
machine is shown
67
c
Figure 38 Flux plot of the 6-phase segmental rotor 12-10 SRM in the aligned position. The energised
coils are shown with the conventional dot / cross representation. Only one half of the machine is
shown
Figure 39 shows a rectilinear representation of the 6-phase 12-10 segmented rotor SRM
with slots labelled as a, b, c, d, e, and f and teeth labelled as 1, 2, 3, 4, 5, and 6. In Figure
39 the arrows pointing up and down from one stator tooth to the next indicate the
direction of flux. The labelling of slots and teeth will be used in Chapter 6.
Figure 39 2D rectilinear representation of six-phase SRM showing the direction of slot MMFs and
teeth fluxes
In the aligned position the overlap angles (X1 and X2 in Figure 36) between the stator
teeth and the rotor segments should be as large as possible to maximise the machine
aligned inductance. In Figure 36, W is the stator tooth width, SD is the active rotor
segment depth (excluding the dovetail section), OutD is the stator outer diameter, InD is
68

the rotor outer diameter, and CBD is the stator core-back depth. The relationships
between the variables given in Figure 35 and Figure 36 are established from Equation 2 to
Equation 7, with dimensions chosen to:
(a) Have all magnetic components carrying a similar peak flux density, so that no one
component creates early onset of magnetic saturation.
(b) Maximising the difference between the unsaturated aligned and unaligned
inductances.
The gap angle y in the unaligned position (shown in Figure 35) is the angle between the
rotor segment tip and one of the energised stator tooth tip and is calculated by Equation 2.
(
y =
360 360
) (Y 2 Y 1)
Nr
Ns
Y 2 Y1
= 3
2
2
Equation 2
where N s is the total number of stator teeth and N r is the total number of rotor segments.
The gap angle z in the unaligned position (shown in Figure 35) is the angle between the
rotor segment tip and the closest non-energised stator tooth tip and is calculated by
Equation 3.
z=
360
Y 2 y = 30 Y 2 y
Ns
X1 =
360
Y 1)
Ns
2
Equation 4
360
Y 2) X 1
Nr
Equation 5
Y2(
X 2 = Y1 (
Equation 3
The stator tooth width - W should be equal to the overlap between stator teeth and rotor
segments in the aligned position. This ensures similar levels of magnetic saturation
throughout the magnetic circuit and yields Equation 6.
69

360
InD
InD
2
)
( X 1 + X 2) 2
(Y 1 + Y 2
2
2
Nr
W=
=
360
360
Equation 6
where W is the stator tooth width and InD is the rotor outside diameter.
From Figure 38, it is evident that the stator core-back and the rotor segment carry similar
levels of flux. The level of flux carried by the stator core-back and the rotor segments is
somewhat less than that carried by the stator teeth. However, inspection of Figure 37 and
Figure 38 reveals that the core-back of the fully energised slot (where coils of both phases
carry current the slot with two dots in Figure 38) is required to carry more flux than the
half energised slots (the slots with a single cross in Figure 38). Therefore, Equation 7 is
used to calculate the stator core-back and the rotor segment depths (excluding the dovetail
section) where they are made larger than half the stator tooth width.
CBD = SD = 0.7 W
Equation 7
In Equation 7, CBD is the stator core-back depth, SD is the segment depth and W is the
stator tooth width. The final values of stator core-back depth and the rotor segment depth
are further increased. The stator core-back is made deeper in order to increase the
stiffness of the machine aiming to reduce the stator vibrations and acoustic noise (but not
too deep which would reduce the available winding area and result in increased copper
losses for the same slot MMF), whereas the rotor segment depth is made bigger to have
better retention against circumferential forces which will allow (mechanically) higher
speed operation.
The electromagnetic limit of torque output of the machine can be increased by changing
the stator tooth width ( W ), the stator tooth tip arc angle ( Y 1 ) and / or the rotor segment
arc angle ( Y 2 ). However the machine torque output per copper loss can be reduced as a
result of the changes in the available slot area, the number of turns, and / or the conductor
diameter. To be able to characterise the torque output vs. total machine copper loss one
needs to calculate the total slot area with a good degree of accuracy. Detailed calculation
steps of the total slot area, conductor diameter and the phase resistance are given in
Appendix D.
70
4.2.1 Determining the Machine Dimensions

Design equations for the prototype 6-phase segmented rotor SRM with 12 stator teeth and
10 rotor segments were listed from Equation 2 to Equation 7. The arc angle of the rotor
segment (Y2 in Figure 35 and Figure 36) was varied from 20o to 30o by steps of 1o. For
each angle the arc angle of the stator tooth tip (Y1 in Figure 35 and Figure 36) was also
varied from 20o to 30o. The designs, which had good overlap (i.e. maximum X1 + X2 in
Figure 35 and Figure 36) in the aligned position (Figure 36) and also had the maximum
values for y and z (Figure 35 and Figure 36) in the unaligned position (Figure 35), were
selected and compared, based upon their torque capability.
Aligned and unaligned flux/MMF curves for this design are presented in Figure 40 and
the mean torque/MMF curve is shown in Figure 41.
Figure 40 Aligned and unaligned average vector potential vs. MMF c urves for the prototype machine
71
Figure 41 Mean torque vs. MMF curve for the prototype machine
Flux plots for this design in the aligned and unaligned positions are given in Figure 42a
and Figure 42b, respectively. The dimensions of this machine were not produced using
formal optimisation, but rather through the use of the design rules discussed earlier and
through comparisons between different similar designs. The final design was judged to
be good, and selected for manufacture.
Figure 42 Magnetic flux plots for the 6-phase 12-10 short-pitched segmented rotor SRM with two
phases excited simultaneously in the aligned position (a) and in the unaligned position (b)
The chosen design dimensions are given in Table 3.
72

Table 3 Full list of design parameters of the prototype machine
Number of phases
Stator bore diameter [mm]
91.40
Air-gap length [mm]
0.30
Rotor outside diameter [mm]
90.80
Number of stator teeth
12
Number of rotor segments
10
Stator tooth tip taper angle [Degrees]
60.00
Rotor segment taper angle [Degrees]
60.00
150.00
Lamination stack length [mm]

Stator Tooth Tip Arc Angle - Y1 [Degrees]
25.00
Rotor Segment Arc Angle - Y2 [Degrees]
27.00
Gap in unaligned position - y [Degrees]
2.00
Gap in unaligned position - z [Degrees]
1.00
Overlap in aligned position - X1 [Degrees]
11.00
Overlap in aligned position - X2 [Degrees]
5.00
Stator Tooth width [mm]
12.68
Stator core-back depth [mm]
11.00
Rotor segment depth [mm]
9.72
273.65
Estimated total slot area [mm^2]
1.40
Total conductor diameter [mm]
44
Number of turns per coil

Slot fill factor (approximately)
0.50
Estimated phase resistance []
0.36
Estimated end-winding length [mm]
45.74
4.2.2 Design Space Analysis

At a later stage of this PhD project a commercially available electro-magnetic finite
element analysis package called MagNet from Infolytica Corporation became available.
In MagNet, by programming at the command level with VBScript or VBA (visual basic
scripting languages) it is possible to automate the process of generating models by using
73

design equations (in this case Equation 2 to Equation 7). The script can create the model,
set model meshing options, run the simulation and extract and save results in spreadsheet
or text format once the simulation is finished. This allows relative comparison of many
different design possibilities within the given design space and requirements in a
relatively short amount of time. This method was used to judge the quality of the design
selected above for manufacture.
The computation time of using 3D FE analysis would be very large, so two dimensional
finite element modelling was performed allowing flux-linkage vs. current characteristics
of the machine to be generated quickly. This is beneficial for a comparative design
process as effects of different machine dimensions can be analysed very quickly and an
optimum design decision can be made. Among others, end winding leakage effects are
not taken into account in 2D FE analysis. End winding leakage effects can become
significant when the machine is in the unaligned position, especially when the machine
has a short stack length. Simulations with two dimensional models will underestimate the
magnetisation curve in the unaligned position. The amount of underestimation will
depend on the ratio of the total end-winding length to the lamination stack length.
However, for comparative analysis of design possibilities within the available design
space, using 2D FE analysis can still be considered a valid approach.
This section summarises the work carried out for machine design space mapping by
varying the stator tooth tip arc angle (Y1 angle in Figure 35) and the rotor segment arc
angle (Y2 angle in Figure 35). As the stator tooth tip and rotor segment arc angles are
varied the following machine parameters change:
The stator tooth width (W in Figure 36).

The stator core-back depth (CBD in Figure 36).
Available stator slot, and, hence, winding area.
The overlap angles (X1 and X2 in Figure 36) between the stator teeth and the rotor
segments in the aligned position.
The separation angles (y and z in Figure 35) between the stator teeth and the rotor
segments in the unaligned position.
Rotor segment depth (SD in Figure 36).
The analysis is based on three factors:
74

1. Magnetisation curves in the aligned and unaligned rotor positions and the area
covered between these two curves. This gives a direct indication of output torque
capability of a particular design.
2. Per phase copper losses. As the stator tooth tip and the rotor segment arc angles are
varied the stator tooth width widens and narrows, changing the available slot area in
the process. For a given conductor diameter this will change the number of turns per
coil and the slot fill factor or for a given slot fill factor it will change the conductor
diameter and the number of turns per coil. In both cases per coil winding resistance
will be changing, directly affecting the copper losses in the process. Comparisons are
made using the assumption of perfect current control.
3. Unaligned magnetisation curve. Although the area between the aligned and
unaligned curves is a good indicator of the output torque capability of the machine,
one cannot ignore the effect of the unaligned curve in the converter VA requirements.
A design with an increased unaligned magnetisation curve and an overall increased
area between the aligned and unaligned curves (aligned curve increases more than the
unaligned) may not always be the preferred choice just because it can generate more
torque. Increased unaligned inductance will make it more difficult to inject the desired
current into the phase windings within the required amount of time. This can increase
the VA requirements of the drive system and make it more difficult to achieve the
desired level of output torque at a given operating speed of the motor.
In order to assess how good the chosen design is, in terms of absolute torque output, total
output torque per total loss and unaligned inductance within the given design space, stator
tooth tip arc angle and rotor segment arc angle have been varied from 20o to 30o in steps
of 1o. This has resulted in 121 different design possibilities. However many of these
possibilities were eliminated before any FE work was carried out, simply due to practical
limits (e.g. too narrow / wide slot area, thin stator core-back depth or narrow slot opening,
etc.). Six potential designs have been chosen for further analysis including the one
corresponding to the design chosen for manufacture.
The following parameters have been kept the same for all six design possibilities:
75

1. Slot fill factor (same as the manufactured design).
2. Number of turns per coil (same as the manufactured design).
3. Lamination stack length.
4. Stator outside diameter.
5. Stator bore diameter.
6. Air-gap length.
7. Number of stator teeth.
8. Number of rotor segments.
9. Stator tooth tip taper angle.
10. Rotor segment taper angle.
The stator tooth tip arc angle and rotor segment arc angle of six possible designs are
tabulated in Table 4.
Table 4 Stator tooth tip and rotor segment arc angles for chosen design possibilities for the 6-phase
segmented rotor SRM for further analysis
Stator Tooth Tip Arc Angle
Rotor Segment Arc Angle
Y1 [Degrees]
Y2 [Degrees]
Option - 1
24
24
Option - 2
25
25
25
27
Option - 4
25
28
Option - 5
25
29
Option - 6
26
27
Option - 3 (prototype
design)
Using Equation 2 to Equation 7, stator tooth width, overlap angles (X1 and X2) in the
aligned position and gap angles (y and z) in the unaligned position have been calculated
for each design. These are listed in Table 5 for each design option.
76

Table 5 Stator tooth width and overlap and gap angles for each design option
Stator
Tooth
Width [mm]
Gap Angle in
Gap Angle in
Overlap Angle
the unaligned
the unaligned
in the aligned
position - y
position - z
position - X1
[Degrees]
[Degrees]
[Degrees]
Overlap
Angle in the
aligned
position - X2
[Degrees]
Option - 1
9.51
Option - 2
11.09
10
12.68
11
Option - 4
13.47
1.5
0.5
11.5
5.5
Option - 5
14.26
12
Option - 6
13.47
2.5
0.5
11.5
5.5
Option - 3
(Mk3 design)
All FE work has been carried out in MagNet .Using a script written in Visual Basic, an
FE model for each design was automatically generated, the 2D FE simulation for each
design was run, and the flux-linkage characteristics in the aligned and unaligned positions
were extracted for each design. A comparison of aligned and unaligned magnetisation
curves for each design option is illustrated in Figure 43, which shows the total fluxlinkage against per coil current, where two adjacent coils (i.e. two adjacent phases) are
energised simultaneously. Unsurprisingly, as the gap angles (y and z) reduce the
unaligned inductance increases. Likewise when the overlap angles (X1 and X2) and the
stator tooth width increase, the aligned magnetisation curve also increases.
In Figure 43, the aligned curves for designs Option-4 and Option-6 are identical. This is
because both designs have the same stator tooth width, stator core-back depth, rotor
segment depth and overlap angles in the aligned position (note that they also have the
same slot MMFs).
77

0.8
Total Flux-Linkage [Wb]
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.000
2.000
4.000
6.000
8.000
10.000
12.000
14.000
16.000
18.000
20.000
Current per Coil [A]

Option-1 Unaligned
Option-5 Unaligned
Option-3 Aligned
Option-2 Unaligned
Option-6 Unaligned
Option-4 Aligned
Option-3 Unaligned
Option-1 Aligned
Option-5 Aligned
Option-4 Unaligned
Option-2 Aligned
Option-6 Aligned
Figure 43 Aligned and unaligned magnetisation curves for each design option where the slot fill factor
and the number of turns per coil are fixed
In practice the design choice depends on the application and can involve factors such as
those listed (but not limited to) below:
Target minimum output torque.
Target cost for the drive system.
Operating environment and available cooling options.
Target maximum output torque ripple.
Target minimum drive efficiency.
The best machine design solution may not satisfy all of the above and is usually a
compromise depending on the design requirements of the application. At the start of this
PhD project, there were no strict design requirements in terms of the machine design. The
comparison between the six design options was based on five factors (objective functions)
listed below.
78

Mean output torque capability: Theoretically, a design option can have a high
torque per copper loss value but it can also have a low average output torque level.
Therefore, as well as the torque per copper loss, the actual average output torque
capability should be considered in the comparison of different design options.
Mean output torque per copper loss: Every machine will have a thermal limit, the
limit where the drive operation becomes unrealistic due to increased temperature of
one or more parts of the drive system. In this context, torque per loss is an important
parameter in electric motor design, as it describes a motors ability to produce torque
as a function of heat. Winding copper loss is generally the predominant factor in
generating the heat in an electric motor unless the machine is running at very high
speeds where the level of iron losses becomes significant due to the increased
operating frequency.
Mean output torque per machine total mass (iron laminations and copper
windings): The weight of the electric motor can also be considered an important
design criterion and is especially significant in application areas such as the
automotive and domestic appliances sectors. Therefore, another comparison factor the torque per total machine mass should be considered.
Size and cost of the power electronics based on estimating the VA requirements
from the linear unaligned inductance: The size and cost of a motor drive is
generally represented by its Volt-Ampere (VA) requirements. In the case of the SRM
the unaligned linear inductance of the machine has a direct impact on the VA
requirements of the drive system (The higher the unaligned inductance the larger the
VA requirements). For this reason, the unaligned inductance of each design option at
the same current level (10A) has also been taken into consideration whilst comparing
the performance of different design options. It must be said that the machine design
with the lowest unaligned inductance does not necessarily result in a reduction in the
cost and the size of the power electronics. The device selection has a quantized
characteristic. There needs to be a step change in voltage and / or current requirements
before a device from the next available family is required (larger or smaller).
79

Relative comparison of iron losses: Copper loss is certainly not the only electrical
loss encountered in an electric machine. The iron core of the machine is also a source
of loss. The absolute values for both the hysteresis and the eddy current losses have
not been calculated, but by using the flux density levels and the loss coefficients for
the eddy current and hysteresis losses in various parts of the magnetic circuit the iron
losses for each design option have been estimated relatively with respect to Design
Option-3.
Table 6 summarises the comparison of the six design options. Results in Table 6 have
been normalised with respect to the prototype design. Except the unaligned inductance
ratio, all results in Table 6 have been generated for a fixed coil current of 20A. The
unaligned inductance ratio has been given at 10A of coil current in order to ensure
comparison is based on the linear part of the unaligned curve.
Table 6 Summary of the performance comparison of the six design options (results normalised with
respect to the prototype design)
Option - 1
Option - 2
-23.71
-11.42
Torque
per
Copper
Loss
Ratio
[%]
5.64
2.87
Option - 3
(prototype
design)
0.00
0.00
0.00
0.00
0.00
0.00
Option - 4
Option - 5
Option - 6
5.01
9.10
4.28
-3.07
-8.66
-6.15
3.06
5.10
3.09
5.94
14.06
8.36
4.25
8.61
3.11
4.29
8.72
3.22
MACHINE
DESIGN
OPTION
Mean
Output
Torque
Capability
Ratio [%]
Torque
per Total
Machine
Mass
Ratio [%]
Unaligned
Inductance
Ratio [%]
Eddy
Current
Loss
Ratio
[%]
Hysteresis
Loss Ratio
[%]
-18.17
-8.04
-16.53
-8.34
-19.27
-10.38
-19.19
-10.24
As the above mentioned objective functions that help with choosing the optimum
machine design can be of conflicting nature, ultimately the design choice turns into a
multi-objective optimisation exercise that aims to maximise the average torque output,
minimise the copper and iron losses for the same output power, increase the average
torque output per unit machine mass and minimise the VA requirements of the drive.
Note that there can be other design requirements depending on the application.
Unfortunately it is unlikely that there is a single solution that simultaneously minimises or
80

maximises each objective to its fullest. There usually is a set of solutions with trade-offs
to be made. For example a design solution might have the highest average output torque
at the expense of increased copper losses.
Given that there were no specific design targets at the beginning of this PhD project, in an
attempt to combine all of the objective functions into a single functional form and aid
with the decision making in choosing a machine design option out of six possibilities,
Equation 8 has been proposed.
OptDesign = ToutRatio + ToutPcuRatio + ToutMassRatio + IndRatio (1)
Equation 8
All values used in Equation 8 are in [%] with respect to Design Option-3, the higher the
OptDesign value the more favourable the design option. In Equation 8 OptDesign
(Optimum Design) is a number that is used to quantify how good a particular machine
design is. ToutRatio is the percentage comparison of the design options in terms of
average output torque, ToutPcuRatio is the percentage comparison of the design options
in terms of average output torque per machine copper loss, ToutMassRatio is the
percentage comparison of the design options in terms of average output torque per total
machine mass, and IndRatio is the percentage comparison of the design options in terms
of unaligned linear inductance (representing VA requirements of the motor drive).
IndRatio has been multiplied by (-1) in order to address the fact that the unaligned
inductance should ideally be minimised. Each variable in Equation 8 can have a
weighting factor depending on the application. For example, if the cost and size of the
power electronics is the primary concern, then the IndRatio value can be multiplied with a
number bigger than 1.0 to increase its influence on the final machine design. Here no
distinction has been made between the four factors (i.e. all factors has been given the
same importance and consequently each uses a weighting factor of 1.0). The results for
OptDesign value for each design option are tabulated in Table 7. Note that the eddy
current and hysteresis losses are not part of Equation 8 and their effects are not reflected
in Table 7, as the actual values for these loss components were not calculated. However,
both core loss components can still be of help in choosing the ultimate machine design.
81

Table 7 Comparison of OptDesign parameters for each design (results are normalised with respect to
Design Option-3)
MACHINE DESIGN
OPTION
OPTIMUM
PERCENTAGE
[%]
Option - 1
Option - 2
-19.7112
-8.2552
Option - 3
(Mk3 design)
Option - 4
Option - 5
Option - 6
0.0000
-0.9366
-8.5256
-7.1397
According to Table 7 Design Option-3 is the best solution but only marginally compared
to Design Option-4. However, considering the results presented for the comparison of the
eddy current and the hysteresis losses in Table 6 (Design Option-3 will potentially
generate approximately 4.3% less eddy current and hysteresis losses compared to Design
Option-4) Design Option-3 becomes the preferred option for the final design. Design
Option-3 was indeed the prototype design.
The flux density plot of the prototype design in the aligned position with 20A of coil
current is depicted in Figure 44.
Figure 44 Flux density plot of Mk3 design in the aligned position with 20A of coil current (adjacent
coils are excited simultaneously)
82
4.3 DESIGN OF THE 6-PHASE TOOTHED ROTOR

SRM
To be able to compare the 6-phase segmented rotor prototype SRM with a conventional
toothed rotor SRM, a 6-phase conventional toothed rotor SRM has been designed. Many
papers have been published outlining the design of conventional toothed rotor SRMs [18,
40, 41].
The basic design ratios are given as:
Tooth width to pole pitch ratio.

Split ratio (ratio of stator outside diameter to rotor outside diameter).
Ratio of stator core-back depth to half of the stator tooth width.
Air-gap length.
A rectilinear model of the conventional design with the critical dimensions listed in Table
8 is given in Figure 45. To be able to carry out a direct comparison with the segmented
design, the machine outer diameter (OUT), outer rotor diameter (R1), the stator core
back depth (CBDS), the air-gap length (Lg) and the lamination stack length are all kept the
same as those of the segmented design. Tooth width to pole pitch ratio (TWS / S) has
been set towards the upper end of the suggested ranges to maximise the torque production
capability of the machine [38]. Thus TWS / S is selected to be 0.4 for this design, which
yields a stator tooth width of 11.41mm.
The recommended ratio of stator core-back depth to half of stator tooth width was in the
range of 1.1 to 1.3, which would give approximately 6.3mm to 7.4mm of stator core-back
depth. However this took no account of the acoustic noise. The rotor tooth width has been
selected to be equal to the stator tooth width. Initially the rotor core-back depth (CBDR)
has been set equal to the stator core-back depth (CBDS). Initial dimensioning of the
toothed rotor design is tabulated in Table 9.
83

Table 8 Critical design parameters for the conventional toothed rotor 6-phase SRM
CBDS Stator Core-Back Depth

TWS
Stator Tooth Width
Stator Pole Pitch
Stator Bore Diameter
R1
Rotor Outside Diameter
R2
Rotor Inside Diameter
shaft
Shaft Diameter
CBDR Rotor Core-Back Depth

TWR
Rotor Tooth Width
Lg
Air-gap Length
OUT
Machine Outside Diameter
hS
Stator Tooth Height
hR
Rotor Tooth Height
Figure 45 Rectilinear representation of the toothed rotor machine design variables
84

Table 9 Machine dimensions for the 6-phase SRM with conventional toothed rotor
CBDS [mm]
11.0
TWS [mm]
11.41
S [mm]
28.525 (Stator Pole Pitch)
S [mm]
91.4
R1 [mm]
90.8
R2 [mm]
58.0
shaft [mm]
36.0
CBDR[mm]
11.0
TWR [mm]
11.41
Lg [mm]
0.3
OUT [mm]
150
hS [mm]
18.3
hR [mm]
16.4
Similar to the method used for the segmented rotor design, the conductor diameter and the
number of turns per coil have been calculated assuming a peak current density of 10.0
[A/mm2] and a peak phase current of 25.0 [A]. The conductor diameter is calculated to be
1.4mm and the number of turns per coil is 50. A slot fill factor of approximately 0.5 can
be achieved.
The design with the machine dimensions listed in Table 9 has been modelled in 2D FE.
Like the segmented rotor design, two adjacent phases have been energised simultaneously
in the FE simulation. Figure 46 illustrates the average vector potential against MMF
curves for the 6-phase toothed rotor design with 11.0mm rotor core-back depth.
85

0.09
0.08
AVERAGE VECTOR POTENTIAL
Unaligned
0.07
Aligned
0.06
0.05
0.04
0.03
0.02
0.01
0
0
200
400
600
800
1000
1200
1400
1600
MMF(At)
Figure 46 Magnetisation curves in the aligned and unaligned positions for the conventional toothed
rotor 6-phase SRM with 11.00mm of rotor core-back depth
Initially the rotor core-back depth (CBDR) is selected to be equal to the stator core-back
depth (CBDS). In order to have a mechanically stiffer rotor design, the core back depth
was increased without affecting the unaligned curve considerably. It was consequently
found that the rotor core-back depth could be increased to as much as 18mm without
affecting the aligned and unaligned magnetisation curves significantly.
As can be seen in Figure 47 the tooth tip contains a tang, the main purpose of which is
to locate a wedge to hold the windings in place. The tang will slightly raise the unaligned
curve as in this position it shortens the gap between the stator and rotor teeth at a time
when they should be as far apart as possible. However, the tang will slightly increase the
aligned curve as well, since the overlap angle between rotor and stator teeth increases in
this position. Therefore these two effects are expected to cancel each other out.
86
Figure 47 A Single stator tooth of the 6-phase conventional toothed rotor SRM showing the stator
tooth tip tang
2D magnetic flux plots of the 6-phase conventional rotor machine in the aligned and
unaligned rotor positions are presented in Figure 48a and Figure 48b, respectively. Figure
49 depicts the magnetisation curves of the 6-phase conventional rotor machine from
unaligned to aligned rotor position.
87
Figure 48 Magnetic flux plots of the 6-phase 12-10 short-pitched conventional toothed rotor SRM
with two phases excited simultaneously in the aligned position (a) and in the unaligned position (b)
0.08
AVERAGE VECTOR POTENTIAL
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
0
200
400
600
800
1000
1200
1400
MMF(At)
Unaligned
10Degrees
2Degrees
12Degrees
4Degrees
14Degrees
6Degrees
16Degrees
8Degrees
Aligned
Figure 49 Magnetisation curves of the 6-phase 12-10 SRM with conventional toothed rotor.
Magnetisation curves from unaligned to aligned position with 2o stepping shown
The comparison of the segmental and toothed rotor designs will be presented in Chapter
10.
88
4.4 SUMMARY
Design processes for the six-phase segmental and toothed rotor 6-phase SRMs have been
presented in this chapter. Initial design for the segmental rotor machine based on simple
design equations has been verified against the design options available within the design
space. The design options have been compared based on several performance parameters
and an equation based on these performance parameters has been recommended in order
to help with choosing the best design option for the segmental rotor machine. The toothed
rotor machine design has been based on well-known design equations sourced from
several publications. The rotor core-back of the toothed rotor has been made as large as
possible without compromising the torque capability of the machine. Stator core-back
depths of both segmental and toothed rotor machines have been intentionally made larger
than the electromagnetically needed values. This was to increase the stiffness of the
machine and consequently reduce the acoustic noise of the designs.
89
Chapter 5 MACHINE CONSTRUCTION TECHNIQUES & TEST RIG

DESCRIPTION
CHAPTER 5
5 Machine Construction
Techniques and Test Rig
Description
5.1 MACHINE with SEPARATE STATOR CORES
Regardless of windings being multi-tooth or single-tooth, the achievable slot fill factor for
conventional stator structures is usually no more than 40%. This is due to the narrow
space available to thread a winding nozzle on the inside of the stator, which imposes
constraints on the tightness of the coils and also on the speed of the winding operation.
The effect gets worse with increased slot number as the dead space rate within a slot
increases. Also with increased slot number, the electromagnetic designer is limited to a
slot opening width of at least the size of the winding nozzle, as otherwise winding
operation cannot be possible. A typical winding process on a conventional stator structure
with a nozzle is depicted in Figure 50.
90

DESCRIPTION
Figure 50 Graphical representation of typical winding operation with a nozzle threaded through the
stator slot opening
For a given number of turns, if the slot fill factor can be increased, one can potentially
increase the conductor diameter and reduce the coil resistance. Having shorter endwindings can also reduce the total length of windings. Compared to single-tooth windings
multi-tooth windings inherently have longer axial overhangs and hence are more resistive.
Therefore using single-tooth windings will also reduce the coil resistance and hence the
copper losses.
Separating the stator core into separate pieces in order to increase the slot fill factor is
certainly not a novelty in the electric motors world. As well as increasing the slot fill
factor, separating the stator core also simplifies the coil winding and insertion process and
reduces the assembly time and cost. Hoefer and Jack [157] have successfully used the
separated stator core principle for a high-speed brushless permanent magnet (PM) motor
for a vacuum pump using soft magnetic composites (SMC) as the stator core material.
This high-speed PM motor used three separate stator teeth, each carrying a separate
winding. A two-piece back iron clamp was used to hold the stator teeth together. This
motor is depicted in Figure 51.
91

DESCRIPTION
3D Shaped
SMC Tooth
SMC Back
Iron Ring
Wound and
Preformed
Coils
Figure 51 Ulrike Hoefer's high speed permanent magnet vacuum pump motor made of SMC core
with three separate teeth, two back-iron rings and preformed windings [157]
Concerning productivity, the non-separate core consists of a small number of parts, thus
assembly is relatively easy. As the separate core consists of a large number of parts,
assembly is difficult, and productivity is unfortunately poor. The number of parts also
influences the core precision [158]. Because a conventional stator core structure is
stamped by one singular punch, the circularity is precise. On the other hand, in the
separate core case, the processing errors in the contact surfaces can accumulate, and as a
result, it is not easy to make the circularity high. In terms of motor characteristics, along
92

DESCRIPTION
with the rigidity, high circularity of the core is important for having small noise and small
vibration. In mass production, keeping the circularity costs a lot for separate core
structures.
Mitsubishi has been successful in utilising separated stator core structures in a number of
their commercial motors. Researchers at Mitsubishi claimed that by winding at highdensity and carefully redesigning the motor electro-magnetically they were able to reduce
the copper losses by 32.2 % and iron losses by 11.9 % [158]. The copper loss reduction
was achieved by having approximately 75 % slot fill factor compared to 59 % attainable
with conventional single punch stator structures thanks to the ingenious joint-lapped
core. Mitsubishi since used the joint-lapped core approach in a wide range of their
commercial products such as air-conditioner compressor motors, elevator motors, and
exhaust gas re-circulation valve motors [159-161]. The joint-lapped core principle has
made it possible to solve the problems of divided cores, securing the same amount of
space for the winding operations as in the divided-core approach without impacting the
ease of assembly and connectivity. With joint-lapping the core, the opening angle of the
teeth is increased. Effectively, joint-lapping allows opening up a core section for high
speed, high density winding operation and once the winding on a tooth is completed
another section of the core is bent back and opened up for winding. Once all teeth are
wound in similar fashion the complete unit is bent back to form a round shape. As a
result, even thick wires can be wound in alignment without coil distortion, which is
common in separate cores. The start and end of the joint-lapped core are connected
together by welding, caulking or shrink fitting into the frame. Moreover joint lapping can
allow the use of narrower slot openings if it is desired for the design as well as increasing
the speed of winding operation by using a circular motion flyer. The winding process on a
joint-lapped core and a picture of such a core are depicted in Figure 52.
93

DESCRIPTION
Figure 52 Winding process with joint lapped core and picture of joint lapped core [158]
5.2 CONSTRUCTION OF THE PROTOTYPE

MACHINE
The prototype machine has six phases with an inner rotor formed by 10 discrete segments
and an outer stator formed by 12 separate cores. Rotor and stator laminations were both
laser cut. Standard 0.5mm lamination steel (M250-35A) was used.
94

DESCRIPTION
5.2.1 Stator Construction

The temperature of the stator windings is the limiting factor for the maximum output
torque of an SRM. This temperature is mainly determined by the copper losses and the
thermal resistance between the windings and the stator laminations, assuming that the
laminations are sufficiently cooled. Furthermore, machine efficiency will be affected
adversely by excessive winding losses [162]
The use of separate tooth laminations greatly increases the possible slot fill factor, but
does introduce a series of joints around the core back of the stator. These joints create
small gaps, which increase the reluctance of the flux path and are detrimental to
performance. In conventional SRMs the magnetic flux has to cross a series of these joints
and the subsequent reduction in performance is often thought to outweigh the benfits of
increased slot fill. However, for segmental rotor SRMs the flux paths are very short and
flux only ever has to cross a single joint. This method of construction was consequently
selected.
The machine stator was constructed from 12 separate cores: no joints were used to hold
the separate stator cores together, as it was believed that the forces exerted by the
machine frame structure would be enough to hold the divided core pieces together. Each
stator core occupied 30 mechanical degrees. A single stator core is illustrated in Figure
53.
Figure 53 Picture of single stator core lamination for the 6-phase 12-10 segmented rotor prototype
machine
95

DESCRIPTION
The stator of the prototype machine was constructed in three stages: Assuming a
lamination packing factor of 97 % (3% occupied by glue, air and iron lamination coating),
each stator tooth of length 150mm required around 291 laminations, each nominally 0.5
mm thick.
Gluing 291 laminations all at once and controlling their radial and horizontal alignment
required a systematic approach. To ease gluing stator laminations and forming of 150mm
stacks, the process was divided into two stages. In the first stage a jig (Figure 54) was
designed to glue together 50mm long stator tooth pieces.
Figure 54 Jig used to produce the 50mm long stator tooth pieces
Each 0.5mm thick lamination was coated with a thin layer of adhesive and then placed
inside the 50mm long jig. Then the glued laminations were cured in an air-circulating
oven at 80oC for about an hour. Pictures of the 50mm piece inside the jig and finished
piece are given in Figure 55.
Figure 55 50mm stator piece inside the jig and a close-up on the glued and cured 50mm piece
To achieve satisfactory bond strength with the 50mm pieces as well as the final 150mm
pieces, the glue used was a formulated epoxy adhesive from Magna-Tac that was ideal
96

DESCRIPTION
for bonding laminations. To form twelve 150mm long stator teeth, 36 of these 50mm
tooth pieces were needed. Before bonding the 50mm laminations to form the stator teeth,
excess glue on the laminations was carefully removed. Another jig (Figure 56) was then
used to produce the 12 stator teeth.
Figure 56 Pictures of the jig used to produce the 150mm long stator tooth pieces
It was required to thoroughly dry the adhesive before curing for optimum bonding. This
was accomplished by forced drying in an air-circulating oven at approximately 90oC for
about 15 minutes. The adhesive could be cured without a prior drying period, however for
optimal results drying was highly recommended [163]. After coating 50mm laminations
with the adhesive and drying thoroughly, three 50mm long lamination stacks were put
together and cured at 170oC for at least two hours. The final product at the end of this
stage is pictured in Figure 57.
Figure 57 Pictures of the complete stator tooth pieces
97

DESCRIPTION
Note that the positioning of the pole face of each lamination was controlled as well as two
side faces (where the teeth would be glued together to form the complete stator
assembly).
5.2.2 Windings
A slot fill factor of 0.5 was thought to be realistic without being too optimistic for
calculating the initial conductor diameter. The maximum permissible current density was
chosen to be 10A/mm2, with a peak current of 25A a wire of diameter of 1.36mm was
required. The nearest wire size to this was 1.4mm, corresponding to 44 turns per coil in a
slot area of 270 mm2.
The above fill factor was achieved with difficulty. Bulging and curving of conductors due
to the relatively long machine axial length and the relatively thick conductor diameter of
1.4mm greatly affected the maximum achievable slot fill factor. It is hypothetically
mentioned in [156] that the curve of coils is caused when the wires are drawn and wound
in such a manner that the residual distortion is not eliminated completely due to the
tension.
In the same paper the reasoning for the bulge in wires is attributed to the hardening due to
continuous plastic deformation during winding process. The curving and bulging of wires
are depicted in Figure 58a and Figure 58b, respectively. The author believes that with a
more careful winding approach the slot fill factor can easily be increased further. A selfbonding enamelled wire with thermosetting bond coat from VonRoll-Isola was selected.
The aim was to increase the retention of windings around separate stator cores and to
keep the bulging of windings to a minimum. Before the winding operation, inside
surfaces of each stator tooth were covered with slot liners to avoid shorting between stator
windings and stator core. Figure 59 pictures a stator tooth covered with slot-liner.
98

DESCRIPTION
Z-axis
Y-axis
X-axis
WINDING
NOZZLE
Z
X
WINDING
Z
Y
CURVE
BULGE
(a)
(b)
Figure 58 Curving and bulging of conductors during winding operation
Figure 59 Stator tooth piece covered with slot liner against winding to lamination shorts
99

DESCRIPTION
A small plastic piece cut to the shape of the stator tooth is glued to both ends of the stator
and can be seen in Figure 59. This is to ease the winding operation and also to protect the
windings against the sharp edges of stator laminations. The windings were hand-wound
using the arrangement pictured in Figure 60.
Figure 60 Manual winding of the stator pieces
Another jig (Figure 61) was used to press the bulging windings from both sides. Once
tightly placed inside this jig, the bonding of the wires (Figure 62) was achieved by
applying energising the coil. The temperature of the wires was raised within 30-60
seconds to about 230oC and kept at this temperature for about 2-3 minutes. The final form
of a stator tooth after winding and curing is pictured in Figure 63.
Figure 61 Jig used to press the bulging windings and make them ready for bonding
100

DESCRIPTION
Figure 62 Bonding of wires through applying a current surge
Figure 63 Stator tooth with bonded windings
Once 12 teeth were wound they were glued and shrink fitted into the aluminium frame of
the machine, pictured in Figure 64. Finally both ends of each coil were brought out of the
frame to allow full flexibility in the way the coils could be connected together (e.g. series,
parallel, series and parallel).
101

DESCRIPTION
Figure 64 Frame and stator assembly - each coil connection brought out to have full flexibility in
connection
5.2.3 Rotor Construction

The rotor construction process was the same as the one used for prototyping the three
phase segmental designs [12, 23]. The segment laminations were mounted on a steel
shaft, which was non-magnetic. Having a non-magnetic steel shaft ensures that there is no
magnetic flux path formed between the rotor segments that can affect the flux-linkage
characteristics and hence the performance of the motor adversely. Circumferential
location was achieved by using flat bottoms on the rotor segment laminations. Radial
retention of the rotor segments was achieved by non-magnetic steel wedges, which were
dovetailed into protrusions at the bottom of each segment, and then bolted on to the rotor
shaft. The construction process is shown pictorially in Figure 65 and the final form of the
rotor after assembling and machining is given in Figure 66. For the same inner and outer
machine dimensions and approximately the same output torque, dovetailing rotor
segments in [12] allowed safe operation up to a speed of 6000 rpm.
Figure 65 Construction of the segmented rotor of the prototype 6-phase 12-10 SR machine
102

DESCRIPTION
Figure 66 Finished and machined segmented rotor assembly of the prototype 6-phase 12-10 SR
machine
This rotor design is also used successfully in [136] as an alternative to Pollocks flux
switching machine.
5.3 TEST RIG DESCRIPTION

The test rig used for the prototype 6-phase segmented rotor SRM comprises the following
main components:
Segmented rotor SRM.
High voltage power converter.
DSP based control electronics.
Mechanical load.
Host PC and DSP emulator.
A schematic representation of the test rig is illustrated in Figure 67. This section briefly
discusses the main parts of the drive system along with the auxiliary equipments used for
measurements and protection.
103

DESCRIPTION
SINGLE-PHASE
VARIAC AND
RECTIFIER UNIT
NONINDUCTIVE
RESISTOR
LOAD BANK
TORQUE
TRANSDUCER
ADJUSTABLE
SUPPORTINGS
FLEXIBLE
COUPLINGS
SWITCHED
RELUCTANCE
MOTOR
DC LOAD MACHINE
ENCODER
BED PLATE
Position
feedback
Emulator and
serial link
connections
RACK SYSTEM
Winding
connections
Current
feedback
Gate
signals
HOST PC
Measurements
HIGH VOLTAGE
INVERTER
POWER ANALYSER
Analogue O/P
OSCILLOSCOPE
Figure 67 Schematic representation of the test rig
5.3.1 Load Arrangement

Throughout this work a separately excited, brushed DC machine was used as the load
machine in parallel with a non-inductive load bank as the load system to test the prototype
machines. The load torque could be adjusted via the armature current or by changing the
stator electric field. The armature was connected via a load bank connected in parallel
with the DC machine; the field excitation was supplied via a variable single-phase
104

DESCRIPTION
transformer and a rectifier. Figure 68 shows the schematic layout of the load arrangement.
A torque transducer capable of measuring rated torque up to 50 Nm was used to measure
the torque output of the electrical machine. To monitor the voltage, current and torque
waveforms 2-channel oscilloscopes were used. Finally for power flow measurements a
power analyser was used as part of the test rig.
Figure 68 Schematic description of the load arrangement dc machine with a resistive load bank
connected in parallel
5.3.2 High Voltage Power Converter

The electrical power converter consisted of an input rectifier, a dc link filter, a dc link
dump and inverter bridges (Figure 69). The input voltage to the power converter was
supplied via a three-phase variac connected to the input of the converter. The three-phase
variac was effectively used to adjust the level of dc-link voltage. A simple three-phase
diode bridge was used to rectify the three-phase supply and to provide the dc-link voltage.
Each leg of the rectifier consisted of a single-phase diode bridge, with diodes connected
in pairs and in parallel, thus increasing the current rating of the device.
105

DESCRIPTION
Figure 69 Electric power converter
A series inductor was used to limit the inrush current and to improve the line current form
factor. In conjunction with the dc-link capacitor, the inductor formed the dc-link filter,
which reduced voltage and current ripple. The main functions of the dc-link capacitor
were to act as an energy storage reservoir and to present low supply impedance to the
inverter bridges.
Diodes (phase diodes) were connected in series with phase windings to eliminate bi-polar
phase currents. These diodes were mounted in an aluminium box near the motor end of
the drive system. The connections were then taken out of this metal box (terminal box Figure 70) to have flexibility in connecting the phase windings. As mentioned earlier both
ends of all machine coils were brought out to have full flexibility in connecting the phase
windings (Figure 71).
106

DESCRIPTION
Figure 70 Terminal box connections - phase diodes are mounted on an aluminium sheet in this box
and all coil connections are brought inside this box
107

DESCRIPTION
Figure 71 Phase connections see Figure 70 in relation to some of the labelling
108

DESCRIPTION
5.3.3 Control Electronics

Control was centred on the Texas Instruments TMS320C31 DSP. This was a 32-bit
floating point processor operating at 40MHz. Other major parts of the control electronics
were the analogue/digital input card, the power device control card, over-current and
over-voltage monitoring cards and the analogue output card. All cards were produced by
previous research projects [89, 165]. The cards communicated via a 2x32 backplane and
were mounted in a 19 rack cabinet, with connections to the host PC, speed encoder and
power converter. Each card was mapped to the main control software as memory blocks
in the DSP. The host computer was used for storing files and writing, compiling and
downloading programs to the control electronics. A serial connection to the DSP allowed
for data to be transferred from the DSP to the PC.
Rotor position was measured using a 12-bit gray code output absolute encoder. Gray code
output of the encoder was decoded into binary format with a look-up table within the
control software.
The controller DSP was supported by a Field Programmable Gate Array (FPGA) based
PWM card. The PWM card was capable of controlling up to 6 pairs of devices
independently with PWM frequencies of up to 40 kHz. PWM control parameters such as
triangular waveform resolution, dead-time, etc. were all software programmable. The
PWM card was also able to produce a synchronisation pulse once per PWM cycle to
allow the phase currents to be sampled at the correct instants. Schematic explanation of
the parts of the control electronics is given in Figure 72.
109

DESCRIPTION
Figure 72 Control electronics
110

DESCRIPTION
Three analogue-to-digital conversion (ADC) cards were used to transform the three-phase
current signals into 10-bit digital code. A Matlab based GUI (Graphical User Interface)
was created to set control variables such as current demand, PID controller gains, and
conduction and advance angles.
Three single and one multiple output digital-to-analogue conversion (DAC) cards were
used to output and monitor signals such as phase current, torque, and flux. Other cards
were: the LED display variable input or output card, the speed encoder card and the
power supply card.
An over-current and over-voltage monitor card worked in parallel with the over-current
and over-voltage detection circuits and were used to protect the system by forcing the
gate commands inactive, and thus forcing some or all of the power devices to turn off.
Over-current detection was achieved by monitoring the analogue output of each current
transducer and continuously comparing it to a threshold value. A similar comparison was
made on the DC link voltage to check for over-voltage conditions caused by excessive
regenerative braking of the machine during test.
5.3.4 Drive Software Explanation

French [166], Clothier [89] and Weiner [137] have all previously contributed to the
programming work of the test-rig. The whole control system was synchronised to the
triangle wave on the PWM card, with synchronisation pulses sent out at the peak of every
triangular wave to initiate the first current control loop in the DSP. The function of this
loop was to sample the encoder position and determine the phase current demands. The
second control loop was the one that actually contained the PID current control. This was
initiated internally in the DSP by a software timer, which in turn was set by the hardware
interrupt supplied by the synchronisation pulse. This software timer also output a
hardware pulse to the ADC cards that initiated the sampling of currents at this point. The
current error(s) could then be calculated, and voltage references generated via the PID
gains. Finally these references were sent to the PWM controller card and were
implemented there at the next synchronisation pulse [89].
The drive system software required the initial unaligned position of the machine. To
determine the initial rotor position for the 6-phase 12-10 machine a pair of phases were
energised when the machine was stationary. The DSP was used to output the rotor
111

DESCRIPTION
position in digits. (between 0 and 4096 (12 bits) for a 12-10 machine). This initial
unaligned position in digits was then entered into the relevant part of the control software.
During operation, the position information was read by the 12-bit gray-code encoder. The
encoder reading then went through a look-up table where the gray-code was converted
into binary. Depending on the initial position value, the number of phases, advance angle
and the rotor position, the software decided which pair of phases needed to conduct. Each
line current was controlled separately and each had its own PID reference voltage and
control routine. Once the pair of phases which needed conducting was determined, each
line current demand was defined and with the user-defined PID gains and measured line
current values, line reference voltages were generated based on each line current error.
These reference voltages (with 8-bit resolution) were then written to the PWM card where
they were compared with an 8-bit 20kHz reference triangular waveform. Depending on
the outcome of this comparison MOSFET gate signals were generated and sent to the
relevant MOSFET modules. This process is explained schematically in Figure 73. The
phases of the 12-10 machine go through energisation, freewheeling and de-fluxing cycles
10 times in each mechanical cycle.
Figure 73 Implementation of current control with the drive software
112

DESCRIPTION
5.3.5 Prototype Testing

This section discusses the static magnetic and thermal characterisation of the prototype 6phase segmented rotor SRM.
5.3.5.1 Flux-Linkage Measurements

Flux-linkage is the product of magnetic flux () and the number of turns (N) linked by
that flux and is used as a key factor in determining the physical characteristics of any
machine, such as the torque capability. The flux-linkage characteristics of a machine also
play an important part during control system development (flux control, current shaping
for low torque ripple operation, etc.) and validating the performance of simulation tools.
It is thus important to obtain a full set of accurate flux-linkage characteristics. Measured
flux-linkage characteristics can also be used for verifying the machine design process.
The prototype motor has 12 stator teeth and 10 rotor segments. Therefore one electrical
cycle lasts 36o mechanical and complete characteristics of the motor can be obtained by
taking measurements within this range. Flux-linkage measurements in this work are based
on the Volt-Seconds method, where stator flux linkage is calculated from voltage and
current measurements on the phase winding circuit [154]. This can be achieved simply by
applying a voltage source to the phase winding, after which the corresponding
instantaneous terminal voltage v(t ) and instantaneous current i(t ) measurements are
taken. The instantaneous flux linkage (t ) can then be calculated by Equation 9.
(t ) = (v(t ) i (t ) R)dt + 0
Equation 9
where R is the measured resistance of the phase windings and 0 represents the initial
flux linkage. The initial value of the flux-linkage 0 is very significant only in
electromagnetic circuits containing remnant fluxes, such as permanent magnets. In the
case of an SRM, this initial value is 0 as there is no magnetic flux in the steady state when
no current is applied.
113

DESCRIPTION
A GUI has been created to assist with the calculation of the flux-linkage characteristics of
the prototype motor from captured voltage and current scope traces. Details of the
operation of the Flux-Linkage Viewer GUI are given in the Appendix A. A full set of
magnetisation curves for the prototype 6-phase segmented rotor SRM is given in Figure
74.
0.7000
Total Flux-Linkage (Wb)
0.6000
0.5000
0.4000
0.3000
0.2000
0.1000
0.0000
0.000
2.000
4.000
6.000
8.000
10.000
12.000
14.000
16.000
18.000
20.000
Current (A) - per phase

Unaligned
10degrees
2degrees
12degrees
4degrees
14degrees
6degrees
16degrees
8degrees
Aligned
Figure 74 Measured flux-linkage curves for the prototype 6-phase 12-10 segmented rotor SRM (Mk3
design) with two phases excited at the same time. Each curve is at a constant position, running from
unaligned to aligned in 2.0 degree steps
Comparison of the aligned and unaligned magnetisation curves obtained from

measurements to those predicted in two-dimensional FE analysis is given in Figure 75.
The differences between the two sets of flux-linkage results are believed to be due to the
following reasons:
The unaligned inductance is higher in reality due to the end winding leakage effects.
When the rotor is at or near the unaligned position the inductance can increase by 2030% due to fringing of magnetic flux in the axial direction. Unfortunately 2D FE
analysis cannot take this into account and results in underestimation of the
magnetisation curve at the unaligned position.
114

DESCRIPTION
The inductance in the unsaturated section of the aligned curve is higher than predicted.
This is thought to be due to the air gap being smaller than the desired value of 0.3mm.
This can be explained by the nature of building the machine. It is believed that the
wire erosion process that was used to form the geometry of the lamination, the
machining process to achieve the correct stator and rotor bore diameters and the
method of locating the rotor within the stator all have an effect on having an air-gap
other than the desired value of 0.3mm.
There is a difference between the measured and theoretical results in the region of
saturation of the teeth and the core back in the aligned magnetisation curve. In this
case the measured results are lower. This can be attributed to two factors. One is that
the achieved lamination packing factor is different to the one assumed in the FE
analysis and the other is that the BH characteristics of the core material in the FEA
package has finite amount of error, especially in the saturation region.
Figure 75 Comparison of 2D FE predicted aligned and unaligned magnetisation curves with those of
measurements
115

DESCRIPTION
5.3.5.2 Static Torque Measurements

Torque characterisation for an SRM can be achieved either indirectly by using other
measured motor parameters or directly by using torque transducers as torque sensing
elements. The former method exploits the close relationship between the flux-linkage, coenergy and torque characteristics of an SRM. Knowing that torque is a function of coenergy, which is in turn a function of flux linkage, it is possible to derive the torque
characteristics from the flux linkage characteristics. However, such indirect torque
characterisation methods can be inaccurate in nature due to errors that arise in the
methods in which the flux-linkage is obtained. This can be experimental or through
mathematical calculations using Finite Element Analysis (FEA) tools. Even though the
errors in flux-linkage data are low, these will be accumulated and further amplified onto
the calculated torque characteristics when integration and differentiation are performed on
the flux-linkage data. Direct measurement techniques do not have errors introduced by
mathematical calculation or approximation. By attaching various sensors to the SRM and
performing measurements at different operating points, it is possible to collect torque,
current, and rotor angle data and form the torque characteristics of an SRM. Of course,
direct measurements will have measurement errors. However, these can be kept to a
minimum by careful consideration of the test set-up and measurements taken.
Windings of two phases of the prototype machine were connected in series and excited
simultaneously. The rotor angle was varied between the unaligned and the aligned
positions for a number of excitation levels, and the torques measured with a torque
transducer coupled to the shaft of the motor. The DSP and the position encoder were once
again used to continuously monitor the rotor position. Using the flux-linkage
characteristics produced by the FE method, torque-position characteristics are predicted
and a comparison made between the predicted and measured results. This is shown in
Figure 76. It can be seen that there is a good correlation between measured and predicted
static torque curves. Full static torque characteristics are given in Figure 77.
116

DESCRIPTION
Figure 76 Measured vs. predicted static torque characteristics of the prototype 6-phase segmented
rotor SRM at various excitation levels
Figure 77 Measured static torque characteristics of the prototype 6-phase segmented rotor SRM up
to 20A of excitation
117

DESCRIPTION
5.3.5.3 Thermal Characterisation of the Prototype Machine

Lastly, the thermal characteristics of the machine have been determined. A temperature
probe was placed on the frame of the machine and another one was embedded within the
phase windings. A DC power supply was used in constant current mode and the current
was set to 6.5A. All six phases were connected in series. Power was injected into the
machine for ninety minutes. The motor was subjected to forced-cooling with a fan.
Temperatures on two probes as well as applied voltage and current were recorded with a
data logger connected to serial port of a PC. The machine exhibited two thermal time
constants: one of approximately thirty minutes associated with the bulk heating of the
entire machine and frame and accounting for 28% of the steady-state temperature rise,
and one of approximately five minutes associated with heating of the winding and
accounting for 72% of the temperature rise. Based on the results it was determined that
100 degrees centigrade rise in winding temperature corresponded to approximately 459W
loss and an RMS current of 13.6A.
Accuracy of the thermal time constants derived from measurements can be observed in
Figure 78, where the lumped parameter curves are superimposed on top of the measured
results.
Figure 78 Measured and predicted temperature measurements of the windings and the machine
frame
118

DESCRIPTION
5.4 SUMMARY
Construction of the prototype segmental rotor 6-phase machine has been explained. The
stator of the prototype was built from separate cores in order to increase the slot fill factor
of the machine; a fill factor of approximately 0.5 has been achieved. The fill factor could
be further improved if the bulging and curving of the conductors could be controlled in a
better way. It is believed that with designs where thin conductors are required and the
machine design has a short stack length the fill factor can be improved to values above
0.6.
The components of the test rig including the control software have been briefly described.
The segmental rotor machine has been built and the static flux-linkage and torque
measurements have been verified against those predicted using 2D FE analysis. The
themal characteristics of the prototype machine have also been determined and presented
in this chapter.
119
Chapter 6 SIMULATION MODELS
CHAPTER 6
6 Simulation Models
6.1 INTRODUCTION
In conjunction with finite element analysis, simulating a drive system is highly desirable
as it permits going through many different design variations such as machine geometries,
inverter topologies and excitation patterns without having to physically construct and test
all of them. A simulation with high results accuracy and short simulation time is ideal. In
many cases though, there is a compromise between the accuracy and the simulation time.
Although every effort is made to achieve a fast simulation, the accuracy of the simulation
has been the main motive for this PhD project. A simulation which is accurate enough to
predict the experimental results will help in:
Eliminating lengthy experimental procedures.

Shortening the product development process.
Determining the optimal system design (including the machine, electronics and the
controller) easily by allowing the assessment of the system behaviour in response to
design modifications such as controller resolution, machine number of turns,
combinations of control angles (advance and total conduction angles), etc.
Reducing the risks associated with prototyping.
The simulation can be divided into distinct sections:
1. Input stage (rectifier and filtering)
120

2. Controller stage
3. Converter stage
4. Machine stage (including the mechanical system)
In this PhD the simulation work was focussed on the controller, converter and machine
modelling.
Simulation of an SRM can be developed with a low level programming language such as
FORTRAN or C/C++. However any additional module requirement will lengthen the
model development greatly. Spice, VisSim, MathCad, and MATLAB / Simulink are
other likely candidates to simulate an electrical machine drive system. Spice has the
disadvantage of being biased towards electronics simulation. MATLAB is a matrix
based programming and simulation environment. It is very flexible and offers different
solver options and has the ability to work with large data sets. These features make
MATLAB a strong candidate for electrical drive system simulation.
The simulation package introduced here is developed in the MATLAB / Simulink
environment. The main part of the simulation is realised in Simulink supported by
scripting in MATLAB m-file format. A Graphical User Interface (GUI) was also
generated to ease the interaction between the user and the simulation models. Simulink
is an environment where mathematical models of systems are represented with a
customisable set of block libraries connected together that establish mathematical
relationships between the blocks (based on the mathematical model of the system),
simulating and testing a wide variety of time-varying systems. Creating a simulation
model is as simple as dragging and dropping the necessary blocks from the library
browser of Simulink onto the graphical editor and connecting them with lines.
Simulink is integrated with Matlab, providing immediate access to an extensive range
of tools for algorithm development, analysis and creating batch processing scripts to run
simulations back to back. As mentioned earlier, Simulink provides different solvers
(computing the system dynamics over time using information contained in the model) to
support the simulation of a broad range of systems, including continuous-time (analog),
discrete-time (digital), hybrid (mixed-signal), and multi-rate systems of any size. Both
fixed step size and variable step size simulations are possible in a Simulink based
simulation.
121
6.2 SRM MODELLING and SIMULATION

6.2.1 Background
In contrast with other electrical machines, SRMs operate in high saturation with a
strongly non-linear current-flux relationship. Therefore the classical equivalent circuit
representations usually applied to the modelling of electrical machines cannot be easily
employed to SRMs as the inductance of the circuit is highly dependant on the rotor
position and the phase current. When modelling short-pitched winding SRMs, the current
in any one phase almost always falls to zero within an electrical cycle. Because this
current is the only source of flux for that phase, the phase flux also falls to zero when the
phase current falls to zero. The states are known at the start and at the end of any
electrical cycle. Thus each electrical cycle can be viewed in isolation. This makes the
simulation of short-pitched SRMs particularly simple, as the initial conditions of phase
flux and phase current are known at the start of fluxing any phase of the machine. This
eliminates the need for long transient simulation. Unfortunately, the same does not apply
to SRMs with mutually coupled phase windings. There are still periods of zero phase
current. However, due to mutual coupling, the flux in any one phase is a function of
currents in all other phases [91]. Therefore a phase with no current flowing in its
windings does not necessarily have zero flux linking them as well. Thus, at the start of
each cycle, the initial conditions are unknown and a number of cycles are needed until
steady-state operation is reached. Although when compared to simple short-pitched SRMs
this might seem as a nuisance, compared to other machine types it is not. For brushless
DC machines the initial conditions are usually unknown. As a result, a number of cycles
of simulation are needed in order to evaluate the performance under steady-state
conditions.
122
6.2.2 Review of SRM Simulation Methods

6.2.2.1 Early Work
In this section some of the early work relating to the modelling and simulation of SR
machines and drives is presented. Methods for the performance prediction of both shortpitched and fully pitched winding SRMs have been extensively developed over the past
two decades. Many methods have been proposed for modelling the torque-current
relationship of SRMs. Almost all of the early work relating to SRMs in general came
from Stephenson, Lawrenson and their colleagues in Leeds University. Work in [18, 167,
168] can all be seen as seminal examples defining the way for modelling and ultimately
simulation of SR machines and drives. Studying the motor's magnetic property is
essential for accurate modelling and control of an SRM. Generally speaking, if the aim is
to assess the motor performance, rather than the performance of an individual converter
topology and the associated controller, then the simulation is aimed at accurate modelling
of the motor magnetic characteristics with a simple but accurate modelling of the
controller.
Real-time simulations of SRMs are given in [169 and 170]. The model in [170] takes the
magnetic non-linearity of the SRM into account. However, the magnetisation curves were
approximated by polynomial functions. In this paper, the authors first obtained the
magnetisation curves from FE analysis and then approximated these curves with
polynomial functions in their Spice simulation. Moreover, only a single example was
given to compare phase current and voltage waveforms of the simulation and real-life
measurements. In [171] the SABER simulation package has been used to generate a nonlinear model of an SR drive system with an asymmetric half bridge inverter. The effects
of magnetic saturation, the results on motor performance with typical changes in supply
voltage, winding turns and switching angles were presented for a particular 2-stator-pole /
2-rotor-pole motor. With the use of SABER, Pollock et al [171] were able to account for
dynamic properties of devices such as voltage spikes, high frequency oscillatory features
caused by reverse recovery and stray capacitance. The claim was the ability to rapidly
change the control strategies and try different circuit topologies and assess their effect on
motor performance in SABER.
In [172] Pulle claims to have improved the method described by Stephenson and Corda
with the use of cubic splines. The method aims to represent the magnetisation curves of
123

an SRM with the use of cubic splines and deals with fitting segments of a curve
representing the flux linkage/current relationship with third order polynomials and
matching the first and second derivatives at the knots where the segments meet. The
generated flux-linkage curves correlate favourably with experimental data and are used to
simulate the motor as per the method described by Stephenson and Corda. Radun
proposed two separate analytical models based on basic magnetic field laws for
computing the magnetisation curves of SRM [173, 174]. In [174] Radun computes the
flux-linkage characteristics of SRMs for the aligned position and for the intermediate
rotor positions. The model proposed is not empirical and does not require any data from
the modelled machine or any FEA results. Radun predicts the unaligned curve of SRM in
[173] using the idea introduced in [174]. Even though the methods are analytically
straightforward and easy to implement with a computer program the methods do not
provide a unified theory of predicting the flux-linkage characteristics of SRMs. It requires
two different methods to generate the full set of magnetisation curves of SRMs.
Torrey et al [175, 176] focused on finding an analytical expression for the flux-linkage /
current / position data which would provide all of the flux-linkage-current information for
every rotor position in one summary equation. They derived a single equation to embody
the flux-linkage characteristics of an SRM and used the Fourier cosine series to represent
the coefficients of the equation. Although they achieved good curve fittings, the method
of representing the fitting coefficients introduced higher order harmonics that added
ripple to the predicted instantaneous torque production because of their rapid spatial
variation.
Miller et al [177] introduced an interesting method for the modelling and design of
SRMs. Their aim was to reduce the computation time substantially and generate a rapid
computer aided design at the expense of accuracy. Instead of representing the
magnetisation data of the machine by flux-linkage as a function of current, they modelled
the SRM by analytical piecewise first or second-order functions of flux-linkage against
rotor position with current as an undetermined parameter. The formulation was empirical
and the data for position other than the aligned and unaligned positions was found by
interpolation. The model forms the basis of the popular PC-SRD program. However, the
idea of modelling by the empirical equations presented by Miller et al in [177] is not easy
to comprehend and incorporate into new methods of modelling SRMs.
124

Xue et al [178] have introduced the application of two-dimensional bi-cubic spline
interpolation functions to the modelling and simulation of SRMs. Compared with the
other techniques they claimed that the new algorithm has given more accurate results.
None of the above mentioned methods for modelling and simulating the SRMs
considered the mutual coupling between phases. Lipo and Moreira [179] argued that the
presence of the mutual coupling between phases of a short-pitched winding SRM has
significant influence on the open circuit voltage waveform across each phase winding and
the turn-off time of the phase current (which would reduce torque and introduce more
losses). They also stated that the mutual coupling can also contribute to unwanted
conduction during the nominal OFF period of a phase winding, hence, also contributing
extra losses together with undesirable torque components. To address the issue, they
introduced a coupled circuit model and simulated a four-phase SRM with short-pitched
windings. Another example where the mutual coupling between machines was taken into
account during the modelling of the machine is given in [180]. Shuyu and Tseng
developed
their
model
using
few
sets
of
FE
computed
phase
flux
linkage/current/position data under the condition of simultaneous excitation of two

neighbouring phases. The combination of piecewise Hermite cubic spline and Fourier
series techniques was then used to represent each FE calculated flux linkage data set in
the dynamic simulation. The model was implemented in SABER for a 4-phase SRM
driven from a split dc-link inverter.
Unlike the short-pitched winding SRM, in a fully pitched winding machine the majority
of torque is produced from the variation of mutual inductance with position. This means
that the mutual coupling must be modelled very accurately, including the full effects of
saturation within these terms. As the flux linkage of any one winding is a non-linear
function of rotor position and currents in all phases, machines with fully pitched windings
are more difficult to model and study. Mecrow et al [91] proposed transformation
matrices to simplify the modelling of fully pitched winding SRMs. They broke down the
instantaneous flux-linkages and currents in each phase to give the flux and magneto
motive force (MMF) in each tooth of the machine. The flux and MMF could then be
linked in the model by a simple look-up table, determined from either simulation or
measurement. By having the transformation matrices, simulation of each phase could be
de-coupled from the other phases of the machine. Barrass [90] described a simulation
model for the fully pitched winding SRM where he utilised the idea of transformation
matrices in his simulation model.
125

Torrey and Xu [181] stated that the approach of Mecrow et al to the modelling and
simulation of fully pitched SRMs did not lead to a deeper understanding of the fully
pitched SRM. They used an FE circuit coupled simulation to eliminate the difficulty of
modelling mutually coupled phases of the fully pitched winding machine. They used an
FE solver (Flux2D) and a circuit simulator (Simplorer) simultaneously. However, their
machine model in FE was solved with a 2D solver ignoring 3D effects. Furthermore, the
overall simulation time was increased by the use of circuit-coupled FE simulation as at
every time step the simulation ran the FE model.
In [182] Torrey and Kokernak used flux tubes to generate a nonlinear magnetic circuit
model for a mutually coupled SRM. The solution was analogous to solving for node
potentials in an electric circuit. The model was claimed to be also suitable for
conventional short-pitched SRM geometries that required consideration of mutual
coupling effects. They added pole tips (that could get saturated) to the model to overcome
the difficulties associated with the rotor and stator pole permeances when they exhibit
partial overlap. The idea of adding saturable pole tips needs further analysis to make the
recommended model suitable for different machine geometries.
Other examples of modelling and simulating the performance of SRMs are: [183] where
Liu and Stiebler suggested an analytical model based on per-unit (p.u.) system and [184]
where Hossain and Husain presented the development of a simple analytical model of
SRMs, which is mostly suitable for real time implementation. In [171] the authors
describe how a decision was made to adopt the SABER simulation package for the
simulation of switched reluctance motor drives.
6.2.2.2 Recent Work

In this section some of the recent work relating to the modelling and simulation of SR
machines and drives is presented. There have been many other simulation methods
recently adopted, including neural networks [185-187], genetic algorithms [188] and
finite element analysis [189]. In [190] Stephenson et al incorporated iron losses and its
effect on the current waveform into the dynamic modelling of the SRM. Mutual coupling
between phases due to shared portions of the magnetic circuit and leakage flux was also
implemented in their modelling. Finally, [191] is a relatively recent example for use of
cubic splines in representing the magnetization characteristics of SRMs.
126
Use of Matlab / Simulink for the realisation of the complete electric drive system
simulation is becoming more and more common. Some of the recent examples relating to
modelling and simulation of SR drives are [150, 192]. In [150] FE generated magnetic
characteristics were used in a Simulink based dynamic simulation of a super-high-speed
4/2 switched reluctance motor. The simulation was used to assist with maximising the
starting torque of the machine. The paper uses static characteristics of the machine in
order to predict the behaviour of the machine in (arguably) the most dynamic operating
region: start-up. Another example for a Matlab / Simulink based simulation is [193],
where magnetisation curves were measured and then stored in look-up tables in the
simulation. In this paper, flux-linkage measurements were taken by locking the rotor,
applying dc volts to the phase windings and measuring this voltage and the resulting
phase current. This information was then used to generate the static torque characteristics
of the motor as well. Flux-linkage and torque characteristics of the motor were then stored
in look-up tables in the dynamic simulation model that were realised in Simulink. In
[194] the authors use Matlab / Simulink in order to model the dynamic behaviour of a
magnetic levitated switched reluctance linear actuator along with the control algorithm.
There are, however, no real measurements to validate and support the conclusions drawn
from the simulation results. In [195] Matlab / Simulink was used to simulate a 3phase asymmetric half-bridge converter in order to determine the optimum snubber
capacitor so that the efficiency of the converter can be improved. However, the method
was only useful for low speed operation where the current was under chopping mode
control. In [196] the authors design and simulate a 3-phase 12-10 switched reluctance
motor with asymmetrical stator poles and short flux paths. In [196] the simulation is used
to analyse and verify the machine design in order to guide the design and the production
of the machine. However, no validation of the machine design and the simulation was
presented to the readers. In [149] authors use Matlab / Simulink to simulate the
dynamic behaviour of an SRM based drive system for an electric vehicle application.
Rather than using FE generated or measured flux-linkage characteristics of an SRM, the
authors chose to estimate the magnetisation curves with analytical methods.
Unfortunately there is not a great deal of explanation as to how this analytical method was
derived and how well it represents the magnetic characteristics of the machine.
Due to the high computational time and resources associated with the FE analysis,
especially during the initial design stage, an analytical approach may be favoured to
narrow down the options within the available design space. FE analysis is favoured at the
127

later stages of the design stage where accurate results are required to verify the machine
design against measurements. Analytical methods aim to generate a Magnetic Equivalent
Circuit (MEC) of the machine for rapid analysis of machine magnetic characteristics.
The advantage of using the MEC approach is the fast solution time, which in turn allows
the designer to examine hundreds of topologies very quickly. Examples of simulation
models using machine characteristics that are generated by using analytical methods are
[197-202]. In [198] authors present a new analytical nonlinear model for rapid SRM
modelling. It is claimed that the method is satisfactory both in accuracy and in calculation
speed. The machine model based on the presented analytical model was then integrated
into the modelling of an Airborne Power System that was implemented in Dymola /
Modelica. In [201] various different control strategies were simulated for a 3-phase 6/4
switched reluctance machine. The model involved a non-linear dynamic model of the
machine based on Fourier series phase self inductance characterisation and a double
closed loop; comprising a chopping controller in the current control loop and fuzzy logic
plus PI controller in the speed control loop. In [203] 2D FEA generated magnetisation
characteristics of an SR machine is used to construct a dynamic equivalent circuit model
in Simplorer to predict the machines dynamic behaviour.
There have been a few examples of real-time simulation of SRMs as well. A d-SPACE
based real-time simulation of a 4-phase 8/6 SR drive in Simulink environment is
presented in [204]. The study ultimately aims to be the stepping stone in developing a
sensorless controller for the SRM by using additional toolboxes that are available in
Matlab / Simulink. Another example of a real-time simulation attempt is [205] where
CPU-based real-time simulation results of a 60-kW current-controlled 6/4 SRM are
presented. The real-time simulation of the drive was conducted on the RT-LAB real-time
simulation platform using SimPowerSystems in the Simulink environment.
6.3 SIMULATION DATABASE

The main parts of a model aiming to simulate the operation of electric machines and
drives can be summarised as follows.
1. The Power electronics: Depending on the detail of the simulation this includes:
128

The front-end bridge rectifier
The front-end filter components, i.e. series inductor and parallel dc-link capacitor
The inverter
2. The Machine: Here, means of representing the electromagnetic characteristics of the

machine is required. More specifically, the flux-linkage-position-current and torqueposition-current relationships are used to model the machines electromagnetic
behaviour.
3. The Control: This can be anything from a control scheme combining PID and PWM
controllers to more complex schemes like neural network based controllers.
Based on the simple partitioning of the simulation model listed above, the simulation
model of an electric machine and its drive can be viewed as a multilevel modelling
problem. All the components of the drive system were considered and it was decided that
a model generated in the Matlab / Simulink environment would give the highest
flexibility in creating different parts of the model to a high enough accuracy. As a result
of this decision, the simulation package presented below is based on several detailed
Simulink models which are controlled by a Graphical User Interface (GUI) and
supported by several m-file scripts (M-files are text files that contain Matlab programs)
and functions. Details of the simulation database are presented in Appendix C.
6.3.1 Machine Model

A numerical approach to the simulation of SRMs is introduced by Stephenson and Corda
[167] that is used to characterise the relationship between flux-linkage, current and rotor
position. This method eliminates the need of using differentiating terms. This way, the
errors introduced by differentiation are eliminated.
Usually for a conventional SRM the only shared parts for flux flowing in the machine
iron are the stator and rotor core-backs. Compared to the saturation levels seen in the
stator teeth the saturation of stator and rotor core-back is small and can usually be
ignored. This means that the mutual coupling introduced by the core-back saturation can
also be ignored. The flux-linkage of any one phase can therefore be assumed to vary with
129

the rotor position and current flowing in that phase only. This greatly simplifies the
simulation and a block diagram can be drawn as the one given in Figure 79. This block
diagram model can simply be copied over for each phase of the machine to model the
machines electromagnetic characteristics.
Using Stephenson and Cordas method, position and flux-linkage information are used to
obtain the phase current from a 2-dimensional look-up table. The data used in the 2dimensional look-up table can be produced by direct measurement of the flux-linkage
characteristics of the machine or by finite element simulation. The calculated phase
current value is then used in the next time step to calculate the resistive voltage drop and,
hence, to update the flux-linkage value by using Equation 10.
n = ( vn R i ( , ) n 1 ) d t
Equation 10
Intermediate values of the flux-linkage are derived by quadratic interpolation. The fluxlinkage characteristics are inverted using successive approximation to give a table of
i ( , ) . Flux-linkage characteristics depicted in Figure 80 are converted into current
characteristics of the form given in Figure 81. The flux-linkage and position signals are
used to calculate the phase current at each simulation time step. Data interpolation and
extrapolation of current/flux-linkage/position data during simulation is based on a cubicspline fit.
130
Figure 79 Schematic representation of modelling one phase of a conventional SRM
Figure 80 Fl ux-linkage - Current - Position characteristics
131
Figure 81 Current - Position - Flux-Linkage characteristics created from Flux-Linkage - Current Position characteristics
The torque is stored in another look-up table and can be obtained either directly by
measurement or indirectly by evaluating the co-energy of the machine. In the latter
method, a table of co-energy is obtained by numerically integrating the flux-linkage with
respect to phase current.
W ( , i ) = ( , i ) di
= const .
Equation 11
The intermediate values of co-energy are again found by quadratic interpolation and the
torque is calculated by a numerical differentiation of the co-energy:
T ( , i ) =
d W ( , i ) i =const .
|
d
Equation 12
Using Equation 11 and Equation 12 the flux-linkage characteristics of the form given in
Figure 82 are converted to static torque characteristics of the form given in Figure 83.
132
Figure 82 3D Flux-linkage characteristics
Figure 83 3D Static torque characteristics
Once again data interpolation and extrapolation of current/flux-linkage/position data

during simulation is based on a cubic spline fit.
Creating the look-up tables used in the Simulink model is done offline, i.e. it is not done
whilst the simulation is running. The transformation starts with flux-linkage
133

characteristics of the machine and using Matlab scripting functionality torque-current
and current-position relationships are created as 2D look-up tables. The screenshots of
look-up tables used in simulation model are given in Figure 84.
Figure 84 (a) 2D torque-current-position look-up table in Simulink

(b) 2D current-flux linkage-position look-up table in Simulink
6.3.2 Control Modelling

A considerable number of Simulink models for control modelling have been built. The
internal structure of these models varies, depending on machine type, power circuit
configuration and control algorithm. An in-depth description of the internal structure of
all the different models would exceed the objectives of this chapter. However, all drive
systems are constructed from basic inverter and control blocks which are summarised in
the following sections. Delays associated with signal processing in real drive systems,
such as the time spent during analogue to digital sampling and conversion and the time
delay between generating the PID reference voltage and updating of this reference voltage
in the PWM controller, are not modelled in the simulation. DC-link voltage ripple and the
voltage drops across the power devices are negligible for most high voltage, high current
drives and have not been factored in the simulation.
6.3.2.1 Angle and Speed Calculations

Phase selection is dependant on rotor position and control angles (i.e. advance angle and
total conduction angle). Zero rotor position is defined as the instance where a phase stator
pole opposes a rotor segment as shown in Figure 85, i.e. unaligned position. All
discussion in this chapter relating to rotor position is based on electrical degrees rather
134

than mechanical. Rotor position with respect to the first phase (Phase-A) is calculated.
This signal is in the form given in Figure 86. Rotor position with respect to the rest of the
machine phases is calculated simply by phase shifting the position signal of Phase-A. The
simulation always starts at the unaligned position for Phase-A.
Figure 85 Magnetic flux plots for a 12-10 single tooth winding segmental rotor machine with a single
phase excited illustrating the zero position implementation in simulation
The position signal given in Figure 86 is calculated from the rotor speed by utilising
Equation 13.
Pos rotor = Speed x
360
x N r [ Elect . Degrees ]
60
where
Posrotor : Rotor position with respect to a phase,

Speed : Rotor speed of rotation,
Nr
: Total number of machine rotor poles.
135
Equation 13
Figure 86 Rotor position signal with respect to Phase-A of the machine in simulation
Once the rotor position is determined for a phase then the energisation signal, i.e. when
that phase is switched on and off with respect to rotor position, is calculated. This is done
by comparing the rotor position signal for that phase with the advance and conduction
angles requested by the user. Advancing is when a phase energisation starts before the
unaligned position for that phase is reached and retarding is when a phase energisation
starts after the unaligned position for that phase is passed.
Phase energisation signal(s) not only define the energisation region for the machine
phase(s) but are also used to inhibit bridge legs and calculate current demand signal for
certain converter topologies and switching strategies.
136
6.3.2.2 Current Sampling

In the real drive system there are three 160A rated current transducers available to
measure drive currents. Three analogue-to-digital conversion (ADC) cards are used to
transform the measured current signals into 10-bit digital code. The sampling rates of the
ADC cards were 20 kHz. Digital sampling is modelled by using zero-order hold blocks in
Simulink. The sampled currents also have a conversion ratio. This ratio is simply
modelled by a gain block in Simulink. In the real drive, the line current demands are
software defined (for the 3-phase full bridge circuit) and are based on the commutation
points calculated from a 12-bit encoder position feedback signal.
6.3.2.3 PID Controller

PID controllers are commonly implemented in microcontroller or DSP based drive
systems to maintain the current within a desired level. The controlled current is fed back
into the controller where the sampling of the current and comparing to the reference value
are implemented. Once the current error signal is calculated it is then sent to the PID
controller. Based on the error signal the PID controller generates the reference voltage(s).
Depending on the type of the converter and the switching strategy, the PID controller
either generates a single reference voltage or two reference voltages, one of them being
the complimentary of the other. An integral anti windup is also included in the PID
modelling of the simulation.
The value of the reference voltage generated by the PID controller is limited to the
resolution of the PWM triangular waveform. The PWM waveform has a peak value of
255 (2 8 1) and so has the PID controller reference voltage. If complimentary switching
is to be employed, the PID controller also generates a negative reference voltage (in
digital terms a complimentary signal with respect to controller resolution). An example of
the operation of the PID controller in simulation is given in Figure 87.
137
Figure 87 Operation of PID controller in simulation of a 2-phase machine driven by an h-bridge

converter shown for a few electrical cycles
In this example, the simulation was of a two-phase machine driven by an h-bridge

converter where phases were connected back-to-back (see Chapter-3 for a description of
this circuit topology). Motor speed is 300rpm, current demand is 10A, and phases are
energised at the unaligned position for 120 electrical degrees. The top trace in Figure 87
shows the demand current in blue, measured current in red and current error signal in
green. The bottom trace shows the positive and negative reference voltages. When there is
positive demand current the PID controller saturates at its maximum value and this can be
seen from the generated reference voltages. As the current gets closer to the demanded
level the error current reduces and the PID controller tries to keep the current at the
desired level. The digitised nature of the measured current can be seen in Figure 88.
138
Figure 88 Zoomed version of measured and demanded current signals (along with the error signal in
green) showing the digitised nature of the measured current
6.3.2.4 PWM Controller

Fixed frequency PWM controllers are usually implemented along with PID control. The
reference voltages generated by the PID controller are fed into the fixed frequency PWM
controller where it is compared to a triangular wave and the inverter switches are turned
on or off accordingly.
The PWM controller converts the voltage reference signals (Vref: the voltage reference
value defining the required duty cycle to be applied to the winding) fed by the PID
controller into gate signals for the transistors. In [137] the PWM controller is simply
modelled as an amplifier. Even though modelling the PWM controller as an amplifier
simplifies the model and reduces the simulation time, it makes the switching patterns of
power devices unclear to the user. This no doubt complicates the part of the simulation
139

where the device losses and voltage drops are calculated. With the method followed in
this PhD project, the switching states of each power device are readily available. The
model can then simply use the device voltage drops and loss models to calculate the
voltage drops and power losses across each device.
The PWM reference voltage is a triangular one. The resolution of this voltage and the
reference voltage(s) - Vref generated by the PID controller both have a resolution of 8
bits. An example of how the gate signals are generated for an H-bridge (see Figure 89)
converter from the triangular PWM reference voltage and PID Vref voltage is depicted in
Figure 90.
Figure 89 Back-to-back connected two-phase SRM driven by an H-Bridge inverter
For leg 1 of the bridge, top device (T1) is switched on when Vref (generated by the PID
controller) is greater than Vtri (PWM carrier signal) and bottom device (T2) is switched
on when Vref is smaller than Vtri. For leg 2 of the bridge top device (T3) is switched on
when Vref (again, generated by the PID controller) is greater than Vtri and bottom
device (T4) is switched on when Vtri is greater than Vref. The PWM strategy ensures
that two switches of the bridge will always be conducting at any one time and both
switches of any one leg cannot be on at the same time. A schematic representation of the
PWM controller for an H-bridge converter in Simulink is given in Figure 91.
140

Vtri
+Vref.
0
-Vref.
T1
T2
T3
T4
Va
Figure 90 PWM control of H-bridge inverter. +Vref and Vref are the voltage control signals
generated by the PID controller. T1, T2, T3, T4 are the transistor gate signals, Va is the voltage
subsequently imposed on the winding
Figure 91 Screenshot of the PWM controller model in Simulink
An example of the PWM controller operation in simulation is given in Figure 92. The top
trace in Figure 92 contains the triangular PWM reference voltage in red, positive
141

reference voltage in green and negative reference voltage in blue. The other four traces
show the switching states of each power device of an h-bridge converter. In this example
the top left and bottom right devices appear to be switching and conducting the phase
current. It is also obvious from this example that the switching strategy balances the
switching of top and bottom devices. Therefore the device losses are balanced between
the top and bottom devices of the circuit. In Figure 93 the waveforms in Figure 92 are
shown for a few PWM cycles for clarity.
Figure 92 PWM controller operation in simulation - Simulation inputs are rotor speed = 300rpm,
advance angle = 0 degrees, conduction angle = 120 degrees, current demand = 10A
142
Figure 93 Traces in Figure 92 shown for a few PWM cycles
As mentioned earlier in the chapter, time lags between the output of the PID controller
and the input of the PWM controller and also between the current sampling and PID
controller are not modelled in the simulation. In reality the voltage reference output
generated by the PID controller is fixed for one entire PWM period. Also there is a finite
amount of time between the sampling of the phase current(s) and PID output(s). It was
believed that omission of these time delays did not spoil the performance analysis based
on simulation results and kept the simulation model simple but accurate.
6.3.2.5 Converter Modelling

The voltage applied to each phase is determined by the states of the devices in the power
electronic converter. A simulation database is able to model several different converter
topologies, namely asymmetric half-bridge, h-bridge and 3-phase full-bridge converters.
143

Rather than modelling the dynamic operation of each power device the states of each
device are determined.
A device is either in a high state (i.e. able to conduct current) or in a low state (i.e. does
not conduct current). Any one device can be in a high state but this on its own does not
suggest that it actually conducts current. Depending on the states of other devices, rotor
position and phase or line currents, the simulation decides whether a power device
conducts current or not. Once the conduction state of each power device is determined,
the voltage applied to each phase can then be decided. For example, in Figure 89 the
phases of a two-phase SRM are connected back-to-back and operated from a PID
controlled H-bridge inverter. The diodes (phase diodes) connected in series with the phase
windings are used to achieve unidirectional current flow. Control can be achieved by
sampling the phase currents (IphaseA and IphaseB in Figure 89) or the line current
(I_Line in Figure 89) and comparing to a reference value (current demand). Sampling the
line current is shown here. Line current control requires only one current sensor as
opposed to two needed for controlling the phase currents separately. Therefore it is likely
to be the preferred method in real-time applications. The PID controller supplies positive
Vref to control leg 1, comprising switches T1 and T2 and negative Vref to control leg 2 of
the bridge, comprising switches T3 and T4. As a result of complimentary switching,
conduction and switching losses are shared equally between each transistor. The
schematic explanation of the controller is given in Figure 94.
+Vref.
Idemand
I_err
PID
PWM
-Vref.
GATE
SIGNALS
IphaseA
IphaseB
I_Link
T1
T2
T3
PHASE
VOLTAGES
Figure 94 Schematic of controller for the two-phase drive given in Figure 89
144
T4

All possible switching states and corresponding operating conditions are summarised in
Table 10.
Table 10 All possible switching states of the transistors in the H-Bridge inverter of the 2-phase SRM
model and operating conditions depending on the states of the power switches
T1
T2
T3
T4
OPERATING CONDITION
NOT POSSIBLE (two switches should be on at any one time)
NOT POSSIBLE (two switches should be on at any one time)
If I_Line > 0 (Ia>Ib) D2 DA D3 conducts and phaseA

freewheels. 0 volts to both phases
NOT POSSIBLE (switches of any one leg cannot be on at the

same time)
NOT POSSIBLE (two switches should be on at any one time
If Ia>Ib T4 - DA - D2 conducts and phaseA freewheels. If Ia<Ib
T2 DB D4 conducts and phase B freewheels. 0 volts applied

to both phases.
Phase B energised and receives positive volts whereas Phase A is
NOT POSSIBLE (too many switches on)
NOT POSSIBLE (two switches should be on at any one time
de-fluxed and receives negative volts.
Phase A energised and receives positive volts whereas Phase B is

de-fluxed and receives negative volts.
If Ia>Ib T1 - DA D3 conducts and phaseA freewheels. If Ia<Ib
T3 DB D1 conducts and phase B freewheels. 0 volts applied

to both phases.
NOT POSSIBLE (switches of any one leg cannot be on at the

same time)
145

A schematic of the subsystem in Simulink is given in Figure 95 where the switching
states of power devices and applied phase voltages are determined.
Figure 95 Screenshot of simulation subsystem calculating phase voltages and power device switching
states
Dead-time is required to prevent turning on the upper and lower devices on the same leg
of the converter simultaneously. Here, the simulation assumes that the switching of power
devices is ideal, i.e. when a switching command is generated; the device reacts and turns
on or off immediately. Therefore there is no need to model the dead-time effects. In
reality the dead-time will affect the available voltage applied to the on-coming phase and
off-going phase during commutation and somewhat distort the waveforms (during deadtime negative voltage is applied to the commutation spike instead of the tail current that
results in a shorter commutation spike but a longer tail current). This will result in an
overshoot in phase current when the inverter is controlled by line-current sensors rather
than phase current sensors [89].
146
6.3.2.6 Phase Diode Modelling

For models where a diode is connected in series with the phase windings to avoid bi-polar
current flow through the phase windings, the operation of this diode (called the phase
diode in the simulations) needs to be modelled correctly. In the simulation, the phase
diode should operate, i.e. block phase current flow in reverse direction if the phase
energisation signal is OFF (phase conduction period has completed) and the phase current
has decayed to zero. The simulation detects the positive to negative zero-crossing of the
phase current and simultaneously checks the phase energisation signal. If the phase is in
the OFF state then the Phase Diode operates and blocks the phase current flowing in the
reverse direction. Figure 96 is a schematic of the phase diode modelling in Simulink.
Figure 96 Phase Diode Modelling in Simulink (Inputs are phase current and phase energisation
signal, output is HIGH when Phase Diode operates, i.e. blocks the phase current reverse direction)
6.3.2.7 Electronics Loss Modelling

If the simulation is to be used to determine the performance of the power devices and
some device related design parameters such as peak device voltage, current stresses,
switching losses and electromagnetic interference (EMI) due to switching action, then it is
desirable to model the device dynamically. That is to use the device transient equivalent
models in the simulation environment [142].
The aim of the inverter loss calculation was to be able to perform comparative studies
between different drives and assess their impact on power electronic ratings in terms of
147

device loss as well as device peak/rms current. For this reason, devices were characterised
only in a reasonably simple way as it was the comparative performance of the drive rather
than the comparative performance of individual devices that were being assessed.
Nevertheless it is best to use the loss curve illustrating the highest junction temperature as
this will give the most accurate representation of the device in real life.
Conduction loss of a diode, IGBT or MOSFET was based on a linear model in the form
of y = mx + c equation. Therefore, in the case of an IGBT, three parameters were entered
as shown in Figure 97. Vce (on) and Ic at (V2, I2) and Vce (on) at (V1, I1). I1 is always 0
amps [89].
Ic
(V2,I2)
(V1,I1)
V(ce)on
Figure 97 IGBT conduction loss characterisation Red trace is a representation of the forward VI
curve from the datasheet and green trace is the linear approximation within the operating current
range of the application
This is a first order approximation to the non-linear characteristic which is typically of the
form given in Figure 97. If the device operates at a limited current range then the linear
approximation is such that only the operating range of the device in the application is
taken into account. This way, the accuracy of the linear approximation is maximised.
A choice of the method to calculate switching loss of an IGBT or MOSFET is offered.
One way is to enter the energy per switch at turn on (Eon) and the energy per switch at
turn off (Eoff). The device current and volts that this energy is based on is also entered.
E(on) and E(off) are then scaled proportionally in the simulation to suit the actual currents
and voltages being used. This method is particularly suitable for IGBTs and the energy
148

per switch value quoted in the data sheet includes tail losses and the reverse recovery
current effect of the diode.
The model is able to detect the switching instances of any device. Also the type of
switching is distinguishable in the simulation so that the coefficients for turn-on and turnoff losses can be applied correctly to the right instance of switching.
6.3.3 Inverter Modelling

6.3.3.1 Asymmetric Half-Bridge Inverter Drive
This requires knowledge of the voltage across the phase at all times. During the turn off
process, the two transistors controlling one particular phase are simply turned off. With
positive current still continuing to flow, the diodes conduct the current and the phase
windings see a negative voltage bringing the phase current down to zero. However, once
at zero, the current does not reverse direction. Therefore, the voltage that is applied to the
winding by the diodes is not known and hence flux linkage cannot be calculated. It is
therefore necessary in the simulation to control the current to zero by setting the current
demand to zero during the off period. This way, the current controller knows the voltage
being applied and therefore the problem is overcome.
6.3.3.2 Six-Phase Delta Connected Drive

The delta connected drive (Figure 33 repeated in Figure 98) presents no problem to the
simulation as unlike the star connected inverter, the controller directly controls the
voltage across each phase winding. Therefore the phase voltage is known at every time
step for a delta connected inverter, so solving for phase currents and flux linkages is
relatively straightforward. The only difficulty comes with the methodology used in the
current controller itself. In the case of modelling a six-phase SRM where phases are
paired and connected back-to-back through so-called phase diodes, the voltage applied to
any one module directly affects the voltage available to be supplied to the other two
modules. This will affect the ability of the converter to supply voltage to phases on
different modules.
149
The phase voltages of the delta connected 6-phase machine driven from the 3-phase full
bridge circuit are calculated as:
Vx Vy = VphaseA = VphaseD
Vy Vz = VphaseE = VphaseB
Equation 14
Vz Vx = VphaseC = VphaseF
Simulation results of the 6-phase delta-connected SRM with magnetically de-coupled

phase windings working under current and voltage control are presented in Chapter 9.
6.3.3.3 Six-Phase Star Connected Drive

The star connected drive (Figure 32 repeated in Figure 99) is an unusual case for the
simulation. Phase currents must sum to zero at the star point. Moreover, the star point is
not directly controlled by the current controller, and hence, its voltage is unknown to the
controller. Unlike the delta-connected drive, the controller in simulation requires the star
point voltage so that it can calculate the phase voltages and consequently flux-linkages,
currents and torques. In [89] Clothier uses an iterative method to calculate the star point
voltage at every time step of the simulation. The principle is based on the fact that the
sum of the currents flowing into the star point must be zero. With the star point voltage
from the previous time step, the new time step is entered. Inevitably an error current is
produced at the star point. The star point voltage is then adjusted in the appropriate
direction by a calculated amount and the whole time step restarted until the sum of the
currents is zero (or within a specified error band).
150
The method applied in this project also aims to minimise the current error at the star
point. However, this is achieved with a slightly different method. All phase flux-linkages
and phase currents along with the dc-link voltage, phase resistance and simulation timestep are all transferred into an m-file at each time step during the simulation. Within this
m-file the phase inductance is assumed to be linear. The star point voltage is varied
between Vdclink / 2 and + Vdclink / 2 . The value which minimises the star point current is
selected as the star point voltage and used in the next time step of the simulation. An
example of star-point voltage calculated in the simulation is given in Figure 100. In
Figure 100 the blue trace is the star-point voltage calculated during the simulation and the
red trace is a filtered version of the same voltage waveform. The filtering is done by a
simple moving average filter. The depth of the filter was selected so that there was no
significant loss to the waveform (both the phase and the amplitude of the filtered
waveform are believed to be relatively accurate). The star-point current calculated (with
the star-point voltage in Figure 100) is given in Figure 101. The current demand was set
to 10A for this simulation.
151
Figure 100 Raw and filtered star point voltage waveforms from simulation of 6-phase SRM driven by
a 3-phase full-bridge inverter Blue trace: Raw star-point voltage; Red trace: Filtered star-point
voltage
Looking at Figure 101 the error in star point current is approximately 0.3% of the current
demand at any time step. The controller works the same as for line current controlled
delta-connected drive. The PID controller generates positive reference voltages for each
line of the bridge. These values are then compared to the PWM triangular reference
voltage and the switching states of the power devices are determined. Then depending on
the switching states of the devices and the direction of current, the devices which actually
conduct current are determined. Once the current conducting power devices are known
the line voltages can be determined easily. Once the star-point voltage and the line
voltages (controlled by the controller) are known, module voltages and, hence, phase
voltages can be calculated. Simulation results of the 6-phase line current controlled starconnected SRM with magnetically de-coupled phase windings working under current and
voltage control are presented in Chapter 9.
152
Figure 101 Raw and filtered star point current waveforms from simulation of 6-phase SRM driven by
a 3-phase full-bridge inverter (Star-point voltage in Figure 100 Blue trace: Raw star-point current;
Red trace: Filtered star-point current
6.4 MODELLING OF SRMs WITH MUTUALLY

COUPLED PHASE WINDINGS
Conventional SRMs with short-pitched windings and some segmental rotor SRMs are
assumed to be magnetically de-coupled and the mutual saturation of the core-back is
considered to be of no significance. The flux-linkage of any one phase is therefore
assumed to be solely a function of phase current and rotor position of that phase
(operation of any one phase has no effect on other phases of the machine). This
assumption allows modelling of each phase separately. When an SRM has a segmental
rotor construction, the magnetic flux paths enclose a single slot and the magnitude of the
flux is a function of the segment position and the total MMF in the slot. For many designs
each slot contains two coils of adjacent phases consequently there is strong mutual
coupling between phases. Moreover, the prototype 6-phase segmental rotor machine that
153

is presented in this thesis will always have two adjacent phases conducting at the same
time during operation, when the machine is connected in either a star or delta
configuration. Mutual coupling means that the flux linking each coil is a non-linear
function of rotor position and currents flowing in all other phases. Therefore, the circuit
equation of each phase cannot be solved independently, from a single non-linear
differential equation. Modelling becomes complex because flux-linkage is a function of
multiple phase currents and rotor position, resulting in a three or more dimensional lookup table.
In other words, the techniques used to model the magnetically de-coupled SRMs cannot
be used to simulate the 6-phase machine with mutually coupled phases. To ease the
difficulty faced with the simulation of the machine with mutually coupled phases, an
alternative network relating flux and MMF in the magnetically coupled 6-phase machine
to those in an equivalent magnetically de-coupled 6-phase machine is needed. Use of
transformation matrices to de-couple the mutual coupling between phases has previously
been successfully used for conventional toothed rotor SRMs with fully pitched windings
[91] The idea is simple: Use a set of so-called transformation matrices in order to decouple the mutually coupled nature of phase quantities such as MMF and flux-linkage
into equivalent but mutually de-coupled quantities so that the analysis, understanding and
the simulation of the machine are simplified [91].
The flux-linkage characteristics obtained from 2-D FE analysis (the end-winding length to
stack length ratio is relatively small) can be used to characterise the machine at different
rotor positions. It then becomes a relatively simple problem to solve for the phase current,
flux-linkage and torque values. In the case of a short-pitched 6-phase segmented rotor
SRM, where phases are connected back-to-back and then connected in star or delta, the
flux-linkage in a phase is a non-linear function of rotor position and the currents in all
phases. It is rather difficult to solve for each phase current, flux-linkage and torque due to
the mutual coupling between phases. It makes the machine both very difficult to simulate
and difficult to understand.
A set of transformation matrices has been generated to convert current and flux-linkage
values into corresponding values in a magnetically de-coupled machine. A more
appropriate terminology for magnetically de-coupled machine parameters, in this case
is perhaps slot parameters. For a segmented rotor SRM the torque is produced as the
rotor segments modulate the stator slot permeance and magnetic flux paths enclose stator
slots, therefore it is more appropriate to think in terms of permeance variation per slot
154

rather than the permeance variation per tooth, which is evaluated in conventional SRMs.
Once the transformations are completed it is then possible to apply the techniques used to
simulate a conventional SRM with de-coupled phases to the 6-phase delta / star connected
segmented rotor SRM with magnetically coupled phases.
The aim is to:
1. Develop a set of equations that convert teeth MMFs to slot MMFs.
2. Check that the 6x6 matrix formed by combining the transformation equations that are
found in 1 has an inverse (i.e. determinant of the matrix is non zero). This is to be able
to convert slot MMFs into teeth MMFs in the simulation. The magnetic characteristics
that will be used in the simulation will be relating flux circulating around one of the
machine slots against the MMF in that slot for varying rotor positions. However, the
ultimate aim of the simulation is to calculate the phase currents at every time step.
Therefore the slot MMFs determined from the magnetisation characteristics of the slot
(flux and rotor position required) will need to be converted to teeth MMFs so that the
phase currents can be determined (using number of turns). This is demonstrated with
Equation 15.
V PHASE
[ ]PHASE
[ ]SLOT
POSITION
[MMF ]SLOT
[i ]PHASE
Equation 15
3. Develop a set of equations that convert slot fluxes to teeth fluxes.
4. Check that the 6x6 matrix formed by combining the transformation equations that are
found in 3 has an inverse. This is to be able to convert teeth fluxes (relating to phase
flux-linkage via the number of turns) into slot fluxes. Slot fluxes along with the rotor
position are required to be known at every time step of the simulation so that the slot
MMFs can be determined (Equation 16). Next, the slot MMFs and the rotor position
can be used to calculate the torque contribution of each slot. Slot MMFs are then
converted into teeth MMFs and finally phase currents to be used at the next time step
of the simulation are determined.
155
[ ]SLOT = [C ]FLX [ ]TOOTH
Equation 16
It would be a good idea at this stage to present the schematic explanation of the
simulation for the 6-phase prototype with mutually coupled phases. This is done in Figure
102 for the simulation of the star-connected 6-phase segmental rotor SRM. The same
schematic can be used to represent the delta-connected drive as well. The only
modification to the schematics in Figure 102 would then be the method used to calculate
the phase voltages from line voltages (Vx, Vy and Vz).
[MMF ]TOOTH = [C ]MMF [MMF ]SLOT

[ ]TOOTH
Equation 17
in Equation 16 and [MMF ]TOOTH in Equation 17 can be linked to [ ]PHASE and
[i ]PHASE in Equation 15 via the number of turns per tooth.

By examining Figure 103 (Figure 39 repeated here for clarity) the slot currents can be
written as
Ia 1
Ib 0

Ic 0
=
Id 0
Ie 0

If 1
1
1
0
0
0
0
0
1
1
0
0
0
0
0
1
1
0
0
0
0
0
1
1
0
0 I 1
0 I 2
0 I 3

0 I 4
1 I 5

1 I 6
Equation 18
156

3x1
Demand Currents
3x1
PID
Vref.
3x1
Line Currents
6x1
PWM
BRIDGE
LEGS
6x1
POTENTIALS
@ U, V, W
3x1
DIRECTION
3x1
STAR POINT
VOLTAGE
PHASE
VOLTAGES
ELECTRICAL
TORQUE PER
SLOT
Phase Currents
Phase
Resistances
2D LOOK-UP
TABLE
SLOT MMFs
TO
PHASE CURRENTS
s
2D LOOK-UP
TABLE
ROTOR
POSITION
TOOTH FLUX-LINKAGES
TO
SLOT FLUXES
PHASE
FLUX-LINKAGES
SLOT
FLUXES
Figure 102 Schematic explanation of simulation of six-phase SRM with mutually coupled phase
windings
Figure 103 2D rectilinear representation of six-phase SRM showing the direction of slot MMFs and
teeth fluxes
157

From Equation 17 and Equation 18 the transformation matrix for converting teeth
currents into slot currents is given in Equation 19.
[C ]MMF
1
0
0
=
0
0
1
1
0
0
0
0
0
1
1
0
0
0
0
0
1
1
0
0
0
0
0
1
1
0
0
0
0
0
1
Equation 19
In order to convert slot MMFs to teeth MMFs, [C ]MMF in Equation 19 must be a nonsingular matrix (the determinant of [C ]MMF must be non-zero). However, it is possible to
obtain any one row or any one column of [C ]MMF from a linear combination of other rows
or columns of the same matrix. This means that the determinant of [C ]MMF is 0, and,
hence [C ]MMF does not have an inverse matrix. The transformation matrix in Equation 19
is of no use on its own and additional equations are required in order to overcome the
problem of matrix singularity.
In Figure 103 the net slot MMF should equal zero. Therefore,
Ia Ib + Ic Id + Ie If = 0
Equation 20
However, Equation 20 can be solved implicitly from Equation 18.
Additionally for a star connected machine as shown in Figure 32, module currents must
sum up to zero at the star point.
I1 + I 3 + I 5 I 2 I 4 I 6 = 0
Equation 21
As current in any one phase can be calculated from currents in the other five phases, the
first row and column of the matrix in Equation 19 can be removed (this is not a rule, it
can be the second row and column, etc.). However, the first column cannot be removed
directly as the current in Phase-F (i.e. If ) cannot be calculated without the first column of
the matrix. However from Equation 21 I1 can be rewritten as
158
I1 = I 2 + I 4 + I 6 I 3 I 5
Equation 22
Removing the first column and row from Equation 18 and using Equation 22, the new
transformation matrix can be written as in Equation 23.
Ib 1 1
Ic 0 1

Id = 0 0

Ie 0 0
If 1 1
0 0
1 0
1 1
0 1
1 1
0 I 2
0 I 3
0 I 4

1 I 5
2 I 6
Equation 23
The 5 by 5 matrix in Equation 23 is non-singular and has an inverse of the form given in
Equation 24.
0 1 1 1 Ib
I 2 0
I 3 1 0
0 1 1 Ic

I 4 = 1 1 0
0 1 Id

0 Ie
I 5 1 1 1 0
I 6 0 2 0 2 1 If
Equation 24
Using Equation 24 and Equation 22 all phase currents can be calculated.
The above transformation is valid for the star-connected drive.

Once again by examining Figure 103 the teeth fluxes can be written as in Equation 25.
1 1
2 1

3 0
=
4 0
5 0

6 0
0
1
1
0
0
0
0
0
1
1
0
0
0
0
0
1
1
0
0
0
0
0
1
1
1 a
0 b
0 c

0 d
0 e

1 f
Equation 25
159

Similar to the transformation matrix in Equation 19, any one row or column of the
transformation matrix in Equation 25 can be calculated by linear combination of others,
i.e. there is no inverse of the matrix. For the machine in Figure 39, the net flux flowing
through the stator teeth should sum up to zero. This can be represented by Equation 26.
1 + 3 + 5 2 4 6 = 0
Equation 26
However, Equation 26 can be solved implicitly from Equation 25.
The net flux enclosing the stator slots should also sum up to zero. This necessitates
Equation 27 to be satisfied.
a + b + c + d + e + f = 0
Equation 27
From Equation 27 the flux circulating around slot-a (i.e. a ) can be rewritten as given in
Equation 28.
a = b c d e f
Equation 28
Similar to the process of calculating the current transformation matrix, the first row and
column of the 6 by 6 matrix in Equation 25 are removed. From Equation 25 and Equation
28 the following transformation matrix is devised (Equation 29).
2 0 1 1 1 1 b
3 1 1
0
0
0 c

4 = 0 1
1
0
0 d

1
0 e
5 0 0 1
6 0 0
0 1
1 f
Equation 29
The 5 by 5 matrix in Equation 29 is non-singular and has an inverse of the form given in
Equation 30.
160

b 1 2 2 2 2 2
c 0
0 1 1 1 3

d = 1 0
0 1 1 4

0 1 5
e 1 1 0
f 1 1 1 0
0 6
Equation 30
Using Equation 28 and Equation 30 all slot fluxes can be determined. As the
transformation matrix in Equation 25 is the transpose of the transformation matrix in
Equation 18, the transformations are power invariant (flux transformation relates directly
to voltage transformation and power is the product of current and voltage). Once the
transformations are completed, it is then possible to apply the techniques used to simulate
a segmented rotor SRM with de-coupled phases to the 6-phase delta / star connected
segmented rotor SRM.
As mentioned earlier, due to the mutual coupling between phases, flux-linkage in a phase
is not necessarily zero even if the current in that phase is zero. In the simulation of the
machines with mutually de-coupled phases the phase flux-linkages are all forced to stay at
zero until the beginning of the next energisation cycle of the phases once the phase
currents fall to zero. However, with the machine where the phases are mutually coupled,
any change in a phase current will result in a change to the flux linking all six phases.
Therefore, the phase flux-linkages should not be zeroed when the phase currents fall to
zero. It should be noted that at the periods where the phase current is zero and the phase
energisation cycle is finished the phase diode still operates and stops the current reversing
direction. Moreover, unlike the short-pitched SRMs with mutually de-coupled phases
where the flux-linkage is always positive, the flux-linkage of a phase may need to go
negative during the off period of a phase. This is again due to the mutual coupling effects.
It has been already mentioned that the simulation model requires the knowledge of the
flux vs. MMF characteristics of one of the machine slots at different rotor positions as
well as the transformation matrices. These characteristics used in the simulation are
obtained in the FE package. The method used to generate these characteristics is
explained in detail in Chapter 9.
The transformation matrix used to transform tooth fluxes into equivalent slot fluxes of a
star-connected machine can also be utilised for the delta-connected machine. However the
transformation matrix derived for transforming the slot MMFs into phase currents cannot
be used for the delta-connected drive as the module currents do not necessarily sum up to
161

zero. Therefore another equation is needed for the delta-connected machine to create the
MMF/Current transformation matrix. At the time of writing this thesis the transformation
matrices for the delta-connected drive were not completed.
6.5 PERFORMANCE CALCULATIONS

At the end of a simulation, the results are transferred into a script by feeding the required
signals into individual output ports at the highest level of the Simulink model. Each
output port is given a name (Figure 104). The names given to the output ports are then
input into the Data Import / Export part of the Configuration Parameters (Figure 105 the order of occurrence should be kept the same as the one in the Simulink model).
Figure 104 Data Exporting for Offline Statistical Calculations Once the Simulation Has Completed
(Output Ports Circled in Red)
These results are then used to carry out statistical calculations such as inverter losses,
average machine torque, torque ripple, inverter VA ratings and machine copper losses.
Once the simulation completes, statistical calculations are carried out over the last three
162

complete electrical cycles. For this reason, the M-file script first extracts the last three
electrical cycles of every simulation output waveform.
Figure 105 Configuration parameters pane used for data importing / exporting in Simulink
Power electronics loss calculations vary depending on the converter topology used.
However the principle of the performance calculations is the same for all of the converter
topologies. IGBT conduction, switching losses and diode conduction losses are all
calculated at each time step within the simulation. The averaging of these losses, on the
other hand, is done offline once the simulation finishes and these waveforms are available
for the post-processing code.
Instantaneous torque is calculated at every time step of the simulation for each phase of
the machine. Total electric machine torque is then calculated by summing instantaneous
torque values of each phase of the machine.
6.6 SUMMARY
A simulation database has been developed that is able to model SR machines for different
machine geometries and phase numbers by employing the machines flux-linkage
characteristics, either obtained from measurements or from finite element analysis. As the
flux-linkage characteristics accurately describe the way flux varies with position and
current, they can be used to correctly model the magnetic saturation in any switched
reluctance machine. Simulation models for two-phase, three-phase and six-phase SRMs
have been generated. Asymmetric half-bridge, 3-phase full bridge and h-bridge converters
have also been modelled in the simulation package. The simulation model is flexible - any
163

simulation model can be changed to run with a different converter or controller without
having to re-create the entire model.
For six phase segmental rotor machine modelling, an approach has been taken which is
similar to the one employed for modelling of fully-pitched winding machines. For this, a
set of transformation matrices have been generated in order to ease the modelling of six
phase segmental rotor machines by de-coupling phase quantities into slot quantities.
Magnetic saturation is modelled by using a combination of the machines flux-linkage /
current / angle relationship and a suitable time-stepping solution.
For the time-stepping solution, a fourth order Runge-Kutta solver readily available in
Simulink has been chosen which generates accurate simulation results without
compromising the speed of simulation. The time stepping routine is able to determine
how the current in the machine is affected at any instance in time by the switching states
of the inverter and the position of the rotor. A GUI has also been developed to ease the
running of any simulation for the user. All of the essential parameters needed for the
simulation are presented in a window so that the user does not have to know the details of
the simulation to be able to drive it. The details of the GUI are presented in Appendix C.
The user interface is also able to carry out statistical calculations including inverter losses,
copper losses and torque ripple based on simulation results, in addition to producing
waveforms such as instantaneous output torque and phase current.
In Chapters 7, 8, and 9 simulation results along with measurements will be presented for
the 2-phase, 3-phase and 6-phase drives, respectively.
164
Chapter 7 MACHINE OPERATION AND COMPARISON TO SIMULATION

TWO PHASE DRIVES -
CHAPTER 7
7 Machine Operation and
Comparison to Simulation
- Two Phase Drives 7.1 INTRODUCTION
This chapter contains the dynamic measurement and simulation results of two-phase
segmental rotor machine drives. A two-phase machine was not designed; instead the
phases of the six-phase segmental machine were connected such that it could be run as a
two phase machine. The operational principles of the asymmetric half-bridge and hbridge converters were both presented in Chapter 3; the description of the test rig was
given in Chapter 5. The inverter of the test rig was configured to drive:
The two-phase segmental rotor machine from the asymmetric half-bridge converter.
The two-phase segmental rotor machine from the h-bridge converter with series
diodes.
In the next section, the simulated results will be compared against those measured for the
h-bridge and asymmetric half-bridge driven two-phase machine. Matlab based scripts
were used to analyse the measured waveforms in order to predict some of the
performance parameters of the drive systems and ease the comparison against simulated
165

TWO PHASE DRIVES waveforms and simulated performance parameters. Examples at low and high speeds are
presented.
Following on from the comparison of simulated results against measurements, section 7.3
focuses on the relative performance comparison of the asymmetric half-bridge and hbridge driven 2-phase machine in the dynamic simulation environment. The section looks
at how well the two topologies compare in terms of:
The total inverter loss at high and low speeds.
Overall drive efficiency at high and low speeds.
7.2 MEASUREMENTS AND SIMULATION

RESULTS
A segmental rotor machine with two-phases has not been designed; instead the phases of
the six-phase segmental machine were connected such that it could be run as a two phase
machine. A schematic description of the six-phase machine configured as a two-phase
machine is given in Figure 106.
Phases A and F and phases C and D of the six phase machine have been paired together to
form Phase-1 and Phase-2 of the two-phase segmental machine, respectively. Phase B and
phase E of the six-phase machine have been left unconnected. There is no need to
measure the flux-linkage characteristics of the two-phase machine as the flux-linkage
measurements taken from the six-phase machine by connecting two adjacent phases in
series are equivalent to the per phase flux-linkage characteristics of the newly configured
two phase machine.
The inverter legs of the drive system (explained in Chapter 5) have been configured to
drive the two phase machine from an h-bridge and an asymmetric half bridge inverter.
The operating principles of both asymmetric half-bridge and h-bridge inverters have been
given in Chapter 3. Only the measured and simulated waveforms are presented in this
section. In the h-bridge driven case the phases of the two-phase machine were connected
back-to-back through phase diodes (which are used to force the phase currents to be
unipolar). Instead of controlling phase currents separately, the line current was controlled
166

TWO PHASE DRIVES with a single current transducer. This is thought to be beneficial in terms of drive
simplicity and cost.
PHASE-1
IN
+
-
PhA+
PhF-
PhB+
PHASE-2
OUT
PhC+
PhE-
PHASE-2
IN
PhD-
PhD+
PhC-
PhE+
PhB-
PhF+
PhA-
PHASE-1
OUT
Figure 106 Six-phase segmental rotor machine configured to run as a two-phase machine
A Matlab based script was written to analyse the oscilloscope captured running
waveforms and indirectly calculate performance parameters such as total output torque
and copper losses. The script takes the phase current waveform, switch-on angle, phase
resistance, number of phases and estimated winding temperature during test as inputs and
is able to identify / calculate the following parameters:
The total duration of phase energisation i.e. conduction angle

Torque generated by a phase
Total machine output torque
167

TWO PHASE DRIVES Torque ripple
Running speed
Shaft power
Peak phase current
RMS phase current
Copper loss
Flux-linkage locus
Line current waveform for h-bridge connected drive
The test rig had a torque-meter coupled to the shaft of the machine in test. The torquemeter was capable of providing average shaft torque measurement. However, the
instantaneous torque waveform analysis was not possible with the torque-meter, except at
very low speed. The above mentioned Matlab script is very useful for deriving the torque
waveforms. The script uses the flux-linkage-current-position (-i-) characteristics of the
machine to calculate the phase flux-linkage and the torque-current-position (T-i-)
characteristics to calculate the torque generated by the phase. Both (-i-) and (T-i-)
characteristics can be obtained by measurements or by FE simulations. Phase quantities
are then phase shifted (depending on the number of phases and the number of rotor poles)
to calculate the contribution of all phases to torque production. The total instantaneous
electromagnetic torque for the machine is then the sum of the torques for each phase. The
torque calculated by this technique does not include rotational losses, nor does it include
the magnetic saturation of one phase affecting the characteristics of another. The result
generated by the script can therefore be higher than the shaft torque actually delivered by
the machine. Rotational losses increase with speed. As a result, the discrepancy is
expected to be larger at higher speeds.
7.2.1 Two-Phase Machine Driven by H-Bridge Inverter

Simulating the H-bridge drive shown in Figure 107 is beneficial in that it can be used as a
building block for the more complicated six-phase drive system simulation. Both the sixphase drive and the two-phase h-bridge drive have two phases connected back to back
through phase diodes.
168

TWO PHASE DRIVES -
Figure 107 H-bridge converter driving a back-to-back connected two-phase machine
Simulation techniques used to model the two-phase h-bridge drive can then ease the
modelling of the six-phase drive.
To explain the operation of the H-bridge inverter driving the 2-phase SRM with its phases
connected back-to-back through phase diodes, the simulated waveforms in Figure 108
and Figure 109 are used. In Figure 108 and Figure 109 the machine is under current
control and voltage control, respectively. In both Figure 108 and Figure 109 the
conduction angles are 120o (electrical), the phases are energised at the unaligned position,
the line current demands are 30A and the dc-link voltages are 175V. The phase voltage
trace in both figures depicts two plots: in red is the actual phase voltage applied during
simulation and the plot in blue is the filtered (with a moving average filter) version of this
simulated phase voltage.
169

TWO PHASE DRIVES -
Figure 108 Simulation results of the two-phase machine driven from an H-bridge inverter under
current control. Conduction angle is 120 degrees (electrical), no advancing, current demand is 30A
and the dc-link voltage is 175V. Top trace = phase voltages red for L1, blue for L2, black for
filtered L1 voltage; middle trace = phase currents red for L1, blue for L2; bottom trace = line
current. Speed of operation is 350RPM
With respect to Figure 107, in Figure 108:
Point A: Phase L2 starts to turn off. Phase L1 is not turned on at this point. However, the
negative voltage across phase L2 (Figure 108 waveform in blue in top trace) to de-flux
the phase is seen as positive voltage across phase L1 (Figure 108 waveforms in red and
black in top trace). The line current falls (Figure 108 bottom trace) to zero whereas the
current in phase L1 (waveform in red in middle trace) rises until point B.
Point B: Line current falls to zero. The controller in the simulation aims to keep the line
current at zero until point C (where phase L1 is turned on) by applying positive and
negative dc-link voltage across both phases. The current in phase L1 drops slightly due to
the resistive voltage drop across phase windings.
170

TWO PHASE DRIVES Point C: Phase L1 is turned on. Both the line current and the phase L1 current rises
rapidly with the positive voltage across phase L1 windings.
Point D: The current demand is reached. The machine is under current control. The
controller in the simulation aims to keep the current at the demand level by applying
positive and zero volts across phase L1 windings. The positive voltage across phase L1 is
seen as a negative voltage across the phase L2 windings. The diode (Db in Figure 30) is
reverse-biased and does not allow the current in phase L2 to flow in the reverse direction.
Point E: Phase L1 starts to turn off. The controller applies negative volts across phase L1
windings and its current falls to zero. Phase L2 current (similar to the rise in phase L1
current between points A and B) starts rising due to positive volts seen across phase L2
windings.
Point F: Line current falls to zero once again. The controller works similar to the way it
did between points B and C.
Point G: Phase L2 is energised and its current rises rapidly until the line current demand
is reached at point H, as between points C and D.
Point H: Line current demand is reached and controlled, in an identical manner to point
D.
Point J: Back to point A.
The only difference in traces in Figure 109 compared to those in Figure 108 is the fact
that the machine is running under voltage control. The line current demand of 30A is not
reached and the controller applies positive volts for the entire duration of the conduction
period. Similar to Figure 108, when the line current falls to zero, the controller aims to
keep the current at zero by applying positive and negative volts across phases until the
next energisation point of either phase. From Figure 110 to Figure 113 the simulation
results of the h-bridge driven two-phase machine are compared to the measurements
taken from the test rig under various operating conditions. The red trace in each figure is
171

TWO PHASE DRIVES the measured waveform and the blue trace is the simulated waveform. The instantaneous
measured torque waveform is not actually measured. It is calculated with the Matlab
script by using the rotor position information (which is also calculated) and the phase
current. Negative advance angles in the figures are when the phase is energised after the
unaligned position.
Figure 109 Simulation results of the two-phase machine driven from an H-bridge inverter under
voltage control. Conduction angle is 120 degrees (electrical), no advancing, current demand is 30A
and the dc-link voltage is 175V. Top trace = phase voltages (red for L1, blue for L2); middle trace =
phase currents (red for L1, blue for L2); bottom trace = line current. Speed of operation is 1000RPM
172

TWO PHASE DRIVES -
Figure 110 Phase current and machine total output torque comparison of simulated and measured
results for h-bridge inverter with Advance Angle = 10Degrees; Conduction Angle = 115.4Degrees;
Speed = 1013.5RPM; Current Demand = 30A; Vdclink = 175V
173

TWO PHASE DRIVES -
174

TWO PHASE DRIVES -
175

TWO PHASE DRIVES -
176

TWO PHASE DRIVES The Matlab script mentioned at the beginning of this section is used to calculate
performance parameters. The results of the Matlab script for each of the comparison
figures (i.e. Figure 110 to Figure 113) are compared to those of the measurements and the
simulation in Table 11. Test-1, Test-2, Test-3 and Test-4 in Table 11 correspond to Figure
110, Figure 111, Figure 112 and Figure 113, respectively.
TEST-2
TEST-3
TEST-4
MEASURED
Speed [RPM]
1017
365
1179
387
Torque [Nm]
4.10
20.60
4.20
23.00
Iphase RMS [A]
6.30
17.80
6.90
18.30
Speed [RPM]
1014
366
1186
388
Torque [Nm]
4.60
22.60
4.77
23.20
Iphase RMS [A]
6.16
17.63
6.25
18.70
Error-1 [%]
12.20
9.71
13.57
0.87
Speed [RPM]
1013.50
366.20
1181.70
387.50
Torque [Nm]
4.90
21.90
4.47
23.00
Iphase RMS [A]
6.23
17.57
6.45
19.00
19.51
6.31
6.31
0.00
SIMULATED
TEST-1
SCRIPT
Table 11 Comparison of results obtained from direct measurements, predictions from the Matlab
script and the simulation for the h-bridge driven two-phase machine
Error-2 [%]
Error-1 in Table 11 is the percentage difference between the measured total average
output torque of the machine and the predicted average output torque of the machine from
the Matlab script. Likewise, Error-2 in Table 11 is the percentage difference between
the measured total average output torque of the machine and the simulated average output
torque of the machine from the dynamic simulation in Simulink.
177

TWO PHASE DRIVES Overall, the simulated and measured waveforms correlate favourably, which is also
backed-up by the performance parameter calculations from the Matlab script and the
dynamic simulation in Simulink. In Table 12 some of the Matlab script calculated
performance parameters are presented for the measurements given in Table 11.
Table 12 Script calculated performance parameters for the 2-phase 12-10 segmental rotor machine
driven from the asymmetric half-bridge inverter
Advance Angle [Elect.Deg.]

Conduction Angle [Elect. Deg.]
Current Demand [A]
Speed [RPM]
RMS Phase Current [A]
Peak Phase Current [A]
Average Total Torque [Nm]
Maximum Total Torque [Nm]
Minimum Total Torque [Nm]
Torque Ripple - MEAN [%]
Torque Ripple - MAX [%]
TEST-1
TEST-2
TEST-3
TEST-4
10
115.4
30
1013.5
6.16
10.04
4.58
11.08
-0.14
244.7
101.3
0
127.7
30
366.2
17.63
32.23
22.6
39.8
-0.071
176.6
100.2
20
175.6
30
1181.7
6.25
9.72
4.8
8.99
-0.11
190.7
101.2
40
176.8
30
387.5
18.69
30.90
23.2
37.50
-3.33
176.0
108.9
Torque Ripple MEAN [%] in Table 12 is calculated by

(T max T min )
Tripple MEAN = 100
Tave
Equation 31
Where TrippleMEAN is the torque ripple based on peak-to-peak instantaneous torque

variation and the average torque; T max is the maximum output torque; T min is the
minimum output torque and Tave is the average output torque.
Torque Ripple MAX [%] in Table 12 is calculated by

(T max T min )
Tripple MAX = 100
T max
Equation 32
Where TrippleMAX is the torque ripple based on peak-to-peak instantaneous torque

variation and the maximum output torque; T max is the maximum output torque and
178

TWO PHASE DRIVES T min is the minimum output torque. In Table 12 the Torque Ripple MAX [%] values
are all more than 100%. This can only happen if the T min value is negative (Minimum
Total Torque in Table 12).
7.2.2 Two-Phase Machine Driven by Asymmetric Half-Bridge

Inverter
The two-phase machine in section 7.2.1 has also been driven by an asymmetric halfbridge inverter of the form given in Figure 114. This section summarises the
measurements taken from the drive system and the results obtained from the dynamic
simulation.
Figure 114 Asymmetric half-bridge converter driving a two-phase machine
From Figure 115 to Figure 118 the simulation results of the asymmetric half bridge driven
two-phase machine are compared to the measurements taken from the test rig under
various operating conditions. Once again the Matlab script is used to calculate
performance parameters. The results of the Matlab script for each of the comparison
figures (i.e. Figure 115 to Figure 118) are compared to those of the measurements and the
simulation in Table 13. Test-1, Test-2, Test-3 and Test-4 in Table 13 correspond to Figure
115, Figure 116, Figure 117 and Figure 118, respectively.
179

TWO PHASE DRIVES -
SIMULATED
SCRIPT
MEASURED
Table 13 Comparison of results obtained from direct measurements, predictions from the Matlab
script and the simulation for the asymmetric half bridge driven two-phase machine
TEST-1
TEST-2
TEST-3
TEST-4
Speed [RPM]
937.50
802.14
400.00
364.96
Torque [Nm]
4.40
4.70
14.80
14.30
Iphase RMS [A]
8.50
5.40
15.40
18.70
Speed [RPM]
945.00
805.40
400.30
364.70
Torque [Nm]
5.00
5.21
14.80
15.62
Iphase RMS [A]
7.90
5.34
15.21
18.46
Error-1 [%]
13.52
10.85
0.00
9.23
Speed [RPM]
945.00
805.40
400.30
364.70
Torque [Nm]
5.41
5.82
14.30
15.77
Iphase RMS [A]
7.95
5.82
14.90
17.97
22.95
23.83
-3.38
10.28
Error-2 [%]
180

TWO PHASE DRIVES -
results for asymmetric half-bridge inverter with Advance Angle = 12Degrees; Conduction Angle =
175.75Degrees; Speed = 945RPM; Current Demand = 30A; Vdclink = 175V
181

TWO PHASE DRIVES -
114.8Degrees; Speed = 805.4RPM; Current Demand = 30A; Vdclink = 175V
182

TWO PHASE DRIVES -
183

TWO PHASE DRIVES -
results for asymmetric half-bridge inverter with Advance Angle = -20Degrees; Conduction Angle =
184

TWO PHASE DRIVES Once again, the correlation between the simulated and measured waveforms is good. The
performance parameter predictions from the Matlab script and the dynamic simulation
are also good. In Table 14 some of the Matlab script calculated performance parameters
are presented for the measurements given in Table 13.
Table 14 Script calculated performance parameters for the 2-phase 12-10 segmental rotor machine
driven from the h-bridge inverter

Current Demand [A]
Speed [RPM]
TEST-1
TEST-2
TEST-3
TEST-4
12
175.75
30
945
7.90
12.12
4.99
11.25
-1.72
259.6
115.3
0
114.8
30
805.4
5.34
11.03
5.21
12.63
-0.08
243.8
100.6
40
124.6
30
400.3
15.21
29.43
14.80
37.46
-4.48
283.4
111.9
-20
176.2
30
364.7
18.46
30.28
15.62
37.35
-17.96
354.1
148.1
Torque Ripple MAX [%] and Torque Ripple MEAN [%] in Table 14 are
calculated by using Equation 31 and Equation 32, respectively. It is apparent from
Equation 31 and Equation 32 that Torque Ripple MEAN [%] will always be greater
than Torque Ripple MAX [%].
The Matlab script is also able to generate the flux-linkage locus for a given set of
measurements. An example is given in Figure 119. The flux-linkage locus in Figure 119
was generated for the test condition given in Figure 113.
185

TWO PHASE DRIVES -
Figure 119 Estimated flux-linkage locus for h-bridge inverter driven system with Advance Angle =
40Degrees; Conduction Angle = 176.8Degrees; Speed = 387.5RPM; Current Demand = 30A; Vdclink
= 175V
The discrepancies between measured and simulated results can be attributed to several
factors:
1. Simulated results were based on FE simulated flux-linkage characteristics of the

machine. The torque characteristics of the machine were also not measured but
derived from the flux-linkage characteristics of the machine.
2. The real drive itself suffers from the discrete nature of its digital controller. A
100usec control cycle equates to an electrical rotation of 6.0o at 1000 rpm
operation (for the machine with 10 rotor segments). This means that the angle at
which commutation takes place can be in error up to this amount.
3. It is likely that the position encoder of the drive system also suffers from a certain
amount of lag due to the flexible nature of the coupling connecting it to the
186

TWO PHASE DRIVES machine. As a result, the sampled current waveform may be phase shifted towards
the aligned position by a small angle.
7.3 INVERTER VA-RATING COMPARISON

The main cost of an SR drive system is its power electronic converter. Therefore, the
performance of the converter plays a significant role in cost and performance of the
overall drive system. After correlating the simulation models for the h-bridge and the
asymmetric half-bridge inverter driven two-phase machine, comparison has been made
between the performances of the two drive circuits. It must be underlined that the main
aim here is not assessing how well a particular IGBT or MOSFET operates with one of
the two inverter circuits, but is to analyse how the asymmetric half-bridge and h-bridge
inverters perform compared to each other driving a two-phase machine.
The size and cost of an inverter drive are usually assessed by determining its rating. There
are two commonly used techniques to evaluate the inverter rating of a drive system [206]:
1. The peak voltage and the current that each power electronic device must withstand are
calculated. The peak voltage and current are then multiplied with the number of
devices in the circuit. Inverter rating is then expressed in terms of kVA of inverter
power per kW of motor output power.
2. A repeat of the first technique but instead of using the peak device current, the rms
device current is used. The inverter rating is again expressed in terms of kVA per kW
of motor output power.
The above methods result in different comparison results and the decision as to which
method is more appropriate depends on the application. Ultimately, the main factor used
to determine a suitable device for an application is the calculation of its peak junction
temperature. This is determined by the losses in the device and its thermal resistance to
the heatsink. The device junction temperature should be calculated at low and high
speeds. In case of high speed operation, the device junction temperature is dominated by
the average device losses. On the other hand, when the motor is operating at low speeds,
the device junction temperature will reach its steady-state value within an electrical cycle
187

TWO PHASE DRIVES when the current is at its peak (assuming that the motor speed is low enough to reach this
steady-state condition. Otherwise, more complex calculations involving the thermal time
constant of the device are required.). This suggests that at low speeds it will be more
appropriate to use the inverter rating comparison based on peak device current whereas at
high speeds this should be based on rms device current. Clearly, the application
conditions need to be known before the decision is made in choosing the method for
determining the inverter rating.
In practice, the total cost and the size of an inverter drive requires additional
considerations other than the peak / rms VA rating calculations. These can be summarised
as follows:
1. Number of Switches: The cost of one device is not necessarily twice the cost of two
devices of half the VA rating. The assessment should also include the available
packaging space, cost relating to component assembly and additional circuitry
required, e.g. the number of drive and control circuits. In general, an inverter with a
lower switch count is preferable.
2. Circuit Topology: Indirect savings can be made by reducing the size of the PCB
(through reduction in material cost), heatsink mounting area and assembly cost by
using a different circuit topology. For example, if the freewheel diodes are in parallel
with the power devices they can easily be placed within the same package and hence
the component count can be reduced (note that the total cost of the power devices is
not necessarily reduced). An inverter solution can also be favourable in terms of
number of connections required between the motor and the inverter.
3. Total Inverter Loss: In a situation where the device rating is the same between two
inverter topologies (based on the peak voltage and current requirements of individual
devices) but the number of active devices (i.e. loss generating devices) at any given
time is different, the heatsink requirements will be different. For example, in one
drive two phases may be conducting at a time, and in the other only one phase,
resulting in twice the inverter losses and hence twice the heatsink rating (which will
cost more and take up more space).
188

TWO PHASE DRIVES The following two sections detail the VA-rating comparison of the two phase drives
driven from the h-bridge and asymmetric half-bridge inverters using the two assessment
techniques described above. It should be noted that the results are obtained from the
validated dynamic simulation of the machine and controller. The use of the simulation is
justified with the results presented in the preceding sections of this chapter where very
good correlation has been demonstrated between the simulated and measured waveforms
for both drives at high and low speeds. The first section below is the comparison of
inverter VA-ratings when both drives operate under full voltage control (high speed
operation). The second section follows on from the first one but looks at the comparison
when both drives operate under current control (low speed operation). Negative advance
in the following sections is when the phase is energised before the unaligned position and
positive advance is when the phase is energised after the unaligned position. For all
simulations the dc-link voltage is fixed at 175V. It should be noted that the drive
efficiency calculations in the following two subsections do not include the machine iron
losses. However, it is believed that the exclusion of iron loss values does not affect the
relative comparison results severely.
7.3.1 Drive under Voltage Control

For the purposes of this comparison, both drives have been simulated at 1200rpm and the
current demands were set to 50A in order to ensure operation under full voltage control
with both drives. It should be noted that the current demand for the h-bridge circuit is set
for the line current rather than the phase current. With both drives the machine rated
output torque target was set to 10Nm. In order to maximise the output torque at high
speed, the conduction angle was set to 180o for both drives. The number of turns of the
machine was then adjusted (by using the simulation GUI) in order to achieve 10Nm at
1200rpm with both drives. The aim here is to maximise the number of turns for the given
speed and torque requirements in order to minimise the current in the circuit. This way,
the inverter losses are minimised for the given speed and torque. The number of turns of
the machine was reduced if the machine was not able to achieve the required torque and
increased if the machine was more than able to achieve it. The final number of turns was
chosen when the machine was just able to achieve the required torque at the specified
speed when the advance angle was optimised for the specified conduction angle. It should
189

TWO PHASE DRIVES be noted that any change in the number of turns does not affect the machine copper
losses. The same amount of stator slot area with the same slot fill factor is available to fill
with copper. For example, if the number of turns was doubled the current demand would
be halved and as a result the copper losses would remain constant because the conductor
cross sectional area per turn would need to be halved for the same slot area and fill factor.
Table 15 summarises the simulation results for the asymmetric half-bridge and h-bridge
driven two phase machine when the drive is under voltage control at 1200rpm aiming to
achieve 10Nm of output torque.
It is not surprising to see that both drives result in similar phase currents and consequently
similar winding copper losses (number of turns is the same for both drives). This is due to
the fact that the conduction angle is set to 180o and with this conduction angle both drives
work in a very similar way and generate the same phase currents. However the devices
and the diodes of the h-bridge circuit do not see all of the phase current (due to circulation
of current around the phase windings and the diodes) whereas all the phase current goes
through the devices and diodes of the asymmetric half-bridge circuit. This can be
observed in Figure 120. As a result, inverter ratings of the h-bridge topology compares
favourably against those of the asymmetric half-bridge topology. The difference is
especially large when the VA rating comparison is based on the RMS device current
(almost 20% improvement). This is also reflected in the total inverter losses as the hbridge generates more than 16.0W less inverter losses. However, the gain in the inverter
losses is somewhat spoiled by the additional conduction losses generated by the diodes in
series with the phases. As a result total power electronic losses of the h-bridge drive are
only 6.0W less than those of the asymmetric half-bridge drive. In terms of overall drive
efficiency both drives perform similarly. The h-bridge drive is less than 1.0% more
efficient than the asymmetric half-bridge drive. In Table 15 torque ripple results are
virtually the same, as the phase currents are the same for both drives.
190

TWO PHASE DRIVES Table 15 Summary of comparison of the performances of h-bridge and asymmetric half-bridge
driven two-phase machine drive systems under voltage control
Vdc_Link [V]
Number of Motor Phases
Speed [RPM]
Average Output Torque [Nm]
Output Power [W]
Demand Current [A]
Nturns per coil - ORIGINAL
Nturns per coil - FINAL
R_Phase [Ohms]
Advance Angle [Deg.]
Conduction Angle [Deg.]
H-Bridge
Asymmetric
Half-Bridge
175
2
1200
10.0
1257.0
50
88
87
0.8582
-83.5
180
175
2
1200
10.0
1257.4
50
88
87
0.8582
-83.5
180
Percentage
Difference [%]
Iphase_RMS [A]
10.85
10.85
Winding Cu Losses [W]
202.0
202.2
Iphase_PEAK [A]
Total Inverter Losses [W]
Total Phase Diode Losses [W]
Total Power Electronic Losses [W]
Drive Efficiency [%]
18.82
50.7
10.0
60.8
82.7
18.82
66.4
0.00
66.4
82.4
Swt_VARating_Peak [kVA / kW]
9.8
10.5
Swt_VARating_RMS [kVA / kW]
5.0
5.6
Dio_VARating_Peak [kVA / kW]
9.6
9.6
Dio_VARating_RMS [kVA / kW]
1.8
2.3
Total Peak VA Rating [kVA / kW]
19.4
20.1
Total RMS VA Rating [kVA / kW]
6.7
7.9
Torque Ripple Pk-2-Pk [%]
114.1
114.1
Torque Ripple Ave-2-Pk [%]
257.4
257.3
0.09
6.91
12.69
-0.06
28.17
3.46
16.78
0.00
-0.04
Some of the abbreviations in Table 15 are described in Table 16. The Percentage
Difference [%] column in Table 15 is calculated by dividing the value under the
Asymmetric Half-Bridge column with that under the H-Bridge column.
191

TWO PHASE DRIVES Table 16 Explanation of abbreviations used in Table 15 and Table 17
Vdc_Link
Nturns
R_Phase
Iphase_RMS
Iphase_PEAK
Cu
Available DC-Link Voltage

Number of turns
Phase resistance at room temperature
RMS phase current
Peak phase current
Copper
Swt_VARating_Peak
VA rating of IGBTs based on Peak Current
Swt_VARating_RMS
VA rating of IGBTs based on RMS Current

VA rating of Recovery Diodes based on Peak
Current
VA rating of Recovery Diodes based on RMS
Current
Dio_VARating_Peak
Dio_VARating_RMS
Figure 120 Phase and IGBT currents for the h-bridge and asymmetric half-bridge inverter two-phase
drives at 1200 rpm with 175V dc-link and 50A of current demand
192

TWO PHASE DRIVES -
7.3.2 Drive under Current Control

The motor speed was reduced to 100rpm. The number of turns determined when the
drives were under voltage control (in the previous section) is used. The controllers now
work under current control. Both drives use the same PID controller parameters. It should
be noted that the h-bridge drive is line controlled, i.e. the sum of the two phase currents is
controlled by the controller. In order to maximise the torque per unit copper loss the
conduction angle was limited to 120o for both asymmetric half-bridge and h-bridge driven
machine. The current demands and the advance angles for both drives were then
optimised in order to achieve the required output torque (same as the previous section
10Nm). Table 17 tabulates the comparison based on the simulation results for the
asymmetric half-bridge and h-bridge driven two phase machine when the drive is under
current control at 100rpm, aiming to achieve 10Nm of output torque.
From Table 17, in terms of VA ratings, both inverters perform similarly. This is reflected
in the total inverter losses as well, as the switches and diodes do not experience the
circulating phase current in the h-bridge driven machine (see Figure 121 and Figure 122).
However, the total power electronic losses are higher with the h-bridge topology as the
diodes in series with phase windings incur additional conduction losses.
Furthermore, due to the circulating currents and the long phase tail currents with the hbridge circuit, the rms phase current is higher with the h-bridge drive (see Figure 121).
This results in increased copper losses. To understand this, one needs to go back to Figure
108 and analyse the region between Points B and C. In this region the line current is zero
and the controller aims to keep this current at zero by applying positive and negative dclink voltage across the phases. As no voltage is applied across the phases on average, the
phase current only drops slightly due to resistive voltage drop and results in longer tail
current (although the line current is zero, the phase currents circulate around the phases
through the phase windings and the phase diodes). Moreover, phases start conducting
before their energisation cycles begin. This can again be explained by the help of Figure
108 (region between points A and B). Negative voltage applied across the turning-off
phase results in positive voltage applied across the off-phase. This positive voltage results
in the off phase start conducting long before the start of its actual energisation cycle.
Consequently the rms phase current is increased with the h-bridge inverter, resulting in
more copper losses.
193

TWO PHASE DRIVES Table 17 Summary of comparison of the performances of h-bridge and asymmetric half-bridge
driven two-phase machine drive systems under current control
Vdc_Link [V]
Speed [RPM]
Output Power [W]
Demand Current [A]
R_Phase [Ohms]
Advance Angle [Deg.]
Conduction Angle [Deg.]
P
I
D
H-Bridge
Asymmetric
Half-Bridge
175
2
100
10.0
104.7
13.8
88
87
0.8582
40.0
120
10
0.5
0
175
2
100
10.0
104.7
14
88
87
0.8582
40.0
120
10
0.5
0
Percentage
Difference [%]
Iphase_RMS [A]
9.12
8.26
142.8
117.1
Iphase_PEAK [A]
17.15
77.9
8.35
86.2
31.4
15.24
78.5
0.00
78.5
34.9
101.6
101.8
41.1
41.8
101.6
101.81
35.3
36.0
203.3
203.6
76.5
77.8
106.9
100.0
191.9
192.9
-17.98
0.18
1.66
0.18
1.90
0.18
1.77
-6.43
0.47
Combination of extra diode losses and higher copper losses result in the h-bridge drive
operating approximately 3.5% less efficient.
194

TWO PHASE DRIVES -
Figure 121 Phase and IGBT currents for the h-bridge and asymmetric half-bridge inverter twophase drives at 100 rpm with 175V dc-link
Figure 122 Zoomed-in IGBT currents for the h-bridge and asymmetric half-bridge inverter twophase drives at 100 rpm with 175V dc-link
195

TWO PHASE DRIVES -
7.4 SUMMARY
Before the more complicated simulation of the 6-phase drives, two phase drives were
analysed in this chapter. In terms of circuit topology, the h-bridge inverter has similarities
with the 6-phase drive topologies studied in this thesis. Using the test rig explained in
Chapter 5 Machine Construction Techniques and Test Rig Description and rearranging
the phase winding connections, the six phase segmental rotor SRM has been successfully
driven from the asymmetric half-bridge and h-bridge inverters as a 2-phase machine.
With the h-bridge circuit, the line current was the controlled parameter, whereas both
phase currents were controlled with the asymmetric half-bridge inverter. Both drives were
run under voltage and current control and measurements under various operating points
from both 2-phase drives have been presented. Matlab scripts were written to analyse
captured waveforms from both 2-phase drives. These scripts were able to estimate some
of the performance parameters such as phase RMS currents, conduction angles, running
speed (from the period of phase current), instantaneous phase torques, average total
output torque, torque ripple, shaft power and total machine copper losses (assuming
winding temperature of 100oC during steady-state operation). Measured conditions were
then simulated in the dynamic simulation package in the Matlab environment (see
Chapter 6 Simulation Models). The simulated waveforms were compared with the
measured ones and good correlation was achieved with both 2-phase drives under voltage
and current controlled conditions.
Once the simulations of both the h-bridge and asymmetric half-bridge driven 2-phase
machines were verified against the measurements, further comparison between the two
drive circuits was carried out by using the dynamic simulation package. Performance of
the 2-phase machine driven by the asymmetric half-bridge and h-bridge inverters was
then compared.
It has been discussed that many factors (the number of switches, circuit topology,
indirectly affecting the size of the PCB and the heatsink requirements) affect the size of
the inverter of a drive system. Two methods of calculating inverter size were presented in
this chapter VA rating and inverter losses. VA rating comparison in itself can be done in
two different ways one considering peak device current and the other considering the
RMS device current. It is worth noting that the relevance of each inverter size comparison
method changes with application.
196

TWO PHASE DRIVES To enable meaningful comparison of inverter VA ratings of the 2-phase drives, two points
in the operating range were selected for comparison. The drives were first compared at
1200rpm when the controllers of both inverters were under voltage control. The number
of turns and the advance angles were adjusted to achieve the target output torque of
10Nm. The number of turns was then fixed and the current demands and the advance
angles were adjusted at 100rpm where both controllers were under current control.
The overall conclusion from Table 15 and Table 17 is that the h-bridge and asymmetric
half-bridge driven two-phase segmental rotor SR drives perform very similarly under
voltage and current control. There is a slight gain with the asymmetric half-bridge
topology when the drive is under current control in terms of overall drive efficiency.
However, with h-bridge circuit, the number of required current transducers is reduced
from two to one. Furthermore, the total number of connections between the inverter and
the motor is reduced from four to two with the h-bridge inverter (assuming phase diodes
are mounted at the motor end). Under both voltage and current control, the VA rating
comparison based on peak device and diode currents suggest that the inverter topologies
behave in a similar manner. However, when the comparison is based on the rms device
and diode currents, the h-bridge topology is favourable (only very slightly under current
control). On the negative side, the h-bridge topology requires two additional phase diodes
in order to achieve unipolar phase currents. However, this is somewhat compensated as
the bridge recovery diodes can be part of the power devices with the h-bridge circuit,
whereas the asymmetric half-bridge topology requires separate recovery diodes. It must
be said that the phase diodes of the h-bridge arrangement do not need to be fast diodes
and as a result, can be relatively cheap. Although with 1/3 of an electrical cycle
conduction both drives perform similarly, the h-bridge drive outperforms the asymmetric
half-bridge drive (in terms of inverter VA ratings especially rms current based
comparison) at higher speeds when the conduction period is extended to 1/2 of an
electrical cycle. Therefore, the overall cost of the h-bridge circuit is likely to be less than
the asymmetric half-bridge circuit.
For both the asymmetric half-bridge and h-bridge inverters the drive system performance
can be optimised by adjusting the advance and conduction angles as the speed of the
machine changes. The relative importance of inverter size and machine efficiency
depends on the specific application. In terms of assessing the size of the inverter, it also
197

TWO PHASE DRIVES depends on the individual application that will determine the most appropriate method:
inverter loss or peak / rms VA.
Finally, torque ripple results have shown that there is not a significant difference between
the asymmetric half-bridge and h-bridge driven two-phase SR drives.
198

THREE PHASE DRIVES -
CHAPTER 8
Comparison to Simulation
Three Phase Drives 8.1 INTRODUCTION
This chapter contains the dynamic measurement and simulation results of the three-phase
single tooth winding (12-10) segmental rotor machine drive. The results of the 3-phase
drive have been presented in this chapter as they will form the basis of comparison
between the 3-phase drive and the 6-phase one in one of the later chapters. The results of
this comparison will be presented in Chapter 10. The 3-phase segmental rotor machine
was designed as part of a previous PhD project in the Newcastle University, U.K [97].
The 3-phase segmental rotor machine has been used as the baseline (throughout this
thesis) for comparing the performance of the 6-phase segmental and conventional rotor
SRMs designed as part of this PhD project. The inverter of the test rig has been
configured to drive the three phase machine from a 3-phase asymmetric half-bridge
inverter.
In the next section, the simulated results will be compared against those measured. As for
the two-phase drives, Matlab based scripts were used to analyse the measured
waveforms in order to predict some of the performance parameters of the drive system
and ease the comparison against simulated waveforms and performance parameters.
Examples at low and high speeds are presented. Inverter VA-rating assessment (based on
199

THREE PHASE DRIVES rms and peak device currents) under various operating conditions for the three-phase
drive is also presented in this chapter. Dynamic torque-speed characteristics of the threephase segmental rotor machine have been measured and presented in section 2.2 of this
chapter.
8.2 MEASUREMENTS AND SIMULATION

RESULTS
A dynamic simulation model for the three-phase asymmetric half-bridge inverter driven
machine has been created and the model has been correlated against the measurements
taken from the three-phase single tooth winding (12-10) segmental rotor SRM. The
measured and simulated instantaneous phase current and total output torque waveforms
were plotted on top of each other. The results of the Matlab script were compared to
those of the measurements and the simulation in the following sub-sections for several
tests. Similar to the analysis of the 2-phase drive measurements, a Matlab based script
was written to analyse the scope-captured data and generate instantaneous torque
waveforms based on phase currents and positions using FE generated flux-linkage
characteristics.
8.2.1 Comparison of Measurements against Simulations

Four different operating conditions are presented here. Test-1 and Test-4 are when the
drive is under full voltage control and Test-2 and Test-3 are when the drive is under
current control. Table 18 summarises the operating conditions of the measurements
presented in the proceeding sections.
Table 18 Tabulated summary of test conditions for the measurements presented in the proceeding
sections

Current Demand [A]
Speed [RPM]
TEST-1
TEST-2
TEST-3
TEST-4
12
119.2
10
978.8
-8.8
164.6
10
1047
0
118.6
10
588.2
-10
154.4
10
1109.1
200

THREE PHASE DRIVES When the drive is under full voltage control the Matlab script is additionally able to
determine the power device and diode currents from the measured line current and
calculate the VA rating of the inverter. The same could not be done for the drive under
current control as neither the switching states of the devices nor the actual device currents
were measured. However, for all of the examples below, the script determines the
advance angle and conduction angle applied by the controller and the resulting operating
speed and average output torque. The script is also able to determine performance
parameters, such as the rms and peak phase currents, and the maximum and minimum
instantaneous total torque. The advance angle, conduction angle, and speed determined by
the Matlab script and the drive current demand are then input into the dynamic
simulation. Power device and diode VA ratings are determined, based on rms and peak
values of currents as well as total inverter losses and total machine copper losses. The
drive efficiency quoted under the Simulated column in the tables within the following
sections is calculated by Equation 33.
= 100
(Tave (2 N s / 60))
(Tave (2 N s / 60)) + Pinv + Pcu
Equation 33
Where Tave is the average total output torque in [Nm], N s is the operating speed in [rpm],
Pinv is the total inverter losses (conduction + switching) in [W], and Pcu is the total
machine copper losses in [W]. For the total machine copper losses, it was assumed that
the windings were running at 100oC. Equation 33 does not include the machine iron
losses. However, it was believed that Equation 33 could be still be used as an indicator for
the overall drive efficiency.
201

8.2.1.1 Test-1 (Voltage Control)
Figure 123 Phase voltage waveform with Advance Angle = 12o; Conduction Angle = 119.2o; Speed =
978.8RPM; Current Demand = 10A; Vdclink = 562V
Figure 124 Comparison of simulated and measured phase current and torque with Advance Angle =
12o; Conduction Angle = 119.2o; Speed = 978.8RPM; Current Demand = 10A; Vdclink = 562V
202

Figure 125 IGBT and recovery diode currents with Advance Angle = 12o; Conduction Angle =
119.2o; Speed = 978.8RPM; Current Demand = 10A; Vdclink = 562V
During Test-1, the machine is running relatively fast. As the turn-on point of the phase is
delayed past the unaligned rotor position and the conduction angle is limited to
approximately 1/3 of an electrical cycle, the current never reaches the required level. The
resistive voltage drop across the phase windings becomes negligible and all of the
available voltage is consumed by the change in the phase flux-linkage. At the end of the
energisation cycle the power devices are turned-off and the phase current decays to zero.
However, the current decay gets slower and slower as the phase inductance rises when the
rotor position approaches alignment. This is reflected in the output torque waveform
(Figure 124) as a relatively high torque ripple is experienced under Test-1.
203

THREE PHASE DRIVES Table 19 Summary of measured, script and simulation determined results for Test-1 where Advance
Angle = 12o; Conduction Angle = 119.2o; Speed = 978.8RPM; Current Demand = 10A; Vdclink =
562V

Current Demand [A]
Speed [RPM]
RMS Device Current [A]
Peak Device Current [A]
RMS Diode Current [A]
Peak Diode Current [A]
Total Device RMS VA [kVA / kW]
Total Device PEAK VA [kVA /
kW]
Total Diode RMS VA [kVA / kW]
Total Diode PEAK VA [kVA / kW]
Total Copper Losses [W]
Shaft Power [W]
Measured
Script
Simulated
10
969
3.64
7.46
12
119.2
10
978.8
3.64
7.46
12
119.2
10
978.8
3.58
7.5
3.2
7.5
1.71
7.5
13.83
19.81
5.66
102.31
71.43
7.51
14.6
14.26
19.2
6.63
88.15
65.47
7.2827
17.04
4.03
16.8
1481.51
1461.65
17.8
4.0
17.8
26.83
99.0
1418
91.9
0.1402
0.1102
Torque per Copper Loss [Nm/W]

Torque per Total Loss [Nm/W]
204

8.2.1.2 Test-2 (Current Control)
Figure 126 Phase voltage waveform with Advance Angle = -8.8o; Conduction Angle = 164.6o; Speed =
1047RPM; Current Demand = 10A; Vdclink = 567V
Figure 127 Comparison of simulated and measured phase current and torque with Advance Angle = 8.8o; Conduction Angle = 164.6o; Speed = 1047RPM; Current Demand = 10A; Vdclink = 567V
205

Angle = -8.8o; Conduction Angle = 164.6o; Speed = 1047RPM; Current Demand = 10A; Vdclink =
567V

Current Demand [A]
Speed [RPM]
kW]
Shaft Power [W]
Measured
Script
Simulated
10
1061
5.74
11.35
-8.8
164.6
10
1047
5.74
11.35
-8.8
164.6
10
1047
5.93
11.15
5.55
11.15
3.0497
11.13
23.24
25.73
18.6
30.68
27.71
6.92
20.5
2277.71
21.3
25.61
18.07
35.40
29.44
2335.37
14.9
3.84
14.87
79.2
269.4
2548.1
87.97
0.0863
0.0667

During Test-2, the controller enters into current control mode right at the end of the
energisation cycle. This can be clearly seen in the figures for the device and diode
currents of Test-2 in Figure 134 and Figure 135, respectively (red traces in both figures).
This is not as obvious in Figure 126. The reason the simulation predicts a longer chopping
period under Test-2 is believed to be due to the difficulty in ensuring the PID controller
gains in the current controller of both the simulation and the actual drive are identical.
206

8.2.1.3 Test-3 (Current Control)
Figure 128 Phase voltage waveform with Advance Angle = 0o; Conduction Angle = 118.6o; Speed =
Figure 129 Comparison of simulated and measured phase current and torque with Advance Angle =
0o; Conduction Angle = 118.6o; Speed = 588.2RPM; Current Demand = 10A; Vdclink = 575V
207

Angle = 0o; Conduction Angle = 118.6o; Speed = 588.2RPM; Current Demand = 10A; Vdclink = 575V

Current Demand [A]
Speed [RPM]
kW]
Shaft Power [W]
Measured
Script
Simulated
10
586
5.6
10.75
0
118.6
10
588.2
5.6
10.75
0
118.6
10
588.2
5.641
10.73
4.93
10.72
2.87
10.71
26.06
36.9
7.04
114.43
80.90
10.1
27.5
1687.56
26.8
34.4
9.84
91.64
71.40
1650.78
23.03
6.7
23.03
168.4
244.7
1605.3
79.53
0.1065
0.0631

During Test-3, the machine is running relatively slowly compared to other tests presented
in this chapter. The controller has enough time to build the phase current and reach the
required value before the phase inductance starts limiting the rate of change of current.
The current rises rapidly against a small inductance around the unaligned rotor position.
Later in the energisation cycle the rate of rise and fall of current is reduced in the PWM
cycle as the inductance increases. Once the demand current is reached, the controller
enters into current control mode and aims to keep the phase current at the same level until
the end of the energisation period. This can be seen in Figure 134 and Figure 135. Once
the conduction period is elapsed, the controller turns off the devices and negative voltage
is applied across the phase windings to de-flux the phase before the aligned position is
reached.
208

8.2.1.4 Test-4 (Voltage Control)
Figure 130 Phase voltage waveform with Advance Angle = -10o; Conduction Angle = 154.4o; Speed =
Figure 131 Comparison of simulated and measured phase current and torque with Advance Angle = 10o; Conduction Angle = 154.4o; Speed = 1109.1RPM; Current Demand = 10A; Vdclink = 567V
209

Figure 132 IGBT and recovery diode currents with Advance Angle = -10o; Conduction Angle =
154.4o; Speed = 1109.1RPM; Current Demand = 10A; Vdclink = 567V
Similar to Test-1, the machine is running relatively fast during Test-4. Unlike Test-1, the
turn-on point of the phase is in advance of the unaligned rotor position and the conduction
angle approaches 1/2 of an electrical cycle. Although compared to Test-1, the phase
energisation is advanced and the phase is allowed to conduct longer, the current still never
reaches the required level of 10A and the controller is under voltage control.
210

Angle = -10o; Conduction Angle = 154.4o; Speed = 1109.1RPM; Current Demand = 10A; Vdclink =
567V

Current Demand [A]
Speed [RPM]
kW]
Shaft Power [W]
Measured
Script
Simulated
10
1100
5.03
8.81
-10
154.4
10
1109.1
5.03
8.81
-10
154.4
10
1109.1
5.04
9.03
4.6761
9.02
2.37
9.02
20.65
23.2
14.73
41.02
36.51
6.38
23.7
21.16
24.32
15.17
43.24
37.62
6.0943
11.9372
3.0348
11.8196
2730.04
2457.62
12.81
3.22
12.81
42.9
195
2398.4
90.98
0.1059
0.0868

Based on the results obtained from the four test conditions presented above, the
correlation between the simulated and measured waveforms is very good. From Table 19
to Table 22 the performance parameter predictions from the Matlab script and the
dynamic simulation also correlate very well with each other. Phase, device and diode
currents from all four test conditions are plotted in the same figure in Figure 133, Figure
134, and Figure 135, respectively.
211

Figure 133 Simulated phase currents of Test-1, Test-2, Test-3 and Test-4
Figure 134 Simulated device currents of Test-1, Test-2, Test-3 and Test-4
212

Figure 135 Simulated diode currents of Test-1, Test-2, Test-3 and Test-4
The drive efficiency is higher when the machine is running faster and the controller is
under full voltage control. This is typical of SR drives. Although the drive under Test1,
Test2 and Test-4 operates approximately at the same speed (978.8rpm, 1047rpm and
1109.1rpm, respectively) the average output torques and the drive efficiencies are
different. As the conduction angle is increased the motor is able to generate more output
torque. However the machine efficiency is comprised as more current flows through the
power devices and the phase windings, increasing the total inverter and copper losses.
Although the peak and rms device and diode currents are the lowest during Test-1 (see
Figure 134 and Figure 135), the inverter VA rating based on both peak and rms currents is
the second worst after Test-3. This is due to the reduced shaft power during Test-1. A
much better performance is achieved under Test-4 in terms of shaft power, output torque
ripple and inverter VA rating by energising the phase earlier and prolonging the
conduction period. The compromise is the reduced drive efficiency due to the increased
copper and inverter losses. However, the drop in drive efficiency between Test-1 and
Test-4 is only 0.9%.
213

8.2.2 Dynamic Torque vs. Speed Characteristics

Torque/speed characteristics of the 3-phase single-tooth winding segmental rotor SRM
were measured on the test rig for various conduction angles. The advance angle has been
varied for each conduction angle and measurements taken at different speeds for each
advance angle. The speed was varied by varying the load (see Chapter 5 Machine
Construction Techniques and Test Rig Description for the description of the load used in
the test rig). The measurements were taken with a dc-link voltage of 580V. Phase current
was controlled and the demanded current level was fixed at 10A. The results are
presented in Figure 136. The conduction angle was varied from one third of an electrical
cycle (i.e. 120o) to half of an electrical cycle (i.e. 180o). For each conduction angle, the
advance angle was varied from 60o before the unaligned position to 60o after the
unaligned position with 20o steps. The aim was to see the effects of control angles (i.e.
advance and conduction angles) on the torque vs. speed characteristics of the machine.
In general, advancing the phase energisation instance with respect to the unaligned
position of the phase is beneficial in torque production as the speed increases. As the
speed reduces, delaying the phase energisation instance after the unaligned position helps
with torque production, so that, with high conduction angles, a delay angle as high as 60o
is helpful in torque production. However, no matter what the conduction angle is, as the
speed increases the torque production reduces rapidly with high delay angles. This is
because it starts to become too late for the controller to energise the phase windings and
increase the phase current to the required level, as well as keeping the negative torque low
enough to generate high average output torque. As a consequence, the phase current stays
positive for longer and longer. There is still positive current flowing in the phase
windings even when the rate of change of inductance changes direction, resulting in a
large amount of negative torque.
Until approximately 170o of conduction angle, switching the phases 60o before the
unaligned position results in a flat torque-speed characteristics. This means that the
machine output power increases with speed (however, the actual torque level and the
speed range where the power increases is somewhat limited). Although generally with
higher conduction angles more torque is possible, this comes with a significant
compromise in the total machine copper losses.
214

THREE PHASE DRIVES With longer conduction periods the rms phase currents are also increased, resulting in
elevated total machine copper losses. At low speed, it is often more advantageous to
increase the current demand when more torque is required (from the point of view of
copper and inverter losses) rather than extending the phase energisation period. At higher
speeds, the controller operates under full voltage control and increasing the current
demand will not have any effect on the machine output torque. The only way to increase
the torque then, is to increase the conduction angle and to advance the switching-on
instance of the phases before the unaligned rotor position. The general conclusion from
the torque vs. speed curves given in Figure 136 is that there is an optimum energisation
instance with respect to the unaligned position to generate the highest average output
torque for a given operating condition (i.e. conduction angle and speed) without
compromising the drive efficiency and the inverter VA-rating significantly.
215

Figure 136 Measured torque-speed curves of 3-phase single-tooth segmental rotor SRM. Conduction
angle varied from 120 degrees to 180 degrees. Phase current demand was fixed at 10A and dc-link
voltage was 580V
216

8.3 SUMMARY
Several measurements taken from the 3-phase segmental machine, driven from the threephase asymmetric half-bridge circuit, have been presented. The simulated waveforms
were compared with the measured ones and good correlation was achieved throughout,
including current and voltage control modes involving significant magnetic saturation.
Good correlation between the measured and simulated waveforms also suggest that the
subtleties introduced by the digital PWM control system can be ignored and satisfactory
simulation results can still be achieved.
At low speeds, the conduction angle should be limited (e.g. 120o) to minimise the copper
losses. At low speeds, the machine BEMF is relatively low and there is plenty of time to
build the phase current from zero to the required level. As a result, the controller operates
under current control and the applied phase voltage is chopped in order to maintain the
current at the required level. The current demand is usually determined by considering the
load torque. At higher speeds, the machine BEMF is increased and there is a limited
amount of time to increase the phase current to the required level. Therefore, the
conduction angle is maximised (e.g. 180o) in order to fully utilise the available dc-link
voltage and generate the required torque. Furthermore, under voltage control, the phase
energisation is advanced in order to increase the current when the phase inductance is at
its minimum. However, the torque ripple is likely to increase for a given conduction angle
as more negative torque is likely to be generated with increased advance angles. Under
voltage control the drive efficiency is usually higher compared to current control (as
constant output power can be maintained even tough the output torque is reduced). With
increased conduction angle, the total inverter and machine copper losses are
compromised. However, as the region, where the output torque of individual phases
overlap is increased, the output torque ripple is reduced with increased conduction angles.
217
Chapter 9 MACHINE OPERATION AND SIMULATION SIX PHASE DRIVES
CHAPTER 9
Simulation - Six Phase
Drives 9.1 INTRODUCTION
In the two preceding chapters dynamic simulation and measurement results for the 2phase and 3-phase drives were presented. This chapter is dedicated to the dynamic
measurement and simulation results of the prototype six-phase (12-10) segmental rotor
machine drive. Simulation models of the star and delta connected six phase machine were
created. These models are believed to be valid where the phases of the machine were
mutually de-coupled. These simulation results would not be accurate for the six-phase
segmental rotor machine with its strong mutual coupling effects between phases and
hence were not compared against the dynamic measurements taken from the segmental
rotor prototype machine. However, the six-phase 12-10 conventional rotor machine
(designed as part of this PhD project) had mutually de-coupled phases and could be
analysed using these simulation models. The phases of the machine were connected in
pairs and in parallel through phase diodes, these pairs of phases were connected in star.
The inverter of the test rig has been configured to drive the six phase machine from a 3phase full bridge circuit.
218

Matlab based scripts were used to analyse the measured waveforms in order to predict
some of the performance parameters of the drive. Particular attention was paid to the
instantaneous output torque waveform in order to assess the torque ripple performance of
the six-phase drive.
Dynamic simulation results of a six-phase SRM with mutually de-coupled phases,
connected in both star (Figure 137) and delta (Figure 138) configuration, operating under
current and voltage conditions are given in section 9.2.
Section 9.3 comprises the dynamic measurement results of the six-phase prototype
segmental rotor machine. Several examples have been given for low speed and high speed
operation. When the machine is under voltage control at high speeds, the line current
219

waveforms were used to derive the currents flowing through the power devices and
diodes in the 3-phase full bridge circuit. These currents, along with the dc-link voltage
during testing, were used to estimate the VA-rating of the inverter. Finally, in section 9.4,
the findings of the chapter are summarised.
9.2 DYNAMIC SIMULATION RESULTS

The measurements taken from the star-connected six-phase segmental rotor machine
driven from the three phase full bridge circuit are given in section 9.3.
As part of this PhD project a six-phase conventional rotor SRM was also designed. The
phases of this machine were magnetically de-coupled, meaning the phases could be
analysed in isolation in the simulation (Flux linking any phase is assumed to be solely due
to the rotor position and the current in that phase only). Thus, it was still believed to be a
useful exercise to realise the simulation models for the six-phase machine with mutually
de-coupled phases when driven from the 3-phase full bridge circuit and connected in star
and delta configurations. This exercise will be beneficial in understanding the operation
of the 3-phase bridge circuit with the 6-phase machine topology when there were no
mutual coupling effects in the control of the drive. The following two subsections
summarise the simulation results obtained from the dynamic simulation for a six-phase
SRM with mutually de-coupled phases, driven from the 3-phase bridge circuit. FEA
generated flux-linkage characteristics of the prototype 6-phase machine with a
conventional toothed rotor structure were used in the dynamic simulation.
9.2.1 Mutually De-Coupled Star Connected Line Current

Controlled Six-Phase Drive
Previously in [89] a switched reluctance motor driven from a 3-phase bridge circuit has
been demonstrated and various switching waveforms were analysed. In [89] the controller
was able to control the phase currents directly, however, when the phases of the prototype
6-phase machine are connected in star and driven from the 3-phase bridge circuit, phase
currents can no longer be directly regulated by the controller that is linked to the line
currents of the drive. Results in this section were obtained from the dynamic simulation of
220

the six phase conventional machine with the phase windings connected in star
configuration (see Chapter 6 Simulation Models). The model presented is for machines
with mutually de-coupled phases. Examples at low and high speed are given, with the
results summarised in Table 23 and Table 24. Torque ripple values given in Table 23 and
Table 24 are calculated by using Equation 32 in Chapter 7 Machine Operation and
Comparison to Simulation Two Phase Drives. Negative advance angle denotes
energising before the unaligned rotor position and zero advance angle stands for
energising at the unaligned rotor position. For explanation of some of the variables in
Table 23 and Table 24 see Table 16 in Chapter 7 Machine Operation and Comparison
to Simulation Two Phase Drives.
9.2.1.1 Drive under Current Control

Here, the six-phase machine is connected in star configuration and driven from the 3phase full bridge converter. The three line currents were the controlled parameters and
each line had its own PID and PWM controllers, with the current demand set to 30A and
the machine rotating at 250rpm. At this speed, the machine BEMF is low and the
controller has enough time to build the current from zero to the demanded level. Thus, the
controller works in the current control mode. The phases were energised at the unaligned
rotor position. The conduction period was fixed to 120o and the dc-link voltage was set to
175.0V. The results from the dynamic simulation are depicted from Figure 139 to Figure
148. At this speed, the line currents are very close to those expected from the ideal 120o
conduction. In Figure 139, the black trace is the instantaneous phase voltage waveform
obtained directly from the simulation. However, it is quite difficult to use the simulated
phase voltage waveform to explain the operation of the drive. For this purpose the red
trace (Vphase Filtered) in Figure 139 is generated after post-processing the simulated
phase voltage via a simple moving average based filter.
221
Figure 139 Simulated phase voltage (in black), phase-A current (x5 in blue), phase-D current (x5 in
magenta), line-X current (x5 in green) and filtered phase voltage (in red) for the 3-phase bridge
circuit driven star connected 6-phase conventional rotor SRM with mutually de-coupled phase
windings running at 250rpm
In Figure 139 it is quite clear that the phase currents have a two-stage tail-off period.
There are two reasons for this:
Not all the dc-link voltage is available to de-flux the phase windings. This is a virtue
of the star-connected three phase system and can be explained with the help of Figure
140. Across a pair of parallel connected phases (phaseA / phaseD, phaseB / phaseE or
phaseC / phaseF in Figure 140), the controller is only able to control the potential at
one end of the circuit (Vx, Vy or Vz in Figure 140), the other end is the uncontrolled
star-point potential (Vs in Figure 140).
222
Figure 140 Six phase machine is star configuration
The controller is not able to control the phase currents directly. Instead, the line
currents are controlled. When the line current falls to zero the controller aims to keep
it at zero. This effectively means no voltage is applied across the phase windings and
the decay in the phase current is very small.
As the controller is only able to regulate the line potential directly in order to control the
line current, it is hard to see that the controller is under current control by observing the
phase voltage waveform in Figure 139. With respect to Figure 140, Figure 141 depicts
the line-X potential (Vx) and current (IlineX). In Figure 141, the controller is clearly
under current control and the Vx potential swings between positive and negative Vdclink
/ 2 (i.e. 87.5V). All of the available line potential is applied at point Vx in Figure 141.
This continues until the line current reaches the demanded level. Once the demanded
level is reached the controller chops the voltage between positive and negative values
until the end of the conduction period. Once the conduction period elapses the controller
applies the full negative voltage at point Vx to reduce the line current back down to zero.
223
Figure 141 Simulated line potential (in black) and current (x5 in green) for the 3-phase bridge circuit
driven star connected 6-phase conventional rotor SRM with mutually de-coupled phase windings
running at 250rpm
To clarify the operation of the star-connected drive under current control Figure 142 has
been generated.
224
Figure 142 Simulated phase-A current (in red), phase-D current (in blue), line-X current (in green)
for the 3-phase bridge circuit driven star connected 6-phase conventional rotor SRM with mutually
de-coupled phase windings running at 250rpm
The operating states in Figure 142 are summarised below with respect to the circuit
diagram in Figure 137:
Point A : Phase-A is turned on.

Point B : Phase-A is turned off.
Point C : Line current (IlineX) falls to zero.
Point D : Phase-D is turned on.
Point E : Phase-D is turned off.
It is quite clear from Figure 142 that the peak phase current is higher than the demanded
current level of 30A. However, this current is not seen by the power devices in the full
bridge as it circulates around the parallel connected phases (note the balancing current
flowing in phase-D between points B and C in Figure 142 as well). Once the current in
phase-D falls to zero the line current equals phase-A current (the phase current that is in
its conduction period). As mentioned earlier, when the line current falls to zero, the
controller aims to keep this at zero. The currents in phases A and D circulate around the
225

phase windings and the phase diodes. Between points C and D in Figure 142, the Line-X
and the star point potentials are zero (see bottom trace of Figure 146).
Figure 143 Simulated phase-A current (x5 in red), phase-D current (x5 in blue), line-X current (x5 in
green) and star-point potential (in black) for the 3-phase bridge circuit driven star connected 6-phase
conventional rotor SRM with mutually de-coupled phase windings running at 250rpm
From Figure 143, there are six distinct periods of star point potential, each corresponding
to a phase commutation point. This ensures that the currents flowing into the star point
sum up to zero. The effect of this star point potential can be seen in the phase and line
current waveforms, as a dip appears in the figures approximately half way through the
conduction period. The positive star point potential effectively reduces the available
voltage across phase-A windings when the line current is flowing in the positive direction
and the Line-X potential is connected to the positive dc-link rail (Figure 144-a). The
reduction in the available phase voltage can also be seen in Figure 139. The same applies
when phase-D conducts. Negative star point potential reduces the voltage available to
drive the phase-D current, as in this period the Line-X is connected to the negative dc-link
rail (Figure 144-b). These waveforms are very similar to those experienced with the h-
226

bridge driven two-phase drive, apart from the dip during phase commutation. This is of
no surprise, as the only difference between the two drives is the fact that the six phase
drive can only control the potential at one end of the module (e.g. comprising phases A
and D with the phase diodes in series with the phase windings). The other end is the
uncontrolled star point potential that regulates the star point current.
T1
D1
PhA
PhD
Vx+
Vs
IlineX
PhB
T2
D2
PhE
PhF
PhC
(a)
T1
D1
PhA
PhD
Vx-
Vs
IlineX
PhB
T2
D2
PhE
PhF
PhC
(b)
Figure 144 Schematic explanation of inverter states for the 3-phase bridge circuit driven star
connected 6-phase conventional rotor SRM: a-) PhA conducting and line current in positive
direction, b-) PhD conducting and line current in negative direction. Conductive paths in red
In Figure 145, the simulated instantaneous star point potential and current are depicted.
The black traces are taken directly from the dynamic simulation and the green traces are
simply the filtered versions of these simulated waveforms. As before, filtering is achieved
by a simple moving average method. It is obvious from the bottom trace in Figure 145
that the error in the current flowing into the star point voltage (i.e. sum of all phase
currents) is at its maximum during phase commutation. However, even during phase
commutation the error is only 0.1% of the demand current level. It can be said that the
227

calculation of the star point potential based on the trial of a set of predetermined values
explained in Chapter 6 Simulation Models works adequately.
Figure 145 Simulated star-point potential (top trace) and star-point current (bottom trace) for the 3phase bridge circuit driven star connected 6-phase conventional rotor SRM with mutually de-coupled
phase windings running at 250rpm
In Figure 139, it is also interesting to note that the filtered phase voltage never reaches the
full dc-link rail voltage. This can be explained with the help of the two traces given in
Figure 146. The top trace in Figure 146 depicts the phase voltage, line potential and the
star point potential that are generated by the simulation. The bottom trace contains the
filtered (with a simple moving average method) versions of the same waveforms. As
mentioned earlier, with the machine connected in the star configuration, the controller is
unable to regulate the phase voltage. The phase voltage depends on the star point
potential, which in turn varies in order to keep the sum of the currents flowing into it at
zero. As the filtered star point potential never reaches the full negative voltage neither
does the phase voltage.
228
Figure 146 Simulated raw phase, line and star point voltages (top trace) and filtered phase, line and
star point voltages (bottom) for the 3-phase bridge circuit driven star connected 6-phase conventional
rotor SRM with mutually de-coupled phase windings running at 250rpm
In Figure 147, the phase currents are displayed along with all three line currents.
229
Figure 147 Simulated phase currents (top trace) and line currents (bottom trace) for the 3-phase
bridge circuit driven star connected 6-phase conventional rotor SRM with mutually de-coupled phase
Figure 148 depicts the contribution of each machine phase to the total output torque. The
dip during phase commutation approximately half-way through a phases conduction
period is visible in the phase torque waveforms in Figure 148.
Table 23 summarises some of the simulation input parameters along with the results of
some of the performance parameters generated by the simulation for the star connected
drive running at 250rpm under current control.
230
Figure 148 Simulated phase torques and resulting total output torque for the 3-phase bridge circuit
running at 250rpm
Table 23 Summary of simulated performance parameters for the 3-phase bridge circuit driven star
connected 6-phase conventional rotor SRM with mutually de-coupled phase windings running at
250rpm.
Speed [RPM]
Current Demand [A]
Advance Angle [Deg]
Conduction Angle [Deg]
DC-Link Voltage [V]
Average Output Power [Nm]
Maximum Output Torque [Nm]
Minimum Output Torque [Nm]
Output Torque Ripple - Pk2Pk [%]
Total Machine Copper Losses [W]
250
30
0
60
175
42.8
1120.4
47.4
36.3
23.33
18.79
34.28
961.7
240.5
50
47.2
231

29.37
13.69
29.32
8.85
58.69
22.54
9.2.1.2 Drive under Voltage Control

The current demand was kept at 30A, but this time, the machine was run at 1200rpm. At
this speed, the machine BEMF is relatively high and the controller does not have enough
time to build the line current from zero to the demanded level against the rapid change in
the machine inductance between the unaligned and aligned rotor positions. Thus, the
current demand is never reached and the controller works under full voltage control
mode. In this case, the values for the PID gains are irrelevant. The energisation of the
phases was advanced by 40o in order to help with the torque production. Similar to the
low speed operation, the conduction period was fixed to 120o and the dc-link voltage was
set to 175.0V. The results from the dynamic simulation are depicted from Figure 149 to
Figure 156.
circuit driven star connected 6-phase conventional rotor SRM with mutually de-coupled phase
232

It is clear from Figure 149 that the line current waveform deviates significantly from the
ideal 120o as it is almost continuous with a very narrow zero period (between points B
and C in Figure 151).
Figure 150 clearly demonstrates that the controller is under full voltage control and the
Vx potential swings between positive and negative dc-link voltage (i.e. 87.5V). All of the
available line potential is applied at point Vx (see Figure 140) until the end of the
conduction period. Once the conduction period elapses the controller applies the full
negative voltage at point Vx to reduce the line current back down to zero.
running at 1200rpm
Once again, to help with the understanding of the operation of the star-connected drive
under voltage control, Figure 151 has been generated. States in Figure 151 are
summarised below:
Point A : Phase-A is turned off.

Point B : Line current (IlineX) falls to zero.
233

Point C : Phase D is turned on.
Point D : Phase D is turned off.
Point E : Phase A is turned on.
Figure 151 Simulated phase-A current (in red), phase-D current (in blue), line-X current (in green)
for the 3-phase bridge circuit driven star connected 6-phase conventional rotor SRM with mutually
de-coupled phase windings running at 1200rpm
Once again the filtered phase voltage fails to reach the full dc-link rail voltage in Figure
149. The situation is similar to the current controlled drive and illustrated in Figure 152.
In the bottom trace of Figure 152 the potential at point X (Vx) has a triangular
characteristic for short periods of time. This triangular characteristic is a result of moving
average filtering as the filter does not produce good results with signals which have very
quickly changing sharp edges.
234
Figure 152 Simulated raw phase, line and star point voltages (top trace) and filtered phase, line and
star point voltages (bottom) for the 3-phase bridge circuit driven star connected 6-phase conventional
rotor SRM with mutually de-coupled phase windings running at 1200rpm
From Figure 153, there are six distinct periods of star point potential, each corresponding
to a phase commutation point. Similar to the current controlled situation, significant
changes in the star point potential appear around the phase commutation points.
235
Figure 153 Simulated phase-A current (x5 in red), phase-D current (x5 in blue), line-X current (x5 in
green) and star-point potential (in black) for the 3-phase bridge circuit driven star connected 6-phase
conventional rotor SRM with mutually de-coupled phase windings running at 1200rpm
From Figure 154, unlike the drive under current control, the (filtered) star point potential
varies continuously with the drive under voltage control. Once again the error in the star
point current is small compared to the demand current level, which suggests that the
simulation is behaving well in controlling the line currents under full voltage control.
236
Figure 154 Simulated star-point potential (top trace) and star-point current (bottom trace) for the 3phase bridge circuit driven star connected 6-phase conventional rotor SRM with mutually de-coupled
Figure 155 depicts the phase and line currents when the drive is under voltage control in
the simulation.
Figure 156 suggests that - compared to current controlled drive - a larger proportion of the
instantaneous torque per phase is negative. This is also apparent when comparing the
torque ripple figure in Table 24 to that in Table 23. This is typical of high speed operation
where the peak to peak variation of output torque is compromised for a greater average
value.
237
bridge circuit driven star connected 6-phase conventional rotor SRM with mutually de-coupled phase
Table 24 summarises some of the simulation input parameters along with the results of
some of the performance parameters generated by the simulation for the star connected
drive running at 1200rpm under full voltage control.
238
running at 1200rpm
Table 24 Summary of simulated performance parameters for the 3-phase bridge circuit driven star
1200rpm
Speed [RPM]
Current Demand [A]
Advance Angle [Deg]
DC-Link Voltage [V]
1200
30
-40
60
175
10.6
1336.8
12.7
9.2
27.06
8.66
16.23
204.6
60.5
21
82.4
239

12.75
5.85
11.26
2.46
24.01
8.31

From Table 24, compared to the drive under current control, the inverter VA rating is
significantly reduced when the drive is under full voltage control. This is due to the
reduced currents flowing in the system, which also results in reduced total machine
copper losses. The drive efficiency is significantly improved. This is typical of SR drives,
where the output power is kept almost constant even though the phase currents and the
total average output torque are reduced with increasing operating speeds. As mentioned
earlier, compared to the star connected drive under current control, the torque ripple is
increased, which is due to the peaky nature of individual phase torque waveforms (see
Figure 156). Nevertheless, the torque ripple is still relatively low at approximately 26.7%.
9.2.2 Mutually De-Coupled Delta Connected Line Current

Controlled Six-Phase Drive
The control of the 6-phase machine can also be achieved when the phase connection is
changed to a delta configuration. Similar to the simulation where the phases were starconnected, the control was based on the three line currents (IlineX, IlineY and IlineZ in
Figure 157). Separate PID and PWM controllers were needed for controlling each line
current. Unlike the star-connected drive where module currents must sum up to zero at the
star point, with the delta configuration module voltages must sum up to zero (once again a
module is formed by two back-to-back connected phases e.g. PhA and PhD in Figure
157). Like the star-connected drive two examples are given; one for low, and the other for
high speed operation. Some of the performance results obtained from the simulations are
tabulated inTable 25 and Table 26 for low speed and high speed operation, respectively.
240
Figure 157 Graphical illustration of the line, link and phase currents of the 6-phase delta connected
machine with back-to-back connected phase windings driven from 3-phase full bridge circuit
Line currents (IlineX, IlineY, and IlineZ), phase currents (IphA, IphB, IphC, IphD, IphE,
and IphF) and link currents (Ilink1, Ilikn2, and Ilink3) in Figure 158 to Figure 170 are all
illustrated in Figure 157. In order to clearly demonstrate the mode of operation for the
controller, Figure 159 and Figure 166 are presented for the drive under current and full
voltage control, respectively.
9.2.2.1 Drive under Current Control

The same simulation conditions explained in section 2.1.1 of this chapter are used here.
The results from the dynamic simulation are depicted from Figure 158 to Figure 164.
Clearly visible in Figure 157, unlike the star-connected drive, the line currents are a result
of two link currents. This can be formulated as follows:
241

IlineX = Ilink1 Ilink 3
IlineY = Ilink 2 Ilink1
Equation 34
IlineZ = Ilink 3 Ilink 2
circuit driven delta connected 6-phase conventional rotor SRM with mutually de-coupled phase
Unlike the machine connected in star configuration, it is clear from Figure 158 that the
filtered phase voltage is now able to reach the full dc-link voltage as the controller can
control the two potentials at each end of a phase connection (e.g. Vx and Vy through
IlineX and IlineY in Figure 157).
242
driven delta connected 6-phase conventional rotor SRM with mutually de-coupled phase windings
running at 250rpm
In helping to explain Figure 158, Figure 160 and Figure 161 have been generated. In
Figure 158, until the line current (IlineX) reaches the demanded level, the full dc-link
voltage is available across the phase windings (Vphase Filtered). Once the line current
reaches the demand, the controller aims to keep it at this level. As a result, the phase
currents (phase-A and phase-D) remain approximately the same, i.e. the current circulates
around the parallel connected phases and the phase diodes. The slight drop in both phase
currents is due to the resistive voltage drop across each phase. Once phase-B is turned on
(around 0.004sec in Figure 158 and Figure 161), the controller keeps phase-C off by
applying positive potentials at point X and Z (Figure 160) and negative potential at points
Y, which result in positive voltage across phase-A. Consequently, current in phase-A rises
again until the current flowing in line-Y reaches its demand level. Comparing Figure 158
to Figure 139 reveals that the phase current tail-off period is significantly longer than with
the star connected drive. From Figure 161 and Figure 162, it is apparent that the link
current has a six step characteristic. Compared to the star-connected drive, where the
demand current was set to the same value, the peak phase current is significantly lower in
243

the delta-connected drive. This is due to the fact that two phase currents formed one line
current in the star-connected drive whereas four phase currents are needed in the deltaconnected drive.
X
T1
D1
T3
D3
T5
D5
Ph
D
F
Ph
Ph
A
C
Ph
PhE
T2
D2
T4
D4
T6
D6
Z
PhB
Figure 160 Delta-connected six-phase machine driven from 3-phase full bridge inverter
Figure 161 Simulated phase-A current (in red), phase-D current (in blue), line-X current (in green),
link-1 current (in magenta) and link-3 current (in black) for the 3-phase bridge circuit driven delta
250rpm
244

The operating states in Figure 161 are summarised below with respect to the circuit
diagram in Figure 160:
Point A : Phase-A is turned on.
Point B : Phase-A is turned off.
Point C : Phase-D is turned on.
Point D : Phase-D is turned off.
Point E : Line current (IlineX) falls to zero. Current in link-1 (Ilink-1) equals that in link3 (Ilink-3).
Between points D and E, phase-A current rises even though it is in its off period. This is
due to the negative voltage applied across the windings of phase-D appearing as positive
voltage across the windings of phase-A. The series connected phase diode becomes
forward biased and phase-A starts conducting current. This lasts until the line current falls
to zero. The current in phase-A gradually falls back to zero before the start of the next
energisation cycle (point A in Figure 161).
Figure 162 Simulated link currents (top trace) and line currents (bottom trace) for the 3-phase bridge
245
bridge circuit driven delta connected 6-phase conventional rotor SRM with mutually de-coupled
As mentioned earlier, with the delta-connected drive the phases conduct current for long
periods although the actual energisation period for each phase is relatively low (i.e. 120o).
Having long periods of current in each phase results in negative torque generated by each
phase as there is still current flowing in the phase windings even when the rate of change
phase inductance changes its sign, i.e. the phase inductance starts to reduce once the rotor
passes the aligned position. However, the generated negative torque is small, as the phase
current has fallen to a small value by the time the rotor passes the aligned rotor position.
As each phase conducts current for longer periods there is significant overlap between
phase output torque waveforms. This results in long periods of flat total output torque that
is visible in Figure 164.
246
running at 250rpm
Although the simulations for the star and delta configured drives were run at the same
speed, when both were under current control they resulted in different average output
torques. This makes it hard to compare the output torque ripple values and draw a
conclusion. Although both the star and delta connected drives have similar amount of line
currents, i.e. the current flowing through devices is similar and the dc-link voltage is kept
the same, the inverter VA rating of the delta-connected drive is significantly higher. This
is due to the fact that lower levels of currents circulate in the phases and consequently the
output torque is reduced. As both drives run at the same speed the reduced output power
results in increased inverter VA. Comparison of the inverter ratings of the star and delta
connected drives will be a topic of the next chapter On the other hand, the voltage that
can be applied to each phase of the delta connected drive is higher, so the peak speed of
the machine can be higher with the delta connected drive. The efficiency of the deltaconnected drive is slightly higher than the star-connected drive. This is due to the reduced
ohmic losses as a result of lower current levels in the delta-connected drive.
247

Table 25 Summary of simulated performance parameters for the 3-phase bridge circuit driven delta
250rpm
Speed [RPM]
Current Demand [A]
Advance Angle [Deg]
DC-Link Voltage [V]
250
30
0
60
175
23.7
621.7
25.3
20.4
19.53
11.79
22.66
379
242.6
31.4
48.8

53.81
22.97
53.74
18.88
107.55
41.85
9.2.2.2 Drive under Voltage Control

In order to achieve an operation that is under full voltage control, the speed of the motor
has been increased to 1500rpm. This is 300rpm faster than the speed setting for the
simulation with the star configuration. This is due to the fact that there is more voltage
available across the phase windings with a delta configuration than a star configuration.
All other simulation inputs are kept the same as those used for the high speed simulation
of the star configuration. Similar to operation in the current control mode, the phase
voltage is able to reach the full dc-link voltage with voltage control. Figure 166 confirms
that the drive is under full voltage control.
248
249
running at 1500rpm
The operating states in Figure 167 are summarised below:

Point A : Phase-A is turned off.
Point B : Phase D is turned on.
Point C : Phase D is turned off.
Point D : Line current falls to zero. Current in link-1 current equals that in link-3
Point E : Phase A is turned on.
Similar to the drive under current control, the two-stage rise in the phase currents is still
visible in Figure 167. However, the rate of change of current is slower in the second stage
although the same amount of voltage is available across the phase windings. This is
believed to be due to the rapid increase in phase inductance as the rotor position
approaches alignment.
250
Figure 167 Simulated phase-A current (in red), phase-D current (in blue), line-X current (in green),
link-1 current (in magenta) and link-3 current (in black) for the 3-phase bridge circuit driven delta
connected 6-phase conventional rotor SRM running at 1500rpm
Figure 168 Simulated link currents and line currents for the 3-phase bridge circuit driven delta
connected 6-phase conventional rotor SRM running at 1500rpm
251
bridge circuit driven delta connected 6-phase conventional rotor SRM running at 1500rpm
driven delta connected 6-phase conventional rotor SRM running at 1500rpm
252

In Figure 165, Figure 167 and Figure 169 it is interesting to note that the phase current is
almost continuous. This is believed to be a result of the anti-parallel connection of phase
windings. Negative voltage applied across one of the phases of a module appears as
positive across the anti-parallel phase in the same module. This results in the off-phase in
a module conducting current even out of its conduction period. Moreover, compared to
low speed operation, at high speed there is even less time to de-flux a phase completely
before the next energisation cycle. At point C in Figure 167, before the phase-A current
completely falls to zero the anti-parallel connected phase (phase-D in this case) is turned
off and negative voltage appears across this phase. Negative voltage across phase-D
appears as positive across phase-A and the phase diode connected in series with the phase
windings becomes forward biased. Consequently at point C in Figure 167, the current in
phase-A starts rising again. The current in phase-A eventually falls to zero once the line
current is zero and (in simulation) the current decays due to resistive voltage drops across
the windings of phase-A and phase-D.
Table 26 Summary of simulated performance parameters for the 3-phase bridge circuit driven delta
1500rpm
Speed [RPM]
Current Demand [A]
Advance Angle [Deg]
DC-Link Voltage [V]
Average Output Power [W]
1500
30
-40
60
175
18
2826.2
21
15.9
24.29
10.63
19.2
307.8
130.2
30.1
85.8
253

10.54
5.12
10.54
1.93
21.08
7.05

In simulations of both the star and delta connected drives presented in the preceding
sections, the phase currents conduct current even when they are not in their respective
conduction periods. This is similar to what was seen with the 2-phase machine driven
from the h-bridge inverter. The duration of time a phase is conducting current (even if it is
not in its conduction period) seems to be longer when the machine is in delta
configuration. The star connected drive suffers from higher machine copper losses
compared to the delta connected drive when under current control. This is simply because
of the increased phase currents in the star connected configuration even when the
conduction period and the line current demand are the same as those of the delta
connected drive. This can be related to the fact that the phases forming a module in the
star connected drive also make up the line current, whereas two modules make up a line
current in the delta configuration.
9.3 DYNAMIC MEASUREMENTS OF THE

PROTOTYPE SIX-PHASE SEGMENTAL ROTOR
SRM
9.3.1 Dynamic Torque against Speed Characteristics
Torque/speed characteristics of the prototype 6-phase segmental rotor SRM were
measured on the test rig for different conduction angles. The machine phases were
connected in star and driven from a 3-phase bridge circuit (Figure 137). The advance
angle has been varied and measurements taken at different speeds for each advance angle.
The speed was varied by varying the load. The measurements were taken with a dc-link
voltage of 175V. Line currents were controlled and the demand current level was set to
15A and then 30A. The results are presented in Figure 171. The conduction angle was
fixed at 120o. Adv in Figure 171 represents switching angle that is before the unaligned
rotor position and Del represents switching angle that is after the unaligned rotor
position. The unaligned position in the case of the 6-phase machine driven from the 3phase bridge circuit where two adjacent phases conduct current simultaneously is
254

somewhat different to the 3-phase machine. This will be explained in detail in the next
section. There are no unexpected conclusions from Figure 171 and Figure 172.
Figure 171 Torque-Speed curves of the prototype segmental rotor star-connected 6-phase SRM
driven from a 3-phase bridge circuit. Top trace is when the line current demand is 15A and bottom
trace is when the line current demand is 30A
For each point in the curves presented in the top and bottom traces of Figure 171 the total
machine copper losses were also calculated (rms phase current was measured via an
oscilloscope). In Figure 172 each curve is plotted for a fixed energisation angle (advanced
or delayed with respect to the unaligned rotor position) and calculated total machine
255

copper losses were given against the total machine output torque. Each point in the top
and bottom traces of Figure 172 is therefore at a different operating speed. The measured
results are for two different current demands: 15A and 30A. In general, advancing the
phase energisation instance with respect to the unaligned position of the phase is
beneficial in torque production as the speed increases. As the speed reduces, delaying the
phase energisation instance after the unaligned position helps with torque production.
Figure 172 Total machine copper loss vs. Total mean output torque curves of the prototype segmental
rotor star-connected 6-phase SRM driven from a 3-phase bridge circuit. Top trace = 15A current
demand and bottom trace = 30A current demand
256

Mechanical losses (friction and windage) of the prototype machine were also measured
and the results are presented in Figure 173. During this test, the phases of the machine
were not energised. The machine was run from the load (dc-machine with controlled field
and armature windings). The dc voltage to the field terminals of the dc-machine was
adjusted to increase and reduce the speed of the dc-machine and hence the speed of the
prototype 6-phase SRM. The speed and torque were measured by a handheld tachometer
and a torque transducer coupled in between the dc-machine and the SRM, respectively.
Before the measurements were taken the machine was run for 15 minutes to allow for the
warming of the couplings and bearings. The measurements were taken up to around
1800rpm as this was the rated speed of the load machine.
Figure 173 Mechanical losses vs. speed of the prototype 6-phase segmental rotor machine
9.3.2 Dynamic Measurements

This section summarises the measurements taken from the six-phase star connected drive
system and the estimated drive performance derived from the captured phase current and
rotor position information. Unlike the 2-phase and 3-phase drives analysed in previous
chapters where magnetic coupling between phase windings was negligible (and hence
257

was ignored) the phase current and static flux-linkage characteristics of the 6-phase
segmental machine cannot be used to predict the instantaneous torque waveforms and
hence performance parameters relating to the machine torque i.e. average torque, torque
ripple and shaft power cannot easily be determined. The phase windings of the prototype
6-phase segmental rotor SRM are connected in star and driven from the 3-phase bridge
inverter whereby two adjacent phases are energised simultaneously, resulting in
magnetically coupled phases. As the phases are magnetically coupled, the flux linking a
phase is a non-linear function of seven variables: the six phase currents, and the rotor
position. However, by using the equivalent slot quantities it is possible to predict the
instantaneous output torque.
By using the transformation matrices presented in Chapter 6 Simulation Models along
with the flux-linkage characteristics per machine slot, it is possible to calculate the
instantaneous flux-linkage and torque contribution of each slot. The total output torque is
then the sum of all slot torques. A section of the FE model of the 6-phase segmental
machine that is used to generate the Flux vs. MMF characteristics of a slot is depicted in
Figure 174. In Figure 174, phases A and F are energised simultaneously. The motor is in
the aligned position with respect to phases A and F and the Slot1 permeance is at its
maximum. Unlike the 3-phase segmental machine where the aligned and unaligned
positions of the motor can be related directly to the phases of the machine it is more
appropriate for the 6-phase segmental machine to refer to the aligned and unaligned
positions with respect to stator slots. In Figure 174 MMFs of Slot6 and Slot2 are not zero
but the motor is clearly not in the aligned position with respect to either of these slots. By
using the core-back flux density (along the black line in Figure 174), the core-back depth
and the lamination packing factor, Flux vs. MMF characteristics for a single slot can be
generated.
258
e
as
Ph
A
1
ot
Sl
Figure 174 Flux density plot of the 6-phase segmental rotor SRM in the aligned position
Current per coil was varied from 0A to 40A with 1A steps and the rotor position was
varied from the unaligned to aligned position with 2 degree steps. The average flux
density across the core-back of the slot (along the black line in Figure 174) was then
calculated with the FE simulation at every position and phase current. Using the number
of turns per coil, the core-back depth and the stator lamination packing factor the Flux vs.
MMF characteristics of the slot were generated. The generated Flux vs. MMF
characteristics for Slot1 (in Figure 174) are given in Figure 175. Flux-MMF-Position
characteristics of the slot were then used to calculate the Torque-MMF-Position
characteristics. This is given in Figure 176.
259
Figure 175 Slot (Slot1 in Figure 18) Flux vs. MMF characteristics from unaligned position to aligned
position with 1 degree steps up to 3520 ampere-turns
The 6-phase machine was connected in star and driven from the 3-phase bridge circuit
(see Figure 137). Phase-A current, rotor position with respect to Slot1, line current, lineto-line voltage and line-to-star point voltage were all captured with an oscilloscope. The
slot position was captured by using the DSP measured encoder feedback and the digitalto-analog conversion card of the drive system.
260
40A
0A
Figure 176 Torque-Position-MMF characteristics per slot generated from the Flux-MMF-Position
characteristics per slot
Similar to the 2-phase and 3-phase drives a Matlab based script was written to analyse
the captured waveforms to determine some of the machine performance parameters, such
as copper losses and the instantaneous torque waveform. Instantaneous output torque was
then used to determine the machine output torque ripple. The phase current was used to
calculate the copper losses and the speed of operation. The speed was then used to
calculate the required amount of phase shift to generate all the other five phase current
waveforms. Similarly by phase shifting, all slot positions were determined. The slot
positions and the phase currents were then used along with the transformation matrices to
calculate the instantaneous machine torque. The instantaneous torque waveform was then
further analysed to estimate the machine average torque and the torque ripple. To validate
the Matlab script, the average torque and the running speed estimations obtained from
the script were compared to those measured. The results are divided into two sections:
low speed measurements (current controlled drive) and high speed measurements (voltage
controlled drive). Several examples are given for each control mode (current or voltage
control). Three figures are presented for each example for measurements of both control
modes. These are:
261

Six phase currents (top trace of first figure) and three line currents (bottom trace
of first figure).
Six slot torques and the total machine output torque (second figure).
One of the line currents (top trace of third figure), line-to-line voltage (middle
trace of third figure) and line-to-star point voltage (bottom trace of third figure).
Red traces in the middle and bottom figures are the filtered waveforms for the
line-to-line and line-to-star point voltages. (A simple moving average filter was
used).
A table where some of the measured and script estimated performance parameters are
presented is also given for each example (Table 27 to Table 36). In these tables;
Adv Angle [Deg.] is the point where the phase is energised in advance with respect
to the unaligned rotor position.
Idemand [A] is the demand current. In case of the six-phase machine driven from
the 3-phase bridge circuit, this is the line current demand.
Speed [RPM] is the speed of operation at which the measurement was taken.
Taverage [Nm] is the average total machine output torque
IphRMS [A] is the root mean square of the phase current.
IphPEAK is the peak value of the phase current.
Cu Loss [W] is the total machine copper losses. 100oC was assumed for the
winding operating temperature for copper loss calculations.
Tripple MEAN [%] is calculated by Equation 31 in Chapter 7. This is the torque
ripple based on peak-to-peak instantaneous torque variation and the average
output torque.
Tripple MAX [%] is calculated by Equation 32 in Chapter 7. This is the torque
ripple based on peak-to-peak instantaneous torque variation and the maximum
output torque.
Pshaft [W] is the total machine shaft power calculated by Equation 35.
Efficiency [%] is the machine efficiency calculated by 100 (Pshaft / (Pshaft +
Cu Loss))
262

Pshaft =
2
Speed Torque
60
Equation 35
9.3.2.1 Low Speed Measurements of the Prototype Six-Phase StarConnected Segmental Rotor Machine Driven by the Three-Phase Bridge
Circuit
9.3.2.1.1 Low Speed Measurement-1:

Captured results for this low speed measurement are presented from Figure 177 to Figure
179. Table 27 summarises the measured and script estimated performance parameters
based on the captured data.
Figure 177 Phase currents (top trace) and line currents (bottom trace) with Advance Angle = -30o;
Line Current Demand = 30A; Speed = 335.57RPM; DC-Link Voltage = 175V
263
Figure 178 Instantaneous slot torques and total machine output torque (in red) with Advance Angle =
-30o; Line Current Demand = 30A; Speed = 335.57RPM; DC-Link Voltage = 175V
Figure 179 Line current (top trace), line-to-line voltage (middle trace) and line-to-star point voltage
(bottom trace) with Advance Angle = -30o; Line Current Demand = 30A; Speed = 335.57RPM; DCLink Voltage = 175V
264

Clearly, the motor is running slowly and the controller has enough time to build the line
current from zero to the demanded level and therefore is under current control. Although
there is strong mutual coupling between machine phases, at this speed, the line currents
are very close to those expected from the ideal 120o conduction. It is fair to say that the
controller works reasonably well controlling a machine with mutually coupled phase
windings.
It is apparent from Figure 177 that the phase currents have two-stage rise and fall periods.
This is similar to what was seen with the simulated results of the six phase machine with
no mutually coupled phases. The reason for the rise at the beginning of the phase current
waveform is the same: the negative voltage applied across the phase that is switched-off
appears as positive voltage across the phase that is not in its conduction period. The phase
diode becomes forward biased and the phase starts conducting current even though it is
not in its conduction period. Once the line current falls to zero, there is still current in
both of the phases in a module (formed by a pair of phases connected back to back with
series connected phase diodes) that circulates around the phase diodes and the phase
windings. In this period where the line current is zero, the phase currents have a slow
decay due to resistive voltage drop across the phases and the forward voltage drops across
the phase diodes.
When the phase energisation cycle starts positive voltage is applied across the newly
conducting phase. The phase current has a second rise period until the line current hits the
demand current level and the controller starts chopping the line voltage. During the
chopping period, one of the phase currents has already fallen to zero. The line current
equals the phase current that is in its conduction period. Once the conducting phase gets
to the end of its energisation cycle, it is turned off and both the phase and line current
decay. Once the line current hits zero, the controller reacts to keep it at this level until the
beginning of the next energisation cycle of one of the anti-parallel connected phases. It is
also quite apparent from the individual torque waveforms (Figure 178) that the phases go
through two level energisation and de-fluxing periods.
265

Table 27 Summary of measured and predicted performance parameters for the star-connected 6phase segmental rotor SRM when driven from 3-phase bridge circuit with Advance Angle = -30o;
MEASURED
ESTIMATED
Adv Angle [Deg.]
-30
Idemand [A]
30
Speed [RPM]
335.57
333.3
Taverage [Nm]
37.02
38.36
IphRMS [A]
17.8
IphPEAK [A]
32.3
Cu Loss [W]
1095
Tripple - MEAN [%]
27.0
Tripple - MAX [%]
24.2
Pshaft [W]
1300.915
1339.1
55.02
Efficiency [%]
According to Table 27, measured and estimated values for the average torque and shaft
power are in good correlation. Therefore the torque ripple estimation can be viewed as
quite reasonable. Although the drive works well and the motor is able to generate
approximately 38.0Nm of average output torque, the machine copper losses are
considerably high compared to the generated shaft power. This is believed to be due to
the two-stage rise and fall periods visible in phase current waveforms that result in
increased rms phase currents. The efficiency value of 55.02% in Table 27 is reasonable
for an SR machine under current control running at low speed. Moreover, in the case of
Low Speed Measurement-1 the machine is pushed close to its limits of torque output
affecting the drive efficiency. Even with this high advance angle, torque ripple is quite
favourable at 27.0% with plenty of overlapping between adjacent slot torque waveforms.
For the second example, the current demand is kept at 30A and the machine was run at
approximately 342 rpm with a dc-link voltage of 175V. The phase energisation instance
was increased to 50o before the unaligned rotor position. The captured results are
presented from Figure 180 to Figure 182.
266
267
Figure 182 Line current(top trace), line-2-line voltage (middle trace) and line-2-star point voltage
From the above figures, it is clear that the controller is still under current control. The
two-stage rise and fall time periods in the phase current waveforms are also visible. Table
28 tabulates the measured against estimated performance parameters. The correlation
between measured and estimated torques is within 1%. Still operating at a reasonably low
speed and generating similar amounts of output torque, the efficiency of Low Speed
Measurement-2 is similar to that of Low Speed Measurement-1.
268

MEASURED
ESTIMATED
Adv Angle [Deg.]
-50
Idemand [A]
30
Speed [RPM]
341.7
342.8
Taverage [Nm]
37.2
37.4
IphRMS [A]
17.5
IphPEAK [A]
32.2
1058.8
Cu Loss [W]
Tripple - MEAN [%]
27.6
Tripple - MAX [%]
24.7
Pshaft [W]
1331.8
Efficiency [%]
1344.4
55.9

In the final low speed example the current demand was reduced to 15A. The machine was
run slower at approximately 255.3rpm with a dc-link voltage of 175V. The phases were
energised 40o before the unaligned rotor position was reached. The captured results are
presented from Figure 183 to Figure 185. The controller is again under current control
and the operation is similar to that with 30A current demand with the two-stage rise and
fall periods in all phase currents. It is interesting to note that the behaviour of the phase
currents in the first part of the rise period is very similar to that seen with the examples
with 30A current demand. The phase currents initially rise to approximately 6.0A and
then fall to around 5.0A. This is due to the fact that the first rise period in the phase
currents is not controlled by the controller but is a result of negative voltage applied
across the anti-parallel phase appearing as positive voltage across the phase windings and
forward-biasing of the phase diode. The rate of change of phase current is faster with the
15A current demand as the controller only needs to elevate the line current to 15A rather
than 30A with the same dc-link voltage available. This is reflected in the individual phase
torque waveforms. As a result the total output torque waveform has more torque ripple
(see Table 29).
269
270
MEASURED
ESTIMATED
Adv Angle [Deg.]
-40
Idemand [A]
15
Speed [RPM]
255.3
258
Taverage [Nm]
19.7
18.8
IphRMS [A]
10.1
IphPEAK [A]
Cu Loss [W]
18.1
350
Tripple - MEAN [%]
42.9
Tripple - MAX [%]
37.2
Pshaft [W]
525.6
Efficiency [%]
271
508.9
59.2

As the demand current is lower (15A as opposed to 30A) it is not surprising to see that the
average output torque and the total machine copper losses are reduced. At the same time
the efficiency is slightly higher than those experienced during Low Speed Measurement-1
and Low Speed Measurement-2.
The flux vs. MMF loci of all three low speed measurements are displayed against the
aligned and unaligned magnetisation curves (see Figure 175) of a single slot in Figure
186.
Figure 186 Flux vs. MMF loci for the low speed measurements
9.3.2.2 High Speed Measurements of the Prototype Six-Phase StarConnected Segmental Rotor Machine Driven by the Three-Phase Bridge
Circuit
With the examples in the next four subsections, the machine BEMF is relatively high and
the controller does not have enough time to achieve the demanded current level. The
272

machine therefore operates under full voltage control. This is apparent from the line-toline and line-to-star point voltage waveforms given in Figure 189, Figure 193, Figure 197
and Figure 201.
The current demand was set to 30A for High Speed Measurement-1 and High Speed
Measurement-2 and to 15A for High Speed Measurement-3 and High Speed
Measurement-4. A summary of the results will be given at the end of this subsection.
9.3.2.2.1 High Speed Measurement-1:

Here the machine was run at approximately 1271rpm. The phases were energised at the
unaligned rotor position. The results are presented from Figure 187 to Figure 189. Table
30 summarises the predicted and measured performance parameters such as average
output torque for the High Speed Measurement-1. The power device and diode current
waveforms of the measurement are illustrated in Figure 190. Based on the measured line
currents, the VA ratings of the power devices and diodes have been estimated and the
results are presented in Table 31.
Figure 187 Phase currents (top trace) and line currents (bottom trace) with Advance Angle = 0o; Line
Current Demand = 30A; Speed = 1271.19RPM; DC-Link Voltage = 175V
273
0o; Line Current Demand = 30A; Speed = 1271.19RPM; DC-Link Voltage = 175V
Figure 189 Line current (top trace), line-2-line voltage (middle trace) and line-2-star point voltage
(bottom trace) with Advance Angle = 0o; Line Current Demand = 30A; Speed =1271.19RPM; DCLink Voltage = 175V
274

Table 30 Summary of measured and predicted performance parameters for the star-connected 6phase segmental rotor SRM when driven from 3-phase bridge circuit with Advance Angle = 0o; Line
MEASURED
ESTIMATED
Adv Angle [Deg.]
Idemand [A]
30
Speed [RPM]
1271.2
1263.2
7.3
7.7
Taverage [Nm]
IphRMS [A]
7.0
IphPEAK [A]
Cu Loss [W]
13.5
170.5
Tripple - MEAN [%]
62.7
Tripple - MAX [%]
46.2
Pshaft [W]
974.4
Efficiency [%]
1023.1
85.7
Figure 190 IGBT currents (top trace) and reverse recovery diode currents (bottom trace) based on
the measured line currents presented in Figure 187
275

Table 31 Summary of estimated power electronic performance parameters based on measured line
currents with Advance Angle = 0o; Line Current Demand = 30A; Speed = 1271.19RPM; DC-Link
Voltage = 175V
Top Device RMS Current [A]
Top Device Peak Current [A]
Bottom Device RMS Current [A]
Top Diode RMS Current [A]
Top Diode Peak Current [A]
Bottom Diode RMS Current [A]
Bottom Diode Peak Current [A]
Device VA Rating - PEAK [kVA / kW]
Device VA Rating - RMS [kVA / kW]
Diode VA Rating - PEAK [kVA / kW]
Diode VA Rating - RMS [kVA / kW]
5.48
11.02
5.42
11.11
1.83
10.14
2.05
10.77
11.35
5.6
10.73
2.0

Here the machine was run at approximately 1342rpm. The phases were energised 30o
prior to the unaligned rotor position. The results are presented from Figure 191 to Figure
193.
Figure 191 Phase currents (top trace) and line currents (bottom trace) with Advance Angle = 30o;
276
277

MEASURED
ESTIMATED
Adv Angle [Deg.]
30
Idemand [A]
30
Speed [RPM]
1342.3
1336.3
9.1
9.8
Taverage [Nm]
IphRMS [A]
11.4
IphPEAK [A]
Cu Loss [W]
22.3
449.2
Tripple - MEAN [%]
57.2
Tripple - MAX [%]
41.9
Pshaft [W]
1281.9
Efficiency [%]
1373.1
75.35
278

Voltage = 175V
10.00
20.83
9.93
20.86
3.88
18.85
4.07
19.20
15.37
7.35
14.03
2.93

Here the machine was run at approximately 1470rpm with a reduced current demand
value of 15A. The phases were energised 20o prior to the unaligned rotor position. The
results are presented from Figure 195 to Figure 197.
Figure 195 Phase currents (top trace) and line currents (bottom trace) with Advance Angle = 20o;
279
280

MEASURED
ESTIMATED
Adv Angle [Deg.]
20
Idemand [A]
15
Speed [RPM]
1474.2
1496.3
7.0
7.4
Taverage [Nm]
IphRMS [A]
8.3
IphPEAK [A]
Cu Loss [W]
16.2
239.6
Tripple - MEAN [%]
71.5
Tripple - MAX [%]
49.5
Pshaft [W]
1074.5
Efficiency [%]
1155.8
82.8
Compared to the previous high speed measurement results, significant negative torque is
present in the individual phase torque waveforms (see Figure 196). This is reflected in the
torque ripple values presented in Table 34. This can be related to the speed of operation
and the selected advance angle. As the operating speed is faster there is even less time for
the controller to de-flux the phases completely before the aligned rotor position is reached
and the machine starts generating negative torque.
281

Voltage = 175V
7.05
14.30
7.00
14.37
2.53
12.96
2.68
13.21
14.41
7.06
13.16
2.62

Here the machine was run at approximately 1357rpm with reduced current demand value
of 15A. The phases were energised at the unaligned rotor position. The results are
presented from Figure 199 to Figure 201.
Figure 199 Phase currents (top trace) and line currents (bottom trace) with Advance Angle = 0o; Line
282
Figure 201 Line current (top trace), line-2-line voltage (middle trace) and line-2-star point voltage
283

MEASURED
ESTIMATED
Adv Angle [Deg.]
Idemand [A]
15
Speed [RPM]
1357.5
1348.3
6.4
6.0
Taverage [Nm]
IphRMS [A]
6.5
IphPEAK [A]
Cu Loss [W]
12.6
147.3
Tripple - MEAN [%]
45.3
Tripple - MAX [%]
36.2
Pshaft [W]
912.6
Efficiency [%]
851.0
85.2
284

currents with Advance Angle = 0o; Line Current Demand = 15A; Speed =1357.47RPM; DC-Link
Voltage = 175V
5.00
10.22
5.10
10.24
1.745
9.92
1.41
8.72
12.62
6.24
11.5
1.95
Compared to the low speed measurements, the motor efficiency was significantly
improved throughout the high speed measurements. This is again of no surprise for an SR
machine running under voltage control at high speeds. Although the output torque is
reduced the output power remains relatively high with speed resulting in improved motor
efficiency. As the machine was running under full voltage control with all the high speed
examples presented in the preceding subsections, the currents flowing through the IGBTs
and the reverse recovery diodes could be determined from the line currents. Once the
instantaneous IGBT and diode currents were determined, the peak and rms IGBT and
diode currents and the inverter VA ratings could be calculated. These are presented in
Table 31, Table 33, Table 35 and Table 37 for High Speed Measurement-1, High Speed
Measurement-2, High-Speed Measurement-3, and High Speed Measurement-4,
respectively. The VA ratings for the IGBTs and the diodes are very similar when the
calculations are based on the device peak currents. However, there is a significant
difference in VA ratings of IGBTs and recovery diodes when the rms device currents are
used. This can easily be explained by analysing the device current waveforms in Figure
190, Figure 194, Figure 198, and Figure 202. The peak currents flowing through the
IGBTs and the diodes are very similar in all of the four figures. However, the conduction
periods of the diodes are significantly lower than IGBTs. This results in significantly less
rms diode currents and hence significantly less VA ratings based on rms currents. It can
be said that the correlation between the measured and estimated average torque output is
very good in Table 30, Table 32, Table 34, and Table 36. The Matlab script used for post
processing the measured data is still able to predict the instantaneous output torque
285

reasonably well, even when the operating speed of the machine is increased and the drive
enters into full voltage control.
The periods where the phase currents are zero are shorter than under current control due
to a reduced amount of time available to de-flux the phase and at the same time control
the line currents. The torque waveforms appear to be peaky with more negative torque
generated. As a result, the torque ripple is higher at high speeds than low speeds. This is
typical of high speed operation where the torque ripple is compromised in order to
achieve high average output torque. The flux vs. MMF loci of all four high speed
measurements are displayed against the aligned and unaligned magnetisation curves (see
Figure 175) of a single slot in Figure 203.
Figure 203 Flux vs. MMF loci for the high speed measurements
Attempts for simulating the six-phase drives with mutually coupled phase windings have
been unsuccessful. The following points have yet to be resolved:
Switching states and phase voltages / phase diode states when the phase currents
fall to zero.
286

Star point voltage. When there is no mutual coupling a vector definition of the star
point voltage has been successful. However, the same approach has not been
successful when simulating the machine with mutually coupled phase windings.
Transformation matrices. As the simulation model has not been successful it is

unknown how well the transformation matrices represent the machine dynamics.
9.4 SUMMARY
A method of keeping the size and cost of the power electronics down and increase the
number of phases in an attempt to reduce the output torque ripple of SRM drives is to
make use of the 3-phase full-bridge circuit and drive a 6-phase machine. To achieve
unipolar phase currents additional phase diodes were connected in series with each phase
winding. A pair of phases was connected back-to-back to form a module of phases. Three
current transducers were required to achieve control of the drive system based on line
currents. Each line current control required separate PID and PWM controllers.
Simulation of the six-phase SRM from a 3-phase full bridge circuit has been successfully
achieved when the phase windings were magnetically de-coupled. Unfortunately the
simulation of the 6-phase machine with magnetically coupled phase windings could not
be finalised and verified against measurements. Although the waveforms from the
simulation models of the star and delta connected 6-phase drives with magnetically decoupled phases looked favourable they need to be validated against measurements in
future work.
Using the test rig explained in Chapter 5 Machine Construction Techniques and Test
Rig Description the star-connected six phase segmental rotor SRM has been successfully
driven from the 3-phase full bridge circuit. The line currents were the controlled
parameters.
The machine was run under voltage and current control and the measurements have been
presented. Matlab scripts were written to analyse captured waveforms. The scripts were
able to estimate some of the performance parameters such as phase RMS currents,
running speed (from the period of phase current), instantaneous phase torques, average
total output torque, torque ripple, shaft power and total machine copper losses (assuming
287

steady-state winding temperature of 100oC). Although estimated performance parameters
were based on the 2D FE simulated flux-MMF characteristics of a slot of the machine, the
predicted values for the average total output torque correlated reasonably well with the
measurements when the machine was run at both high and low speeds.
It is believed that the higher-than-expected torque ripple values experienced with the 6phase segmental rotor SRM drive were due to the slot MMF pattern resulting in peaky
output torque characteristics. Unlike the conventional toothed rotor SRMs, where stator
tooth permeance is regulated by the positioning of the rotor teeth, for the segmental rotor
SRM the slot permeance is regulated by the position of the rotor segments. Therefore in
terms of machine torque, the total output torque is the sum of the individual slot torques
as opposed to the sum of the individual tooth torques in the conventional toothed rotor
SRM.
288
Chapter 10 COMPARISON OF MACHINES & DRIVES
CHAPTER 10
10 Comparison of
Machines and Drives
10.1 ELECTROMAGNETIC PERFORMANCE
COMPARISON
This section summarises the comparison made between the 6-phase segmental and
toothed rotor SRMs with 12 stator teeth and 10 rotor poles and the 3-phase single-tooth
winding (also a 12-10 machine) segmental rotor SRM. Comparisons given in this section
of the chapter are based on finite element analysis obtained for each machine and are
based on torque per copper loss. For the sake of comparison it is assumed that the phase
currents are under perfect current control for each machine (illustrated for 6-phase
machines in Figure 204 and for 3-phase machine in Figure 205).
289
290
10.1.1 Number of turns per slot, slot fill factor and conductor
diameter as designed
In this section comparison will be made based on the ability to produce torque for a given
total copper loss. Measured phase resistances are used throughout except for the 6 phase
conventional SRM, which was not completed. Total average torque has been calculated
by
Nph Nr
360 o
Torque = A _ Pos ( X ) A _ Pos ( X

)
Nph * Nr
2
Equation 36
where A _ Pos ( X ) is the area below the magnetisation curve at X degrees after the
360o
unaligned position, A _ Pos ( X
) is the area below the magnetisation curve at
Nph * Nr
(X
360o
) after the unaligned position, Nph is the number of phases, and Nr is the
Nph * Nr
number of rotor poles.
Torque per unit copper loss for each machine has been tabulated in Table 38.
Table 38 Summary of torque per copper loss as per machine design for the 6-phase and 3-phase
segmental and 6-phase conventional machines
6-phase
3-phase
6-phase
Segmental
Segmental
Conventional
Conductor Diameter [mm]
1.40
1.00
1.40
Slot Fill Factor [%]
49.51
40.98
48.86
Number of Turns per Slot
88.00
135.00
100.00
Phase Resistance []
0.44
2.56
0.45
Total Copper Loss [W]
300.00
300.00
300.00
Iphase [A]
18.48
10.83
18.17
Max Torque [Nm]
20.99
28.72
22.81
Torque per Copper Loss [Nm / W]
0.0700
0.0957
0.0760
291

In Table 38 each machine is given as they are designed and it is obvious that each has a
different slot fill factor and number of turns. According to Table 38, the 3-phase
segmental machine is able to produce approximately 36.8% more torque per copper loss
than the 6-phase segmental machine. Also the 6-phase conventional machine is able to
generate approximately 8.7% more torque per copper loss than the 6-phase segmental
machine.
10.1.2 Number of turns per slot, slot fill factor kept the same
In this case the number of turns per slot and the slot fill factor of all three machines have
been fixed to those of the 3-phase segmental machine (135 turns per slot and 41% slot fill
factor). Table 39 summarises the torque per copper loss comparison.
Table 39 Comparison of torque per copper loss of 6-phase and 3-phase segmental and 6-phase
conventional machines for fixed number of turns per slot and slot fill factor
6-PHASE
SEGMENTAL
3-PHASE
SEGMENTAL
6-PHASE
CONVENTIONAL
273.59
258.74
315.06
136
135
136
Intended Slot Fill Factor
0.41
0.41
0.41
Conductor CSA [mm ^ 2]
0.82
0.79
0.95
1.0248
1.0003
1.0997
Realistic Dcond [mm]
1.00
1.00
1.10
Achieved Slot Fill Factor
0.39
0.41
0.41
PHASE RESISTANCE []
1.37
2.56
1.02
TOTAL COPPER LOSS [W]
300.00
300.00
300.00
PHASE CURRENT [A]
10.46
10.83
12.12
MAXIMUM TORQUE [Nm]
17.59
28.72
19.63
0.0586
0.0957
0.0654
Total Available Slot Area

[mm ^ 2]
Nturns
Calculated Dcond [mm]
TORQUE per unit COPPER

LOSS [Nm]
292

According to Table 39, the 3-phase segmental machine is able to produce approximately
63.3% more torque per copper loss than the 6-phase segmental machine. The 6-phase
segmental machine is still inferior to the 6-phase conventional machine, this time, by
approximately 11.6%.
The performances of the machines can also be compared using MMF vs. flux
characteristics. The numbers of turns per coil, the number of active coils and the stator
tooth widths for all three prototype machines are tabulated in Table 40.
Table 40 Number of turns, number of active coils and stator tooth widths of all three machines
Number of Turns per Coil

Active Number of Coils
Stator Tooth Width - Narrow [mm]
Stator Tooth Width - Wide [mm]
6-phase
Segmental
68
4
n/a
12.68
3-phase
Segmental
135
2
10
20
6-phase
Conventional
68
4
n/a
11.41
As both the segmental and conventional rotor 6-phase machines have two phases
conducting at any one time, the number of active machine coils is 4 for both machines.
Only the wide teeth of the 3-phase segmental rotor machine are wound resulting in 2
active coils at any one time. In order to have the same MMF in each machine, the
numbers of turns per coil of both 6-phase machines are adjusted to 68. The coils of each
machine are energised in 2D FE and the MMF vs. flux curves in the aligned and
unaligned rotor positions are plotted for each machine in Figure 206.
293
Figure 206 MMF vs. flux plots of the 3-phase segmental (in black), 6-phase segmental (in blue) and 6phase conventional (in red) rotor machines in the aligned and unaligned (dotted curves) positions
From Figure 206, the area between the aligned and unaligned curves can be calculated
and used to assess the output torque capability of each machine. This is tabulated in Table
41.
Table 41 Table of the areas covered by the magnetisation curves of all three machines
AREA between ALIGNED and

UNALIGNED curves
6-phase Seg
3-phase Seg
6-phase Conv
8.54
14.4
8.05
According to Table 41, the 3-phase segmental machine is approximately 68.0% better
than the 6-phase segmental and the 6-phase segmental is only about 6.0% better than the
6-phase conventional machine. In Figure 206, it is clear that the saturation point of the
aligned magnetisation curve of the 3-phase machine is considerably higher than those of
the 6-phase machines. This is due to the difference in the stator tooth widths of the
machines (Table 40). The linear inductances of the 3-phase segmental machine in the
aligned and unaligned positions are also considerably higher than those of the 6-phase
machines. This can be explained with the help of Figure 207. In Figure 207, the flux plots
294

of all three machines are presented in the aligned position. Clearly, the overlap between
the stator teeth (slots) and the rotor segments of the 3-phase segmental machine is
considerably larger than that seen in the segmental and conventional rotor 6-phase
machines resulting in increased linear inductance in the aligned and unaligned rotor
positions for the 3-phase machine.
Figure 207 2D flux plots of a-) 3-phase segmental rotor b-) 6-phase segmental rotor c-) 6-phase
conventional rotor machines in the aligned position
It is apparent from the analysis above that the 6-phase segmental machine is not
significantly better than its conventional counterpart. This is believed to be due to the
machine design; having short pitched coils wound around each stator teeth. Note that a
machine design approach similar to that of the 3-phase single-tooth segmental rotor
design could be undertaken with the 6-phase segmental machine, where the machine
stator could be formed by narrow and wide teeth and coils wound only around the wide
ones. However, this would result in unbalanced forces during machine operation,
increasing noise and vibration of the machine and reducing bearing life.
295
10.2 TORQUE RIPPLE COMPARISON OF THREE

AND SIX PHASE DRIVES
In chapters 7, 8 and 9 measured waveforms for the 2-phase, 3-phase and 6-phase drives
were used to estimate the instantaneous total machine output torque waveforms. The
estimated waveforms were verified against measurements by comparing the values for
speed and total average torque output. The aim of this PhD project was to achieve a low
torque ripple SRM drive without the need for complicated control algorithms and / or
expensive and complicated power electronics. It is quite apparent (and as expected) from
the waveforms and tables given in the preceding chapters (7, 8 and 9) that the two-phase
drive has the highest torque ripple. However comparing the instantaneous torque
waveforms of the 3-phase segmental machine driven from the asymmetric half-bridge
inverter to those of the 6-phase segmental machine connected in star and driven from the
3-phase bridge circuit the conclusion is not as apparent as one expects. Out of the four
examples given for the 3-phase segmental rotor drive driven from the asymmetric halfbridge inverter the one with the highest conduction angle has the lowest torque ripple.
This is not unexpected, as an increased conduction angle results in increased overlapping
between torque contributions of individual phases. Higher conduction angles are
advantageous for reducing the machine output torque ripple but result in increased copper
losses and potentially reduced average torque output per copper loss. The other two
examples given for the 3-phase drive where the torque ripple is relatively high are when
the conduction angles are as low as (approximately) 120o.
The star connected 3-phase full bridge circuit driven 6-phase machine is restricted to 120o
conduction angle. Comparing the 3-phase drive example with the high conduction angle
of 164.6o to the high and low speed examples given for the 6-phase star connected drive
examples suggest that in terms of torque ripple, no considerable gain has been achieved
with designing and using a 6-phase machine. Arriving to a final conclusion based on one
example would not be justifiable. Therefore, theoretical analysis of the average total
torque, torque ripple and copper loss of the 3-phase and the 6-phase 12-10 segmental
rotor machines has been carried out in Matlab by a simple m-file script simulation. The
analysis was extended to include the 6-phase 12-10 machine with conventional toothed
rotor. The analysis assumes that all of the machines are under perfect current control and
have the pattern shown in Figure 208.
296
Figure 208 Assumed phase current pattern for the torque ripple comparison study
In Figure 208:
rise is the angle corresponding to the rise time for the phase current from zero to the
demand level. This is set to 20o for all of the machines.
conduction is the conduction angle. This is fixed at 120o for the six phase drives driven from
the 3-phase full bridge circuit.
fall is the angle corresponding to the fall time for the phase current from the demand
level back down to zero. This is again set to 20o for all of the machines.
As introduced in the previous chapter, the total output torque of the 6-phase segmental
drive is the sum of the contribution of each slot torque. Therefore the slot MMF pattern of
the 6-phase segmental drive has a direct contribution to the instantaneous output torque
waveform and hence the machine output torque ripple. The stator slots of the 6-phase
segmental rotor machine share windings of two adjacent phases. Once the phase current
demand is defined as given in Figure 208 the slot MMF pattern has the form depicted in
Figure 209. This is significantly different to the phase current pattern in Figure 208 that is
useful for the 6-phase conventional and 3-phase segmental rotor machines. In Figure 209,
as each phase is energised for 120o, each slot is energised for 200o in total.
297
Figure 209 Graphical explanation of the total conduction period for the slot MMF
The Matlab script was configured to go through batch-runs where advance angle,
conduction angle and current demand could be varied over a wide range of values for
each machine. The set-ups for batch-runs for each machine are given below:
1. 6-phase 12-10 Segmental Rotor Machine:

Conduction angle was fixed at 120o. Advance angle was varied from -60o to +60o
and current demand from 5A to 40A, respectively.
2. 3-phase 12-10 Segmental Rotor Machine:

Conduction angles of 120o, 150o and 180o were analysed. For each conduction
angle advance angle was varied from -60o to +60o and the current demand from
2A to 28A.
3. 6-phase 12-10 Conventional Toothed Rotor Machine:

Conduction angle was first fixed at 120o and then 150o was analysed in order to
asses the improvement in output torque ripple with higher values of conduction
angle. Similar to the 6-phase segmental machine, advance angle was varied from 60o to +60o and the current demand from 5A to 40A.
The results heavily depend on the magnetic characteristics of the machine. In order to
have a fair comparison, the magnetic characteristics for each machine should be generated
by using the same method, i.e. FE or measurements (assuming FE models and
measurements for each machine have similar accuracy). The analysis presented in this
298

section uses the magnetic characteristics for all machines generated from 2D FE analysis.
Also it is assumed that from one machine to the next, the errors in magnetic
characteristics are not a significant factor on the final results, i.e. all FE models have
sufficient detail to achieve similar amounts of accuracy. Although already presented in
the relevant chapters, it is beneficial to have the static torque characteristics of each
machine in this chapter as well in order to aid with the understanding of the torque ripple
comparison analysis. The static torque characteristics of the 6-phase segmental, 6-phase
conventional and 3-phase segmental rotor machines are presented in Figure 210, Figure
211 and Figure 212, respectively. The static torque characteristics of the 6-phase
segmental rotor machine in Figure 210 are therefore given for a single slot.
Figure 210 Static torque characteristics of the 6-phase 12-10 segmental rotor machine from 0At to
3520At
299
Figure 211 Static torque characteristics of the 6-phase 12-10 conventional rotor machine from 0A to
45A
Figure 212 Static torque characteristics of the 3-phase 12-10 single-tooth winding segmental rotor
machine from 0A to 28A
300

It is clear from the figures presented above that the 6-phase machine with a conventional
rotor has more sinusoidal static torque characteristic compared to both the 3-phase and 6phase segmental rotor machines. Moreover, at low excitation levels, as the magnetic
permeance rises almost linearly with angle of overlap of the teeth, torque is almost
constant. In the segmental machines, as the area of overlap on one side of the magnetic
circuit increases, it decreases on the other resulting in non-constant torque even at low
excitation levels [12]. At high excitation levels, the difference in the static torque
characteristics of a segmental and a conventional machine is small.
The results are then used across a range of torque values to find the minimum torque
ripple which can be achieved with each drive. The process is explained in a simple flowchart format in Figure 213.
Define Torque Range

(e.g. 1Nm to 40Nm)
(Tmin Tmax)
YES
FINISHED!
Define Torque
Stepping (e.g. 1Nm)
(Tstep)
Target
Torque Range
Finished?
(Tmax)
NO
Load Target Torque
Value (e.g. 1Nm
starting with Tmin)
Define Target Band

(e.g. 0.2Nm)
(Tband)
Generate Target Torque

Band (e.g. 1Nm Tband)
Load Batch-Run
Results EXCEL File
Find Batch-Run Results

within Desired Torque Band
Extract the Batch-Run

Result with the Minimum
Torque Ripple
Figure 213 Simple flow-chart explanation of the Matlab script analysing the batch-run results of the
perfect current control simulation
301

In Figure 214 a graphical comparison of torque ripple against average output torque for
the 6-phase segmental, 3-phase segmental and 6-phase conventional 12-10 SRM drives is
given. The results are shown from 6.0Nm up to 43.0Nm of average output torque. In
Figure 215 and Figure 216, the variations of demand current and the advance angle
corresponding to each point in Figure 214 are given. In Figure 216 negative advance
angle corresponds to energisation before the unaligned rotor position whereas positive
advance angle corresponds to energisation after the unaligned rotor position.
Figure 214 Torque ripple vs. average output torque characteristics for the 6-phase conventional
toothed rotor drive, 6-phase segmental rotor drive and 3-phase segmental rotor drive (with three
different conduction angles)
302
Figure 215 Current demand vs. average output torque characteristics for the 6-phase conventional
Figure 216 Advance angle vs. average output torque characteristics for the 6-phase conventional
303

In terms of output torque ripple, the following can be concluded from the above figures:
The 6-phase conventional machine suffers from restricted conduction angle of

120o. If the conduction angle could be further increased the torque ripple would be
further reduced for the 6-phase conventional drive. Of course increased
conduction angle will inevitably result in increased copper losses and
consequently reduce the average output torque per unit machine copper loss.
The 6-phase segmental machine suffers from the two-stage slot MMF pattern
which results in peaky output torque characteristics. This is demonstrated with a
theoretical example in Figure 217.
The 3-phase segmental machine with 120o conduction angle is significantly worse
than both of the 6-phase drives. However, when the conduction angle of the 3phase drive is increased beyond 120o the torque ripple characteristics of the 3phase segmental drive improve dramatically. When the conduction angle is set to
1/2 of an electrical cycle for the 3-phase drive, it is able to compete very well
against both 6-phase drives throughout the torque range demonstrated in Figure
214.
In reality, higher torque can be achieved under current control at low speeds and
lower torque under voltage control at high speeds. The conduction angle is
generally limited to 120o under current control in order to maximise the torque per
copper loss. This can compromise the torque ripple characteristic of the 3-phase
drive as it is likely to have more torque ripple than that shown in Figure 214 at
higher torque levels.
In the first part of this chapter, it was shown that both the conventional and the
segmental rotor 6-phase drives are significantly inferior to the 3-phase segmental
drive in terms of torque per copper loss. As a result, the 3-phase drive is able to
achieve the same output torque with significantly less current (see Figure 215)
even with 120o conduction. If all three drives were designed to have similar output
torque capability the author believes that the torque ripple difference between the
6-phase and 3-phase drives would be more pronounced even when the phases of
the 3-phase drive were energised for longer periods of time. Of course, this would
result in a significantly smaller 3-phase machine design.
304
The 6-phase segmental rotor machine still performs better than the 3-phase drive
with 180o conduction up to 20Nm. This is believed to be the result of having 200o
of effective slot energisation due to 120o of phase energisation, which results in
significant overlap between individual slot torques. On the other hand, the 6-phase
conventional machine is better than the 3-phase drive with 180o conduction only
up to 11Nm.
As seen in Figure 214, the torque ripple characteristics of the 6-phase and 3-phase
segmental drives (at 150o and 180o) first reduce with increased average output
torque and then increase. This is believed to be a result of the static torque
characteristics of both of the segmental machines. The trend is more pronounced
with the 6-phase segmental drive but can also be seen in the 3-phase drive,
especially when the conduction angle is changed to 180o (only the beginnings of
this characteristic can be seen in the 3-phase drive with 180o).
Figure 217 a-) Instantaneous slot torque (top trace) and slot MMF (bottom trace) waveforms for the
6-phase segmental rotor drive with ~47% torque ripple
The idea of what would happen to the torque ripple if the conduction angle of the 6-phase
drives could be increased beyond 120o was investigated using the 6-phase conventional
toothed rotor machine. For this purpose, the conduction angle of this machine was further
increased to 150o. The results are illustrated from Figure 218 to Figure 220.
305
Figure 218 Torque ripple vs. average output torque characteristics for the 6-phase conventional
toothed rotor drive. Blue trace when conduction angle = 120o and red trace when conduction angle =
150o
Figure 219 Current demand vs. average output torque characteristics for the 6-phase conventional
150o
306
Figure 220 Advance angle vs. average output torque characteristics for the 6-phase conventional
150o
It is quite clear from Figure 218 that increasing the conduction angle beyond 120o
improves the torque ripple performance of the 6-phase conventional toothed rotor drive. It
is not surprising to see that the current demand and the advance angle profiles are shifted
downwards with increased conduction angle in Figure 219 and Figure 220, respectively.
It must be stressed here that in practice, it is not possible to achieve perfect current control
and the shape of the currents will deviate significantly from those presented in Figure 208
and Figure 209, especially at high speeds when the controller cannot achieve the desired
current level and enters into full voltage control. The required advance angle and the
converter topology will also significantly influence the shape of the output torque.
Therefore the practical limitations will affect the torque ripple for all of the drives.
307
10.3 COMPARISON OF INVERTER VA-RATINGS

Measurements and simulation results of the 3-phase segmental rotor machine, driven
from the asymmetric half-bridge inverter were presented in Chapter 8. Measurements
taken when the drive was operating under full voltage control were used to estimate the
converter VA rating of the 3-phase drive. Similarly, measurements of the 6-phase
segmental rotor machine driven from the 3-phase full bridge circuit when connected in
star configuration were used to estimate the converter VA rating of the 6-phase segmental
rotor drive. However, no direct comparison could be made between the 3-phase and 6phase drives in terms of converter VA ratings as the measured operating speed and the
average output torque values were different.
Unfortunately, as far as the 6-phase segmental rotor machine is concerned, analysis
concerning converter VA ratings could not be advanced any further as the efforts to
generate dynamic simulation models of the drive were unsuccessful (see Chapter 9).
Although not verified against measurements, the six-phase drive simulations with star and
delta configurations were generated and run successfully when the machine with no
mutual coupling was modelled. It was believed that a comparison of converter VA ratings
between the 3-phase and 6-phase drives would be beneficial. For this purpose, the
dynamic simulation models for the 3-phase and 6-phase drives were used. The 3-phase
12-10 single-tooth winding segmental rotor and the 6-phase 12-10 conventional rotor
machine with no mutually coupled phases were modelled. The 6-phase machine was
connected in star and delta configurations and driven from the 3-phase full bridge circuit,
whereas the 3-phase machine was driven from the asymmetric half-bridge circuit. The
comparison was divided into two sections. The first section below is the comparison of
inverter VA-ratings of the 3-phase and 6-phase drives when both drives were operating
under full voltage control (high speed operation). The second section follows on from the
first one but looks at the comparison when both drives were operating under current
control (low speed operation). For all simulations the dc-link voltage is fixed at 175V.
308
10.3.1 Drive under Voltage Control

For the purposes of this comparison, all three drives have been simulated at 1200rpm and
the current demands were set to 50A in order to ensure operation under full voltage
control. It should be noted that the current demand for the 3-phase drive is for the phase
current and for the 6-phase drives it is the line current demand. The numbers of turns of
the machines and the advance angles were then adjusted (by using the simulation GUI) in
order to achieve 20Nm at 1200rpm with all three drives. The conduction angles of the 6phase drives were limited to 120o, however, in order to maximise the output torque at
high speed, the conduction angle was set to 180o for the 3-phase drive. As explained in
Chapter 7, the aim here is to maximise the number of turns for the given speed and torque
requirements in order to minimise the current in the circuit. This way, the inverter and
machine losses are minimised for the given operating point. Once again, any change in
the number of turns does not affect the machine copper losses (the same amount of stator
slot area with the same slot fill factor is available to fill with copper). Table 42
summarises the simulation results for the 6-phase delta connected drive, 6-phase star
connected drive and 3-phase drive under voltage control.
For the definition of the parameters in Table 42 readers are referred to Chapter 7.
309

Table 42 Summary of dynamic simulation results with the delta and star connected six-phase
conventional rotor SRM driven from the 3-phase full bridge inverter and the three-phase single-tooth
winding segmental rotor SRM driven from the asymmetric half-bridge inverter at 1200 rpm
generating approximately 20Nm average output torque
Six-Phase
Delta
Connected
3-Phase Full
Bridge
Six-Phase
Star
Connected
3-Phase Full
Bridge
ThreePhase
Asymmetric
Half-Bridge
Vdc_Link [V]
Speed [RPM]
Tave [Nm]
Output Power [W]
Demand Current [A]
Nturns per coil FINAL
R_Phase [Ohms]
Advance [Deg.]
Conduction [Deg.]
175
6
1200
20.01
2514.25
50
50
64
0.744
-59.9
120
175
6
1200
20.00
2512.80
50
50
39
0.276
-38.2
120
175
3
1200
20.01
2513.93
50
135
66
0.612
-85
180
Iphase_RMS [A]
Iphase_PEAK [A]
9.71
420.55
19.29
128.89
25.81
154.69
81.38
16.24
436.67
29.97
136.75
41.11
177.86
80.35
12.26
275.83
22.58
117.00
0.00
117.00
86.49

11.71
5.68
11.65
2.36
23.36
8.04
12.53
5.89
11.73
2.44
24.26
8.33
9.43
4.89
6.83
1.52
16.26
6.41

27.1
32.6
32
39.5
85.2
143.8
310
10.3.2 Drive under Current Control

Here, the motor speed was reduced to 100rpm. The numbers of turns determined when
the drives were under voltage control (in the previous section) were used in the
simulations for each drive. The controllers now work under current control. All three
drives use the same PID controller parameters. In order to maximise the torque per unit
copper loss the conduction angle of the 3-phase drive is now reduced to 120o. The current
demands and the advance angles for all three drives were then optimised in order to
achieve the required output torque (same as the previous section 20Nm). Simulation
results for the 6-phase delta and star connected and the 3-phase drive are summarised in
Table 43.
311

Table 43 Summary of dynamic simulation results with the delta and star connected six-phase
conventional rotor SRM driven from the 3-phase full bridge inverter and the three-phase single-tooth
winding segmental rotor SRM driven from the asymmetric half-bridge inverter at 100 rpm
generating approximately 20Nm average output torque
Six-Phase
Delta
Connected
3-Phase
Full Bridge
Six-Phase
Star
Connected
3-Phase Full
Bridge
ThreePhase
Asymmetric
Half-Bridge
Vdc_Link [V]
Speed [RPM]
Tave [Nm]
Output Power [W]
Demand Current [A]
R_Phase [Ohms]
Advance [Deg.]
Conduction [Deg.]
175
6
100
20.06
210.08
21.5
50
64
0.744
3
120
175
6
100
20.06
210.04
20.5
50
39
0.276
30
120
175
3
100
20.00
209.42
14.4
135
66
0.612
30
120
Iphase_RMS [A]
Iphase_PEAK [A]
8.24
302.87
15.39
163.91
20.72
184.63
30.12
13.84
317.30
26.05
158.95
37.79
196.74
29.01
8.54
133.98
15.78
122.29
0.00
122.29
44.97

114.95
47.42
114.93
41.30
229.88
88.71
109.80
45.29
109.61
38.80
219.41
84.09
79.09
32.54
79.09
32.54
158.18
65.08
25.7
27.9
25.8
29.6
75.8
99.3

312
10.3.3 Discussion on Converter VA-Rating Comparison

In the preceding two sections the simulation results of the star and delta connected sixphase drives and the three-phase drive under voltage and current control were given in
tabulated formats. In this section, these simulation results will be scrutinised.
Although the rms phase current seen in the star connected six-phase drive is
approximately 1.67 times more than the delta connected six-phase drive, the total
machine copper losses of both drives are very similar. This is due to the difference in
machine phase resistances between the star and delta connected drives. Phase resistance
of the star connected six-phase machine is almost 1/3 of the phase resistance of the delta
connected six-phase machine.
The six-phase delta connected drive has approximately 1.32 times the power electronic
losses of the three-phase drive and the star connected drive has approximately 1.52 times
the power electronic losses of the three-phase drive. According to the inverter VA ratings
comparison (based on peak currents), under voltage control, the six-phase delta connected
drive requires approximately 1.43 times and the star connected drive requires an
approximately 1.5 times bigger inverter than the three-phase drive. When the same
comparison is based on the rms currents the difference is smaller. Inverter VA rating
comparison based on rms currents reveals that the inverter of the star connected drive is
1.25 times and the inverter of the delta connected drive is 1.3 times larger than the
inverter of the three-phase drive under voltage control. Similar results are observed when
the drives are under current control. The main difference under current control in terms of
inverter VA ratings is that the star-connected drive outperforms the delta-connected drive.
However, the difference is only about 5.0% and is unlikely to result in a bigger inverter
requirement for the delta connected drive as the ratings of the devices will have quantised
characteristics.
Overall, the three-phase drive is approximately 6% more efficient than the six-phase star
connected drive when both drives are under full voltage control. The difference is larger
when the drives are under current control. The three-phase drive is approximately 15%
more efficient than the six-phase star connected drive when both drives are under current
control. The increase in the efficiency difference under current control is mainly due to
the increased circulating currents in the six-phase drives. The efficiencies of the star and
delta connected six phase drives are very similar, both under current and voltage
controlled situations. In both cases the delta connected drive is approximately 1.0% more
313

efficient than the star connected drive. It is interesting to note that the delta connected
drive generates more inverter losses compared to the star connected drive when the drives
are under current control but the total power electronics losses generated by the delta
connected drive is still lower than the star connected drive. The difference is due to the
additional phase diode losses. The delta connected drive is more efficient in terms of the
phase diode losses. The star connected drive generates approximately 16.0W more phase
diode conduction losses compared to the delta connected drive under voltage and current
control situations. This can be related to the reduced rms phase currents achieved with the
delta connected drive.
10.4 DISCUSSION ON TORQUE RIPPLE

COMPARISON
As a result of the analysis presented in section 10.2, it was concluded that compared to
the 3-phase segmental rotor drive, there was no real advantage of employing either of the
6-phase drives in terms of output torque ripple across a wide range of average torque
levels. However, at the end of the same section, it was also briefly mentioned that due
to practical limitations- the phase current shape would deviate significantly from the
ideal case affecting the output torque waveform. Also the advance angle and the desired
current level used in practice would also affect the overall output torque ripple of the
drive system.
In terms of the aim of this PhD project, it is certainly worth mentioning that unlike the
results obtained with the theoretical analysis in section 10.2 of this chapter, it is obvious
from Table 42 and Table 43 that both the star and delta connected six phase drives have
significantly better output torque ripple characteristics compared to the three-phase drive
under both voltage and current control modes: peak-to-peak torque ripple performances of
the six-phase drives are approximately 1/3 of the three-phase drive under both control
modes. It is also interesting to note that in Table 42, despite the fact that the 3-phase
segmental drive operates with a conduction angle of 180o, the generated output torque has
significantly more ripple compared to those generated by both of the 6-phase drives. The
reasons can be explained with the help of Figure 221 to Figure 224. Note that the current
and torque waveforms in Figure 221 and Figure 223 are scaled to the level of rotor
position for visual clarity.
314
Figure 221 Phase current, phase torque and rotor position waveforms for drives tabulated in Table
42
Figure 222 Phase torques and total torque waveforms for drives tabulated in Table 42
315
Figure 223 Phase current, phase torque and rotor position waveforms for drives tabulated in Table
43
Figure 224 Phase torques and total torque waveforms for drives tabulated in Table 43
316

It is not surprising to see that the output torque of the 3-phase drive has significant torque
ripple component when the conduction angle is set to 120o. This can also be observed
from Figure 214 in the theoretical torque ripple analysis work carried out in section 10.2.
As already discussed, this is because of the relatively low overlap angle between
individual phase torque waveforms (see bottom trace in Figure 224). Even though the
conduction angles of both of the 6-phase drives are limited to 120o both maintain a high
degree of overlap between individual phase torque waveforms which results in relatively
low output torque ripple under both voltage and current control modes. This is due to a
combination of increased number of phases and phases conducting current for long
periods of time over an electrical cycle (see first two traces in Figure 221 and Figure 223).
It should also be noted that the static torque characteristics that are used to generate the
results in section 10.2 are now different due to the changes made to the numbers of turns
for each drive. When the drives are under voltage control, the phase current waveforms
are significantly different to those simulated in the theoretical torque ripple analysis
presented in section 10.2. Another reason for the high torque ripple observed with the 3phase drive is believed to be the required advance angle and the current demand for the
given torque level to achieve minimum converter VA-rating.
10.5 SUMMARY
In this chapter both 6-phase designs have been compared against the 3-phase, single-tooth
winding segmental rotor design, based on torque per copper loss. The comparison
revealed that both 6-phase designs were significantly inferior to the 3-phase design. This
was attributed to two factors; firstly, the stator teeth of the 3-phase machine that carry the
windings of the coils are considerably wider than those of the 6-phase machines.
Secondly, the overlap between the stator teeth and the rotor segments of the 3-phase
machine is larger than that seen in the 6-phase machines.The comparison based on the
torque per unit copper loss suggested a slight advantage for the 6-phase toothed rotor
machine compared to the 6-phase segmental rotor machine.
In the second part of this chapter, theoretical analysis of torque ripple vs. average output
torque of the 3-phase and the 6-phase 12-10 segmental rotor machines and the 6-phase
conventional rotor machine has been carried out in Matlab by a simple m-file script
317

simulation. For the six-phase drives, the conduction angles were fixed to 120o, current
demands and the advance angles were varied. For the three-phase drive, the conduction
angle was also varied. Once the average torques for all of the three drives were calculated,
the script then searched for the result with the lowest torque ripple for each average
output torque level from 6.0Nm to 43.0Nm. It was shown that with the same conduction
angle, the output torque ripple could be significantly improved with both conventional
and segmental rotor 6-phase drives. However, as the 3-phase drive can operate with
conduction angles of up to 180o, the output torque ripple of the 3-phase drive can be
improved to a level comparable to that of both 6-phase drives throughout the torque range
(that was analysed in this chapter). It was also shown that the 6-phase segmental rotor
machine can still outperform the 3-phase drive with 180o of conduction angle up to 20Nm
mean output torque.
The VA-ratings of the 6-phase conventional rotor machine connected in star and delta
configurations and driven from the 3-phase full bridge circuit has been compared to that
of the 3-phase segmental rotor machine driven from the asymmetric half-bridge circuit in
the dynamic simulation. The numbers of turns of the machines were adjusted in order to
achieve 20Nm at 1200 rpm when all of the drives were under full voltage control. The
same numbers of turns were then used at 100 rpm and the current demands were adjusted
in order to achieve the same output torque level of 20Nm from all of the drives. The
three-phase drive outperformed both the star and delta connected six-phase drives.
Comparison of the star and delta connected six-phase drives revealed that similar VAratings can be achieved. However, the delta connected drive was 1.0% more efficient than
the star connected drive under both voltage and current controlled situations.
Although the VA comparison of the drives was in favour of the 3-phase one, the same
comparison was also beneficial in realising that both the star and delta connected six
phase drives significantly outperformed the 3-phase drive in terms of output torque ripple.
This was a result of the practical phase current waveforms that deviated significantly from
the idealised ones used in the theoretical torque ripple analysis that was presented in the
second part of this chapter. In terms of the torque ripple comparison, the simulated results
obtained from the dynamic simulation package are believed to be more relevant and they
clearly show the inadequacy of the theoretical approach presented in section 10.2.
318
Chapter 11 CONCLUSIONS AND FUTURE WORK
CHAPTER 11
11 Conclusions and
Future Work
11.1 CONCLUSIONS
The principle aim of the work presented here was to design a 6-phase segmental rotor
machine drive capable of generating torque ripple levels comparable to a 3-phase AC
machine without the need for complicated circuit topologies or control methods. For this
purpose a prototype 6-phase segmental rotor machine was built and tested. The static
torque and flux-linkage characteristics were correlated against FE predictions and were in
good agreement. For direct comparison, a 6-phase conventional toothed rotor machine of
the same inner diameter, outer diameter and stack length was also designed. The
segmental machine design was based on the principles that were used to design the 3phase segmental rotor designs. On the other hand the conventional toothed rotor 6-phase
machine was designed using simple and well-known conventional SRM design equations.
Both 6-phase machines were also compared to a 3-phase single-tooth winding segmental
rotor machine of the same outer diameter, inner diameter and stack length.
The optimal machine design for the 6-phase segmental rotor machine was chosen using
an equation that was based on the average machine output torque, output torque per
copper loss, output torque per machine mass and the inverter VA rating that was
indirectly derived from the unaligned inductance of the machine. 2D FE analysis was
used for the initial and refined designs for both toothed and segmental rotor machines. In
terms of comparative study of different design options, use of 2D FE was believed to be
319

valid. Furthermore, the stack length of the machine is believed to be large enough to keep
the 3D end effects low, hence permitting the use of 2D FE analysis to predict the
measured values. The discrepancies between the measured and 2D FE generated results
are more obvious in the unaligned position where the axial fringing flux is at its highest
level. In the aligned position the saturation characteristic is predicted well with 2D
studies. The difference in the linear region of the magnetisation curves in the aligned
position is believed to be due to the difference in the final machine air-gap compared to
the design intent value of 0.3mm.
11.1.1 Machine Performance

In terms of machine performance comparison between the 3-phase segmental and the 6phase segmental and conventional machines, the results are summarised in the following
three sub-sections.
Based on all three machines having the same total slot MMF, the 3-phase segmental rotor
machine requires the least amount of copper for a given average output torque. For a
kilogram of copper, the 3-phase machine is able to generate twice the amount of torque
compared to 6-phase segmental rotor machine whereas the 6-phase machine with the
segmental rotor is only marginally better than the 6-phase machine with the conventional
rotor.
For a given amount of total machine copper losses, the 3-phase machine is found to be
more favourable than both 6-phase machines. The 3-phase machine is able to generate
approximately 50% more torque compared to the 6-phase machines for 300W of machine
copper loss. The comparison between the two 6-phase designs is unlike that experienced
with the three phase machines, where the 3-phase segmental rotor design proved to
generate significantly more torque for a given copper loss compared to a 3-phase
conventional rotor machine of similar dimensions. In the case of the 6-phase designs, the
conventional rotor machine is able to generate marginally more torque than the segmental
rotor machine.
Similar conclusions can be drawn from the comparison of the MMF vs. flux curves of
each machine in the aligned and unaligned rotor positions. The 3-phase segmental rotor
machine is more than 50% better than both 6-phase machines and there is no significant
difference between the segmental and conventional rotor 6-phase machines.
320
11.1.2 Machine Construction Techniques

The stator of the 6-phase segmental machine was constructed from separate cores. This
was somewhat different to the conventional single lamination construction. The use of
separate cores for the stator construction has allowed an easier winding operation and
consequently improved the machine slot fill factor beyond the typical values of 0.4
achieved with conventional single lamination designs. Although higher than usual slot fill
factor was achieved by the separate stator core construction, the author believes that there
is scope to further improve the slot fill factor with a more careful winding operation that
focuses on reducing the bulging of long and thick conductors. Also for designs where
conductors with small diameters and a short machine stack length are required, the
winding operation is expected to be easier with less bulging and curving of conductors,
which in turn is expected to further improve the slot fill factor.
It should be mentioned that the increased component count associated with using separate
stator cores is likely to limit the use of this construction technique. However, in this work,
increasing the slot fill factor has been the main focus. Furthermore, the separation of
cores can be achieved in a similar fashion to the technique used by the Mitsubishi Motors
group [156], where the entire stator can be opened up to ease the winding operation and to
increase the slot fill factor.
11.1.3 Drive Modelling

A set of transformation equations allowing phase quantities in the 6-phase segmental
rotor SRM related to the machine slot quantities has been developed in order to aid with
the understanding and simulation of the drive system. By de-coupling the windings of the
prototype machine, these transformations allow the machine to be modelled in a
conventional way (i.e. each phase can be analysed separately), and are a very useful tool
for simulation and performance analysis of the drive system. However, because the
terminal equations for all phases are effectively coupled with the introduction of the
transformation matrices, all phase flux-linkages are needed to be solved simultaneously.
Compared to the machines with naturally de-coupled phase windings, the computation
time is substantially increased, but this is unavoidable.
321
11.1.4 Inverter VA-Ratings and Drive Testing

Using a power circuit which (by control of device signals) can emulate several common
configurations and by implementing control functions in software, it has been possible to
explore several circuit configurations and control strategies without extensive hardware
modifications. To aid with the simulation and understanding of the more complex 6-phase
drive, the 6-phase machine was also configured to run as a 2-phase machine. In both cases
standard recovery type diodes were connected in series with each phase winding to
prevent phase currents becoming bipolar. In this case, the standard rectifier grade
recovery diodes are low cost. It should be noted that these diodes never experience a large
di/dt as they are placed in series with the phase windings. It is also believed that the phase
diodes do not contribute to turn on losses in the active power devices due to stored charge
effects. The drive circuitry was configured to run the two-phase machine from both an
asymmetric half-bridge and h-bridge circuit configurations. Performance comparison
between the two inverter topologies for the two-phase machine has been carried out in the
dynamic simulation, once the simulations of both topologies were correlated with the
measurements.
With the H-bridge circuit there is a restriction that the sum of the voltages applied to the
two phases is equal and opposite. A scheme is adapted where the on-coming phase
current is used to determine the voltage demand. This reduces the voltage applied to the
phase switching off and consequently gives a long current tail during a period when the
phase produces negative torque. This is a fundamental disadvantage of the H-bridge
arrangement, resulting in a small reduction in torque and a small increase in winding loss.
However, there is a reduction in power device losses because current in an off-going
phase is transferred to the on-coming phase without it flowing through the power
switches.
The two-phase drives were compared in the dynamic simulation at two different operating
speeds (1200 rpm and 100 rpm) where the drives were under voltage and current control,
respectively. When the drives were under voltage control with phases conducting half the
electrical cycle, both drives operate at similar efficiencies (around 82-83%). However,
under current control the asymmetric half-bridge driven drive is slightly more efficient
than the h-bridge driven drive. This is due to increased copper losses and additional phase
diode losses of the h-bridge drive. In terms of total inverter losses, the h-bridge drive
322

performs better than the asymmetric half-bridge drive. However, this is somehow
counterbalanced by extra conduction losses incurred due to the use of the phase diodes in
the h-bridge drive for ensuring unipolar phase currents. Even under current control, the
total inverter losses are not worse with the h-bridge drive as the power devices in the hbridge inverter experience the difference between two phase currents. In terms of the total
inverter VA-ratings (including reverse recovery diodes and IGBTs), the h-bridge drive is
only slightly favourable under current control. However, based on rms device currents,
the advantage of the h-bridge drive in terms of total VA-ratings is elevated when both
drives are under full voltage control. The other advantages of the h-bridge driven twophase drive over the asymmetric half-bridge driven two-phase drive can be summarised
as follows:
The number of connections between the motor and the converter is reduced from 4 to
2. This assumes that the phase diodes can be mounted at the motor end of the drive
system.
Adequate control can be achieved by a single current transducer.
The reverse recovery diodes needed for the h-bridge drive can be part of the power
devices, potentially reducing the required board space for the electronics. Once again
this assumes that the phase diodes are located at the motor end of the drive system.
Furthermore, machines with two or less phases are preferred for high speed applications.
At high speeds the drive is more than likely to be operating under full voltage control
where use of h-bridge topology might be advantageous.
Switched reluctance machines suffer from high torque ripple. A major cause of this is that
each phase produces one torque pulse per electrical cycle, whereas other machine types
produce two. Hence, to reduce the torque ripple of an SRM towards that of a three phase
a.c. machine, it is necessary to have a six-phase SRM. Existing SRM concepts have not
been able to do this in an economical fashion because there are two wires per phase
connecting the machine to the power converter and two controlled power electronic
switches per phase. This research has shown that a six phase SRM can be successfully
driven from a three phase bridge with only three connections between the inverter and the
motor. The number of controlled devices per phase has been reduced from two to one and
323

the number of connections between inverter and machine has been reduced from twelve
to three.
Prior to this PhD work it was shown that the utilisation of basic star and delta connections
were somewhat limited for the 3-phase SR machines. For the star connection each phase
was required to conduct for 2/3 of the electrical cycle resulting in significant periods of
negative torque. The extended period of conduction also increased the machine copper
losses. For the delta connected machine there cannot be a net voltage applied around the
delta. There is a circulating current which cannot be maintained and the phase currents are
required to be bipolar, which results in the same problems as the star connection.
Experimental operation of the 6-phase segmental rotor switched reluctance drive has been
explored with star-connected windings driven from a 3-phase bridge circuit.
In terms of inverter VA-rating and total power electronics losses (in the case of the 6phase drives total power electronics losses include the conduction losses incurred across
the series connected phase diodes) of the six-phase and three phase drives, both the delta
and star connected six-phase drives are inferior to the three-phase drive under voltage and
current control conditions. The delta connected drive was found to be slightly better than
the star-connected one. The gain with the 3-phase drive was smaller when the comparison
was based on the rms device currents.
A set of transformation matrices has been used to determine the torque produced by a
segmental rotor SRM with mutual coupling between phases. Considering the likelihood
of a difference between the dynamic and static torque characteristics and the strong
mutual coupling experienced in this machine, the good correlation achieved between the
measured and predicted average torque values based on the estimated instantaneous
torque waveforms suggest that the transformation matrices work well and are useful in
explaining / predicting the performance of the 6-phase segmental rotor machine with
mutually coupled phase windings.
324
11.1.5 Torque Ripple

Torque ripple results based on the limited number of measurements made upon the 3phase and 6-phase segmental rotor machines were not conclusive. As one might expect
the 3-phase machine driven from the asymmetric half-bridge circuit has the advantage of
being able to run with conduction angles exceeding 1/3 of an electrical cycle. As the
conduction angle is increased towards 1/2 of an electrical cycle the overlap angle between
individual phase torques increases, resulting in reduced torque ripple at the expense of
increased machine copper losses. For this reason a theoretical torque ripple comparison
was carried out between the 3-phase segmental rotor and the 6-phase segmental and
conventional rotor machines. The analysis was based on perfect current control and the
FE generated torque characteristics were used to calculate the instantaneous torque values
for many combinations of advance angle, conduction angle and current demand.
Comparisons were made between 6.0Nm and 43.0Nm of average output torque. For every
value of average output torque, the solution with the lowest torque ripple was chosen for
each machine. The relatively marginal improvement made with the 6-phase segmental
machine was attributed to the slot MMF pattern and the way the torque is generated in the
6-phase segmental machine. For the segmental rotor machine the slot permeances are
regulated by the rotation of the rotor segments and the slot torque is generated as a result
of the slot MMF. As each slot carries windings of two adjacent phases and in the star
connected configuration two adjacent phases are energised simultaneously the MMF
pattern has a double step characteristic. As a result the torque generated per slot also takes
a profile with two steps. The resulting output torque waveform has more peaky
characteristics, which consequently means increased torque ripple. On the other hand, the
output torque ripple of the 6-phase conventional rotor machine driven from the 3-phase
full bridge converter suffers from the limited conduction angle of 120o. Overall, when all
three drives were limited to 120o of phase conduction periods, both 6-phase drives
significantly outperform the 3-phase drive in terms of output torque ripple. However,
when the conduction angle of the 3-phase drive was increased towards 1/2 of an electrical
cycle, the output torque ripple of all three drives was comparable. Nevertheless, compared
to the 3-phase drive, the 6-phase segmental drive was still able to generate torque with a
lower ripple component up to 20Nm.
325

During the inverter VA-rating comparison, it was found that both star and delta connected
6-phase drives with a conventional rotor outperform the 3-phase segmental rotor drive in
terms of output torque ripple; generating only 1/3 of the torque ripple seen in the 3-phase
drive under voltage and current control conditions. This was due to a combination of
factors; required current demand and phase commutation angle for the given output
torque in order to keep the VA-rating of the drive to a minimum, the changes made to the
numbers of turns of each drive and hence the static torque characteristics and the resulting
non-ideal phase current waveforms. In the case of the current control mode, the majority
of the difference is due to the low conduction angle used for the 3-phase drive in order to
maximise the torque per copper loss of the machine. The analysis in the dynamic
simulation shows the inadequacy of the theoretical approach where the ideal current
waveforms are used.
Considering the aim of this PhD work, it can be concluded that the torque ripple of an
SRM can be reduced by increasing the number of phases to six. In doing so, the number
of power electronic devices, current transducers and the number of connections between
the machine and the converter can still be kept to a minimum. This can be achieved by
driving the machine from a conventional 3-phase bridge converter and connecting the
phase windings of the machine back-to-back through diodes. Compared to a conventional
toothed rotor machine topology, segmental rotor SRM results in a smaller reduction in
the torque ripple. This is believed to be due to the way the phase windings are arranged
within the machine slots and the way the torque is generated in segmental rotor SRMs.
11.2 FUTURE WORK

Although several simulation models have been successfully generated throughout the
course of this project, without a doubt one of the first steps of a future project would be
the generation of a working dynamic simulation model of the six-phase machine with
mutually coupled phase windings. Once this dynamic simulation model is correlated
against the measurements, a much more detailed analysis of the 6-phase segmental rotor
SR drives can be investigated.
326

The aim was to have an SR drive capable of achieving torque ripple levels that are
comparable to the 3-phase brushless dc machines. Due to time restrictions and
unavailability of the characteristics of such a 3-phase BLDC machine the comparison has
not been done in this thesis. A direct comparison of the torque ripple with a 3-phase
brushless dc machine is therefore something to pursue in the future.
The 6-phase conventional toothed rotor design, once built, will have mutually de-coupled
phases. The simulation models of the star and delta connected 6-phase machine driven
from the 3-phase bridge circuit with mutually de-coupled phase windings can then be
correlated against the measurements made upon this machine. A direct comparison can
then be made between the performances of the 6-phase conventional toothed and
segmental rotor machines when driven from a 3-phase bridge inverter in star and delta
connected configurations.
Instead of trying to control the phase currents to a fixed current demand, it is possible
instead to control the line currents to a sinusoidal shape. Sinusoidal phase current
excitation was previously used for driving a 3-phase conventional toothed rotor machine
with fully-pitched phase windings. The machine was connected in star and hence the
phase currents were also the line currents. The sinusoidal excitation pattern resulted in
reduced peak-to-peak torque ripple in the 3-phase conventional machine with fullypitched windings. This brings the SR drive another step closer to the inverter fed
induction motor in terms of the hardware and software required to drive the motor [89].
The same sinusoidal excitation pattern can be used with the 6-phase machine drive. This
time the controlled currents are the line currents. It would be interesting to see the
resulting shape of the phase currents and the effect of the sinusoidal excitation pattern on
the machine output torque for a given loss and more importantly on the ripple component
of the output torque.
The ability to use a 3-phase full bridge circuit to drive an SRM has many advantages,
such as reduced number of power connections to the machine, ability to use devices with
internally packaged diodes and three current transducers (or even only two if the current
transducers can be placed at the power converter end of the drive) instead of six but
perhaps the most promising benefit comes with the volume of production of standard 3phase drives. With minimal modification to the standard drives the integration of SRMs
327

into the drives market will be considerably eased if the use of 3-phase bridge circuitry is
made a reality without compromising the performance of the machine. It would therefore
be an interesting and a very useful study to use a commercially available off-the-shelf 3phase drive to run the 6-phase conventional toothed and segmental rotor machines.
328
Appendix A FLUX-LINKAGE MEASUREMENTS & POST-PROCESSING OF

DATA
APPENDIX A
A. Flux-Linkage
Measurements & PostProcessing of Data
A.1 FLUX-LINKAGE MEASUREMENTS
Locked rotor tests were used to determine the flux-linkage/current/position characteristics
of the prototype machine. As mentioned in Chapter 4, Flux-linkage measurements in this
project are based on the Volt-Seconds method.
In order to collect the flux-linkage characteristics with the volt-seconds method, it is
necessary to supply the machine phase windings with a known voltage for a specified
length of time. A signal generator can supply a voltage pulse and this would work for a
small machine. However for larger machines the current supplied by a signal generator
would be rather small to obtain any sensible readings from the machine. Here, output
from a signal generator is used to switch on a high voltage high current power MOSFET
which in turn turns a larger power supply on. The signal generator is protected by an
opto-isolator against any faults at the high voltage high current end of the test set-up. Full
rated voltage is required over the entire time period of the voltage pulse. Such a supply
can be produced by either having an ac rectifier or having bench-top dc power supply
being used to charge some large capacitors. Due to the size and cost of the required
rectifier and time associated with the building of such a rectifier a bench-top dc power
supply along with two parallel connected high-voltage electrolytic capacitors (4700uF
329

DATA
400V) are used. The pulse width of the signal generator is used to adjust the achievable
peak current. A 100 resistor is also placed in the circuit to safely discharge the large
capacitors once the testing is finished. Two phases of the prototype are connected in
series and excited simultaneously whilst the rotor of the machine is locked at a desired
position. The frequency of the signal generator output is adjusted until the required peak
current is attained. Voltage and current waveforms are monitored and recorded in excel
spreadsheet format for a series of rotor positions. The DSP of the drive system was
arranged so that it could process the 12- bit encoder readings and output the rotor position
to a display on the PC. It must be reminded here that the rotor clamping action must be
strong enough to hold the rotor through the impulse torque which is the force trying to
align the rotor pole(s) with the nearest stator pole(s), when current is applied. This will
ensure errors in rotor position are minimised. As the prototype machine had 10 rotor
segments, measurements are taken for 18 mechanical degrees to obtain the full fluxlinkage characteristics from unaligned to aligned position. A position stepping of 2
degrees was decided to be sufficient enough to predict the performance of the motor
either manually or in simulation with a high accuracy.
In the proceeding section, the GUI that processes the voltage and current data in order to
generate the set of flux-linkage characteristics is explained.
A.2 FLUX-LINKAGE VIEWER GUI FOR DATA

POST-PROCESSING
To overcome problems of noise on the measured data and twisting of shafts, with large
torques, the measured data is post processed in MATLAB environment. As both voltage
and current affect the flux-linkage calculation, any noise or offset present in these
waveforms will be reflected into the flux-linkage as well. The large number of data points
captured by the oscilloscope meant that minute variations in current being captured were
also magnified through the calculation process. Therefore, the raw flux-linkage
characteristics were found to be erratic and a method of filtering was applied to the
current data. A GUI has been created to assist with the calculation of the flux-linkage
characteristics of the prototype motor from captured voltage and current scope traces. The
GUI is able to remove offsets from measured waveforms, apply a moving average filter to
330

DATA
remove any noise effects and plot and save the flux-linkage characteristics of an SRM.
The GUI can also generate torque characteristics based on calculated flux-linkage
characteristics and plot and save this data as well. A screen-shot of the GUI is presented
in Figure A. 1.
Figure A. 1 Screenshot of the GUI used to post-process the voltage and current measurements and
generate flux-linkage characteristics of the prototype machine
The user can change the phase resistance, number of rotor poles, voltage scaling, current
scaling, and filter depth in the GUI window. Voltage and current scaling are there to scale
the captured waveforms to get the real values (due to current transducer and the voltage
probe settings). Filter depth denotes the order of the moving average filter. The GUI is
also able to plot offset values for each set of measurements (each set meaning, a set of
current and voltage measurements at a specific rotor position) for current and voltage
separately. Once the folder containing all captured waveforms is loaded the GUI
calculates the flux-linkage characteristics for each rotor position and plots the results in
the axis provided within the GUI window. Users can print the contents of GUI window by
pressing the Print pushbutton at the right hand corner of the GUI window. By pressing
331

DATA
the XLS button users can also save the calculated flux-linkage data in excel spreadsheet
format. The GUI not only computes the flux-linkage for the filtered voltage and current
waveforms but also for the unfiltered (raw) voltage and current waveforms as well. Both
flux-linkage results for each rotor position are plotted in the axis provided in the GUI
window. In Figure A. 2 the effects of unfiltered voltage and current waveforms can
clearly be seen. If the user wants to see the effect of phase resistance value or wants to use
a different filter depth, he can do so by changing the corresponding values under the
VARIABLES panel of the GUI window and pressing the RECALCULATE
pushbutton on the left hand side of the GUI window just above the flux-linkage plots.
Figure A. 2 Effects of undesirable voltage and current waveforms on the calculated flux-linkage
characteristics
332
Appendix B LAMINATION DRAWINGS OF SIX PHASE SEGMENTAL AND

CONVENTIONAL TOOTHED ROTOR MACHINES
APPENDIX B
B. Lamination Drawings
of the Six Phase
Segmental and
Conventional Rotor
Machines
This appendix gives mechanical details of the six-phase 12-10 switched reluctance
machine prototypes with conventional toothed and segmental rotor structures. The
information here is intended to compliment the details of the design stages for both
prototype machines in Chapter 4 Machine Design.
Note that the drawings of both prototypes in the following sections are not to scale. The
drawings were generated using AutoCad technical drawing software package. The first
two sections below detail the lamination drawings for the stator and the rotor of the
segmental rotor machine and the last two sections detail those for the stator and rotor of
the conventional toothed rotor machine.
333

B.1 STATOR SEGMENT LAMINATION DRAWING

OF THE SEGMENTAL ROTOR SIX-PHASE
MACHINE
Figure B. 1 Stator segment lamination of the segmental rotor six-phase switched reluctance machine
334

B.2 ROTOR SEGMENT LAMINATION DRAWING

OF THE SEGMENTAL ROTOR SIX-PHASE
MACHINE
Figure B. 2 Rotor segment lamination of the segmental rotor six-phase switched reluctance machine
335

B.3 STATOR SEGMENT LAMINATION DRAWING

OF THE TOOTHED ROTOR SIX-PHASE MACHINE
Figure B. 3 Stator segment lamination of the toothed rotor six-phase switched reluctance machine
336

B.4 ROTOR LAMINATION DRAWING OF THE

TOOTHED ROTOR SIX-PHASE MACHINE
Figure B. 4 Rotor lamination of the toothed rotor six-phase switched reluctance machine
337
Appendix C SIMULATION PACKAGE & GRAPHICAL USER INTERFACE
APPENDIX C
C. Simulation Package
and Graphical User
Interface
C.1 INTRODUCTION
This appendix details the Graphical User Interface (GUI) created for the simulation
package (details of which was presented in Chapter 6- Simulation Models). The
motivation to create a GUI for the simulation was to simplify the entry points to the
simulation and encapsulate simulations functional behaviour. Effectively, interfacing the
simulation models generated in the Simulink environment with a GUI makes it easier
for end-users to change key parameters. Having a GUI also helps to minimise the users
efforts to run simulations and view results.
The GUI for the simulation package presented in this thesis was created in Matlab
environment. GUIs in Matlab are created using a tool called Guide, Java-based GUI
development environment. This tool allows the programmer to layout the GUI, to select
and align the components to be placed in the GUI. Once the components are in place, the
programmer can then edit their properties such as name, colour, size, font, text to display,
etc. When Guide saves the GUI, it creates a working program (in m-file format) including
skeleton functions that the programmer can modify to implement the behaviour of the
GUI. A screenshot to show how Guide in Matlab looks like along with some of its key
parts is depicted in Figure C. 1.
338
Figure C. 1 A screenshot of GUIDE in Matlab environment for creating GUIs
The three principal elements required to create a Matlab GUI are:
1. Components: Each item on a Matlab GUI is a graphical component. The types

of components include graphical controls (pushbuttons, edit boxes, lists, sliders,
etc.), static elements (frames and text strings), menus, and axes.
2. Figures: The components of a GUI must be arranged within a figure, which is a

window on the computer screen.
3. Callbacks: A mouse click or a key press is an event, and the Matlab program
must respond to each event if the program is to perform its function. The code
executed in response to an event is known as a callback. There must be a callback
to implement the function of each graphical component on the GUI. Once a
callback has been assigned to an object, the control flow of the program will be
passed to the assigned callback. If a button has been assigned the callback named
close, then once that button is pressed, the code under close will be executed.
339

As a rule of thumb, if a user is able to use an interface without intervention and
understand what happens behind the GUI and is able to interpret the results then the GUI
is said to have accomplished its goal.
C.2 SIMULATION DATABASE

GUI created for this project is conveniently named as SRMSim and it is flexible
enough to accommodate new functionalities if required in the future. GUI used here has a
consistent appearance and armed with intuitive controls like pushbuttons and edit boxes,
and, behaves in an understandable and predictable manner. Data flow between the
functions of the main simulation GUI, the main simulation GUI window and the postprocessing GUI windows like the Results GUI and the Flux-Linkage Locus GUI is
transferred through the use of handles structure. In Matlab, handle is the unique identifier
that gives direct access to each instance of any graphics objects properties. By providing
access to each graphics object, handles enable the designer/user to access, explore, and
modify the characteristics that define said objects. These characteristics are called
properties. A graphics objects properties control many aspects of its appearance and
behaviour. Properties include general information such as the objects type, its parent and
children, and whether it is visible, as well as information unique to the particular class of
object.
The variables for the Simulink models need to be identified before any of the simulation
models can be run. This is done via the GUI. GUI passes the simulation parameters
(defined as variable names rather than actual numbers) into the model workspace rather
than the base workspace. One can have access to the model workspace and check the
variables and make sure the desired values are being used by the simulation. On start-up
the main GUI window is displayed as in Figure C. 2.
340
Figure C. 2 SRMSim Database Main GUI Window Initial State
The main GUI window is divided into several sub-sections. These are:
1. SIMULATION INPUTS: This is where information relating to the rotor speed,

current demand, advance and conduction angles and dc-link voltage are entered. Here
the user can also choose the number of electrical cycles the simulation is required to
run for. Depending on the speed of the machine and the number of electrical cycles in
the simulation the total length of each simulation run will vary.
2. PWM / PID: This is where parameters relating to the PID and PWM controllers are
entered.
3. MOTOR: This is where basic information for the motor being simulated is displayed.
This is more of an informative section. Manually changing values in this section will
not change the parameters used in the simulation model.
341

4. SIMULINK MODEL: This is an informative section where the name of the simulink
model being simulated is displayed. Again manually changing this will affect the
simulation results.
5. MOTOR DATA: This is again an informative section where the name of the excel
spreadsheet for the motor parameters such as flux-linkage / current / position and
phase resistance is displayed.
The user can choose the simulation model and the motor parameters associated with that
model by clicking the File menu at the top left corner of the main GUI window. This is
shown in Figure C. 3.
Figure C. 3 SRMSim Database Main GUI Window File Menu List
342

Once the user selects a spreadsheet that stores the flux-linkage characteristics of a
particular motor, the GUI then calculates the torque / current / rotor position information.
Algorithms employed for generating the torque characteristics from the flux-linkage
characteristics are explained in Chapter 6. Both the flux-linkage (axis on the left) and the
torque (axis on the right) characteristics are displayed in 3D surface plot format within the
GUI window. The designer of the motor, the number of phases, rotor poles and the stator
teeth along with the phase resistance are displayed in the MOTOR sub-section. All of
these are depicted in Figure C. 4.
Figure C. 4 SRMSim Database Main GUI Window Once Motor Data Loaded
Once the simulink model is selected the GUI window displays the name of the model to
be simulated. This is shown in Figure C. 5.
343
Figure C. 5 SRMSim Database Main GUI Window Once Simulink Model Loaded
Once the simulation is completed, the RESULTS GUI is loaded. This is shown in Figure
C. 6. RESULTS GUI displays the results of the performance calculations carried out once
the simulation is completed such as total copper losses, inverter VA-rating, total inverter
losses, output torque and torque ripple, etc. In RESULTS GUI there are two more GUIs
available both of which can be loaded by clicking the pushbuttons; Vstar and
FlxLkgLOCUS. FlxLkgLocus GUI window displays the flux-linkage locus of the motor
during simulation for one electrical cycle. This is displayed in Figure C. 7. Vstar
pushbutton only becomes available if the motor is connected in star configuration. Once
the Vstar pushbutton is pressed the Vstar GUI is displayed. In this GUI the user can view
the simulated and filtered waveforms for the star point potential and the current flowing
into the star point to further validate the simulation. This is displayed in Figure C. 8.
344
Figure C. 6 SIM RESULTS GUI Window (Left-Hand Side = Statistics Calculation Results; First
Figure = Filtered Phase Voltage; Second Figure = Phase Current; Third Figure = Total Machine
Torque)
345
Figure C. 7 FlxLkgLocus GUI Window depicting Phase Flux-Linkage vs. Phase Current
346
Figure C. 8 STAR POINT VOLTAGE GUI Window
347
Appendix D ESTIMATION OF MACHINE SLOT AREA & PHASE

RESISTANCE
APPENDIX D
D. Estimation of Machine
Slot Area and Phase
Resistance
This appendix details the derivation of mathematical equations that were used to calculate
the slot area and consequently, estimate the phase resistance of the six phase 12-10
segmental rotor prototype switched reluctance machine. The information presented here
was used to generate some of the results and discussions detailed in Chapter 4 Machine
Design. This appendix has been divided into two parts. The first part details the
calculation of the slot area using machine design variables. The second part, on the other
hand, details the estimation of the machine phase resistance using the calculated slot area
from the first part.
D.1 CALCULATION OF MACHINE SLOT AREA

In Figure D. 1 a stator tooth segment of the six phase segmental rotor machine is outlined
by the vertex points S_V0, S_V1, S_V2, S_V3, S_V4, S_V5, S_V6, S_V7, S_V8, and
S_V9. It is a 30o section of the full machine.
Coordinates of each vertex point in Figure D. 1 can be calculated by using the machine
design variables. Machine design variables are explained graphically in Figure D. 2 and
Figure D. 3. In Figure D. 2, BCD is the z-axis insulation layer thickness (separating the
conductors from touching the inner surface of the stator tooth) and BCW is the x-axis
348

RESISTANCE
insulation layer thickness. Obvious from its definition, BCD will affect the total axial
length of the machine. The insulation layer was formed by wrapping slot liners around
each stator tooth. Table D. 1 lists the machine design variables depicted in Figure D. 2
and Figure D. 3. Coil gap angle in Table D. 1 (5 in Figure D. 3) defines the gap between
the two coils (of neighbouring phases) wrapped around two adjacent stator teeth. In
reality, coil gap is a linear dimension in [128]. However, during the modelling of the
machine in the FE package it was represented in mechanical degrees in order to ease the
generation of the machine in the FE environment.
Coordinates for each vertex point defining the area of a single stator tooth segment are
calculated by using the equations listed below (Eq.D. 1 to Eq.D. 20).
A3
A4
A2
R3
S_V7
S_V5
h2
S_V2
S_V6
S_V4
S_V3
S_V8
S_V1
h1
S_V9
A1
S_V0
A5
R1
1
R2
(0,0)
Figure D. 1 2D illustration of the stator segment area and the total slot area
349

RESISTANCE
Insulation
Layer
BCD
Phase
Conductors
Stator Tooth
BCW
Figure D. 2 Figure showing dimensions of insulation layer between the stator core and the phase
conductors (View from the top positive y-axis)
Table D. 1 Table of machine design variables
CBD
SD
BCH
BCW
BCD
TW
Lg
1
2
3
1
2
3
4
5
Ns
Nr
Stator core-back depth

Rotor segment depth
y-axis insulation layer thickness
x-axis insulation layer thickness
z-axis insulation layer thickness
Stator tooth width
Air-gap length
Stator outside diameter
Stator bore-diameter
Rotor outside diameter
Stator tooth tang angle
Rotor segment angle
Stator tooth tip arc angle
Rotor segment arc angle
Coil gap angle
Total number of stator teeth
Total number of rotor segments
350

RESISTANCE
Figure D. 3 Rectilinear representation of the segmented rotor machine design variables
S _ V 0( x ) =
S _ V 0( y ) =
S _ V 1( x) =
S _ V 1( y ) =
2
2
cos(90
2
2
sin(90
3
2
3
2
Eq.D. 1
Eq.D. 2
TW
2
Eq.D. 3
S _ V 0( y ) + (( S _ V 0( x)
tan(1 )
TW
))
2
S _ V 2( x) = S _ V 1( x)
S _ V 2( y ) = ((
S _ V 3( x) = (
1
2
1
2
CBD) 2 (
Eq.D. 4
Eq.D. 5
TW 2
) )
2
CBD) cos(90
Eq.D. 6
360
)
2 Ns
Eq.D. 7
351

RESISTANCE
S _ V 3( y ) = (
S _ V 4( x) = (
1
2
S _ V 4( y ) = (
1
2
CBD ) sin(90
360
)
2 Ns
Eq.D. 8
) cos(90
360
)
2 Ns
Eq.D. 9
) sin(90
360
)
2 Ns
Eq.D. 10
S _ V 5( x) = S _ V 4( x)
Eq.D. 11
S _ V 5( y ) = S _ V 4( y )
Eq.D. 12
S _ V 6( x) = S _ V 3( x)
Eq.D. 13
S _ V 6( y ) = S _ V 3( y )
Eq.D. 14
S _ V 7 ( x ) = S _ V 2( x )
Eq.D. 15
S _ V 7 ( y ) = S _ V 2( y )
Eq.D. 16
S _ V 8( x) = S _ V 1( x)
Eq.D. 17
S _ V 8( y ) = S _ V 1( y )
Eq.D. 18
S _ V 9( x ) = S _ V 0( x )
Eq.D. 19
S _ V 9( y ) = S _ V 0( y )
Eq.D. 20
The total slot area can now be calculated by using the equations listed below (Eq.D. 21 to
Eq.D. 32).
h 2 = S _ V 2( y ) S _ V 1( y )
Eq.D. 21
where h 2 is the stator body height and, S _ V 2( y ) and S _ V 1( y ) are the stator vertex
coordinates given in Eq.D. 6 and Eq.D. 4, respectively. Using Eq.D. 21, the stator body
area can be calculated by Eq.D. 22.
352

RESISTANCE
Abody = h 2 TW
Eq.D. 22
where Abody is the stator body area enclosed by the vertex points S_V1, S_V2, S_V7, and
S_V8.
h1 = S _ V 1( y ) S _ V 0( y )
Eq.D. 23
where h1 is the stator tooth tip height and, S _ V 1( y ) and S _ V 0( y ) are the stator vertex
coordinates given in Eq.D. 4 and Eq.D. 2 , respectively.
A5 = 0.5 (
2
2
)2 (
3
180
sin( 3 ))
Eq.D. 24
where A5 is the stator tooth tip segment area as shown in Figure D. 1. 2 (R2 in Figure D.
1) and 3 (2 in Figure D. 1) are illustrated in Figure D. 3.
Atrap = 0.5 h1 (
TW
+ ( S _ V 0( x) S _ V 9( x)))
2
Eq.D. 25
where Atrap is the area of the stator tooth tip trapezoid enclosed by the vertex points S_V0,
S_V1, S_V8, and S_V9 in Figure D. 1. h1 is the stator tooth tip height shown in Figure D.
1. Definitions for S _ V 0( x) and S _ V 9( x) are given in Eq.D. 1 and Eq.D. 19,
respectively.
A4 =
Ns
((
1
2
)2 (
1
2
CBD) 2 )
Eq.D. 26
where A4 is the area of the stator core-back segment as shown in Figure D. 1.
A3 = 0.5 (
1
2
CBD ) 2 (
3
180
sin( 3 ))
Eq.D. 27
where A3 is the area of the stator core-back sector as shown in Figure D. 1. 3 is the
core-back sector area angle depicted in Figure D. 1 and is calculated by Eq.D. 28.
353

RESISTANCE
3 = 2 arctan(
S _ V 2( x )
)
S _ V 2( y )
Eq.D. 28
The total stator tooth area can be calculated by Eq.D. 29.

ATstator = A3 + A4 + Abody + Atrap A5
Eq.D. 29
where ATStator is the total stator tooth area. The area enclosed by points (0,0), S_V4 and
S_V5 in Figure D. 1 is called the total segment area and is calculated by Eq.D. 30.
ATSegment =
( 1 )2
Ns 2
Eq.D. 30
The area enclosed by points (0,0), S_V0 and S_V9 in Figure D. 1 is called the inner
segment area and is calculated by Eq.D. 31.
AISegment =
Ns
2
2
)2
Eq.D. 31
Finally the total slot area is calculated by Eq.D. 32.

ATSlot = A1 + A2 = ATSegment AISegment ATStator
Eq.D. 32
Areas A1 and A2 are both illustrated in Figure D. 1. If the total number of turns per coil
and the total conductor diameter (including conductor insulation layer) are known, the
slot fill factor can be calculated by Eq.D. 33.
D
2 N turns cond
4
SFF =
ATSlot
Eq.D. 33
where SFF is the slot fill factor, N turns is the number of turns per coil, Dcond is the total
conductor diameter (including insulation layer), and ATSlot is the total slot area calculated
by Eq.D. 32. Estimation of phase resistance is given in the following section.
354

RESISTANCE
D.2 ESTIMATION OF MACHINE PHASE

RESISTANCE
Coil resistance can be calculated by Eq.D. 34.
Rcoil =
Ltot
Acond
Eq.D. 34
where Rcoil is the coil resistance in [], is the electrical resistivity in [.m], Ltot is the
total length of the conductors forming the coil in [m], and Acond is the bare conductor
cross-sectional area in [m2]. Bare conductor cross-sectional area is given by Eq.D. 35.
Acond =
( Dtotal 2 Linsulation ) 2
4
Eq.D. 35
where Dtotal is the total conductor diameter and Linsulation is the conductor insulation layer
thickness. Apparent from Eq.D. 34, an accurate estimate for the total length of the
conductors forming the coil is required for an accurate estimation of coil resistance.
The total length of conductors (forming a coil) can be estimated by Eq.D. 36.
Ltot = N turns MPL
Eq.D. 36
where N turns is the total number of turns per coil and MPL is the mean path length of a
single coil turn.
The mean path length of a single coil turn can be estimated by Eq.D. 37.
MPL = 2 ( Lstk + Lendwinding )
Eq.D. 37
where Lstk is the total lamination stack length and Lendwinding is the end-winding length. To
ease the calculation process the winding area is assumed to be a square as in Figure D. 4-a
where in fact it is more of the form given in Figure D. 4-b.
355

RESISTANCE
Figure D. 4 Rectilinear view of the winding area looking from the top (towards y axis) and the front
(towards z axis) of the machine showing vertex points that outline the winding area
The end-winding length can be estimated by Eq.D. 38.

Lendwinding = TW + C
Eq.D. 38
where TW is the stator tooth width and C is the length of the arc, both depicted in Figure
D. 5. The length of the arc C can be approximated by a straight line (Z in Figure D. 5)
and a correction factor as in Eq.D. 39.
C = kZ = k
(X
+Y 2
Eq.D. 39
where k is the correction factor, and X and Y are the straight lines shown in Figure D. 5.
Correction factor k in Eq.D. 39 is used to account for the linearization of the arc length C
into the straight line Z (Figure D. 5).
356

RESISTANCE
C
Y
Z
X
BCD
Stator Tooth Width

Winding Width
BCW
Figure D. 5 Graphical explanation of end-winding length estimation in rectilinear form (Top view of
machine stator tooth and coil windings)
To be able to calculate X and Y , the coordinates of vertex points defining the winding
area need to be defined. Figure D. 4-b is the graphical representation of the winding area
vertex points showing V_W1, V_W2, V_W3 and V_W4. The coordinates of these vertex
points can be calculated by using the following equations (Eq.D. 40 to Eq.D. 47). Note
that equations from Eq.D. 40 to Eq.D. 47 are not for the rectangular winding area model
given in Figure D. 4-b but they are for the actual machine geometry.
357

RESISTANCE
W _ V 1( x) =
TW
+ BCW
2
W _ V 1( y ) = ((
W _ V 2( x ) =
1
2
CBD ) 2 (
Eq.D. 40
TW 2
) )
2
Eq.D. 41
TW
+ BCW
2
Eq.D. 42
TW
W _ V 2( y ) = 1 CBD BCH
+ BCW
2
2
2
W _ V 3( x) = (
W _ V 3( y ) = (
W _ V 4( x ) =
1
2
1
2
CBD BCH ) cos(90
CBD BCH ) sin(90
360
+ 5 )
Ns 2
360
+ 5 )
Ns 2
W _ V 1( y )
360
tan(90
+ 5 )
Ns 2
Eq.D. 43
Eq.D. 44
Eq.D. 45
Eq.D. 46
W _ V 4( y ) = W _ V 1( y )
Eq.D. 47
X and Y in Eq.D. 39 can now be calculated by using Eq.D. 48 and Eq.D. 49, respectively.
X =
W _ V 3( x) W _ V 2( x)
+ BCD
2
Eq.D. 48
W _ V 3( x) W _ V 2( x)
+ BCW
2
Eq.D. 49
and
Y=
The correction factor k in Eq.D. 39 is set to 1.05. Equations from Eq.D. 38 to Eq.D. 49
assume that the windings cover a rectangular area (A1 + A2 in Figure D. 4-b). However,
in reality, windings cover the area A1 in Figure D. 4-b. Therefore once the coil resistance
358

RESISTANCE
is calculated a correction factor must be added to account for this fact. This is achieved by
using Eq.D. 50.
R _ Coil = Area _ Factor R _ Coil Re ct
Eq.D. 50
where R _ Coil is the actual coil resistance, R _ Coil Re ct is the coil resistance calculated
by using equations from Eq.D. 38 to Eq.D. 49 and Area _ Factor is the ratio of the areas
A1 and A2 in Figure D. 4-b. Area _ Factor is calculated by Eq.D. 51.
Area _ Factor =
A1
A1 + A2
Eq.D. 51
Once the machine dimensions and the bare conductor diameter are known the coil
resistance can be calculated. Depending on the coil connections (either parallel or series)
estimation for the phase resistance can be made. The following steps explain how the
conductor diameter has been determined for the prototype 6-phase segmented rotor
machine.
Peak current density of 10.0 [A/mm2] and peak phase current of 25.0 [A] are assumed.
From Eq.D. 52 (any one phase of the machine phase is assumed to be conducting for one
third of the electrical cycle) 25.0 [A] of peak phase current equates to approximately 14.4
[A] of RMS phase current.
I phPEAK
I phRMS =
Eq.D. 52
The required conductor cross sectional area ( Acond ) can then be calculated by Eq.D. 53
Acond =
I phRMS
Eq.D. 53
J PEAK
where J PEAK is the maximum current density. Acond is calculated to be approximately

1.443 [mm2] which equates to a conductor diameter of approximately 1.36[mm]. The
closest available conductor diameter was 1.4 [mm] so this was chosen as the conductor
diameter.
359
LIST OF REFERENCES
List of References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
Cardenas, R.; Pena, R.; Perez, M.; Clare, J.; Asher, G.; Wheeler, P.;, "Control of a
switched reluctance generator for variable-speed wind energy applications,"
Energy Conversion, IEEE Transactions on, vol. 20, pp. 781-791, 2005.
Cardenas, R.; Ray, W.F.; Asher, G.M.;, "Switched reluctance generators for wind
energy applications," in Power Electronics Specialists Conference, 1995. PESC
'95 Record., 26th Annual IEEE, 1995, pp. 559-564 vol.1.
Mese, E.; Sozer, Y.; Kokernak, J.M.; Torrey, D.A.;, "Optimal excitation of a high
speed switched reluctance generator," in Applied Power Electronics Conference
and Exposition, 2000. APEC 2000. Fifteenth Annual IEEE, 2000, pp. 362-368
vol.1.
M. A. Mueller, "Design of low speed switched reluctance machines for wind
energy converters," in Electrical Machines and Drives, 1999. Ninth International
Conference on (Conf. Publ. No. 468), 1999, pp. 60-64.
"Offshore Wind Power Market Potential," Energy Business Reports, April 2008.
"U.S., China Lead Global Wind Installation," North American Wind Power,
February 2009.
Powell, D.J.; Jewell, G.W.; Calverley, S.D.; Howe, D.;, "Iron loss in a modular
rotor switched reluctance machine for the "More-Electric" aero-engine,"
Magnetics, IEEE Transactions on, vol. 41, pp. 3934-3936, 2005.
Hall, R.; Jack, A.G.; Mecrow, B.C.; Mitcham, A.J.;, "Design and initial testing of
an outer rotating segmented rotor switched reluctance machine for an aero-engine
shaft-line-embedded starter/generator," in Electric Machines and Drives, 2005
IEEE International Conference on, 2005, pp. 1870-1877.
S. R. MacMinn and W. D. Jones, "A very high speed switched-reluctance startergenerator for aircraft engine applications," in Aerospace and Electronics
Conference, 1989. NAECON 1989., Proceedings of the IEEE 1989 National,
1989, pp. 1758-1764 vol.4.
E. Richter, "High Temperature, Lightweight, Switched Reluctance Motors and
Generators for Future Aircraft Engine Applications," in American Control
Conference, 1988, 1988, pp. 1846-1851.
S. Shoujun, L.Weiguo, D. Peitsch, U. Schaefer;, "Detailed Design of a High Speed
Switched Reluctance Starter/Generator for More/All Electric Aircraft," Chinese
Journal of Aeronautics, vol. 23, pp. 216-226, 2010.
Mecrow, B.C.; Finch, J.W.; El-Kharashi, E.A.; Jack, A.G.;, "Switched reluctance
motors with segmental rotors," Electric Power Applications, IEE Proceedings -,
vol. 149, pp. 245-254, 2002.
A. C. Clothier and B. C. Mecrow, "Inverter topologies and current sensing
methods for short pitched and fully pitched winding SR motors," in Applied
Power Electronics Conference and Exposition, 1999. APEC '99. Fourteenth
Annual, 1999, pp. 416-423 vol.1.
Mecrow, B.C.; Jack, A.G.; Atkinson, D.J.; Green, S.R.; Atkinson, G.J.; King, A.;
Green, B.;, "Design and testing of a four-phase fault-tolerant permanent-magnet
machine for an engine fuel pump," Energy Conversion, IEEE Transactions on,
vol. 19, pp. 671-678, 2004.
360
LIST OF REFERENCES
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
D. Kennedy, "Rare Earth Permanent Magnet Raw Materials Supply," magnews The international publication of the UK Magnetics Society, UKMAG one-day
seminar Cutting Costs by Optimised Machine Design, April 2009, 2009.
B. D. Bedford, ""Compatible Permanent Magnet or Reluctance Brushless Motors
and Controlled Switch Circuits", US Patent No 3678352, 18 July 1972.."
B. D. Bedford, ""Compatible Brushless Reluctance Motors and Controlled Switch
Circuits", US Patent No 3679953, 25 July 1972.."
Lawrenson, P.J.; Stephenson, J.M.; Fulton, N.N.; Blenkinsop, P.T.; Corda, J.;,
"Variable-speed switched reluctance motors," Electric Power Applications, IEE
Proceedings B, vol. 127, pp. 253-265, 1980.
L. J. G. Byrn J. V., "Characteristics of saturable stepper and reluctance motors," in
Small Electrical Machines, 1976 IEEE International Conference on., 1976.
W. F. Ray and R. M. Davis, "Inverter Drive for Doubly Salient Reluctance Motor:
Its Fundamental Behaviour, Linear Analysis and Cost Implications," Electric
Power Applications, IEE Journal on, vol. 2, pp. 185-193, 1979.
R. T. Fingers and C. S. Rubertus, "Application of high temperature magnetic
materials," Magnetics, IEEE Transactions on, vol. 36, pp. 3373-3375, 2000.
Mecrow, B.; Bennett, J.; Jack, A.; Atkinson, D.; Freeman, A.;, "Very high
efficiency drives for solar powered unmanned aircraft," in Electrical Machines,
2008. ICEM 2008. 18th International Conference on, 2008, pp. 1-6.
Mecrow, B.C.; El-Kharashi, E.A.; Finch, J.W.; Jack, A.G.;, "Segmental rotor
switched reluctance motors with single-tooth windings," Electric Power
Applications, IEE Proceedings -, vol. 150, pp. 591-599, 2003.
T. J. E. Miller, "Electronic Control of Switched Reluctance Machines", Newnes
Press, Oxford, 2001.
Yong Kwon Choi; Hee Sung Yoon; Chang Seop Koh;, "Pole-Shape Optimization
of a Switched-Reluctance Motor for Torque Ripple Reduction," Magnetics, IEEE
Transactions on, vol. 43, pp. 1797-1800, 2007.
Johnson, J.P.; Rajarathnam, A.V.; Toliyat, H.A.; Gopalakrishnan, S.; Fahimi, B.;,
"Torque optimization for a SRM using winding function theory with a gapdividing surface," in Industry Applications Conference, 1996. Thirty-First IAS
Annual Meeting, IAS '96., Conference Record of the 1996 IEEE, 1996, pp. 753760 vol.2.
T. A. Lipo, "Novel reluctance machine concepts for variable speed drives," in
Electrotechnical Conference, 1991. Proceedings., 6th Mediterranean, 1991, pp.
34-43 vol.1.
Van der Giet, M.; Lange, E.; Correa, D.A.P.; Chabu, I.E.; Nabeta, S.I.; Hameyer,
K.;, "Acoustic Simulation of a Special Switched Reluctance Drive by Means of
Field-Circuit Coupling and Multiphysics Simulation," Industrial Electronics,
IEEE Transactions on, vol. 57, pp. 2946-2953, 2010.
Pengov, W.; Hendershot, J.R.; Miller, T.J.E.;, "A new low-noise two-phase
switched reluctance motor," in Electric Machines and Drives, 2005 IEEE
International Conference on, 2005, pp. 1281-1284.
Moallem, M.; Ong, C.-M.; Unnewehr, L.E.;, "Effect of rotor profiles on the torque
of a switched-reluctance motor," Industry Applications, IEEE Transactions on,
vol. 28, pp. 364-369, 1992.
D. P. Tormey and D. A. Torrey, "A comprehensive design procedure for low
torque-ripple variable-reluctance motor drives," in Industry Applications Society
Annual Meeting, 1991., Conference Record of the 1991 IEEE, 1991, pp. 244-251
vol.1.
361
LIST OF REFERENCES
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
Ozoglu, Y.; Garip, M.; Mese, E.;, "New pole tip shapes mitigating torque ripple in
short pitched and fully pitched switched reluctance motors," in Industry
Applications Conference, 2002. 37th IAS Annual Meeting. Conference Record of
the, 2002, pp. 43-50 vol.1.
R. T. Naayagi and V. Kamaraj, "Optimum Pole Arcs for Switched Reluctance
Machine with Reduced Ripple," in Power Electronics and Drives Systems, 2005.
PEDS 2005. International Conference on, 2005, pp. 761-764.
N. K. Sheth and K. R. Rajagopal, "Optimum pole arcs for a switched reluctance
motor for higher torque with reduced ripple," Magnetics, IEEE Transactions on,
vol. 39, pp. 3214-3216, 2003.
Bhiwapurkar, N.; Jain, A.K.; Mohan, N.;, "Study of new stator pole geometry for
improvement of SRM torque profile," in Electric Machines and Drives, 2005
IEEE International Conference on, 2005, pp. 516-520.
Krishnan, R.; Arumugan, R.; Lindsay, J.F.;, "Design procedure for switchedreluctance motors," Industry Applications, IEEE Transactions on, vol. 24, pp.
456-461, 1988.
Anwar, M.N.; Husain, I.; Radun, A.V.;, "A comprehensive design methodology
for switched reluctance machines," Industry Applications, IEEE Transactions on,
vol. 37, pp. 1684-1692, 2001.
J. Faiz and J. W. Finch, "Aspects of design optimisation for switched reluctance
motors," Energy Conversion, IEEE Transactions on, vol. 8, pp. 704-713, 1993.
Finch, J.W.; Faiz, J.; Metwally, H.M.B.;, "Design study of switched reluctance
motor performance," in Industry Applications Society Annual Meeting, 1992.,
Conference Record of the 1992 IEEE, 1992, pp. 242-248 vol.1.
T. J. E. Miller, "Optimal design of switched reluctance motors," Industrial
Electronics, IEEE Transactions on, vol. 49, pp. 15-27, 2002.
A. V. Radun, "Design considerations for the switched reluctance motor," Industry
Applications, IEEE Transactions on, vol. 31, pp. 1079-1087, 1995.
Vijayakumar, K.; Karthikeyan, R.; Paramasivam, S.; Arumugam, R.; Srinivas,
K.N.;, "Switched Reluctance Motor Modeling, Design, Simulation, and Analysis:
A Comprehensive Review," Magnetics, IEEE Transactions on, vol. 44, pp. 46054617, 2008.
R. M. Davis, "A comparison of switched reluctance rotor structures," Industrial
R. M. Davis, "Variable reluctance rotor structures-their influence on torque
production," Industrial Electronics, IEEE Transactions on, vol. 39, pp. 168-174,
1992.
Neagoe, C.; Foggia, A.; Krishnan, R.;, "Impact of pole tapering on the
electromagnetic torque of the switched reluctance motor," in Electric Machines
and Drives Conference Record, 1997. IEEE International, 1997, pp. WA1/2.1WA1/2.3.
A. Pelikant and S. Wiak, "The influence of the rotor pole shape on the static and
dynamic characteristics of the switched reluctance motor," Magnetics, IEEE
N. K. Sheth and K. R. Rajagopal, "Torque profiles of a switched reluctance motor
having special pole face shapes and asymmetric stator poles," Magnetics, IEEE
H. C. Lovatt and J. M. Stephenson, "Influence of number of poles per phase in
switched reluctance motors," Electric Power Applications, IEE Proceedings B,
vol. 139, pp. 307-314, 1992.
362
LIST OF REFERENCES
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
T. J. E. Miller, "Switched Reluctance Motors and Their Controls", Clarendon

Press, Oxford, 1993.
B. C. Mecrow, "Fully pitched-winding switched-reluctance and stepping-motor
arrangements," Electric Power Applications, IEE Proceedings B, vol. 140, pp. 6170, 1993.
Feng Liang; Yuefeng Liao; Lipo, T.A.;, "A new variable reluctance motor
utilizing an auxiliary commutation winding," in Industry Applications Society
Annual Meeting, 1992., Conference Record of the 1992 IEEE, 1992, pp. 219-225
vol.1.
Arumugam, R.; Lowther, D.; Krishnan, R.; Lindsay, J.;, "Magnetic field analysis
of a switched reluctance motor using a two dimensional finite element model,"
V. Kamaraj and C. Aravind Vaithilingam, "Modelling and simulation of switched
reluctance machine (SRM) using MAGNET6.0," in Power Electronics and Drive
Systems, 2003. PEDS 2003. The Fifth International Conference on, 2003, pp. 480484 Vol.1.
Koibuchi, K.; Ohno, T.; Sawa, K.;, "A basic study for optimal design of switched
reluctance motor by finite element method," Magnetics, IEEE Transactions on,
vol. 33, pp. 2077-2080, 1997.
Murthy, S.S.; Singh, B.; Sharma, V.K.;, "Finite element analysis to achieve
optimum geometry of switched reluctance motor," in TENCON '98. 1998 IEEE
Region 10 International Conference on Global Connectivity in Energy, Computer,
Communication and Control, 1998, pp. 414-418 vol.2.
Ohdachi, Y.; Kawase, Y.; Miura, Y.; Hayashi, Y.;, "Optimum design of switched
reluctance motors using dynamic finite element analysis," Magnetics, IEEE
Wei Wu; Dunlop, J.B.; Collocott, S.J.; Kalan, B.A.;, "Design optimization of a
switched reluctance motor by electromagnetic and thermal finite-element
analysis," Magnetics, IEEE Transactions on, vol. 39, pp. 3334-3336, 2003.
A. M. Michaelides and C. Pollock, "Effect of end core flux on the performance of
the switched reluctance motor," Electric Power Applications, IEE Proceedings -,
vol. 141, pp. 308-316, 1994.
Fahimi, B.; Suresh, G.; Ehsani, M.;, "Large switched reluctance machines: a 1
MW case study," in Electric Machines and Drives, 1999. International
Conference IEMD '99, 1999, pp. 84-86.
Lobo, N.S.; Hong Sun Lim; Krishnan, R.;, "Comparison of Linear Switched
Reluctance Machines for Vertical Propulsion Application: Analysis, Design, and
Experimental Correlation," Industry Applications, IEEE Transactions on, vol. 44,
pp. 1134-1142, 2008.
M. F. Momen and I. Husain, "Design and performance analysis of a switched
reluctance motor for low duty cycle operation," Industry Applications, IEEE
K. M. Rahman and S. E. Schulz, "Design of high-efficiency and high-torquedensity switched reluctance motor for vehicle propulsion," Industry Applications,
S. S. Ramamurthy and J. C. Balda, "Sizing a switched reluctance motor for
electric vehicles," Industry Applications, IEEE Transactions on, vol. 37, pp. 12561264, 2001.
363
LIST OF REFERENCES
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
Schofield, N.; Long, S.A.; Howe, D.; McClelland, M.;, "Design of a Switched
Reluctance Machine for Extended Speed Operation," Industry Applications, IEEE
J. M. Stephenson and G. C. Jenkinson, "Single-phase switched reluctance motor
design," Electric Power Applications, IEE Proceedings -, vol. 147, pp. 131-139,
2000.
J. R. Suriano and O. Chee-Mun, "Variable reluctance motor structures for lowspeed operation," Industry Applications, IEEE Transactions on, vol. 32, pp. 345353, 1996.
T. Suzuki, Ito, S., Tanaka, N., Chiba, A., Fukao, T. and Ninomiya, H.,
"Development of high-efficiency switched reluctance motor," Electrical
Engineering in Japan, vol. 162, pp. 7382, 2008.
K. S. Chai and C. Pollock, "Evolutionary computer controlled design of a
reluctance motor drive system," in Industry Applications Conference, 2003. 38th
IAS Annual Meeting. Conference Record of the, 2003, pp. 1480-1487 vol.3.
Kosaka, T.; Pollock, C.; Shikayama, T.; Nakagami, T.; Kano, Y.; Matsui, N.;,
"GA-based computer aided autonomous electromagnetic design of switched
reluctance servomotor drives," in Industry Applications Conference, 2004. 39th
IAS Annual Meeting. Conference Record of the 2004 IEEE, 2004, p. 423 Vol.1.
Mirzaeian, B.; Moallem, M.; Tahani, V.; Lucas, C.;, "Multiobjective optimization
method based on a genetic algorithm for switched reluctance motor design,"
N. R. Matveev A., Undeland T., "Design optimization of switched reluctance
drives using artificial neural networks," in Power Electronics and Motion Control,
2002. EPE-PEMC 2002. 10th International Conference on., Cavtat & Dubrovnik,
2002.
S. Brisset and P. Brochet, "Optimization of switched reluctance motors using
deterministic methods with static and dynamic finite element simulations,"
J. R. Hendershot. (1989, Short flux loops cool SR motors. Machine Design, pp.
106-111.
A. Michaelides and C. Pollock, "Design and performance of a high efficiency 5phase switched reluctance motor," in Electrical Machines and Drives, 1995.
Seventh International Conference on (Conf. Publ. No. 412), 1995, pp. 266-270.
A. M. Michaelides and C. Pollock, "Modelling and design of switched reluctance
motors with two phases simultaneously excited," Electric Power Applications,
IEE Proceedings -, vol. 143, pp. 361-370, 1996.
Lobo, N.S.; Swint, E.; Krishnan, R.;, "M-Phase N-Segment Flux-Reversal-Free
Stator Switched Reluctance Machines," in Industry Applications Society Annual
Meeting, 2008. IAS '08. IEEE, 2008, pp. 1-8.
O. Seok-Gyu and R. Krishnan, "Two-Phase SRM With Flux-Reversal-Free Stator:
Concept, Analysis, Design, and Experimental Verification," Industry Applications,
W. Pengov, ""Staggered Pole Switched Reluctance Motor", US Patent No
6060809, 9 May 2000.."
Cheewoo Lee; Krishnan, R.; Lobo, N.S.;, "Novel Two-Phase Switched Reluctance
Machine Using Common-Pole E-Core Structure: Concept, Analysis, and
Experimental Verification," Industry Applications, IEEE Transactions on, vol. 45,
pp. 703-711, 2009.
364
LIST OF REFERENCES
[80]
[81]
[82]
[83]
[84]
[85]
[86]
[87]
[88]
[89]
[90]
[91]
[92]
[93]
[94]
[95]
C. Pollock and M. Wallace, "The flux switching motor, a DC motor without

magnets or brushes," in Industry Applications Conference, 1999. Thirty-Fourth
IAS Annual Meeting. Conference Record of the 1999 IEEE, 1999, pp. 1980-1987
vol.3.
Pollock, C.; Pollock, H.; Barron, R.; Coles, J.R.; Moule, D.; Court, A.; Sutton, R.;,
"Flux-Switching Motors for Automotive Applications," Industry Applications,
J. D. Wale and C. Pollock, "Novel converter topologies for a two-phase switched
reluctance motor with fully pitched windings," in Power Electronics Specialists
Conference, 1996. PESC '96 Record., 27th Annual IEEE, 1996, pp. 1798-1803
vol.2.
Corda, J.; Tataru, A.M.; Rasmussen, P.O.; Ritchie, E.;, "Analytical estimation of
torque enhancement of the SR machine with saw-shaped (shark) pole surfaces,"
Electric Power Applications, IEE Proceedings -, vol. 151, pp. 223-229, 2004.
A. M. T. Kjaer, ""Shark, New Motor Design Concept for Energy Saving Applied
to Switched Reluctance Motor", Thesis submitted for Doctor of Philosophy,"
Institute of Energy Technology, 2004., Aalborg University.
Rasmussen, P.O.; Blaabjerg, F.; Pedersen, J.K.; Jensen, F.;, "Switched reluctanceshark machines-more torque and less acoustic noise," in Industry Applications
Conference, 2000. Conference Record of the 2000 IEEE, 2000, pp. 93-98 vol.1.
Faiz, J.; Harris, M.R.; Finch, J.W.;, "Two-dimensional finite element analysis of
multi-tooth per stator pole switched reluctance motor-four teeth per stator pole," in
Industry Applications Society Annual Meeting, 1990., Conference Record of the
1990 IEEE, 1990, pp. 234-240 vol.1.
J. R. Hendershot, ""A five phase switched reluctance brushless DC motor with a
low loss magnetic circuit"," ed: Magna Physics Corp., 1989., 1989.
Faiz, J.; Raddadi, J.; Finch, J.W.;, "Spice-based dynamic analysis of a switched
reluctance motor with multiple teeth per stator pole," Magnetics, IEEE
A. C. Clothier, ""Switched Reluctance Motor Drives with Fully Pitched
Windings", Thesis submitted for Doctor of Philosophy," School of Electrical,
Electronic and Computer Engineering, University of Newcastle upon Tyne, 2001.
P. G. Barrass, ""High Performance Switched Reluctance Drives", Thesis
submitted for Doctor of Philosophy," School of Electrical, Electronic and
Computer Engineering, University of Newcastle upon Tyne, 1995.
Mecrow, B.C.; Weiner, C.; Clothier, A.C.;, "The modeling of switched reluctance
machines with magnetically coupled windings," Industry Applications, IEEE
Li, S.-Y.H.; Feng Liang; Yifan Zhao; Lipo, T.A.;, "A doubly salient doubly
excited variable reluctance motor," Industry Applications, IEEE Transactions on,
vol. 31, pp. 99-106, 1995.
B. C. Mecrow, ""Switched Reluctance Motor", European Patent No
EP1359660A2, 5 November 2003."
Mecrow, B.C.; El-Kharashi, E.A.; Finch, J.W.; Jack, A.G.;, "Performance
evaluation of switched reluctance motors with segmental rotors," in Electric
Machines and Drives Conference, 2003. IEMDC'03. IEEE International, 2003,
pp. 568-574 vol.1.
Xu, L.; Lipo, T.A.; Rao, S.C.;, "Analysis of a new variable-speed singly salient
reluctance motor utilizing only two transistor switches," Industry Applications,
365
LIST OF REFERENCES
[96]
[97]
[98]
[99]
[100]
[101]
[102]
[103]
[104]
[105]
[106]
[107]
[108]
[109]
[110]
[111]
G. A. Horst, ""Isolated Segmental Switched Reluctance Motor", US Patent No

5111096, 5 May 1992.."
E. A. El-Kharashi, ""Segmented Rotor Switched Reluctance Motors", Thesis
submitted for Doctor of Philosophy," Thesis submitted for Doctor of Philosophy,
School of Electrical, Electronic and Computer Engineering, University of
Newcastle upon Tyne, 2003.
Oyama, J.; Higuchi, T.; Abe, T.; Tanaka, K.;, "The fundamental characteristics of
novel switched reluctance motor with segment core embedded in aluminum rotor
block," in Electrical Machines and Systems, 2005. ICEMS 2005. Proceedings of
the Eighth International Conference on, 2005, pp. 515-519 Vol. 1.
Vattikuti, N.; Rallabandi, V.; Fernandes, B.G.;, "A novel high torque and low
weight segmented switched reluctance motor," in Power Electronics Specialists
Conference, 2008. PESC 2008. IEEE, 2008, pp. 1223-1228.
J. C. Zolper, "Emerging silicon carbide power electronics components," in Applied
Power Electronics Conference and Exposition, 2005. APEC 2005. Twentieth
Annual IEEE, 2005, pp. 11-17 Vol. 1.
C. Pollock and W. Chi-Yao, "Acoustic noise cancellation techniques for switched
reluctance drives," Industry Applications, IEEE Transactions on, vol. 33, pp. 477484, 1997.
I. Husain, "Minimization of torque ripple in SRM drives," Industrial Electronics,
Sahoo, N.C.; Xu, J.X.; Panda, S.K.;, "Determination of current waveforms for
torque ripple minimisation in switched reluctance motors using iterative learning:
an investigation," Electric Power Applications, IEE Proceedings -, vol. 146, pp.
369-377, 1999.
Ilic-Spong, Marija; Miller, Timothy J. E.; Macminn, Stephen R.; Thorp, James S.;,
"Instantaneous Torque Control of Electric Motor Drives," Power Electronics,
IEEE Transactions on, vol. PE-2, pp. 55-61, 1987.
R. S. Wallace and D. G. Taylor, "Low-torque-ripple switched reluctance motors
for direct-drive robotics," Robotics and Automation, IEEE Transactions on, vol. 7,
pp. 733-742, 1991.
R. S. Wallace and D. G. Taylor, "A balanced commutator for switched reluctance
motors to reduce torque ripple," Power Electronics, IEEE Transactions on, vol. 7,
pp. 617-626, 1992.
Mir, S.; Elbuluk, M.E.; Husain, I.;, "Torque-ripple minimization in switched
reluctance motors using adaptive fuzzy control," Industry Applications, IEEE
M. S. Islam and J. Husain, "Torque-ripple minimization with indirect position and
speed sensing for switched reluctance motors," Industrial Electronics, IEEE
Russa, K.; Husain, I.; Elbuluk, M.E.;, "Torque-ripple minimization in switched
reluctance machines over a wide speed range," Industry Applications, IEEE
P. G. Barrass and B. C. Mecrow, "Flux and torque control of switched reluctance
machines," Electric Power Applications, IEE Proceedings -, vol. 145, pp. 519527, 1998.
Xue, X.D.; Cheng, K.W.E.; Ho, S.L.;, "A Control Scheme of Torque Ripple
Minimization for SRM Drives Based on Flux Linkage Controller and Torque
Sharing Function," in Power Electronics Systems and Applications, 2006. ICPESA
'06. 2nd International Conference on, 2006, pp. 79-84.
366
LIST OF REFERENCES
[112] Reay, D.S.; Mirkazemi-Moud, M.; Green, T.C.; Williams, B.W.;, "Switched
reluctance motor control via fuzzy adaptive systems," Control Systems Magazine,
IEEE, vol. 15, pp. 8-15, 1995.
[113] Rochford, C.; Kavanagh, R.C.; Egan, M.G.; Murphy, J.M.D.;, "Development of
smooth torque in switched reluctance motors using self-learning techniques," in
Power Electronics and Applications, 1993., Fifth European Conference on, 1993,
pp. 14-19 vol.6.
[114] L. Henriques, Branco, P. J. C., Rolim, L., Suemitsu, W., "Automatic Learning of
Pulse Current Shape for Torque Ripple Minimisation in Switched Reluctance
Machines," in European Control Conference, ECC'01, Porto, Portugal, 2001.,
2001, pp. pp. 232-237.
[115] J. C. Moreira, "Torque ripple minimization in switched reluctance motors via bicubic spline interpolation," in Power Electronics Specialists Conference, 1992.
PESC '92 Record., 23rd Annual IEEE, 1992, pp. 851-856 vol.2.
[116] Cailleux, H.; Le Pioufle, B.; Multon, B.; Sol, C.;, "A precise analysis of the phase
commutation for the torque nonlinear control of a switched reluctance motor torque ripples minimization," in Industrial Electronics, Control, and
Instrumentation, 1993. Proceedings of the IECON '93., International Conference
on, 1993, pp. 1985-1990 vol.3.
[117] Ilic'-Spong, M.; Marino, R.; Peresada, S.; Taylor, D.;, "Feedback linearizing
control of switched reluctance motors," Automatic Control, IEEE Transactions on,
vol. 32, pp. 371-379, 1987.
[118] H. C. Lovatt and J. M. Stephenson, "Computer-optimised smooth-torque current
waveforms for switched-reluctance motors," Electric Power Applications, IEE
Proceedings -, vol. 144, pp. 310-316, 1997.
[119] Malesani, L.; Leonardi, F.; Speranza, R.; Scandellari, A.;, "Switched reluctance
motor: a reduced torque ripple drive," in Industrial Electronics, Control, and
Instrumentation, 1993. Proceedings of the IECON '93., International Conference
on, 1993, pp. 1764-1769 vol.3.
[120] Xue, X.D.; Cheng, K.; Cheung, N.C.;, "Evaluation of torque sharing functions for
torque ripple minimization of switched reluctance motor drives in electric
vehicles," in Power Engineering Conference, 2008. AUPEC '08. Australasian
Universities, 2008, pp. 1-6.
[121] Bizkevelci, E.; Leblebicioglu, K.; Ertan, H.B.;, "A sliding mode controller to
minimise SRM torque ripple and noise," in Industrial Electronics, 2004 IEEE
International Symposium on, 2004, pp. 1333-1338 vol. 2.
[122] I. Husain and M. Ehsani, "Torque ripple minimization in switched reluctance
motor drives by PWM current control," Power Electronics, IEEE Transactions on,
vol. 11, pp. 83-88, 1996.
[123] H. Ishikawa, Kamada, Y. and Naitoh, H., "Instantaneous current profile control
for flat torque of switched reluctance motors," Electrical Engineering in Japan,
vol. 163, pp. 7887, 2008.
[124] T. Kosaka, Matsui, N., Taniguchi, Y.-I. and Do-meki, H., "Some considerations
on torque ripple suppression in reluctance motors," Electrical Engineering in
Japan, vol. 130, pp. 118128, 2000.
[125] N. T. Shaked and R. Rabinovici, "New procedures for minimizing the torque
ripple in switched reluctance motors by optimizing the phase-current profile,"
367
LIST OF REFERENCES
[126] Stephenson, J.M.; Hughes, A.; Mann, R.;, "Torque ripple minimisation in a
switched reluctance motor by optimum harmonic current injection," Electric
Power Applications, IEE Proceedings -, vol. 148, pp. 322-328, 2001.
[127] Zhengyu Lin; Reay, D.S.; Williams, B.W.; Xiangning He;, "Torque Ripple
Reduction in Switched Reluctance Motor Drives Using B-Spline Neural
Networks," Industry Applications, IEEE Transactions on, vol. 42, pp. 1445-1453,
2006.
[128] D. S. Schramm and B. W. Williams, "High bandwidth measurement of SRM
torque ripple," in Electrical Machines and Drives, 1993. Sixth International
[129] Corda, J.; Masic, S.; Stephenson, J.M.;, "Computation and experimental
determination of running torque waveforms in switched-reluctance motors,"
Electric Power Applications, IEE Proceedings B, vol. 140, pp. 387-392, 1993.
[130] A. C. Koenig and S. D. Pekarek, "Measurement of torque ripple in mass-produced
switched-reluctance motor drives," in Electric Machines and Drives, 2005 IEEE
International Conference on, 2005, pp. 6 pp.-509.
[131] Ehsani, Mehrdad; Bass, James T.; Miller, Timothy J. E.; Steigerwald, Robert L.;,
"Development of a Unipolar Converter for Variable Reluctance Motor Drives,"
Industry Applications, IEEE Transactions on, vol. IA-23, pp. 545-553, 1987.
[132] Ehsani, M.; Husain, I.; Ramani, K.R.; Galloway, J.H.;, "Dual-decay converter for
switched reluctance motor drives in low-voltage applications," Power Electronics,
[133] R. Krishnan and P. N. Materu, "Analysis and design of a low-cost converter for
switched reluctance motor drives," Industry Applications, IEEE Transactions on,
vol. 29, pp. 320-327, 1993.
[134] Mir, S.; Husain, I.; Elbuluk, M.E.;, "Energy-efficient C-dump converters for
switched reluctance motors," Power Electronics, IEEE Transactions on, vol. 12,
pp. 912-921, 1997.
[135] Jun-Ho Kim; Eun-Woong Lee; Jong-Han Lee;, "Design of the Starting Device
Installed in the Single-Phase Switched Reluctance Motor," Magnetics, IEEE
[136] B. C. Mecrow, Bedford, T. J., Bennet, J. W., and Celik, T. , "The use of segmental
rotors for 2 phase flux-switching motors," in Electrical Machines, 2006. ICEM
'06. 17th International Conference on. , 2006.
[137] C. Weiner, ""High Performance Switched Reluctance Drives for Electric Vehicle
Applications", Thesis submitted for Doctor of Philosophy," Thesis submitted for
Doctor of Philosophy, School of Electrical, Electronic and Computer Engineering,
University of Newcastle upon Tyne, 2000.
[138] Acarnley, Paul P.; Hill, Roland J.; Hooper, Clive W.;, "Detection of Rotor
Position in Stepping and Switched Motors by Monitoring of Current Waveforms,"
Industrial Electronics, IEEE Transactions on, vol. IE-32, pp. 215-222, 1985.
[139] J. P. Lyons, Macminn, S. R., Preston, M. A., ""Discrete position estimator for
switched reluctance machine using a flux-current map calculator", US Patent No
5140243, 18 August 1992.."
[140] A. Lumsdaine and J. H. Lang, "State observers for variable-reluctance motors,"
Industrial Electronics, IEEE Transactions on, vol. 37, pp. 133-142, 1990.
[141] P. K. Sood, Skinner, J. L., Petty, D. M. , ""Method and apparatus of operating a
dynamoelectric machine using DC bus current profile", US Patent No5420492, 30
May 1995.."
368
LIST OF REFERENCES
[142] R. Krishnan, "Switched Reluctance Motor Drives: Modeling, Simulation, Analysis,
Design, and Applications": CRC Press, Boca Raton, 2001.
[143] Blaabjerg, F.; Kjaer, P.C.; Rasmussen, P.O.; Cossar, C.;, "Improved digital current
control methods in switched reluctance motor drives," Power Electronics, IEEE
[144] Miller, T.J.E.; Bower, P.G.; Becerra, R.; Ehsani, M.;, "Four-quadrant brushless
reluctance motor drive," in Power Electronics and Variable-Speed Drives, Third
[145] Hava, A.M.; Blasko, V.; Lipo, T.A.;, "A modified C-dump converter for variablereluctance machines," Industry Applications, IEEE Transactions on, vol. 28, pp.
1017-1022, 1992.
[146] King-Jet Tseng; Shuyu Cao; Jijiu Wang;, "A new hybrid C-dump and buckfronted converter for switched reluctance motors," Industrial Electronics, IEEE
[147] Barrass, P.G.; Mecrow, B.C.; Clothier, A.C.;, "Bipolar operation of fully-pitched
winding switched reluctance drives," in Electrical Machines and Drives, 1995.
Seventh International Conference on (Conf. Publ. No. 412), 1995, pp. 252-256.
[148] Oliveira, A.C.; Lima, A.M.N.; Jacobina, C.B.; Salvadori, F.;, "Startup and Fault
Tolerance of the SRM Drive with Three-Phase Bridge Inverter," in Power
Electronics Specialists Conference, 2005. PESC '05. IEEE 36th, 2005, pp. 714719.
[149] Sadeghi, S.; Milimonfared, J.; Mirsalim, M.; Jalalifar, M.;, "Dynamic Modeling
and Simulation of a Switched Reluctance Motor in Electric Vehicles," in
Industrial Electronics and Applications, 2006 1ST IEEE Conference on, 2006, pp.
1-6.
[150] Kartono, I.R.; Kajiwara, K.; Dohmeki, H.;, "Dynamic simulation of maximizing
the starting torque for super-high-speed drive of a 4/2 Switched Reluctance
Motor," in Electrical Machines, 2008. ICEM 2008. 18th International Conference
on, 2008, pp. 1-6.
[151] Y. Hayashi and T. J. E. Miller, "A new approach to calculating core losses in the
SRM," Industry Applications, IEEE Transactions on, vol. 31, pp. 1039-1046,
1995.
[152] P. N. Materu and R. Krishnan, "Estimation of switched reluctance motor losses,"
Industry Applications, IEEE Transactions on, vol. 28, pp. 668-679, 1992.
[153] S. Vukosavic and V. R. Stefanovic, "SRM inverter topologies: a comparative
evaluation," Industry Applications, IEEE Transactions on, vol. 27, pp. 1034-1047,
1991.
[154] Z. Jinhui and A. V. Radun, "A New Method to Measure the Switched Reluctance
Motor's Flux," Industry Applications, IEEE Transactions on, vol. 42, pp. 11711176, 2006.
[155] R. Krishnan and P. Materu, "Measurement and instrumentation of a switched
reluctance motor," in Industry Applications Society Annual Meeting, 1989.,
Conference Record of the 1989 IEEE, 1989, pp. 116-121 vol.1.
[156] Akita, H.; Nakahara, Y.; Miyake, N.; Oikawa, T.;, "A new core," Industry
Applications Magazine, IEEE, vol. 11, pp. 38-43, 2005.
[157] U. Hoefer, ""Design and Development of a High-Speed Motor for a Vacuum
Pump", Thesis submitted for Engineering Doctorate," School of Electrical,
Electronic and Computer Engineering, University of Newcastle upon Tyne, 2008.
[158] Akita, H.; Nakahara, Y.; Miyake, N.; Oikawa, T.;, "New core structure and
manufacturing method for high efficiency of permanent magnet motors," in
369
LIST OF REFERENCES
[159]
[160]
[161]
[162]
[163]
[164]
[165]
[166]
[167]
[168]
[169]
[170]
[171]
[172]
[173]
[174]
[175]
Industry Applications Conference, 2003. 38th IAS Annual Meeting. Conference

Record of the, 2003, pp. 367-372 vol.1.
Y. Fujita, Hazeyama, M., ""Electromagnetic Driving Technology of DC Motor for
EGR Valve"," Mitsubishi Electric Technical Reports, December 2006, Japan.
K. Inoue, Miyake, N., ""New Traction Machine for Machine Room-Less
Elevators," Mitsubishi Electric Technical Reports, September 2002, Japan.
H. Kawaguchi, Oikawa, T., ""High-Efficiency Motors for Air-Conditioner
Compressors"," Mitsubishi Electric Technical Reports, September 2003,
JapanSeptember 2003.
C. E. Carstensen, Bauer, S. E., Inderka, R. B. & De Doncker, R. W. , "Efficiency
Comparison of Different Winding Configurations for Switched Reluctance
Vehicle Propulsion Drives," in In 20th International Electric Vehicle Symposium
(EVS-20), Long Beach, USA, 2003.
"Beacon Adhesives Co. Inc.," in Magna-Tac E645, Stack Lamination Epoxy
datasheet, ed.
I. Husain and S. A. Hossain, "Modeling, Simulation, and control of switched
reluctance motor drives," Industrial Electronics, IEEE Transactions on, vol. 52,
pp. 1625-1634, 2005.
J. A. Haylock, ""Fault Tolerant Drives for Safety Critical Applications", Thesis
submitted for Doctor of Philosophy," School of Electrical, Electronic and
Computer Engineering, University of Newcastle upon Tyne, 1998.
C. French, ""Real Time Control of Electric Drives", Thesis submitted for Doctor
of Philosophy," School of Electrical, Electronic and Computer Engineering,
University of Newcastle upon Tyne, 2001.
J. M. Stephenson and J. Corda, "Computation of torque and current in doubly
salient reluctance motors from nonlinear magnetisation data," Electrical
Engineers, Proceedings of the Institution of, vol. 126, pp. 393-396, 1979.
J. M. Stephenson and M. A. El-Khazendar, "Saturation in doubly salient
reluctance motors," Electric Power Applications, IEE Proceedings B, vol. 136, pp.
50-58, 1989.
M. Ehsani, Abourida, S., ""Real time Simulation of Switched Reluctance Motor
Drives"," in Technical Documents, ed: Opal-RT Technologies, Sep 2003.
T. Tsukii, Nakamura, K. and Ichinokura, O., "SPICE simulation of SRM
considering nonlinear magnetization characteristics," Electrical Engineering in
Japan, vol. 142, pp. 50-56, 2003.
Phillips, N.W.; Bolton, H.R.; Lewis, J.D.; Pollock, C.; Barnes, M.;, "Simulation of
switched reluctance drive system using a commercially available simulation
package," in Electrical Machines and Drives, 1995. Seventh International
D. W. J. Pulle, "New database for switched reluctance drive simulation," Electric
Power Applications, IEE Proceedings B, vol. 138, pp. 331-337, 1991.
A. Radun, "Analytical calculation of the switched reluctance motor's unaligned
inductance," Magnetics, IEEE Transactions on, vol. 35, pp. 4473-4481, 1999.
A. Radun, "Analytically computing the flux linked by a switched reluctance motor
phase when the stator and rotor poles overlap," Magnetics, IEEE Transactions on,
vol. 36, pp. 1996-2003, 2000.
D. A. Torrey and J. H. Lang, "Modelling a nonlinear variable-reluctance motor
drive," Electric Power Applications, IEE Proceedings B, vol. 137, pp. 314-326,
1990.
370
LIST OF REFERENCES
[176] Torrey, D.A.; Niu, X.-M.; Unkauf, E.J.;, "Analytical modelling of variablereluctance machine magnetisation characteristics," Electric Power Applications,
IEE Proceedings -, vol. 142, pp. 14-22, 1995.
[177] T. J. E. Miller and M. McGilp, "Nonlinear theory of the switched reluctance motor
for rapid computer-aided design," Electric Power Applications, IEE Proceedings
B, vol. 137, pp. 337-347, 1990.
[178] Xue, X. D.; Cheng, K. W. E.; Ho, S. L.;, "Simulation of switched reluctance
motor drives using two-dimensional bicubic spline," Energy Conversion, IEEE
[179] J. C. Moreira and T. A. Lipo, "Simulation of a four phase switched reluctance
motor including the effects of mutual coupling," Electric Machines & Power
Systems, vol. 16, pp. 281 - 299, 1989.
[180] C. Shuyu and K. J. Tseng, "Evaluation of neighboring phase coupling effects of
switched reluctance motor with dynamic modeling approach," in Power
Electronics and Motion Control Conference, 2000. Proceedings. IPEMC 2000.
The Third International, 2000, pp. 881-886 vol.2.
[181] Y. Xu and D. A. Torrey, "Study of the mutually coupled switched reluctance
machine using the finite element-circuit coupled method," Electric Power
Applications, IEE Proceedings -, vol. 149, pp. 81-86, 2002.
[182] J. M. Kokernak and D. A. Torrey, "Magnetic circuit model for the mutually
coupled switched-reluctance machine," Magnetics, IEEE Transactions on, vol. 36,
pp. 500-507, 2000.
[183] M. Stiebler and L. Ke, "An analytical model of switched reluctance machines,"
Energy Conversion, IEEE Transactions on, vol. 14, pp. 1100-1107, 1999.
[184] S. A. Hossain and I. Husain, "A geometry based simplified analytical model of
switched reluctance machines for real-time controller implementation," Power
[185] Vejian Rajanran, R.; Sahoo, N.C.; Gobbi, R.;, "Mathematical Modeling of FluxLinkage Characteristics of Switched Reluctance Motors Using Polynomial Neural
Networks," in Power and Energy Conference, 2006. PECon '06. IEEE
International, 2006, pp. 378-382.
[186] Yan Cai; Qingxin Yang; Yanbin Wen; Lihua Su;, "Nonlinear modeling and
simulating of switched reluctance motor and its drive," in Computer and
Automation Engineering (ICCAE), 2010 The 2nd International Conference on,
2010, pp. 465-469.
[187] Zhengyu Lin; Reay, D.S.; Williams, B.W.; Xiangning He;, "Online Modeling for
Switched Reluctance Motors Using B-Spline Neural Networks," Industrial
[188] Owatchaiphong, S.; Carstensen, C.; De Doncker, R.W.;, "Optimization of
Predesign of Switched Reluctance Machines Cross Section Using Genetic
Algorithms," in Power Electronics and Drive Systems, 2007. PEDS '07. 7th
[189] Carstensen, C.E.; Fuengwarodsakul, N.H.; De Doncker, R.W.;, "Flux Linkage
Determination for Correct Modeling of Switched Reluctance Machines - Dynamic
Measurement versus Static Computation," in Electric Machines & Drives
Conference, 2007. IEMDC '07. IEEE International, 2007, pp. 1317-1323.
[190] Charton, J.T.; Corda, J.; Stephenson, J.M.; Randall, S.P.;, "Dynamic modelling of
switched reluctance machines with iron losses and phase interactions," Electric
Power Applications, IEE Proceedings -, vol. 153, pp. 327-336, 2006.
371
LIST OF REFERENCES
[191] Khotpanya, S.; Kittiratsatcha, S.; Kazuhisa, I.;, "A Magnetic Model of a ThreePhase Switched-Reluctance Machine using Cubic Spline Interpolation
Technique," in Power Electronics and Drives Systems, 2005. PEDS 2005.
[192] Ling Yuelun; Wang Mianhua; Wang Yan; Wang Fenli;, "Simulation Research on
Switched Reluctance Motor Modeling and Control Strategy Based on ANSOFT,"
in Measuring Technology and Mechatronics Automation (ICMTMA), 2010
[193] Bingni Qu; Jiancheng Song; Hongda Zhang;, "Simulation for Switched
Reluctance Drive Based on the Measured Flux Linkage Characteristics," in Power
and Energy Engineering Conference, 2009. APPEEC 2009. Asia-Pacific, 2009,
pp. 1-4.
[194] Zhengang Sun; Cheung, N.C.; Jianfei Pan; Shiwei Zhao; Wai-Chuen Gan;,
"Design and simulation of a magnetic levitated switched reluctance linear actuator
system for high precision application," in Industrial Electronics, 2008. ISIE 2008.
IEEE International Symposium on, 2008, pp. 624-629.
[195] Woothipatanapan, S.; Chancharoensook, P.; Jangwanitlert, A.;, "Efficiency
improvement of converter for switched reluctance motor drives at low speed
operation," in Power Electronics and Drive Systems, 2009. PEDS 2009.
[196] M. T. Khor and R. Sotudeh, "A 3-phase 12/10 asymmetrical switched reluctance
motor," in Power Electronics and Applications, 2005 European Conference on,
2005, pp. 9 pp.-P.9.
[197] Lin, D.; Zhou, P.; Stanton, S.; Cendes, Z.J.;, "An Analytical Circuit Model of
Switched Reluctance Motors," Magnetics, IEEE Transactions on, vol. 45, pp.
5368-5375, 2009.
[198] Qiong-zhong Chen; Guang Meng; Yu-feng Mo; Ting-xing Wang;, "Analytical
nonlinear modeling of SRM and its system-level simulation with airborne power
system," in Industrial Technology, 2008. ICIT 2008. IEEE International
Conference on, 2008, pp. 1-8.
[199] D. Wen and L. Deliang, "Fourier series and ANFIS-based modeling and
prediction for switched reluctance motor," in Electrical Machines and Systems,
2008. ICEMS 2008. International Conference on, 2008, pp. 3362-3366.
[200] A. Khalil and I. Husain, "A Fourier Series Generalized Geometry-Based
Analytical Model of Switched Reluctance Machines," Industry Applications, IEEE
[201] Z. Suying and L. Hui, "Modeling and Simulation of Switched Reluctance Motor
Double Closed Loop Control System," in Intelligent Control and Automation,
2006. WCICA 2006. The Sixth World Congress on, 2006, pp. 6151-6155.
[202] Xia, C.L.; Xue, M.; Shi, T.N.;, "A New Rapid Nonlinear Simulation Method for
Switched Reluctance Motors," Energy Conversion, IEEE Transactions on, vol. 24,
pp. 578-586, 2009.
[203] Ding Wen; Liang Deliang; Cheng Zhuping;, "Dynamic Model and Simulation for
a 6/4 Switched Reluctance Machine System Assisted by Maxwell SPICE and
Simplorer," in Mechatronics and Automation, 2007. ICMA 2007. International
Conference on, 2007, pp. 1699-1704.
[204] C. Yanbo and K. W. E. Cheng, "Real-time Simulation and Experiment Platform
for Switched Reluctance Motor," in Power Electronics Systems and Applications,
2006. ICPESA '06. 2nd International Conference on, 2006, pp. 244-249.
372
LIST OF REFERENCES
[205] Dufour, C.; Paquin, J.-N.; Blanchette, H.; Belanger, J.;, "Specifications for realtime simulation of switched reluctance drives using microprocessors and FPGAs
as computational engines," in Electric Machines and Drives Conference, 2009.
IEMDC '09. IEEE International, 2009, pp. 750-754.
[206] T. J. E. Miller, "Converter Volt-Ampere Requirements of the Switched Reluctance
Motor Drive," Industry Applications, IEEE Transactions on, vol. IA-21, pp. 11361144, 1985.
373

Celik 11

Uploaded by

Copyright:

Available Formats

Celik 11

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Celik 11

Uploaded by

Copyright:

Available Formats

Segmental Rotor Switched

University of Newcastle upon Tyne

Switched Reluctance Machines In Industry......................................................... 2

Target Industries for SR Motor Drive Technology...................................... 3

Switched Reluctance Machines With Segmental Rotors ..................................... 9

Multi-tooth Winding Segmental Rotor SRM............................................. 10

Converters And Controllers For Switched Reluctance Machines ..................... 11

Switched Reluctance Machines vs. Permanent Magnet Brushless DC Machines

Objectives And Contribution To Knowledge .................................................... 16

Thesis Overview ................................................................................................ 17

SRM TOPOLOGIES .................................................... 19

SRM Torque Ripple ........................................................................................... 20

Torque Ripple Minimisation through Machine Design ............................. 21

SRM Topologies ................................................................................................ 22

Conventional Toothed Rotor SRMs........................................................... 22

2.3.2.1 Segmented Rotor Switched Reluctance Motors with Fully-Pitched

SR DRIVE TOPOLOGIES .......................................... 43

Torque Ripple Minimisation through Drive Control ......................................... 45

Measuring Torque Ripple .......................................................................... 45

Drive Topologies ............................................................................................... 46

Review of Drives ....................................................................................... 47

Hysteresis Current Control .................................................................... 49

Asymmetric Half-Bridge Converter ...................................................... 51

MACHINE DESIGN AND TESTING ........................ 64

Design of the 6-Phase Segmented Rotor SRM .................................................. 65

Determining the Machine Dimensions ...................................................... 71

Design of the 6-Phase Toothed Rotor SRM ...................................................... 83

Machine with Separate Stator Cores .................................................................. 90

Construction of the Prototype Machine ............................................................. 94

Stator Construction .................................................................................... 95

Test Rig Description ........................................................................................ 103

Load Arrangement ................................................................................... 104

Flux-Linkage Measurements ............................................................... 113

Summary .......................................................................................................... 119

SIMULATION MODELS .......................................... 120

Introduction ...................................................................................................... 120

SRM Modelling and Simulation ...................................................................... 122

Background .............................................................................................. 122

Simulation Database ........................................................................................ 128

Machine Model ........................................................................................ 129

Electronics Loss Modelling ................................................................. 147

Modelling Of SRMs with Mutually Coupled Phase Windings ....................... 153

Performance Calculations ................................................................................ 162

Summary .......................................................................................................... 163

Introduction ...................................................................................................... 165

Measurements and Simulation Results ............................................................ 166

Inverter VA-Rating Comparison...................................................................... 187

Two-Phase Machine Driven by H-Bridge Inverter .................................. 168

Drive under Voltage Control ................................................................... 189

Summary .......................................................................................................... 196

Introduction ...................................................................................................... 199

Measurements and Simulation Results ............................................................ 200

Comparison of Measurements against Simulations ................................. 200

Summary .......................................................................................................... 217

Introduction ...................................................................................................... 218

Dynamic Simulation Results............................................................................ 220

Mutually De-Coupled Star Connected Line Current Controlled Six-Phase

Dynamic Measurements Of The Prototype Six-Phase Segmental Rotor SRM .....

Dynamic Torque against Speed Characteristics ...................................... 254

9.3.2.1 Low Speed Measurements of the Prototype Six-Phase Star-Connected