Vector Control of 3-Stage / Slice Axial Flux Permanent Magnet Synchronous Motor

VECTOR CONTROL OF 3-STAGE / SLICE AXIAL
FLUX PERMANENT MAGNET SYNCHRONOUS

MOTOR
Dissertation submitted
in partial fulfilment of the requirements
for the award of the degree of
MASTER OF TECHNOLOGY
IN
POWER ELECTRONICS AND DRIVES
By
Ragam Srikanth
(Regd. No. 14131D5211)
Under the Guidance of
Dr. Sastry V. Vedula Mr. Ramesh Perla
(Senior Professor, EEE) (Senior Manager, HBL)
DEPARTMENT OF ELECTRICAL & ELECTRONICS ENGINEERING

GAYATRI VIDYA PARISHAD COLLEGE OF ENGINEERING
(AUTONOMOUS)
(Approved by AICTE, New Delhi and Affiliated to JNTU-Kakinada)
Re-accredited by NBA& NAAC with “A” Grade with a CGPA of 3.47/4.00
VISAKHAPATNAM - 530 048
DECEMBER - 2016
i
ii
CERTIFICATE OF PLAGIARISM CHECK
This is to certify the M.Tech (Power Electronics & Drives) dissertation
submitted by RAGAM SRIKANTH bearing Roll No. 14131D5211 of Electrical and
Electronics Engineering department under the supervision of Dr. Sastry V. Vedula
and Mr. Ramesh Perla, HBL, Hyderabad has undergone plagiarism check and
found to have similarity index less than 30%. The details of the plagiarism check are
as under:
File Name: PLAGCHECK_14131D5211.DOCX (2.44M)
Dissertation Title: VECTOR CONTROL OF 3-STAGE / SLICE AXIAL FLUX

PERMANENT MAGNET SYNCHRONOUS MOTOR
Date and Time of Submission: 25-DEC-2016 07:24AM
Submission ID: 755978842
Similarity Index: 25%
Dean – Academic Programs (PG & Research)
iii
ABSTRACT
The permanent magnet synchronous motors (PMSM) are more efficient than
induction motors and conventional synchronous motors, as there is no field winding
and slip rings, so that the associated losses are zero in case of permanent magnet
motors. Permanent magnet synchronous motors have become the serious competitors
for the induction motors due to developments in the permanent magnet technology.
The axial flux machines are also known as disc type machines. The axial flux
machines are compact over radial flux machines of the same power rating up to
certain power level. It is possible to keep all advantages of axial flux machines over
radial flux counterpart at medium power ratings by stacking of disc machines on the
same shaft and in the same enclosure; such a machine is known as multi stage axial
flux machine.
Invention of vector control concept for the speed control of AC motors has
resulted in faster dynamic response. So, the most suitable motor for most compact,
high power, highly efficient and high performance applications is the field oriented
controlled multi stage axial flux brushless permanent magnet synchronous motor.
3-slice axial flux permanent magnet synchronous motor is a non salient pole
PMSM and it consists of 3-Stator units and 4-Rotor units. The 4-Rotor units are
supported on the single shaft. Each Stator unit is provided with independent three
phase windings and there is no magnetic coupling among the three stator units.
Permanent magnets are arranged on the rotor core.3-Phase windings of Stator unit-3
are spatially displaced by 40 degrees electrical to those of stator unit-2 and similarly
3-Phase windings of Stator unit-2 are spatially displaced by 40 degrees electrical to
those of stator unit-1.
This dissertation relates to simulation of vector control of 3-stage / slice axial

flux permanent magnet synchronous motor in Saber-RD simulator. Modeling of 3-
phase PMSM has been done. The concept is extended to the modeling of 3-stage, 3-
phase PMSM. Vector control of 3-phase PMSM has been done in the Saber-RD
simulator. The concept is extended to the vector control of 3-stage, 3-phase PMSM in
the Saber-RD simulator. Uni-polar sinusoidal Pulse Width Modulation (PWM)
technique has been implemented to switch the inverter power devices to minimise the
iv
current ripple. H-bridge inverter is used to feed each phase of 3-stage, 3-phase axial
flux permanent magnet synchronous motor for better utilisation of dc link voltage.
Maximum torque/ampere operation has been achieved by forcing the d-axis

current of each stage/slice to zero using PI controller. An outer speed loop (consists of
PI Controller) for better speed tracking is cascaded to 3 inner current loops (each loop
consists of one PI controller); one for each stage/slice for achieving faster dynamic
response under transient as well as steady state conditions.
v
ACKNOWLEDGEMENT
I wish to express my sincere gratitude towards my supervisor Dr. SASTRY V.

VEDULA, Senior Professor for his advice during my project work. His constant
encouragement, observations and comments helped me to establish the overall
direction of the thesis and been a source of knowledge.
I cherish to express my gratitude to HBL Power Systems Limited, Yapral, and

Hyderabad where I had worked as an intern. My special thanks go to Sri Ramesh
Perla Senior Manager, Sri K. Krishna Murthy DGM R&D, HBL Powers Systems
Limited for giving me the direction and advice, Sri M.V. Vidya Sagar, President,
HBL Power Systems Limited for his motivation and having provided funds for the
advanced simulation software components which is helpful for simulation
experiments.
I wish to express my sincere thanks to Dr. A. B. KOTESWARA RAO,
Principal, and Prof. P. Somaraju, Secretary, Gayatri Vidya Parishad College of
Engineering (Autonomous), for their support. I wish to express my heartfelt thanks to
Dr. K. Narasimha Rao, Head of the Department of Electrical and Electronics
Engineering and also Chairman of Departmental Research Committee (DRC)
Member and Dr. C.V.K. Bhanu, Professor, Dean of Academic Programs-PG and
Research for their advice during the departmental reviews
I would like to express my heartfelt thanks to Dr. G. V. E. Satish Kumar,

Associate Professor (EEE) & PG Co-coordinator, Mr. P. K. Das, Associate Professor
& DRC Member for their advice during the departmental reviews. I wish to express
my sincere thanks to all the Faculty, and Staff especially Sri P. Tejeswara Rao of
Department of Electrical and Electronics Engineering, GVP for their kind co-
operation and helpful services.
Finally, I take this opportunity to thank all the people who helped me in the
completion of project work for their timely encouragement and helpful services.
RAGAM SRIKANTH
(Regd. No. 14131D5211)
vi
CONTENTS
Certificate i
Certificate of Plagiarism check iii
Abstract iv
Acknowledgement vi
Contents vii
List of Figures x
List of Tables xiv
List of Acronyms xv
CHAPTER NAME PAGE NO
1. INTRODUCTION (1-10)
1.1 Introduction 1
1.1.1 Synchronous Motor 2
1.1.2 Synchronous Speed 2
1.2 Axial flux machines 2
1.2.1 Drawbacks of Radial Flux Machines 3
1.2.2 Advantages of Axial Flux Machines 3
1.3 Topologies of Axial Flux Permanent Magnet 4
Brushless Machines
1.3.1 Single Sided Axial Flux Permanent Magnet Synchronous 5
Motor
1.3.2 Double Sided Axial Flux Permanent Magnet Synchronous 5
Motor
1.3.3 Power Limitation of Axial Flux Permanent Magnet 6
Machines and Multi Stage Axial Flux Permanent
Magnet Machines
1.4 Speed Control of Permanent Magnet Synchronous Motor 7
1.4.1 Scalar Control Method 7
1.4.2 Vector Control Strategy 8
1.5 Objective of the Dissertation 9
1.6 Literature Review 10
vii
1.7 Organization of the Dissertation 10
2. PROBLEM DESCRIPTION AND PWM METHODS (11-17)
2.1 Description of Problem 11
2.2 Vector Control of Three Stage Axial Flux Permanent Magnet 11
Synchronous Motor
2.3 PWM Methods 14
2.3.1 Bi-Polar Sinusoidal PWM Technique 14
2.3.2 Uni-Polar Sinusoidal PWM Technique 15
2.4 Conclusion 17
3. MODELLING OF 3-STAGE AXIAL FLUX PERMANENT (18-27)
MAGNET SYNCHRONOUS MOTOR
3.1 Introduction 18
3.2 Clarke’s Transformation 19
3.2.1 Power Invariance 20
3.3 Park’s Transformation 20
3.4 Dynamic Modeling of 3-phase PMSM in (Synchronously rotating) 21
D, q-Reference Frame
3.5 Dynamic Modeling of 3-stage, 3-phase PMSM in
(Synchronously rotating) d, q-Reference Frame 23
3.5.1 Mathematical Equations of Stage-1 25
3.6 Summary 27
4. DESIGN OF PI CONTROLLERS FOR VECTOR CONTROL (28-45)
OPERATION OF 3-PHASE PMSM
4.1 Introduction 28
4.2 Design of PI Controller 28
4.3 Vector Control System Development 30
4.3.1 Current PI Controllers Design 30
4.3.1.1 D-axis Current PI controller Design 30
4.3.1.2 Q-axis Current PI controller Design 32
4.3.1.3 Decoupling Network Formulation 34
4.3.2 Design of Speed PI Controller 36
4.4 Simulation Results 40
viii
4.5 Summary 45
5 DESIGN OF PI CONTROLLERS FOR VECTOR CONTROL (46-53)
OPERATION OF 3-STAGE, 3-PHASE PMSM
5.1 Introduction 46
5.2 Design of Speed PI Controller 47
5.3 Inverse Park’s Transformation 52
5.4 Inverse Clarke’s Transformation 52
5.5 Summary 53
6 SIMULATION RESULTS AND ANALYSIS OF (54-67)
3-STAGE, 3-PHASE PMSM
6.1 Introduction 54
6.2 Simulation Circuits 54
6.3 Simulation Results 57
6.3.1 Simulation of Test-1 Speed Profile 57
6.3.2 Simulation of Test-2 Speed Profile 65
6.4 Summary 67
7 CONCLUSIONS AND FUTURE SCOPE (68-70)
7.1 Conclusions 69
7.2 Future Scope 69
APPENDIX 71
REFERENCES 75
ix
LIST OF FIGURES
Fig.1.1 Classification of axial flux permanent magnet synchronous 4

motors
Fig.1.2 Single sided axial flux permanent magnet machine 5
Fig.1.3 Double-sided axial flux permanent magnet synchronous 6

motor
Fig.1.4 Multi Stage axial flux permanent magnet synchronous 7

motor
Fig.1.5 Block diagram for vector control 3-phase PMSM 8
Fig.2.1 Block diagram of vector control of 3-stage, 3-phase 12

PMSM
Fig.2.2 H-bridge inverter 14
Fig.2.3 Bi-polar sinusoidal PWM voltage switching 15
Fig.2.4 Uni-polar sinusoidal PWM voltage switching 16
Fig.3.1 Representation of Clarke’s and Park’s transformations 18
Fig.3.2 Clarke’s transformation 20
Fig.3.3 Park’s transformation 21
Fig.4.1 Simple PI controller closed loop system 29
Fig.4.2 Block diagram of permanent magnet synchronous motor: 31

D-axis
Fig.4.3 Direct-axis current loop of PMSM 32
Fig.4.4 Block diagram of permanent magnet synchronous motor: 33

Q-axis
Fig.4.5 Quadrature-axis current loop of PMSM 33
x
Fig.4.6 The PI controller along with feed forward compensation for 35
d-axis current control loop
Fig.4.7 The PI controller along with feed forward compensation for 36
electromagnetic torue control for three phase PMSM
Fig.4.8 Outer speed loop of 3-phase permanent magnet 39
synchronous motor
Fig.4.9 Simplified outer speed loop of 3-phase permanent magnet 40
synchronous motor
Fig.4.10 Saber sketch schematic diagrams of vector control of 41

3-phase PMSM
Fig.4.11 Parameters of PI controllers used in the simulation 42
Fig.4.12 Set speed command (black) and actual speed (blue) [in 43
electrical rad/sec] of three phase PMSM
Fig.4.13 Actual speed (in mechanical rad/ sec), of 3-phase PMSM 43
Fig.4.14 Load torque and shaft power of 3-phase PMSM 43
Fig.4.15 D-axis, q-axis currents and electromagnetic torque of 44
PMSM at steady state
Fig.5.1 The PI controllers along with feed forward compensations 46
for electromagnetic torue control for 3-stage, PMSM
Fig.5.2 Outer speed loop of 3-stage permanent magnet 50
synchronous motor
Fig.5.3 Simplified outer speed loop of 3-stage permanent magnet 51
xi
synchronous motor
Fig.6.1 Saber sketch schematics of vector control of 3-stage, 3- 56

phase PMSM
Fig.6.2 Parameters of PI controllers 57
Fig.6.3 Set speed command (pink) and actual speed (orange) [in 58
electrical rad/sec] of three stage axial flux PMSM
Fig.6.4 Actual speed (in mechanical rad/ sec) of three stage 59

axial flux PMSM
Fig.6.5 Load torque applied to the three stage axial flux PMSM 59
Fig.6.6 Shaft power of three stage axial flux permanent magnet 60

synchronous motor
Fig.6.7 D-axis, q-axis currents and electromagnetic torque of slice- 60

1 at steady state

2 at steady state

3 at steady state
Fig.6.10 Electromagnetic torque developed by each stage and total 61

electromagnetic torque developed (Te) by 3-slice motor at
steady state
Fig.6.11 Voltage, current and power of each phase and total power 63
drawn by slice-1 of 3 slice PMSM under steady state
xii
Fig.6.14 Parameters of PI controllers and simulation results for ramp 66
type speed command
xiii
LIST OF TABLES
Table 7.1 Summary table of slice-1 of 3-slice 67

PMSM

PMSM

PMSM
xiv
LIST OF ACRONYMS AND SYMBOLS
PMSM : Permanent magnet synchronous motor
PWM : Pulse width modulation
PI : Proportional plus integral
Rs : Stator resistance / phase in Ohm
Ld : Direct axis inductance in Henry
Lq : Quadrature axis inductance in Henry
p : Number of poles
λf : Amplitude of permanent magnetic flux volt-sec/rad
j : Moment of inertia in kg-m2
B : Viscous damping co-efficient in Nm-sec/rad
λd : Direct axis flux linkages
λds1 : Direct axis flux linkages of stator unit-1
λq : Quadrature axis flux linkages
λqs1 : Quadrature axis flux linkages of stator unit-1
ωe* : Reference set speed command in electrical rad/sec
ωe : Actual speed of motor in electrical rad/sec
ωr : Actual speed of motor in mechanical rad/sec
θe : Electrical rotor flux position angle
xv
θ1 : Electrical rotor flux position angle corresponding to stator unit-1
kt : Torque constant of motor
ρ : d/dt
va : Phase-a voltage
vb : Phase-b voltage
vc : Phase-c voltage
va1 : Phase-a voltage of stator unit-1
vb1 : Phase-b voltage of stator unit-1
vc1 : Phase-c voltage of stator unit-1
ia : Phase-a current
ib : Phase-b current
ic : Phase-c current
ia1 : Phase-a current of stator unit-1
ib1 : Phase-b current of stator unit-1
ic1 : Phase-c current of stator unit-1
xvi
ic2 : Phase-c current e of stator unit-2
ic3 : Phase-c current of stator unit-3
vd : Direct axis voltage
vds1 : Direct axis voltage for stator unit -1
vq : Quadrature axis voltage
vqs1 : Quadrature axis voltage for stator unit -1
id : Direct axis current
ids1 : Direct axis current for stator unit -1
iq : Quadrature axis current
iqs1 : Quadrature axis current for stator unit -1
iα : α-axis current for
xvii
iα1 : α-axis current for stator unit -1
iα2 : α -axis current for stator unit -2
iα3 : α -axis current for stator unit -3
iβ : β-axis current for
iβ1 : β-axis current for stator unit -1
iβ2 : β -axis current for stator unit -2
iβ3 : β -axis current e for stator unit -3
va * : Reference phase-a voltage
vb * : Reference phase-b voltage
vc * : Reference phase-c voltage
v*a1 : Reference phase-a voltage of stator unit-1
v*b1 : Reference phase-b voltage of stator unit-1
v*c1 : Reference phase-c voltage of stator unit-1
vd * : Reference direct axis voltage
v*ds1 : Reference direct axis voltage for stator unit -1
xviii
vq * : Reference quadrature axis voltage
v*qs1 : Reference quadrature axis voltage for stator unit -1
i*d : Reference direct axis current
i*ds1 : Reference direct axis current for stator unit -1
i*q : Reference quadrature axis current
i*qs1 : Reference quadrature axis current for stator unit -1
v*α : Reference α-axis voltage
v*α1 : Reference α-axis voltage for stator unit -1
v*β : Reference β-axis voltage
v*β1 : Reference β-axis voltage for stator unit -1
Kp : Proportional gain of PI controller
Ki : Integral gain of PI controller
𝜏i : Integral time constant of PI controller
xix
Kpd : Proportional gain of d-axis PI controller
Kid : Integral gain of d-axis PI controller
𝜏id : Integral time constant of d-axis PI controller
Kpq : Proportional gain of q-axis PI controller
Kiq : Integral gain of q-axis PI controller
𝜏iq : Integral time constant of q-axis PI controller
Kps : Proportional gain of speed PI controller
Kis : Integral gain of speed PI controller
𝜏is : Integral time constant of speed PI controller
ξ : Damping ratio
ωn : Natural frequency
xx
CHAPTER-1
INTRODUCTION
1.1 Introduction
Based on the input and output, the dynamic electrical machines are of two types.
They are 1) Generators 2) Motors. Generator is an Electro-Mechanical energy
conversion device that converts mechanical energy into electrical energy according to
Faradays laws of electromagnetic induction. Motor is an Electro-Mechanical energy
conversion device that converts electrical energy into mechanical energy through the
medium of magnetic field. The same electrical machine can be operated as a motor as
well as generator. For example, in electric traction, the same electric machine can be
operated in both motoring and ‘generating or breaking’ modes.
The motors that are suitable for variable speed applications are as follows.
1) Conventional dc motors
2) Traditional ac motors
3) Permanent magnet motors
a) Brushed motors
b) Brush-less motors
DC motors are widely used in the high performance variable speed

applications. The dynamic response of separately excited DC motors is faster because
of the inherent decoupling facility between the torque and flux producing channels of
the motor. But DC motors had its disadvantages like maintenance, sparking difficulty
during commutation and it is bulky because it consists of commutator. After the
invention of conventional AC motors like induction motors and synchronous motors,
above difficulties are overcome. But their dynamic performance was poor as they
didn’t have de-coupling facility between its torque and flux producing channels. The
technical advancements in the power electronic components along with the invention
of the vector control method in which the torque and flux producing components of
stator current of AC motors can be independently controlled (i.e. both torque and flux
1
of the machine are de-coupled) made the induction motors most popular in industries
and has become work horse in the industry. But the efficiency of these conventional
drives is less and they occupy more space.
Along with the above disadvantages in the conventional drives, the energy
scarcity and increase in the energy prices have motivated to implement variable speed
brush-less permanent magnet synchronous motor drives as the efficiency of the
permanent magnet synchronous motors is more and they are compact compared to
conventional motors because of elimination of rotor winding, slip rings and brushes.
As there is no rotor winding, slip rings and brushes and hence the associated losses
are zero.
1.1.1 Synchronous Motor
Incase of induction motor, as the speed approaches synchronous speed, the

rotor frequency as well as amplitude of currents decreases. At the synchronous speed,
the amplitude of rotor current becomes zero. Hence no torque is developed at
synchronous speed.
But it is possible to operate a poly phase A.C motor at synchronous speed if

the D.C supply is applied externally to rotor winding. Hence, in this case torque is
developed only at synchronous speed and the average torque developed at any other
speeds will be zero. Hence the name of such A.C motor is called as synchronous
motor. If rotor winding, slip rings and brushes of conventional synchronous motor is
replaced with the permanent magnets, the resultant machine is known as permanent
magnet synchronous motor.
1.1.2 Synchronous Speed
The mechanical speed of the space (voltage/current/flux) vector dictated by

the supply frequency according to the machine poles is known as synchronous speed.
1.2 Axial Flux Machines
Broadly, electrical motors are classified into to two types. They are 1) radial
flux motors 2) axial flux motors. For the given rating of the motor, the axial flux
motors are more efficient and smaller in physical size compared to radial flux motors.
2
Each type of electrical rotating machines will have both radial as well as axial flux
versions [1] at least theoretically. The disadvantages of radial flux machines and
advantages of axial flux and are given [1] below.
1.2.1 Drawbacks of Radial Flux Machines
There is an inherent limit in the increasing of power density further of radial flux
machines [1]. Much of the rotor core around the shaft is utilized as a magnetic circuit.
Heat removal from the radial flux machines is poor without forced cooling
arrangements. These limitations can’t be removed unless a new topology is adopted
which provides better ventilation and cooling.
1.2.2 Advantages of Axial Flux Machines
The advantages of axial flux machines over radial flux machines in certain
applications [1] are given as follows.
Diameter to length ratio of axial flux machines is much higher than radial flux
machines. The air gap of axial flux machines is planar and somewhat adjustable [1].
Power density of axial flux machines is higher than that of radial flux machines and
hence axial flux machines are more compact than the radial flux machines for the
same power rating.
The axial flux machines are ideal to design a multi stage machine in which the
number of modules is suitably chosen to get the required power or torque.
As the diameter of the axial flux machines is large, the number of poles that can
be accommodated is more, so that axial flux machines are suitable choice for high
frequency or low speed operations.
From the above discussion, it can be concluded that, the most compact and
most efficient motors are axial flux permanent magnet brushless motors for the same
output rating.
1.3 Topologies of Axial Flux Permanent Magnet Brushless Machines
The classification of different topologies of axial flux permanent magnet

brushless machines is given in [1] and shown in Fig.1.1.
3
Axial flux Permanent magnet synchronous
motor (PMSM)
Single sided Axial flux Double sided Axial flux

Multi stage Axial flux PMSM
PMSM PMSM
With slotted stator

With iron core
With slotted stator stator
With internal With core less

With slot less stator
stator stator
Without both stator

and rotor cores
With salient pole stator

With salient pole
stator
With slotted stator
With internal
rotor
With salient pole

stator
Fig.1.1. Classification of axial flux permanent magnet synchronous motors
1.3.1 Single-Sided Axial Flux Permanent Magnet Synchronous

Motor:
Single-sided axial flux permanent magnet synchronous motors consist of

single stator core and single rotor core and are arranged side by side as shown in the
Fig.1.2. This topology of the axial flux topology is most simple out of all topologies.
But torque producing capability is lower.
4
Fig.1.2. Single sided axial flux permanent magnet machine [1]
1.3.2 Double Sided Axial Flux Permanent Magnet Synchronous

Motor
Double-sided axial flux permanent magnet synchronous motors are of two

types.
1) Internal stator type double-sided axial flux permanent magnet synchronous

motors
2) Internal rotor type double-sided axial flux permanent magnet synchronous motors
Internal stator type double-sided axial flux permanent magnet synchronous

motors consist of single stator core and two rotor cores. The two rotor cores are
arranged on either sides of the stator core as shown in the Fig.1.3a. Similarly, internal
rotor type double-sided axial flux permanent magnet synchronous motors consist of
single rotor core and two stator cores. The rotor core is arranged between the two
stator cores as shown in the Fig.1.3b. The construction of double sided machines is
complex compared to single sided machines. The torque producing capability of
double sided axial flux synchronous machines is higher than that of the single sided
axial flux permanent magnet synchronous machines. Stator is provided with the
armature winding where as rotor is provided with permanent magnets.
5
Fig.1.3. Double-sided axial flux permanent magnet synchronous motor [1]
1.3.3 Power Limitation of Axial Flux Permanent Magnet Machines

and Multi Stage Axial Flux Permanent Magnet Machines
Axial flux permanent magnet disc type brushless machines are available in the
power range of a fraction of watt to sub MW. As the output power of the axial flux
permanent magnet machine increases the contact surface between the rotor and shaft
becomes smaller in comparison with the rotor power. In the higher range of output
power, it is difficult to design a rotor shaft mechanical joint with high mechanical
integrity. A common solution to improvement of mechanical integrity of the rotor
shaft mechanical joint is to design a multi disc (multi stage) / multi slice machine.
Since the scaling of the torque capability of the axial flux permanent magnet
machine as the cube of diameter while that of radial flux permanent magnet machine
is scale as the square of the diameter times the length, the benefits associated with
axial flux geometries may be lost as the power level or geometries ratio of the length
to diameter of the motor is increased. The transition occurs near the point where the
radius equals twice the length of radial flux permanent magnet machine. This may be
the limiting design consideration for power rating of a single stage disc machine.
Hence the power level can always be increased by simply stacking of disc machines
on the same shaft and in the same enclosure. Fig.1.4 shows the multi stage axial flux
permanent magnet machine with 3 stator units and 4 permanent magnet rotor units.
6
Fig.1.4. Multi Stage axial flux permanent magnet synchronous motor [1]
1.4 Speed Control of Permanent Magnet Synchronous Motor
The speed control methods of PMSM are broadly classified into two types 1)
scalar control method 2) vector control method
1.4.1 Scalar Control Method
As its name indicates, in this control strategy the magnitude of the control
variables is controlled and it does not taken care of coupling between the torque and
flux producing channels of the machine. For example, the magnitude of voltage
applied to the machine is controlled to control the flux of the machine, the frequency
of the applied voltage is controlled to control the torque produced in the machine.
Both applied voltage and its frequency can be controlled to control the speed in such
way that their ratio must be constant to keep main field flux constant. This control
strategy is known as volt/Hz method and is widely used still in the applications that
do not require precise speed control. The drawback of such scalar control is its
dynamic performance is poor.
1.4.2 Vector Control Strategy
In this control strategy, the magnitude, frequency and phase of the control
variables are controlled to control the speed of the motor and in this technique, the
torque and flux producing channels are decoupled. The AC motors can be operated
like a separately excited DC motors by using vector control concept. The steady state
7
as well as dynamic performance of the vector controlled motor is good. The vector
control method is also known as field oriented control method.
The block diagram for vector control of 3-phase PMSM is shown in Fig.1.5.
Fig.1.5 Block diagram for vector control of 3-phase PMSM [2]
To control the electrical speed of 3-phase permanent magnet synchronous

motor using vector control strategy, the feedback signals required are actual electrical
speed (ωe), rotor flux position angle (θe), d-axis current or flux producing component
of stator current (id) and q-axis current or torque producing component of stator
current (iq). The feedback signals id, iq are obtained from ia, ib and ic through abc to d-
q transformation block as shown in the Fig.1.5. The abc to d-q transformation block
requires the rotor flux position angle.
The actual speed of the motor is compared with the desired set speed
command and the speed error is given to speed PI controller as shown in Fig.1.5. The
output of the speed PI controller is the desired torque command. The desired torque
command is divided by the torque constant of the machine to get the desired torque
producing component (iq*) of stator current. The desired torque producing current
component (iq*) is compared with actual torque producing current component (iq) as
shown in Fig.1.5. The error signal is given to the PI controller. The output of the PI
controller is given to the decoupling network or non linear feed forward network. The
desired d-axis current is taken as zero and the actual d-axis current of stator is
compared and the error signal is given to PI controller. The output of the PI controller
8
is given to the decoupling network or non linear feed forward network. As the d-axis
current is forced to zero, in this control strategy the armature flux and field fluxes are
oriented in quadrature to each other which results in maximum torque per ampere
operation. The outputs of the decoupling networks are the d-q axes voltages (vd*, vq*)
of stator as shown in Fig.1.5. The desired control voltage signals (va*, vb* and vc*) are
obtained from the d-q axes voltages through dq to abc transformation block. The dq to
abc transformation block requires the rotor flux position angle. The reference control
voltage signals are given to PWM unit to generate the gate trigger pulses to switch the
power switches of H-bridge inverters to get the input phase voltages. The output
voltage of the inverters is applied as input voltage to the motor as shown in the Fig.1.5
to run the motor at desired set speed command.
From the above discussion, it can be concluded that the suitable drive for
most compact, high power, high efficient and high performance applications is the
field oriented controlled multi stage axial flux brushless permanent magnet
synchronous motor drive.
1.5 Objective of the Dissertation
The objective of the dissertation is to achieve the closed loop speed control of
an axial flux 3-stage, 3-phase permanent magnet synchronous motor drive according
to vector control method.
1.6 Literature Review
The various topologies and geometries of axial flux permanent magnet

synchronous machines, details of each topology, and their advantages over radial flux
permanent magnet synchronous machines are presented in [1]. The various
transformations such as Clarke’s, Park’s, inverse Clarke’s and inverse Park’s which
are useful for dynamic modeling of permanent magnet synchronous motor in
synchronously rotating reference frame and for vector control operation of permanent
magnet synchronous motor drive are explained in [2], [3] and [5]. Dynamic modeling
of three phase permanent magnet synchronous motor in synchronously rotating
reference frame is explained in [2], [3], and [6]. The PWM techniques such as bi-polar
sinusoidal PWM technique, uni-polar sinusoidal PWM technique and their
9
comparisons are explained in [4]. The PI controllers design for vector control
operation of three phase permanent magnet synchronous motor is presented in [2].
1.7 Organization of the Dissertation
The rest of the dissertation is structured as follows:

Chapter-2: Chapter-2 deals with the problem description and pulse width modulation
techniques.
Chapter-3: Chapter-3 describes the modeling of the 3-phase permanent magnet
synchronous motor in the synchronous reference frame and Such a dynamic modeling
has been extended to modeling of the 3-stage, 3-phase axial flux permanent magnet
synchronous motor in the synchronously rotating synchronous motor.
Chapter-4: Chapter-4 explains the designing of the inner current loops and outer
speed loop of the vector control of 3-phase PMSM and simulation results of 3-phase
PMSM.
Chapter-5: chapter-5 explains the designing of the inner current loops and outer speed
loop of the vector control of 3-stage, 3-phase PMSM.
Chapter-6: Chapter-6 presents simulation results and analysis of 3-stage, 3-phase
PMSM using Saber-RD simulator.
Chapter-7: The conclusions and future scope of the work is described in this chapter.
Appendix and references follow thereafter.
10
CHAPTER-2
PROBLEM DESCRIPTION AND PWM METHODS
2.1 Description of Problem
Even though the construction of radial flux machines and axial flux machines
are different, the mathematical model for the radial flux PMSM is assumed to be
applicable axial flux PMSM with the axial flux PMSM design parameters.
. The 3-stage axial-flux PMSM is a non salient pole PMSM and it consists of
three stator units and four rotor units. Each stator unit is provided with independent 3-
phase windings and there is no magnetic coupling among the three stator units.
Independent 3-phase winding means, windings are displaced from one another by
1200 electrical in space but the windings are not connected either in star or delta. In
this case, each winding is individually controlled by H-bridge inverters and the
current in one phase is displaced from currents in other phases by 1200 just as in a
conventional three phase system for each stage. The uni-polar sinusoidal PWM
technique is employed to switch the power semiconductor switches of the inverter.
Three phase windings of stator unit-3 are spatially displaced by 400 electrical to those
of stator unit-2 and similarly three phase windings of stator unit-2 are spatially
displaced by 400 electrical to those of stator unit-1. All 4-rotor units are supported on
the single shaft and permanent magnets are arranged on the rotor units.
2.2 Vector Control of Three Stage Axial Flux Permanent Magnet

Synchronous Motor
The block diagram of vector control of 3-stage axial flux permanent magnet
synchronous motor is shown in the Fig.2.1. As the objective is, to control the
electrical speed of three stage axial flux permanent magnet synchronous motor using
vector control strategy.
For this purpose the feedback signals required are actual electrical speed, rotor
flux position angles (θ1, θ2 & θ3) for respective stator units, d-axis current or flux
producing component of current (ids1, ids2 & ids3) of each stator unit and q-axis current
or torque producing component of current (iqs1, iqs2 & iqs3) of each stator unit.
11
q-axis
Speed PI
iqs1* Current PI
+- + V*ds1 V*a1
ωe* Controller - Controller- V*α1
Inverse Inverse
iqs1
1 Decoupl
Park’s Clarke’s V*b1
ωe ing V*β1 Battery
d-axis network V*qs1 transfor transfor
mation mation V*c1
Ids1* = 0 Current PI
Controller-
+
- 1
θ1
ids1 ids1 iqs1 ωe
q-axis
iqs2* Three stage
Current PI *
Independent
+
- Controller- V*ds2 V*α2 Inverse V a2
Inverse Three phase
2 Decoupl Clarke’s V*
iqs2 Park’s b2 PWM PMSM
ing V*β2 transfor
d-axis network V*qs2 transfor mation V* Genera
mation c2 -tor
Current PI H-Bridge
Ids2* = 0 Inverters
Controller-
+
- 2
ids2 θ2
ids2 iqs2 ωe
q-axis
iqs3*
Current PI
+ V*ds3 V*a3
- Controller- V*α3
Inverse Inverse
3 Decoupl
Clarke’s V*b3
iqs3 ing Park’s V*β3
network V*qs3 transfor transfor
d-axis mation mation V*c3
Current PI
+ Controller-3
-
Ids3* = 0
ids3 iqs3 ωe θ3
ids3
θe = θ1 = Rotor
flux position
angle
iα1 ia1
ids1
Park’s Clarke’s
transforma Ttransfor ib1
iqs1 tion iβ1 mation
ic1
θ1
ids2 iα2 ia2

Park’s Clarke’s
ib2
transforma Ttransfor
iqs2 tion iβ2 mation i
c2
θ2
ids3 iα3 ia3

Park’s Clarke’s
transforma Ttransfor ib3
iqs3 tion iβ3 mation ic3
θ3 θ1
ωe
d/dt
θ2
θ2 = (θ1)-(2*pi)/9
θ3
θ3 = (θ1)-(4*pi)/9
Fig.2.1. Block diagram of vector control of 3-stage, 3-phase PMSM
12
The feedback signals ids1, iqs1; ids2, iqs2 & ids3, iqs3 are obtained from ia1, ib1, ic1;
ia2, ib2, ic2 and ia3, ib3, ic3 through Clarke’s and Park’s transformations as shown in the
Fig.2.1. The Park’s transformation requires the rotor flux position angles of respective
stator unit.
The actual speed of the motor is compared with the desired set speed
command and the speed error is given to speed PI controller as shown in Fig.2.1. The
output of the speed PI controller is the desired torque command. As the internal
electromagnetic torque produced by the motor is the sum of the torques produced by
three stator units together, the desired torque command is divided by three to get
desired torque command of each stator unit. The desired torque commands are
divided by the torque constant of the machine to get the desired torque producing
components (i*qs1, i*qs2 & i*qs3) of stator currents. The desired torque producing current
components (i*qs1, i*qs2 & i*qs3) are compared with the respective actual torque
producing current components (iqs1, iqs2 & iqs3) as shown in Fig.2.1. The error signals
are given to the respective current PI controllers. The output of the PI controller is
given to the decoupling network or non linear feed forward network. The desired d-
axis current is taken as zero and the actual d-axis currents of stator units are compared
and the error signals are given to respective current PI controllers. The output of the
PI controller is given to the decoupling network or non linear feed forward network.
As the d-axis current is forced to zero, the armature flux and field fluxes are oriented
in quadrature to each other which results in maximum torque per ampere operation.
The outputs of the decoupling networks are the d-q axes voltages (v*ds1, v*qs1; v*ds2,
v*qs2 & v*ds3, v*qs3) of respective stator units as shown in Fig.2.1. The desired control
voltage signals (v*a1, v*b1, v*c1; v*a2, v*b2, v*c2 & v*a3, v*b3, v*c3) are obtained from the
d-q axes voltages through inverse Park’s and inverse Clarke’s transformations. Inverse
Park’s transformation require rotor flux position angle of the respective stator unit.
The reference control voltage signals are given to uni polar sinusoidal PWM unit to
generate the gate trigger pulses to trigger the power switching devices of H-bridge
inverters to get the input phase voltages to the motor. The output voltage of the
inverters is applied as input voltage to the motor as shown in the Fig.2.1 to run the
motor at desired set speed command. LC filters are used between the H-bridge
inverters and motor phase windings for filtering purpose and to apply better sinusoidal
signals.
13
2.3 PWM Methods
The motor windings are fed from the H-bridge inverters. The H-bridge
inverter is shown in the Fig.2.2. The widely used PWM techniques are 1) sinusoidal
PWM technique 2) third harmonic injected sinusoidal PWM technique 3) space vector
PWM technique. Third harmonic injected sinusoidal PWM technique and space
vector PWM technique are beyond the scope of this dissertation. The sinusoidal PWM
technique is classified in to two types. 1) Bi-polar sinusoidal PWM technique. 2) Uni-
polar sinusoidal PWM technique.
S1 D1 S3 D3
a b
Vout
Vdc
S4 D4 S2 D2
Fig.2.2. H-bridge inverter
2.3.1 Bi-Polar Sinusoidal PWM Technique
The bi-polar sinusoidal PWM switching scheme is given in [4] and is shown in
Fig.2.3. In this scheme, the sinusoidal control voltage signal 𝑣𝑐𝑜𝑛 is compared with
the triangular carrier wave signal 𝑣𝑡𝑟𝑖 as shown in the Fig.2.3. When the control
voltage is greater than the triangular wave, then a pulse will generated and this pulse
will given to turn on the switches s1 and s2 of the H-bridge inverter and no pulse will
given to s3 and s4 of the inverter so that the output voltage across the terminals a-b is
+vd. Similarly when the control voltage is less than the triangular carrier wave, then a
pulse will generated and this pulse will given to turn on the switches s3 and s4 of the
H-bridge inverter and no pulse will given to s1 and s2 of the inverter, so that the output
14
voltage across the terminals a-b are -vd. So that, in case of bi-polar voltage switching
scheme, always, the output voltage across the inverter pulsates between +vd and –vd
as shown in the Fig.2.3. In this scheme, top switch of first leg and bottom switch of
second leg are switched simultaneously similarly top switch of second leg and bottom
switch of first leg are switched simultaneously.
Fig.2.3 Bi-polar sinusoidal PWM voltage switching [4]
In this technique, the inverter output voltage varies from +vd to –vd as shown
in the Fig.2.3b. It is simple and hardware required is less. The harmonics and ripple
content in the output current is high as the output voltage varies from +v d to –vd in
each carrier wave cycle.
2.3.2 Uni-Polar Sinusoidal PWM Technique
The principle of uni-polar sinusoidal PWM switching scheme is given in [4] and
is shown in Fig.2.4. In this scheme, the two legs of the inverter shown in Fig.2.2 are
not controlled simultaneously; instead they are controlled separately by comparing
sinusoidal control voltage signal 𝑣𝑐𝑜𝑛 and its inverse signal −𝑣𝑐𝑜𝑛 with the triangular
carrier wave signal 𝑣𝑡𝑟𝑖 as shown in the Fig.2.4. When 𝑣𝑐𝑜𝑛 is compared with
15
𝑣𝑡𝑟𝑖 triangular wave, will result in the following logic signals to control the power
switches of first leg of inverter.
𝑣𝑐𝑜𝑛 > 𝑣𝑡𝑟𝑖 : trigger pulse is generated for S1 and van = vd
𝑣𝑐𝑜𝑛 < 𝑣𝑡𝑟𝑖 : trigger pulse is generated for S4 and van = 0
Similarly, when −𝑣𝑐𝑜𝑛 is compared with 𝑣𝑡𝑟𝑖 , will results in the following logic
signals to control the power switches of second leg of inverter.
−𝑣𝑐𝑜𝑛 > 𝑣𝑡𝑟𝑖 : trigger pulse is generated for S3 and vbn = vd
−𝑣𝑐𝑜𝑛 < 𝑣𝑡𝑟𝑖 : trigger pulse is generated for S2 and vbn = 0
Fig.2.4 Uni-polar sinusoidal PWM voltage switching [4]
16
The output voltage for the various combinations of switch-on states is given
below.
When trigger pulses are generated for S1, S2: van = vd, vbn = 0 and vout = vd
When trigger pulses are generated for S3, S4: van = 0, vbn = vd and vout = -vd
When trigger pulses are generated for S1, S3: van = vd, vbn = vd and vout = 0
When trigger pulses are generated for S4, S2: van = 0, vbn = 0 and vout = 0
It is observed that in this scheme, when ever, the trigger pulses are generated
either only to upper switches or only to lower switches of the inverter legs, the output
voltage is zero. The load current circulates through S1 and D3 or S3 and D1 depending
on the current direction when trigger pulses are generated only for both the upper
switches. Similarly the load current circulates through the loop S4 and D2 or S2 and D4
depending on the current direction when trigger pulses are generated only for both the
lower switches. During these intervals the output current drawn from source is zero.
This PWM technique is known as uni polar PWM technique because in this scheme,
the output voltage changes between zero and +vd or between zero and –vd as shown in
Fig.2.4d during each switching. In this scheme, the output voltage jump is reduced to
vd compared to 2vd in case of bi polar voltage switching during each switching.
In this technique, the inverter output voltage varies as zero to +vdc to zero during
positive half cycle and zero to –vdc to zero during negative half cycle as shown in the
Fig.2.4d. It is complex and requires more hardware compared to bi-polar sinusoidal
PWM technique. As the output voltage varies from zero to +vdc to zero or zero to –vdc
to zero, the harmonics and ripple content in the output current is less compared to bi-
polar switching.
2.4 Conclusion
Both bi-polar and uni-polar switching schemes produces the same

fundamental output voltages for given modulation index ma but the harmonic content
is less in uni-polar switching compared to bi-polar switching. Therefore, for better
sinusoidal output wave form, uni-polar PWM switching scheme is preferred
compared to bi polar switching scheme.
17
CHAPTER-3
MODELLING OF 3-STAGE, 3-PHASE PERMANENT
MAGNET SYNCHRONOUS MOTOR
3.1 Introduction
Even though the construction of radial flux machines and axial flux machines
is different, the mathematical models for the radial flux PMSM is assumed to be
applicable for axial flux PMSM with the given design parameters.
The mathematical model of axial flux permanent magnet synchronous motor

in the synchronously rotating reference frame is chosen. As the synchronously
rotating reference frame is chosen, the stator quantities are transformed to
synchronously rotating reference frame through Clarke’s and Park’s transformations.
(a) Various reference frames
(b) Combined Vector diagram

Fig.3.1.Representation of Clarke’s and Park’s transformations [5]
18
Clarke’s transformation is used to convert the balanced three phase quantities
in the three phase stationary reference frame into balanced two phase quantities in the
two phase stationary reference frame as shown in Fig.3.1. Park’s transformation
converts balanced 2-phase quantities in the two phase stationary reference frame to dc
quantities in the synchronously rotating reference frame as shown in the Fig.3.1. The
Park’s transformation requires the rotor flux position angle.
3.2 Clarke’s Transformation
The dynamic model of the 3-phase motor can be converted into a dynamic
model of the equivalent 2-phase motor by using Clarke’s transformation. The
equivalence is based on the equality of mmf produced in two phase windings and
three phase windings and equal current magnitudes. Assuming that each of the 3-
phase windings has T turns and equal current magnitudes, the two phase windings
will have 3T/2 turns per phase for M.M.F (magneto motive force) equality.
The common term, the number of turns in the winding, is cancelled on either
of the equations, leaving the current equalities, the relation between the α-β to abc
currents is as follows.
2 1
iα = 3 [𝑖𝑎 − (𝑖𝑏 + 𝑖𝑐 )] (3.1)
2
2 √3
iβ = [ (𝑖𝑏 − 𝑖𝑐 )] (3.2)
3 2
1 1
1 −2 −2 𝑖𝑎
𝑖α 2
[𝑖 ] = [ ] [ 𝑖𝑏 ] (3.3)
β 3 √3 √3
0 2
− 2 𝑖𝑐
Where
𝑖𝑎 , 𝑖𝑏 𝑎𝑛𝑑 𝑖𝑐 are the three phase currents
𝑖α , 𝑖β are the stationary orthogonal reference frame currents
When 𝑖α is superposed with 𝑖a , and , 𝑖𝑎 , 𝑖𝑏 𝑎𝑛𝑑 𝑖𝑐 can be transformed to 𝑖α , 𝑖β as
𝑖α = 𝑖 a (3.4)
19
1
iβ = [ (𝑖𝑎 + 2𝑖𝑏 )] (3.5)
√3
Where 𝑖𝑎 + 𝑖𝑏 + 𝑖𝑐 = 0 (3.6)
The above equations are also applicable to voltages and fluxes as well.
3.2.1 Power Invariance
The power input to the three phase machine has to be equal to the power input
to the 2-phase machine to have the meaningful interpretation in the modeling and
analysis. For a balanced three phase machine, the power input is compactly
represented as
3
𝑃𝑖𝑛 = 2 ( 𝑣𝛼 𝑖𝛼 + 𝑣𝛽 𝑖𝛽 ) (3.7)
Fig.3.2 Clarke’s Transformation

3.3 Park’s Transformation (α, β to d, q)
The two-axis orthogonal stationary reference frame quantities are transformed
into rotating reference frame quantities using Park’s transformation as shown in
Fig.3.3. The Park’s transformation is expressed by the following equations
i𝑑=i𝛼∗cos (𝜃e) + i𝛽∗sin (𝜃e) (3.8)

i𝑞=i𝛽∗cos (𝜃e) − i𝛼∗sin (𝜃e) (3.9)
𝑖𝑑 cos 𝜃𝑒 sin 𝜃𝑒 𝑖𝛼
[𝑖 ] = [ ][ ] (3.10)
𝑞 − sin 𝜃𝑒 cos 𝜃𝑒 𝑖𝛽
The same transformation is also applicable for voltages and fluxes.
Where,
id, iq are rotating reference frame quantities
iα, iβ are orthogonal stationary reference frame quantities
20
𝜃𝑒 is the electrical rotor angle as defined in Fig.3.3
Fig.3.3 Park’s transformation
3.4 Dynamic Modeling of 3-phase PMSM in (Synchronously rotating)

d, q-Reference Frame
The model of PMSM without damper winding on synchronously rotating

reference frame with the following assumptions is presented below.
1) Saturation is neglected.
2) The induced EMF is sinusoidal.
3) Eddy currents and hysteresis losses are negligible.
4) There are no field current dynamics.
The dynamic modeling of PMSM at synchronously rotating reference frame is
presented in [2], [6] and [3]. Synchronously rotating reference frame is chosen, as
there is a simple relation between the electromagnetic torque (Te) and quadrature axis
current (iq) and more over, the voltage and current variables at the synchronously
rotating reference frame are no longer sinusoidal quantities, instead they are DC
quantities.
Voltage equations of the 3-phase PMSM in the synchronously rotating
reference frame are given by
𝑣𝑞 = 𝑅𝑠 𝑖𝑞 + 𝜔𝑒 𝜆𝑑 + 𝜌𝜆𝑞 (3.11)
𝑣𝑑 = 𝑅𝑠 𝑖𝑑 − 𝜔𝑒 𝜆𝑞 + 𝜌𝜆𝑑 (3.12)
Where
vd = Direct axis voltage
21
vq = Quadrature axis voltage
id = Direct axis current
iq = Quadrature axis current
ωe = Electrical speed in rad/sec
ρ = d/dt
Flux linkages are given by
𝜆𝑞 = 𝐿𝑞 𝑖𝑞 (3.13)
𝜆𝑑 = 𝐿𝑑 𝑖𝑑 + 𝜆𝑓 (3.14)
Where
λd = Direct axis flux linkages
λq = Quadrature axis flux linkages
Substituting equations (3.13) and (3.14) into equations (3.11) and (3.12)
𝑣𝑞 = 𝑅𝑠 𝑖𝑞 + 𝜔𝑒 𝐿𝑑 𝑖𝑑 + 𝜔𝑒 𝜆𝑓 + 𝜌𝐿𝑞 𝑖𝑞 (3.15)
𝑣𝑑 = 𝑅𝑠 𝑖𝑑 − 𝜔𝑒 𝐿𝑞 𝑖𝑞 + 𝜌𝐿𝑑 𝑖𝑑 + 𝜌 𝜆𝑓 (3.16)
Arranging equations (3.15) and (3.16) in matrix form
𝑣𝑞 𝑅𝑠 + 𝜌𝐿𝑞 𝜔𝑒 𝐿𝑑 𝑖𝑞 𝜔𝑒
[𝑣 ] = [ ] [ ] + [ 𝜌 ] 𝜆𝑓 (3.17)
𝑑 −𝜔𝑒 𝐿𝑞 𝑅𝑠 + 𝜌𝐿𝑑 𝑖𝑑
The electromagnetic torque developed by the motor is given by
3𝑝
𝑇𝑒 = (𝜆𝑑 𝑖𝑞 − 𝜆𝑞 𝑖𝑑 ) (3.18)
22
3𝑝
𝑇𝑒 = [(𝐿𝑑 𝑖𝑑 + 𝜆𝑓 )𝑖𝑞 − 𝐿𝑞 𝑖𝑞 𝑖𝑑 ] (3.19)
22
3𝑝
𝑇𝑒 = [𝜆𝑓 𝑖𝑞 + (𝐿𝑑 − 𝐿𝑞 )𝑖𝑞 𝑖𝑑 ] (3.20)
22
As the machine is a non salient pole motor, Ld = Lq = Ls

3𝑝
𝑇𝑒 = (𝜆𝑓 𝑖𝑞 ) (3.21)
22
𝟑
The term 𝟐 in the above equation, indicates the power equivalence between the
3-phase and two phase models.

From equation (3.21), it can be concluded that the torque equation of PMSM
is similar to that of torque equation of separately excited DC motor. So that, while
keeping field flux as constant, electromagnetic torque of PMSM can be controlled by
controlling iq.
The dynamical equation for PMSM is
𝑑
𝑇𝑒 = 𝑇𝑙 + 𝐵𝜔𝑟 + 𝐽 𝑑𝑡 (𝜔𝑟 ) (3.22)
Solving for the rotor mechanical speed form equation (3.22)
22
1
𝜔𝑟 = 𝐽
∫(𝑇𝑒 − 𝑇𝑙 − 𝐵𝜔𝑟 ) dt (3.23)
𝑝
𝜔𝑒 = 2 𝜔𝑟 (3.24)
𝑑
(𝜃𝑒 ) = 𝜔𝑒 (3.25)
𝑑𝑡
𝜔𝑒 = Excitation frequency, electrical rad/sec (= PM motor rotor speed)
𝜔𝑟 = Rotor speed in mechanical rad/sec
p = Number poles of the motor
𝜃𝑒 = Electrical rotor flux position angle
3.5 Dynamic Modeling of 3-stage, 3-phase PMSM in (Synchronously

rotating) d, q-Reference Frame
The 3-stage axial-flux PMSM is a non salient pole PMSM and it consists of
three stator units and four rotor units. Each stator unit is provided with independent 3-
phase windings and there is no magnetic coupling among the three stator units. Three
phase windings of stator unit-3 are spatially displaced by 400 electrical to those of
stator unit-2 and similarly three phase windings of stator unit-2 are spatially displaced
by 400 electrical to those of stator unit-1. To represent this, the derivative of electrical
speed is directly taken as the rotor flux position angle (𝜃1 ) for Park’s transformation
used for stator unit-1. The rotor flux position angle (𝜃2 ) for Park’s transformation
used for stator unit-2 is obtained from 𝜃1 by subtracting 400 electrical from it.
Similarly, the rotor flux position angle (𝜃3 ) for Park’s transformation used for stator
unit-3 is obtained from 𝜃1 by subtracting 800 electrical from it. Therefore the three
phase supply of stator unit-3 is displaced by 400 electrical to the three phase supply of
stator unit-2 and similarly three phase supply of stator unit-2 is displaced by 400
electrical to the three phase supply of stator unit-1. This can be achieved by using the
respective rotor flux position angles in the inverse Park’s transformation blocks used
in the controller part. All 4 rotors are supported on the same shaft. The permanent
magnets are provided on the rotor units.
Voltage equations of the axial flux 3-stage PMSM in the synchronously

rotating reference frame are given by
𝑣𝑞𝑠1 = 𝑅𝑠 𝑖𝑞𝑠1 + 𝜔𝑒 𝜆𝑑𝑠1 + 𝜌𝜆𝑞𝑠1 (3.26)
𝑣𝑑𝑠1 = 𝑅𝑠 𝑖𝑑𝑠1 − 𝜔𝑒 𝜆𝑞𝑠1 + 𝜌𝜆𝑑𝑠1 (3.27)
23
Where
vds1 = Direct axis voltage of stator unit-1 in volt
vqs1 = Quadrature axis voltage of stator unit-1 in volts in volt
ids1 = Direct axis current of stator unit-1 in ampere
iqs1 = Quadrature axis current of stator unit-1 in ampere
ωe = Electrical speed in rad/sec
ρ = d/dt
Flux linkages are given by
𝜆𝑞𝑠1 = 𝐿𝑞 𝑖𝑞𝑠1 (3.32)
𝜆𝑑𝑠1 = 𝐿𝑑 𝑖𝑑𝑠1 + 𝜆𝑓 (3.33)
𝜆𝑞𝑠2 = 𝐿𝑞 𝑖𝑞𝑠2 (3.34)
𝜆𝑑𝑠2 = 𝐿𝑑 𝑖𝑑𝑠2 + 𝜆𝑓 (3.35)
𝜆𝑞𝑠3 = 𝐿𝑞 𝑖𝑞𝑠3 (3.36)
𝜆𝑑𝑠3 = 𝐿𝑑 𝑖𝑑𝑠3 + 𝜆𝑓 (3.37)
Where
λds1 = Direct axis flux linkages of stator unit-1
λqs1 = Quadrature axis flux linkages of stator unit-1
24
λf =Amplitude of permanent magnet flux
Rs = Stator resistance
Ld = Direct axis inductance
Lq = Quadrature axis inductance
3.5.1 Mathematical Equations of Stage-1

𝑣𝑞𝑠1 = 𝑅𝑠 𝑖𝑞𝑠1 + 𝜔𝑒 𝐿𝑑 𝑖𝑑𝑠1 + 𝜔𝑒 𝜆𝑓 + 𝜌𝐿𝑞 𝑖𝑞𝑠1 (3.38)
𝑣𝑑𝑠1 = 𝑅𝑠 𝑖𝑑𝑠1 − 𝜔𝑒 𝐿𝑞 𝑖𝑞𝑠1 + 𝜌𝐿𝑑 𝑖𝑑𝑠1 + 𝜌 𝜆𝑓 (3.39)
(Since there is no magnetic coupling among the stator units)
𝑣𝑞𝑠1 𝑅𝑠 + 𝜌𝐿𝑞 𝜔𝑒 𝐿𝑑 𝑖𝑞𝑠1 𝜔𝑒
[𝑣 ] = [ ][ ] + [ 𝜌 ] 𝜆𝑓 (3.40)
𝑑𝑠1 −𝜔𝑒 𝐿𝑞 𝑅𝑠 + 𝜌𝐿𝑑 𝑖𝑑𝑠1
The torque developed by the stator unit-1 of the motor is given by
3𝑝
𝑇𝑒1 = (𝜆𝑑𝑠1 𝑖𝑞𝑠1 − 𝜆𝑞𝑠1 𝑖𝑑𝑠1 ) (3.41)
22
3𝑝
𝑇𝑒1 = [(𝐿𝑑 𝑖𝑑𝑠1 + 𝜆𝑓 )𝑖𝑞𝑠1 − 𝐿𝑞 𝑖𝑞𝑠1 𝑖𝑑𝑠1 ] (3.42)
22
3𝑝
𝑇𝑒1 = [𝜆𝑓 𝑖𝑞𝑠1 + (𝐿𝑑 − 𝐿𝑞 )𝑖𝑞𝑠1 𝑖𝑑𝑠1 ] (3.43)
22

3𝑝
𝑇𝑒1 = (𝜆𝑓 𝑖𝑞𝑠1 ) (3.44)
22

[𝑣 ] = [ ][ ] + [ 𝜌 ] 𝜆𝑓 (3.47)
3𝑝
𝑇𝑒2 = (𝜆𝑑𝑠2 𝑖𝑞𝑠2 − 𝜆𝑞𝑠2 𝑖𝑑𝑠2 ) (3.48)
22
3𝑝
𝑇𝑒2 = 22
[(𝐿𝑑 𝑖𝑑𝑠2 + 𝜆𝑓 )𝑖𝑞𝑠2 − 𝐿𝑞 𝑖𝑞𝑠2 𝑖𝑑𝑠2 ] (3.49)
3𝑝
22
25
3𝑝
𝑇𝑒2 = (𝜆𝑓 𝑖𝑞𝑠2 ) (3.51)
22

[𝑣 ] = [ ][ ] + [ 𝜌 ] 𝜆𝑓 (3.54)
3𝑝
𝑇𝑒3 = (𝜆𝑑𝑠3 𝑖𝑞𝑠3 − 𝜆𝑞𝑠3 𝑖𝑑𝑠3 ) (3.55)
22
3𝑝
𝑇𝑒3 = [(𝐿𝑑 𝑖𝑑𝑠3 + 𝜆𝑓 )𝑖𝑞𝑠3 − 𝐿𝑞 𝑖𝑞𝑠3 𝑖𝑑𝑠3 ] (3.56)
22
3𝑝
22

3𝑝
𝑇𝑒3 = (𝜆𝑓 𝑖𝑞𝑠3 ) (3.58)
22
𝟑
The term 𝟐 in the above equation, indicates the power equivalence between the
3-phase and two phase models.

From the equations (3.44), (3.51) and (3.58) it can be concluded that the
torque equation of each stator unit is similar to that of torque equation of separately
excited DC motor. Therefore, while keeping field flux as constant, the
electromagnetic torque developed can be controlled by controlling quadrature axis
current.
The total internal electromagnetic torque developed by three stage permanent
magnet synchronous motor is given by the
𝑇𝑒 = 𝑇𝑒1 + 𝑇𝑒2 + 𝑇𝑒3 (3.59)
The dynamical mechanical equation for PMSM is
𝑑
𝑇𝑒 = 𝑇𝑙 + 𝐵𝜔𝑟 + 𝐽 (𝜔𝑟 ) (3.60)
𝑑𝑡
Solving for the rotor mechanical speed form equation (3.60)
1
𝜔𝑟 = 𝐽
∫(𝑇𝑒 − 𝑇𝑙 − 𝐵𝜔𝑟 ) dt (3.61)
𝑝
𝜔𝑒 = 𝜔𝑟 (3.62)
2
𝑑
(𝜃1 ) = 𝜔𝑒 (3.63)
𝑑𝑡
26
2𝜋
𝜃2 = 𝜃1 − (3.64)
9
4𝜋
𝜃3 = 𝜃2 − (3.65)
9
Where
𝜔𝑒 = rotor speed in electrical rad/sec
𝜔𝑟 = the rotor speed mechanical rad/sec
p = number poles of the motor
𝜃1 = rotor flux position angle with respect to stator unit -1
3.6 Summary
In this chapter making the use of Clarke’s transformation and Park’s

transformation, dynamic modelling of 3-phase permanent magnet synchronous motor
according to synchronously rotating d-q reference frame has been presented
[Appendix A-1]. Such a dynamic model has been extended to 3-stage, three phase
permanent magnet synchronous motor [Appendix A-2].
27
CHAPTER-4
DESIGN OF PI CONTROLLERS FOR VECTOR
CONTROL OPERATION OF 3-PHASE PMSM
4.1 Introduction
Vector control concept is most popular and widely used control algorithm of
the permanent magnet synchronous motor drives. Proper calculation of PI-controller
parameters ensures correct operation of vector controller and high precision of the
drive. Inputs to the control algorithms for calculation of proportional gain (P
component) and integral time constant (I component) are motor parameters together
with damping ratio, natural frequencies of current and speed control loops, which can
be chosen to tune PI controllers.
4.2 Design of PI Controller
Let us consider a simple closed loop PI controller system as shown in Fig.4.1.

Let us assume that the plant transfer function is given by
𝑐
G(s) = 𝑠+𝑏 (4.1)
The transfer function of PI controller is represented as
C(s) = Kp (1+1/𝜏is) (4.2)

C(s) = Kp + Kp/𝜏is (4.3)
C(s) = Kp + Ki/s (4.4)
sKp +Ki
C(s) = (4.5)
𝑠
Where KP is the proportional gain and 𝜏i = Kp/Ki is the integral time constant of PI
controller.
Therefore the closed loop transfer function of the Fig.4.1 is given by
𝑐 sKp +Ki
𝑌(𝑠) 𝐺(𝑠)∗𝐶(𝑠) ∗
𝑠+𝑏 𝑠
= = 𝑐 sK p +Ki
(4.6)
𝑅(𝑠) 1+𝐺(𝑠)𝐶(𝑠) 1+ ∗
𝑠+𝑏 𝑠
28
R(s) Y(s)
input PI Controller Plant output
+
- C(s) = Kp (1+1/𝜏is ) G(s) = c / s+b
Fig.4.1. Simple PI controller closed loop system
𝒄(sKp +Ki )
= (4.7)
𝒔(𝒔+𝒃)+𝒄(sKp +Ki )
From the equation (4.7), the closed loop characteristic equation of the system
shown in fig.4.1 is given by
𝒔(𝒔 + 𝒃) + 𝒄(sK p + K i ) = 0 (4.8)
We know that, the location of the poles of the above transfer function
determines the closed loop stability, disturbance rejection and system response to the
reference signal. In equation (4.8), if the plant parameters are known the free
parameters are the controller parameters KP and Ki. So find the controller parameters,
we can equate the actual closed loop characteristic equation to the desired closed loop
characteristic equation as shown below.
𝒔(𝒔 + 𝒃) + 𝒄(sK p + K i ) = 𝒔𝟐 + 𝟐𝝃𝝎𝒏 𝒔 + 𝝎𝟐𝒏 (4.9)
The above equation is simplified in such a way that the coefficients if s2 on

both sides should be equal to unity. Now, by comparing coefficients of ‘s’ on both
sides of equation (4.9), we can get
b + cKp = 𝟐𝝃𝝎𝒏 (4.10)
By comparing constants on both sides of equation (4.9), we can get
cKi = 𝝎𝟐𝒏 (4.11)
From equation (4.10), the proportional gain of the PI controller is calculated as
2𝜉𝜔𝑛 −𝑏
Kp = (4.12)
𝑐
From equation (4.11), the integral gain of the PI controller is calculated as
29
2
𝜔𝑛
Ki = (4.13)
𝑐
And integral time constant is 𝜏i = Kp/Ki
2𝜉𝜔𝑛 −𝑏
𝜏i = (4.14)
𝝎𝟐𝒏
So from the equations (4.12) and (4.13), it can be concluded that, if the plant
parameters a and b are known, the PI controller parameters are calculated by choosing
the values of ξ (damping ratio) and 𝜔𝑛 (natural frequency) of desired stable closed
loop characteristic polynomial.
4.3 Vector Control System Development
The vector control of 3-phase PMSM consists of two current loops namely
direct axis current loop and quadrature axis current loop and a speed loop. The
quadrature axis current loop is cascaded with the speed loop. Hence, the speed loop is
usually called as outer speed loop and q-axis current loop is called as inner current
loop. The outer speed loop consists of a PI controller for better speed tracking. Each
current loop consists of a PI controller for achieving faster dynamic response. The d-
axis PI controller is used to force the d-axis current to zero which results in maximum
torque/ampere operation.
4.3.1 Current PI Controllers Design

From the above discussion, to design a PI controller, plant transfer function is
required. The plant here is the axial flux 3-stage permanent magnet synchronous
motor. As the mathematical models for both radial and axial flux motors are same and
each stage is a 3-phase one. Design of PI controllers for 3-phase PMSM is presented
in [2] and explained in this section. The concepts are extended for designing the PI
controllers of 3-stage, 3-phase machine.
4.3.1.1 D-axis Current PI Controller Design
The d-axis voltage equation of the 3-phase PMSM is given by
𝑣𝑑 = 𝑅𝑠 𝑖𝑑 − 𝜔𝑒 𝐿𝑞 𝑖𝑞 + 𝜌𝐿𝑑 𝑖𝑑 + 𝜌 𝜆𝑓 (4.15)
30
As the 𝜆𝑓 = Amplitude of permanent magnet flux and as it is constant its
derivative is zero. The equation (4.15) is written as
𝑣𝑑 = 𝑅𝑠 𝑖𝑑 − 𝜔𝑒 𝐿𝑞 𝑖𝑞 + 𝜌𝐿𝑑 𝑖𝑑 (4.16)
By applying the Laplace transform to equation (4.16), we can write
𝑣𝑑 (𝑠) = 𝑅𝑠 𝑖𝑑 (𝑠) − 𝜔𝑒 𝐿𝑞 𝑖𝑞 (𝑠) + 𝑠𝐿𝑑 𝑖𝑑 (𝑠) (4.17)
The above equation can be represented in a block diagram form as shown in

Fig.4.2.
Fig.4.2 Block diagram of permanent magnet synchronous motor: D-axis
Therefore the transfer function of d-axis current loop of PMSM is given by
11 1 1
𝑖𝑑 (𝑠) 𝐿𝑑 𝑠 𝑠𝐿𝑑 1 𝐿𝑑
= 𝑅 1 = 𝑠𝐿𝑑 +𝑅𝑠 = = 𝑅 (4.18)
𝑉𝑑 (𝑠) 1+ 𝑠 𝑅𝑠 +𝑠𝐿𝑑 𝑠+ 𝑠
𝐿𝑑 𝑠 𝑠𝐿𝑑 𝐿𝑑
Therefore by comparing equation (1) with equation (4.18), we can write the
plant parameters as
1 𝑅
c=𝐿 and b = 𝐿𝑠 (4.19)
𝑑 𝑑
And the closed loop d-axis current loop is shown in Fig.4.3
31
Fig.4.3 Direct-axis current loop of PMSM
Therefore, the PI controller gains according to the equations (4.12), (4.13) and
(4.14) are calculated as
𝑅𝑠
2𝜉𝜔𝑛 −𝑏 2𝜉𝜔𝑛 −
𝐿𝑑
Kpd = = 1 (4.20)
𝑐
𝐿𝑑
Kpd = 2𝜉𝜔𝑛 𝐿𝑑 -Rs (4.21)
2
𝜔𝑛 2
𝜔𝑛
Kid = = 1 (4.22)
𝑐
𝐿𝑑
Kid = 𝐿𝑑 𝜔𝑛2 (4.23)
And integral time constant is 𝜏id = Kpd/Kid
2𝜉𝜔𝑛 𝐿𝑑 −Rs
𝜏id = 2
𝐿 𝑑 𝜔𝑛
(4.24)
Where 𝜉 is the damping ratio and 𝜔𝑛 is the natural frequency of d- axis current
loop.
4.3.1.2: Q-axis Current PI Controller Design
The q-axis voltage equation of the 3-phase PMSM is given by
𝑣𝑞 = 𝑅𝑠 𝑖𝑞 + 𝜔𝑒 𝐿𝑑 𝑖𝑑 + 𝜔𝑒 𝜆𝑓 + 𝜌𝐿𝑞 𝑖𝑞 (4.25)
𝑣𝑞 = 𝑅𝑠 𝑖𝑞 + 𝜔𝑒 𝜆𝑑 + 𝜌𝐿𝑞 𝑖𝑞 (4.26)
By applying the Laplace transform to equation (4.24), we can write
32
𝑣𝑞 (𝑠) = 𝑅𝑠 𝑖𝑞 (𝑠) + 𝜔𝑒 𝜆𝑑 (𝑠) + 𝑠𝐿𝑞 𝑖𝑞 (𝑠) (4.27)
The above equation can be represented in a block diagram form as shown in

Fig. 4.4.
Fig.4.4 Block diagram of permanent magnet synchronous motor: Q-axis
Therefore the transfer function of q-axis current loop of PMSM is given by
11 1 1
𝑖𝑞 (𝑠) 𝐿𝑞 𝑠 𝑠𝐿𝑞 1 𝐿𝑞
= 𝑅 1 = 𝑠𝐿𝑞+𝑅𝑠 = = 𝑅 (4.28)
𝑉𝑞 (𝑠) 1+ 𝑠 𝑅𝑠 +𝑠𝐿𝑞 𝑠+ 𝑠
𝐿𝑞 𝑠 𝑠𝐿𝑞 𝐿𝑞
Therefore by comparing equation (4.1) with equation (4.28), we can write the
plant parameters as
1 𝑅
c=𝐿 and b = 𝐿𝑠 (4.29)
𝑞 𝑞
And the closed loop q-axis current loop is shown in Fig.4.5
Fig.4.5 Quadrature axis current loop of PMSM
33
𝑅
2𝜉𝜔𝑛 − 𝐿 𝑠
2𝜉𝜔𝑛 −𝑏 𝑞
Kpq = = 1 (4.30)
𝑐
𝐿𝑞
Kpq = 2𝜉𝜔𝑛 𝐿𝑞 -Rs (4.31)
2
𝜔𝑛 2
𝜔𝑛
Kiq = = 1 (4.32)
𝑐
𝐿𝑞
Kiq = 𝐿𝑞 𝜔𝑛2 (4.33)
And integral time constant is 𝜏iq = Kpq/Kiq
2𝜉𝜔𝑛 𝐿𝑞 −Rs
𝜏iq = 2
𝐿 𝑞 𝜔𝑛
(4.34)
Where 𝜉 is the damping ratio and 𝜔𝑛 is the natural frequency of inner q- axis
current loop. So from the equations (4.31) and (4.33), and (4.34) it can be concluded
that, if the plant parameters Lq and Rs are known, the PI controller parameters are
calculated by choosing the values of ξ (damping ratio) and 𝜔𝑛 (natural frequency) of
desired stable closed loop characteristic polynomial.
4.3.1.3 Decoupling Network Formulation:
There are nonlinear cross coupling terms in the equations (4.17) and (4.27)
which are shown in the block diagrams shown in Fig.4.2 and Fig.4.4. These nonlinear
cross coupling terms are eliminated from the PI controller output to get the actual d-
axis and q-axis voltages. This technique is known as decoupling using feed forward
manipulation and is also known as input and output linearization.
Vd’= (Vd + 𝜔𝑒 𝐿𝑞 𝑖𝑞 ) (4.35)
Vd = Vd’ - 𝜔𝑒 𝐿𝑞 𝑖𝑞 (4.36)
Vq’ = (Vq −𝜔𝑒 𝐿𝑑 𝑖𝑑 −𝜔𝑒 𝜆𝑓 ). (4.37)
34
Vq = Vq’ + 𝜔𝑒 𝐿𝑑 𝑖𝑑 + 𝜔𝑒 𝜆𝑓 (4.38)
Where Vd’= D-axis current PI controller output

Vq’= Q-axis current PI controller output
With the proportional gain and integral time constant, the PI controller for the
d-axis current control loop has the form
𝐾𝑝𝑑
Vd’ = Kpd (id*(t) - id (t)) + ((id*( 𝜏) - id (𝜏)) (4.39)
𝜏𝑖𝑑
From the equation (4.36), the d-axis control voltage signal is calculated as
𝐾𝑝𝑑
Vd = Kpd (id*(t) - id (t)) + ((id*( 𝜏) - id (𝜏)) - 𝜔𝑒 𝐿𝑞 𝑖𝑞 (4.40)
𝜏𝑖𝑑
Similarly With the proportional gain and integral time constant, the PI
controller for the q-axis current control loop has the form
𝐾𝑝𝑞
Vq’ = Kpq (iq*(t) – iq (t)) + ((iq*( 𝜏) – iq (𝜏)) (4.41)
𝜏𝑖𝑞
From the equation (4.38), the q-axis control voltage signal is calculated as
𝐾𝑝𝑞
Vq = Kpq (iq*(t) – iq (t)) + ((iq*( 𝜏) – iq (𝜏)) + 𝜔𝑒 𝐿𝑑 𝑖𝑑 + 𝜔𝑒 𝜆𝑓 (4.42)
𝜏𝑖𝑞
The PI controller along with feed forward compensation for d-axis current
control loop is shown in the Fig.4.6
id* PI Controller Vd’(t) Vd(t)

+ +
- C(s) = Kpd +Kpd/𝜏id ʃ -
id
ωe(t)Lqiq(t)
Fig.4.6 The PI controller along with feed forward compensation for d-axis current
control loop
35
Similarly, the PI controller along with feed forward compensation for q-axis
current control loop is shown in the Fig.4.7
ωe(t)Ldid(t)
Te* Vq’(t) Vq(t)

iq* PI Controller +
1/(3/2*p/2*λf) + C(s) = Kpq +Kpq/𝜏iqʃ
+
- +
iq
ωe(t)λd
Fig.4.7 The PI controller along with feed forward compensation for electromagnetic
torue control for three phase PMSM
4.3.2 Design of Speed PI Controller
For designing the outer speed loop PI controller, it is necessary to write

transfer function between the reference quadrature axis reference current iq*(s) and
actual q-axis current iq(s).
As we have provided the compensation for nonlinear cross coupling terms by

feed forward technique, the nonlinear cross coupling terms are cancelled each other.
So that the q-axis PI controller output is given as
𝐾𝑝𝑞 𝑡
Vqs’ = Kpq (iq*(t) – iq (t)) + 𝜏 𝑞 ∫0 𝑖𝑞∗ (𝜏) – iq(𝜏)) (4.43)
𝑖
𝑑 𝐾𝑝𝑞 𝑡
Lq 𝑑𝑡iq(t) + Rs iq(t) = Kpq (iq*(t) – iq (t)) + 𝜏 ∫0 𝑖𝑞∗ (𝜏) – iq(𝜏)) (4.44)
𝑖𝑞
𝑑 1 𝐾𝑝𝑞 𝑡
𝑑𝑡
iq(t) = 𝐿𝑞
(-Rs iq(t) + Kpq (iq*(t) – iq (t)) + 𝜏𝑖𝑞
∫0 𝑖𝑞∗ (𝜏) – iq(𝜏)) (4.45)
By applying Laplace transform to equation (4.45), we can get
36
𝑅 𝐾𝑝𝑞 𝐾
s iq(s) = - 𝐿 𝑠 iq(s) + (iq*(s) – iq (s)) + 𝑠𝐿𝑖𝑞(iq*( 𝑠) – iq(𝑠)) (4.46)
𝑞 𝐿𝑞 𝑞
From equations (4.31) and (4.33) Kpq = 2𝜉𝜔𝑛 𝐿𝑞 -Rs and Kiq = 𝐿𝑞 𝜔𝑛2
𝑅 𝑅 2
𝜔𝑛
s iq(s) = - 𝐿 𝑠 iq(s) + 2𝜉𝜔𝑛 - 𝐿 𝑠 (iq*(s) – iq (s)) + (iq*( 𝑠) – iq(𝑠)) (4.47)
𝑞 𝑞 𝑠
𝑅𝑠 𝑅 2
𝜔𝑛 𝑅 2
𝜔𝑛
s iq(s) + iq(s) + 2𝜉𝜔𝑛 - 𝐿 𝑠 (iq(𝑠)) + iq(𝑠) = 2𝜉𝜔𝑛 - 𝐿 𝑠 (iq*(𝑠)) + iq*(𝑠)
𝐿𝑞 𝑞 𝑠 𝑞 𝑠
(4.48)
𝑅
(𝑠 2 + 2𝜉𝜔𝑛 𝑠 + 𝜔𝑛2 ) iq(s) = ((2𝜉𝜔𝑛 - 𝐿 𝑠 )s + 𝜔𝑛2 ) iq*(𝑠) (4.49)
𝑞
𝑅
𝑖𝑞 (𝑠) ( 2𝜉𝜔𝑛 − 𝑠 )s + 𝜔𝑛
2 )
𝐿𝑞
= (4.50)
𝑖𝑞∗ (𝑠) 2
𝑠2 +2𝜉𝜔𝑛 𝑠+𝜔𝑛
In order to design the outer speed loop PI controller, we have to write the
transfer function between quadrature axis current iq(s), and electrical speed ωe(s) of
the motor in rad/sec. The mechanical balance equation of the PMSM is given by
𝑑
Te (t) = TL + B ωr (t) + j𝑑𝑡 ωr (t) (4.51)
3𝑝 𝑑
𝜆 𝑖 (𝑡) = TL + B ωr (t) + j𝑑𝑡 ωr (t) (4.52)
22 𝑓 𝑞
3𝑝 𝑑
𝜆 𝑖 (𝑡) = TL + B ωr (t) + j𝑑𝑡 ωr (t) (4.53)
22 𝑓 𝑞
Now multiply with (p/2) on both sides of equation (4.53)
3 𝑝 2 𝑑
( ) 𝜆𝑓 𝑖𝑞 (𝑡) = (p/2)*TL + B *(p/2)*ωr (t) + j𝑑𝑡 (p/2)*ωr (t) (4.54)
2 2
3 𝑝 2 𝑑
( ) 𝜆𝑓 𝑖𝑞 (𝑡) = (p/2)*TL + B *ωe (t) + j𝑑𝑡 ωe (t) (4.55)
2 2
The load torque on the motor is considered as a disturbance in the control

system design and it should be rejected by the outer speed loop. So that
3 𝑝 2 𝑑
( ) 𝜆𝑓 𝑖𝑞 (𝑡)
2 2
= B *ωe (t) + j𝑑𝑡 ωe (t) (4.56)
By applying the Laplace transforms on both sides of equation (4.56)
37
3 𝑝 2
( ) 𝜆𝑓 𝑖𝑞 (𝑠) = B *ωe (s) + js ωe (s) (4.57)
2 2
3 𝑝 2
( ) 𝜆𝑓 𝑖𝑞 (𝑠) = (B + js) ωe (s) (4.58)
2 2
3 𝑝 2
( ) 𝜆𝑓
2 2
𝑖𝑞 (𝑠) = (B /j + s) ωe (s) (4.59)
𝑗
3 𝑝 2
( ) 𝜆𝑓
2 2 ⁄
𝑗 ωe (s)
B = (4.60)
( + s) 𝑖𝑞 (𝑠)
j
By substituting equation (4.50) into (4.60), we have
3 𝑝 2
( ) 𝜆𝑓 𝑅
2 2 ⁄ 2𝜉𝜔𝑛 − 𝑠 )s + 𝜔𝑛
2
ωe (s) 𝑗 𝐿𝑞
=( B )( ) (4.61)
𝑖𝑞∗ (𝑠) (s+ )
2
𝑠2 +2𝜉𝜔𝑛 𝑠+𝜔𝑛
j
The transfer function shown in above equation 4.61 is a third order model. But
B
to design a PI controller, first order model is needed. If the value of 𝜔𝑛 >> , the
j
above third order equation can be approximated as first order equation from the
concept of dominant pole, because 𝜔𝑛 is the natural frequency of the inner current
loop. Therefore, dynamics from the inner current loop can be neglected. Therefore,
simplified first order transfer function is expressed as
3 𝑝 2
( ) 𝜆𝑓
2 2 ⁄
ωe (s) 𝑗
≈ ( B ) (4.62)
𝑖𝑞∗ (𝑠) (s+ )
j
The outer speed loop of 3-phase permanent magnet synchronous motor is

shown in Fig.4.8
38
ωe(s)Ldid(s) 𝑅𝑠
iq* Vq’(s) Te(S) ωe(s)

ωe*(s) Vq(s) iq(s)
speed PI Controller Te* q-axis PI Controller 1 + - 1 1
+ 1/(3/2*p/2*λf) + + + (3/2*p/2*λf) (p/2) / j(s+ (B/j))
- C(s) = Kps (1+1/𝜏iss ) - C(s) = Kpq (1+1/𝜏iqs ) + - 𝐿𝑞 𝑆
ωe(s) iq
ωe(s)λf ωe(s)λd(s)
Fig.4.8 Outer speed loop of 3-phase permanent magnet synchronous motor
39
ωe * speed PI Controller ωe
+ {(3/2*(p/2)2*λf)/j}/(s + B/j)
- C(s) = Kps (1+1/𝜏iss)
ωe
Fig.4.9 Simplified outer speed loop of 3-phase permanent magnet synchronous motor
The simplified first order approximation of speed loop is shown in Fig.4.9.
Therefore by comparing equation (1) with equation (4.62), we can write the
plant parameters as
3 𝑝 2
( ) 𝜆𝑓
2 2 𝐵
c= and b= (4.63)
𝑗 𝑗
𝐵
2𝜉𝜔𝑛 −𝑏 2𝜉𝜔𝑛 −
𝑗
Kps = = 3 𝑝 2
(4.64)
𝑐 ( ) 𝜆𝑓
2 2
𝑗
2𝜉𝜔𝑛𝑗−𝐵
Kps = 3 𝑝 2
(4.65)
( ) 𝜆𝑓
2 2
2
𝜔𝑛 2
𝑗𝜔𝑛
Kis = = 3 𝑝 2 (4.66)
𝑐 ( ) 𝜆𝑓
2 2
And integral time constant is 𝜏is = Kps/Kis
2𝜉𝜔𝑛 𝑗−𝐵
𝜏is = 2
𝑗𝜔𝑛
(4.67)
4.4 Simulation Results
The simulation results of vector controlled 3-phase PMSM are presented in

this section. The Saber sketch schematic diagrams of vector control of independent 3-
phase PMSM is shown in Fig.4.10.
40
(a) Top level Saber sketch schematic diagram of vector control of 3-phase PMSM
(b) Internal diagram of controller for the 3-phase PMSM

Fig.4.10 Saber sketch schematic diagrams of vector control of 3-phase PMSM
The parameters of the PI controllers used in this simulation are shown in the
below Fig.4.11.
41
Fig.4.11 Parameters of PI controllers used in the simulation
In the simulation, non-salient pole three phase permanent magnet synchronous

motor has been used.
Total simulation time is 5 seconds. At t = 0sec a step speed command (ω e*) of

3016 electrical rad/sec applied. Load torque is applied on the motor as function of
square of the motor speed in mechanical rad/sec such that when the motor speed
reached to rated value, the rated load torque is applied on the motor. Therefore, the
motor operates at rated speed when its speed reaches to its rated value. The results of
simulation are given below.
As the rated speed (N) is 1800 r.p.m, the rated speed (ωr) in mechanical
radians / sec is calculated as
2𝜋𝑁 2𝜋∗1800
ωr = = = 188.5 mechanical rad/sec
60 60
As the rated mechanical speed (ωm) is 188.5 rad/sec, the rated electrical speed
(ωe) in electrical radians / sec is calculated as
ωe = p/2 (ωr) = 16 * 188.5 = 3016 electrical rad/sec
Fig.4.12 shows the set speed command and actual speed of the motor in
electrical rad/sec. The actual speed of the machine reaches the set speed command
and the steady state error is zero.
42
Fig.4.12 Set speed command (black) and actual speed (blue) [in electrical rad/sec]
of three phase PMSM
Fig.4.13 show the actual speed of the motor in mechanical rad/sec.
Fig.4.13 actual speed (in mechanical rad/sec) of 3-phase PMSM
Fig.4.14 shows the load torque and shaft power of the 3-phase PMSM at
rated speed.
Fig.4.14 Load torque and shaft power of 3-phase PMSM
43
As the rated output power (Ps) of the motor is 50000 watt and rated speed (𝜔𝑟 )
is 188.5 mechanical rad/sec. the rated load torque is calculated as follows.
Out power = Load torque (TL in Nm) * speed (𝜔𝑟 in mechanical rad/sec)
50000(rated) = TL * 188.5(rated)
TL = 265.252 N-m (rated)
The steady state load torque at rated speed of 188.5 mechanical rad/sec
reaches to 265.25 Nm as shown in Fig.4.14. Fig.4.15 show the d-axis current,
q-axis current and internal electromagnetic torque developed in the PMSM.
Fig.4.15 D-axis, Q-axis currents and electromagnetic torque of PMSM at steady

state
The simulation results shows that ids1 = -0.014694 A, iqs1 = 93.637 A, Te =

266.12 Nm. As ids1 reaches to approximately zero, that gives the maximum torque per
ampere operation. The internal electromagnetic torque developed by each slice is
given by
Te = (3/2) (p/2) λf iqs1 = 1.5 * 16 *0.11842* 93.637 = 266.124 Nm
The mechanical balance equation of the motor under steady state is given by
44
Te = TL + B ωm
From the parameter list we know that B = 0.004987 Nm-sec/ rad, rated speed
ωm = 188.5 rad/sec. rated load torque is calculated as TL = 265.252 N-m. Therefore
the rated internal electromagnetic torque developed during steady state should be
Te = 265.252 + 0.004987 * 188.5 = 266.192 Nm
From the simulation results, the internal electromagnetic torque developed by

the PMSM is given by 266.12 N-m. The simulation results and theoretical values are
matched.
4.5 Summary
In this chapter, the designing of the PI controllers for vector control operation
of a 3-phase PMSM. Simulation results of vector control of 3-phase PMSM are
discussed and from the simulation results, it can be concluded that the actual
speed of the motor has followed the set speed command with almost zero steady
state error. In the next chapter design of PI controllers for the 3-stage, 3-phase
PMSM will be discussed.
45
CHAPTER-5
DESIGN OF PI CONTROLLERS FOR VECTOR
CONTROL OPERATION OF 3-STAGE, 3-PHASE PMSM
5.1 Introduction
As the motor is an axial flux three stage permanent magnet synchronous

motor, and there is no magnetic coupling among the stages, the design of the PI
controllers for current loops is similar to that of 3-phase PMSM.
ωe(t)Ldids1(t)
Te* Vqs1’(t) Vqs1(t)

iqs1* PI Controller1 +
1/3(3/2*p/2*λf) + +
- C(s) = Kpq +Kpq/𝜏iqʃ +
iqs1
ωe(t)λf
ωe(t)Ldids2(t)
Vqs2’(t) Vqs2(t)
iqs2* PI Controller2 +
+ +
iqs2
ωe(t)λf
ωe(t)Ldids3(t)
iqs3* Vqs3’(t) Vqs3(t)

PI Controller3 +
+ +
iqs3
ωe(t)λf
Fig.5.1 The PI controllers along with feed forward compensations for electromagnetic
torue control for 3-stage PMSM
46
But, the total electromagnetic torque developed in the motor is the sum of
electromagnetic torques developed by each stage. Therefore, in order to get the
desired q-axis current for each slice of the machine, the desired torque is divided by
1/3kt. The PI controllers along with the feed forward compensations for the torque
control of three stage permanent magnet synchronous motor shown in Fig.5.1. Hence
the design of parameters of the outer speed loop PI controller is presented below.
Under balanced condition, i.e. when 𝑖𝑞𝑠1 = 𝑖𝑞𝑠2 = 𝑖𝑞𝑠3 = 𝑖𝑞𝑠
5.2 Design of Speed PI Controller
For designing the outer speed loop PI controller, it is necessary to write

transfer function between the reference quadrature axis current iqs*(s) and actual q-
axis current iqs(s).
As we have provided the compensation for nonlinear cross coupling terms by

feed forward technique, it is nonlinear cross coupling terms are cancelled each other.
So that the q-axis PI controller output is given as
𝐾𝑝𝑞 𝑡∗
Vqs’ = Kpq (iqs*(t) – iqs (t)) + ∫0 𝑖𝑞𝑠 (𝜏) – iqs(𝜏)) (5.1)
𝜏𝑖𝑞
𝑑 𝐾𝑝𝑞 𝑡
∗
Lq iqs(t) + Rs iqs(t) = Kpq (iqs*(t) – iqs (t)) + ∫0 𝑖𝑞𝑠 (𝜏) – iqs(𝜏)) (5.2)
𝑑𝑡 𝜏𝑖𝑞
𝑑 1 𝐾𝑝𝑞 𝑡
∗
iqs(t) = (-Rs iqs(t) + Kpq (iqs*(t) – iqs (t)) + ∫0 𝑖𝑞𝑠 (𝜏) – iqs(𝜏) (5.3)
𝑑𝑡 𝐿𝑞 𝜏𝑖𝑞
By applying Laplace transform to equation (5.3), we can get
𝑅𝑠 𝐾𝑝𝑞 𝐾𝑖
s iqs(s) = - iqs(s) + (iqs*(s) – iqs (s)) + (iqs*( 𝑠) – iqs(𝑠)) (5.4)
𝐿𝑞 𝐿𝑞 𝑠𝐿𝑞
From equations (4.31) and (4.33) Kpq = 2𝜉𝜔𝑛 𝐿𝑞 -Rs and Kiq = 𝐿𝑞 𝜔𝑛2
𝑅𝑠 𝑅𝑠 2
𝜔𝑛
s iqs(s) = - iqs(s) + 2𝜉𝜔𝑛 - (iqs*(s) – iqs (s)) + (iqs*( 𝑠) – iqs(𝑠) (5.5)
𝐿𝑞 𝐿𝑞 𝑠
𝑅𝑠 𝑅 2
𝜔𝑛 𝑅 2
𝜔𝑛
s iqs(s) + iqs(s) + 2𝜉𝜔𝑛 - 𝐿 𝑠 (iqs(𝑠)) + iqs(𝑠) = 2𝜉𝜔𝑛 - 𝐿 𝑠 (iqs*(𝑠)) + iqs*(𝑠) (5.6)
𝐿𝑞 𝑞 𝑠 𝑞 𝑠
47
𝑅𝑠
(𝑠 2 + 2𝜉𝜔𝑛 𝑠 + 𝜔𝑛2 ) iqs(s) = ((2𝜉𝜔𝑛 - )s + 𝜔𝑛2 ) iqs*(𝑠) (5.7)
𝐿𝑞
𝑅
𝑖𝑞𝑠 (𝑠) ( 2𝜉𝜔𝑛 − 𝑠 )s + 𝜔𝑛
2 )
𝐿𝑞
∗ (𝑠) = 2 (5.8)
𝑖𝑞𝑠 𝑠 2 +2𝜉𝜔𝑛 𝑠+𝜔𝑛
In order to design the outer speed loop PI controller, we have to write the
transfer function between quadrature axis current iq(s), and electrical speed ωe(s) of
the motor in rad/sec. The mechanical balance equation of the PMSM is given by
𝑑
Te (t) = TL + B ωr (t) + j ωr (t) (5.9)
𝑑𝑡
For 3-stage, permanent magnet synchronous motor, the electromagnetic torque

developed is sum of the electromagnetic torques developed by 3-stator units.
Therefore from equation (3.59)
𝑇𝑒 = 𝑇𝑒1 + 𝑇𝑒2 + 𝑇𝑒3 (5.10)
From equations (3.44), (3.51) and (3.58)
3𝑝 3𝑝 3𝑝
𝑇𝑒 = (𝜆𝑓 𝑖𝑞𝑠1 ) + (𝜆𝑓 𝑖𝑞𝑠2 ) + 2 2 (𝜆𝑓 𝑖𝑞𝑠3 ) (5.11)
22 22
Under balanced condition, i.e. when 𝑖𝑞𝑠1 = 𝑖𝑞𝑠3 = 𝑖𝑞𝑠3 = 𝑖𝑞𝑠
3𝑝
𝑇𝑒 = 3 ∗ (𝜆𝑓 𝑖𝑞𝑠 ) (5.12)
22
3𝑝 𝑑
3∗ 𝜆𝑓 𝑖𝑞𝑠 (𝑡) = TL + B ωr (t) + j ωr (t) (5.13)
22 𝑑𝑡
Now multiply with (p/2) on both sides of equation (5.13)
3 𝑝 𝑑
3* ( )2 𝜆𝑓 𝑖𝑞𝑠 (𝑡) = (p/2)*TL + B *(p/2)*ωr (t) + j (p/2)*ωr (t) (5.14)
2 2 𝑑𝑡
3 𝑝 𝑑
3* ( )2 𝜆𝑓 𝑖𝑞𝑠 (𝑡) = (p/2)*TL + B *ωe (t) + j ωe (t) (5.15)
2 2 𝑑𝑡
The load torque on the motor is considered as a disturbance in the control

system design and it should be rejected by the outer speed loop. So that
48
3 𝑝 𝑑
3* ( )2 𝜆𝑓 𝑖𝑞𝑠 (𝑡) = B *ωe (t) + j ωe (t) (5.16)
2 2 𝑑𝑡
By applying the Laplace transforms on both sides of equation (5.16)
3 𝑝
3* ( )2 𝜆𝑓 𝑖𝑞𝑠 (𝑠) = B *ωe (s) + js ωe (s) (5.17)
2 2
3 𝑝
3* ( )2 𝜆𝑓 𝑖𝑞𝑠 (𝑠) = (B + js) ωe (s) (5.18)
2 2
3 𝑝
3∗ ( )2 𝜆𝑓
2 2
𝑖𝑞𝑠 (𝑠) = (B /j + s) ωe (s) (5.19)
𝑗
3 𝑝
3∗2( 2 )2 𝜆𝑓
⁄
𝑗 ωe (s)
B = (5.20)
( + s) 𝑖𝑞𝑠 (𝑠)
j
By substituting equation (5.7) into (5.20), we have
3 𝑝
3∗ ( )2 𝜆𝑓 𝑅
2 2 ⁄ 2𝜉𝜔𝑛 − 𝑠 )s + 𝜔𝑛
2
ωe (s) 𝑗 𝐿𝑞
∗ (𝑠) =( B )( 2 2 ) (5.21)
𝑖𝑞𝑠 (s+ ) 𝑠 +2𝜉𝜔𝑛 𝑠+𝜔𝑛
j
The transfer function shown in above equation 5.21 is a third order model. But
to design a PI controller using a pole assignment technique, first order model is
B
needed. If the value of 𝜔𝑛 >> , the above third order equation can be approximated
j
as first order equation from the concept of dominant pole, because 𝜔𝑛 is the natural
frequency of the inner current loop. Therefore, dynamics from the inner current loop
can be neglected.
3 𝑝
3∗ ( )2 𝜆𝑓
ωe (s) 2 2 ⁄
𝑗
∗ (𝑠) ≈ ( B ) (5.22)
𝑖𝑞𝑠 (s+ )
j
The outer speed loop of 3-stage permanent magnet synchronous motor is

shown in Fig.5.2.
49
ωe(s)Ldids1(s) 𝑅𝑠
ωe*(s) iqs1*
speed PI Controller q-axis PI Controller Vqs1’(s) Vqs1(s) iqs1(s) Te1(s)
+ Te*1/3(3/2*p/2*λf) + - 1 1
- C(s) = Kps (1+1/𝜏iss ) + 1 + + (3/2*p/2*λf)
- + - 𝐿𝑞 𝑆
ωe(s) C(s) = Kpq (1+1/𝜏iqs)
iqs1
ωe(s)λf ωe(s)λds1(s)
ωe(s)Ldids2(s) 𝑅𝑠
iqs2 * Te2(s)
Vqs2’(s) Vqs2(s) iqs2(s) Te(s) ωe(s)
q-axis PI Controller + - 1 +
+ + 1 (3/2*p/2*λf) + (p/2) / j(s+ (B/j))
+ 2 + - 𝐿𝑞 𝑆 +
-
C(s) = Kpq (1+1/𝜏iqs)
iqs2
ωe(s)Ldids3(s)
𝑅𝑠
* Te3(s)
iqs3 iqs3(s)
Vqs3’(s) Vqs3(s)
q-axis PI Controller
+ - 1 1
+
- 3 +
+
+
-
(3/2*p/2*λf)
𝐿𝑞 𝑆
C(s) = Kpq (1+1/𝜏iqs)
iqs3
Fig.5.2 Outer speed loop of 3-stage permanent magnet synchronous motor
50
The simplified first order approximation of speed loop is shown in Fig.5.3.
ωe * speed PI Controller Te
* ωe
+ {(3*3/2*(p/2)2*λf)/j}/(s + B/j)
- C(s) = Kps (1+1/𝜏iss)
ωe
Fig.5.3 Simplified outer speed loop of 3-stage permanent magnet synchronous motor
Therefore by comparing equation (4.1) with equation (5.22), we can write the
plant parameters as
3 𝑝 2
3∗ ( ) 𝜆𝑓 𝐵
2 2
c= and b= (5.23)
𝑗 𝑗
𝐵
2𝜉𝜔𝑛 −𝑏 2𝜉𝜔𝑛 − 𝑗
Kps = = 3 𝑝 2
(5.24)
𝑐 3∗ ( ) 𝜆𝑓
2 2
𝑗
2𝜉𝜔𝑛𝑗−𝐵
Kps = 3 𝑝 2
(5.25)
3∗ ( ) 𝜆𝑓
2 2
2
𝜔𝑛 2
𝑗𝜔𝑛
Kis = = 3 𝑝 2 (5.26)
𝑐 3∗ ( ) 𝜆𝑓
2 2
And integral time constant is 𝜏is = Kps/Kis
2𝜉𝜔𝑛 𝑗−𝐵
𝜏is = 2
𝑗𝜔𝑛
(5.27)
Where 𝜉 is the damping ratio and 𝜔𝑛 is the natural frequency of outer speed
loop. So from the equations (5.25), (5.26), and (5.27) it can be concluded that, if the
plant parameters j, B, and 𝜆𝑓 are known, the PI controller parameters are calculated
51
by choosing the values of ξ (damping ratio) and 𝜔𝑛 (natural frequency) of desired
stable closed loop characteristic polynomial.
After the PI controllers along with feed forward compensation, the direct axis
voltage and quadrature axis voltages are available, these are the dc quantities. These
dc quantities are converted in to sinusoidal voltages in three phase stationary
reference frame through inverse Park’s and inverse Clarke’s transformations.
5.3 Inverse Park’s Transformation
The inverse Park’s transformation is used to convert the dc quantities in the

synchronously rotating reference frame into two phase quantities in the stationary
reference frame. The inverse Park’s transformation requires rotor flux position angle.
The classical inverse Park’s transformation is expressed in the following

equations.
vα = vd∗cos θ𝑒 - vq∗sin θe (5.28)
vβ = vd∗ sin θe + vd∗cos θ𝑒 (5.29)
𝑣𝛼 cos θ𝑒 −sin θe 𝑣𝑑
[𝑣 ] = [ ][ ] (5.30)
𝛽 sin θe cos θ𝑒 𝑣𝑞
The same transformation is also applicable for voltages and fluxes.
Where,
vd, vq are rotating reference frame quantities
vα, vβ are orthogonal stationary reference frame quantities
θe is electrical rotor flux position angle
5.4 Inverse Clarke’s Transformation
Inverse Clarke’s transformation is used to convert the balanced two phase

quantities in the two phase stationary reference frame into balanced three phase
quantities in the three phase stationary reference frame. The classical inverse Clarke’s
transformation is expressed in following equations.
52
Va = Vα (5.31)
1
Vb = - 2 𝑉𝛼 + 𝑉𝛽 (5.32)
1 √3
Vc = - 2 𝑉𝛼 − 𝑉 (5.33)
2 𝛽
The output of the inverse Clarke’s transformation goes to PWM generator as

control signals, where they will be compared with triangular carrier wave to produce
pulses to switch the power inverter switches.
5.5 Summary
In this chapter, the designing of the PI controllers for vector control operation
of 3-stage, 3-phase PMSM is discussed. In the next chapter, the simulation results of
3-stage, 3-phase permanent magnet synchronous motor for different set speed
commands will be discussed.
53
CHAPTER-6
SIMULATION RESULTS AND ANALYSIS OF 3-STAGE,

3-PHASE PMSM
6.1 Introduction
In order to know the performance of the vector controlled 3-stage permanent

magnet synchronous motor, the tests are done in two cases in the Saber RD simulator.
1) At t = 0sec a step speed command (ωref) of 3016 electrical rad/sec applied, 2)
during the time from t = 0sec to t = 1 sec, the set speed command is varied from zero
to 3016 electrical rad/sec and from t = 1 sec on wards it is maintained at 3016
electrical rad/sec.
6.2 Simulation Circuits
The Saber sketch schematics of vector control of 3-stage, 3-phase permanent magnet
synchronous motor are shown in Fig.6.1.
(a) Top level Saber sketch schematic of vector control of 3-stage PMSM
54
(b) Internal Saber sketch schematic of controller for the 3-stage PMSM
(c) Saber sketch schematic of the controller for each stage
(d) Saber sketch schematic of the uni-polar sinusoidal PWM for each stage
55
(e) Saber sketch schematic of the inverter set up block
Vp – Plus terminal of battery
Vm – minus terminal of battery
a, b – output terminals goes to machine phases
(f) Saber sketch schematic of H-bridge power inverter for a phase
Fig: 6.1 Saber sketch schematics of vector control of 3-stage, 3-phase PMSM
Top level Saber sketch schematic of vector control of 3-stage permanent magnet
synchronous motor is shown in Fig.6.1 (a). It mainly consists of motor model,
controller block, uni polar PWM block, H-bridge inverters block, and etc. Fig.6.1 (b)
56
shows what is there in the controller block. It consists of a speed PI controller and
each stage controllers. Fig.6.1 (c) shows what is there in the each stage controller
block. It consists of d-axis PI controller, d-axis PI controller, decoupling or feed
forward compensation network, Inverse Park transformation block and inverse Clarke
transformation block. Fig.6.1 (d) shows what is there in the uni-polar PWM block.
Fig.6.1 (e) shows what is there in the H-bridge inverters block. Fig.6.1 (f) shows
internal diagram of H-bridge power inverter.
6.3 Simulation Results
In this section, the simulation results of 3-stage, 3-phase PMSM are presented.
6.3.1 Simulation of Test-1Speed Profile
Total simulation time is 4 seconds. At t = 0sec a step speed command (ω e*) of

3016 electrical rad/sec applied. Load torque is applied on the motor as function of
square of the motor speed in mechanical rad/sec such that when the motor speed
reached to rated value, the rated load torque is applied on the motor. Therefore, the
motor operates at rated speed when its speed reaches to its rated value. The results of
simulation are given below. In the simulation, non-salient pole axial flux three stage
permanent magnet synchronous motor has been used. The parameters of the PI
controllers used in this simulation are shown in the Fig.6.2
Fig: 6.2 Parameters of PI controllers
57
As the rated speed (N) is 1800 r.p.m, the rated speed (ωm) in mechanical
radians / sec is calculated as
2𝜋𝑁 2𝜋∗1800
ωm = = = 188.5 mechanical rad/sec
60 60
As the rated mechanical speed (ωm) is 188.5 rad/sec, the rated electrical speed (ωe) in
electrical radians / sec is calculated as
ωe = p/2 (ωm) = 16 * 188.5 = 3016 electrical rad/sec
Fig.6.3 shows the set speed command and actual speed of the motor in
electrical rad/sec. The actual speed of the machine reaches the set speed command
and the steady state error is zero.
Fig.6.3 Set speed command (pink) and actual speed (orange) [in electrical rad/sec]
of three stage axial flux PMSM
Fig.6.4 shows the actual speed of the motor in mechanical rad/sec.
58
Fig.6.4 Actual speed (in mechanical rad/ sec) of three stage axial flux PMSM
As the rated output power (Ps) of the motor is 150000 watt and rated speed
(𝜔𝑟 ) is 188.5 mechanical rad/sec. the rated load torque is calculated as follows.
Output power = Load torque (TL in Nm) * speed (𝜔𝑟 in mechanical rad/sec)
150000(rated) = TL * 188.5(rated)
TL = 795.756 N-m (rated)
Fig.6.5 shows the load torque applied on the motor. The steady state load
torque at rated speed of 188.5 mechanical rad/sec reaches to 795.76 Nm as shown in
Fig.6.5.
Fig.6.5 Load torque applied to the three stage axial flux PMSM
Fig.6.6 shows the shaft power output of the three stage axial flux permanent
59
Fig.6.6 Shaft power of three stage axial flux permanent magnet synchronous motor
Fig.6.7 show the d-axis current, q-axis current and internal electromagnetic
torque developed in the slice-1 of the machine.
Fig.6.7 D-axis, q-axis currents and electromagnetic torque of slice-1 at steady state
The simulation results shows that ids1 = 0.0032491 A, iqs1 = 93.66 A, Te1 = 266.19
Nm. As ids1 reaches to approximately zero, that gives the maximum torque per ampere
operation. The internal electromagnetic torque developed by each slice is given by
Te1 = (3/2)(p/2) λf iqs1 = 1.5 * 16 *0.11842* 93.66 = 266.19 Nm
The d-axis current, q-axis current and internal electromagnetic torque developed
in the slice-2 of the machine is shown in Fig.6.8.
60
The simulation results shows that ids2 = -0.0061167 A, iqs2 = 93.663, Te2 = 266.2.
As ids2 reaches to approximately zero, that gives the maximum torque per ampere
operation. The internal electromagnetic torque developed by each slice is given by
Te2 = (3/2) (p/2) λf iqs2 = 1.5 * 16 *0.11842* 93.663 = 266.2 Nm
The d-axis current, q-axis current and internal electromagnetic torque

developed in the slice-3 of the machine is shown in Fig: 6.9.
The simulation results shows that ids3 = 0.0023404 A, iqs3 = 93.664 A, Te3 =
266.2Nm. As ids3 reaches to approximately zero, that gives the maximum torque per
ampere operation. The internal electromagnetic torque developed by each slice is
given by
Te3 = (3/2) (p/2) λf iqs3 = 1.5 * 16 *0.11842* 93.664 = 266.2 Nm
Fig.6.10 shows the electromagnetic torque developed by each slice and total
electromagnetic torque developed by the motor.
Fig.6.10 Electromagnetic torque developed by each stage and total electromagnetic

torque developed (Te) by 3-slice motor at steady state
61
The simulation results shows that, Te1 = 266.19 Nm Te2 = 266.2 Nm Te3 = 266.2
Nm and Te = 798.57 Nm.
The total torque developed by the 3-slice motor is equal to sum of the torques
developed by individual slice. There fore
Te = Te1 + Te2 + Te3 = 266.19 + 266.2 + 266.2 = 298.59 Nm
The mechanical balance equation of the motor under steady state is given by
Te = TL + B ωm
From the parameter list we know that B = 0.014961 Nm-sec/ rad, rated speed
ωm = 188.5 rad/sec. rated load torque is calculated as TL = 795.756 N-m. Therefore
the rated internal electromagnetic torque developed during steady state should be
Te = 795.756 + 0.014961 * 188.5 = 798.576 Nm
From the simulation results, the internal electromagnetic torque developed by

the whole machine is given by 798.59 N-m. The simulation results and theoretical
values are matched.
Fig.6.11a shows the phase voltages and phase currents of slice-1 and fig.6.11b
show the phase power and total power drawn by the slice-1 of the 3-slice permanent
(a) Phase voltages and phase currents of slice-1 of the 3-slice PMSM at steady
state
62
(b) Each phase and total power drawn by the slice-1 of 3-slice PMSM at steady state
Fig.6.11 Voltage, current and power of each phase and total power drawn by slice-1
of 3 slice PMSM under steady state
Fig.6.12a shows the phase voltages and phase currents of slice-2 and fig.6.12b
state
63
Fig.6.12 Voltage, current and power of each phase and total power drawn by slice-
2 of 3 slice PMSM under steady state
Fig.6.13a shows the phase voltages and phase currents of slice-3 and Fig.6.13b
state
64
Fig.6.13 Voltage, current and power of each phase and total power drawn by slice-
3 of 3 slice PMSM under steady state
From the Simulation results, it can be concluded that the machine operates at
approximately unity power factor at its rated load. The shaft power output is 150 kW.
The total power drawn by the motor is the sum of the powers drawn by the three
slices and is equal to Pin = 51962 + 51965 + 51966 = 155.893 kW. Therefore, overall
efficiency of the motor is 96.22%.
6.3.2 Simulation of Test-2 Speed Profile:

During the time from t = 0sec to t = 1 sec, the set speed command is varied
from zero to 3016 electrical rad/sec and from t = 1 sec on wards it is maintained at
3016 electrical rad/sec. The corresponding parameters of the PI controllers and
simulation results are shown in Fig.6.14.
(a) Parameters of PI controllers
65
(b) Mechanical speed (𝜔𝑟 ), electrical set speed command , actual electrical
speed(𝜔𝑒 ) and speed error of 3-slice PMSM
(c) Actual mechanical speed, load torque and shaft power of 3-slice PMSM
Fig.6.14. Parameters of PI controllers & simulation results for ramp type speed
command
66
6.4 Summary
Simulation results of vector control of 3-slice axial flux PMSM for two
different speed profiles are presented and from the simulation results, it can be
concluded that the actual speed of the motor has followed the set speed command
with almost zero steady state error. In the next chapter, the observations made out
of the simulation results presented for a step speed command of 3016 electrical
radian/sec will be summarised in tabular form, conclusions and future scope of
the work will be explained.
67
CHAPTER-7
CONCLUSIONS AND FUTURE SCOPE
The observations made out of the simulation results presented in the previous
chapter for a step speed command of 3016 electrical rad/sec are summarised in the
following tables.
Table: 7.1
Summary table of slice-1 of 3-slice PMSM
Parameter Value
Va1 (Phase-A voltage of Stage-1) 261.17 V
Ia1 (Phase-A current of Stage-1) 66.206 A
P a1 (Phase –A power of Stage-1) 17260 W
Vb1 (Phase-B voltage of Stage-1) 262.58 V
Ib1 (Phase-B current of Stage-1) 66.303 A
Pb1 (Phase –B power of stage-1) 17388 W
Vc1 ( Phase-C voltage of stage-1) 261.66V
Ic1 (Phase-C current of stage-1) 66.221 A
Pc1 ( Phase –C power of stage-1) 17307 W
Total power drawn by stage-1 (P1) 51962 W
68
Table: 7.2 Summary table of slice-2 of 3-slice PMSM
Parameter Value
Vc2 ( Phase-C voltage of stage-2) 262.48 V
Table: 7.3 Summary table of slice-3 of 3-slice PMSM
Parameter Value
Vc3 ( Phase-C voltage of stage-3) 262.22 V
69
7.1 Conclusions
From the simulation results, it can be concluded that the actual speed of the
motor has followed the set speed command with almost zero steady state error.
The internal electromagnetic torque (Te) developed is 798.576 Nm with a
peak to peak torque ripple of 33.825 Nm (4.236%) approximately for the 3-stage
PMSM.
Shaft power (output power) of three slice PMSM is 150kW and the input
power drawn is 155.893kW at a speed of 188.5 mechanical rad /sec. Hence over-
all efficiency of the motor at full load condition is 96.22%.
7.2 Future Scope

In this dissertation, modeling and simulation of the vector control of 3-slice
axial flux PMSM is done. This can be extended to sensor less vector control of 3-
slice axial flux PMSM.
In this dissertation, the sinusoidal PWM technique is used; this can be
extended to third harmonic injected sinusoidal PWM, and/ or space vector PWM.
So obtained results shall be compared for minimum torque ripple.
In this dissertation, PI controllers are used for both speed tracking as well as
current control. Hysteresis Current controllers can be used for current control.
70
APPENDIX
[A-1] Mast model for 3-phase Permanent magnet synchronous motor
template independentpermanentmagnetmotor na1 na2 nb1 nb2 nc1 nc2 tl wr iqs ids the = rs,
lds, lqs, yf, p, j, b
electrical na1,na2,nb1,nb2,nc1,nc2
input nu tl
output nu wr,iqs,ids,the
# template for independent permanent magnet motor
number rs = 0.1305, lds = 53.9e-6, lqs = 53.9e-6, yf = 0.11842, j = 0.2913, b = 0.004987, p =

32
export val tq_Nm te

export val w_radps we # we = wr * p/2
export val nu yq, yd
export val nu vds,vqs
{
number pi = 22/7
branch va = v(na1,na2), ia = i(na1->na2)
branch vb = v(nb1,nb2), ib = i(nb1->nb2)
branch vc = v(nc1,nc2), ic = i(nc1->nc2)
# making initial condition of theta, ia as zero

control_section {
initial_condition(the, 0)
initial_condition (ia, 0)
}
values {
# transformations
vqs = (2/3)*(va*cos(the) + vb*cos(the-2*pi/3) + vc*cos(the+2*pi/3))

vds = (2/3)*(va*sin(the) + vb*sin(the-2*pi/3) + vc*sin(the+2*pi/3))
te = 1.5*(p/2)*iqs*yf
we = wr * p/2
# state variables
make d_by_dt(the) = wr*p/2

make d_by_dt(yq) = vqs - (rs*iqs) - (wr*yd*p/2)
make d_by_dt(yd) = vds - (rs*ids) + (wr*yq*p/2)
# flux linkages
71
yq = lqs*iqs
yd = lds*ids + yf
# Mechanical equation
make d_by_dt(wr) = (te-tl-b*wr)/j
# input current
ia = iqs*cos(the) + ids*sin(the)
ib = iqs*cos(the-2*pi/3) + ids*sin(the-2*pi/3)
ic = iqs*cos(the+2*pi/3) + ids*sin(the+2*pi/3)
[A-2] Mast model for 3-stage, 3-phase Permanent magnet synchronous motor
template permanentmagnetmotormultistage n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14

n15 n16 n17 n18 tl wr iqs1 ids1 iqs2 ids2 iqs3 ids3 the1 the2 the3 = rs, lds, lqs, yf, p, j, b
electrical n1,n2,n3,n4,n5,n6,n7,n8,n9,n10,n11,n12,n13,n14,n15,n16,n17,n18
input nu tl
output nu wr,iqs1,ids1,iqs2,ids2,iqs3,ids3,the1,the2,the3
# template for permanent magnet motor
number rs = 0.1305, lds = 53.9e-6, lqs = 53.9e-6, yf = 0.11842, j = 0.2913, b = 0.0149602, p =

32
export val tq_Nm te

export val w_radps we
export val nu te1
export val nu te2
export val nu te3 # we = wr * p/2
export val nu yq1, yd1
export val nu vds1,vqs1
{
number pi = 22/7
branch va1 = v(n1,n2), ia1 = i(n1->n2)
branch vb1 = v(n3,n4), ib1 = i(n3->n4)
branch vc1 = v(n5,n6), ic1 = i(n5->n6)
72
# making initial condition of theta, ia as zero
control_section {
initial_condition(the1, 0)
initial_condition(ia1, 0)
}
values {
# transformations
vqs1 = (2/3)*(va1*cos(the1) + vb1*cos(the1-2*pi/3) + vc1*cos(the1+2*pi/3))

vds1 = (2/3)*(va1*sin(the1) + vb1*sin(the1-2*pi/3) + vc1*sin(the1+2*pi/3))
#torque equations
te1 = 1.5*(p/2)*yf*iqs1
te2 = 1.5*(p/2)*yf*iqs2
te3 = 1.5*(p/2)*yf*iqs3
te = te1 + te2 + te3
we = wr*p/2
# state variables
make d_by_dt(the1) = wr*p/2

make d_by_dt(yq1) = vqs1 - rs*iqs1 - wr*yd1*p/2
make d_by_dt(yd1) = vds1 - rs*ids1 + wr*yq1*p/2
# flux linkages
yq1 = lqs*iqs1
yd1 = lds*ids1 + yf
yq2 = lqs*iqs3
yd2 = lds*ids2 + yf
yq3 = lqs*iqs3
yd3 = lds*ids3 + yf
the2 = the1 - (2*pi/9)
the3 = the1 - (4*pi/9)
# Mechanical equation
73
make d_by_dt(wr) = (te-tl-b*wr)/j
# input current
ia1 = iqs1*cos(the1) + ids1*sin(the1)

ib1 = iqs1*cos(the1-2*pi/3) + ids1*sin(the1-2*pi/3)
ic1 = iqs1*cos(the1+2*pi/3) + ids1*sin(the1+2*pi/3)
}
Note:
rs : Stator resistance / phase in Ohm
lds : Direct axis inductance in Henry
lqs : Quadrature axis inductance in Henry
p : Number of poles
yf : Amplitude of permanent magnetic flux volt-sec/rad
j : Moment of inertia in kg-m2
B : Viscous damping co-efficient in Nm-sec/rad
yd : Direct axis flux linkages
yq : Quadrature axis flux linkages
yd1 : Direct axis flux linkages of stator unit-1
yq1 : Quadrature axis flux linkages of stator unit-1
we : Actual speed of motor in electrical rad/sec
wr : Actual speed of motor in mechanical rad/sec
74
REFERENCES
[1] Jacek F.Gieras, Rong-jie Wang and Marteen J.Kamper., “Axial Flux Permanent Magnet Brushless
machines” Kluwer Academic Publishers, newyork, Boston, Dordrecht, London, Moscow, 2005.
[2] Liuping Wang, Shan Chai, Dae Yoo, Lu Gan and Ki Ng, “PID AND PREDICTIVE CONTROL OF
ELECTRICAL DRIVES AND POWER CONVERTERS USING MATLAB®/SIMULINK”, John
Wiley & Sons Singapore Pte.Ltd, 2015.
[3] R. Krishnan, “ELECTRIC MOTOR DRIVES Modeling, Analysis and Control” Virginia Tech,
Blacksburg. VA., Prentice Hall, upper Saddle River, New Jersey 07458 2001.
[4] Ned Mohan, Tore M. Undeland, William P. Robbins., “Power electronics, converters, applications
and design”, 2nd edition, John Wiley & Sons, Inc. New York Chichester Brisbane Toronto Singapore,
1995.
[5] Park, Inverse Park and Clarke, Inverse Clarke Transformations MSS Software implementation user
guide, Microsemi. http://www.microsemi.com/documentportal/doc_view/132799-park-inverse-park-
and-clarke-inverse-clarke-transformations-mss-software-implementation.
[6] Araz Darba, Mohammad Esmalifalak, Ebrahim Sarbaz Barazandeh, “Implementing SVPWM
technique to Axial Flux Permanent Magnet Synchronous Motor Drive with Internal Model Current
Controller”., The 4th International Power Engineering and Optimization Conference (PEOCO2010),
Shah Alam, Selangor, MALAYSIA. 23-24 June 2010.
[7] http://www.synopsys.com/Prototyping/Saber/eUpdate/Pages/SaberRD-Desktop-Env-Jun10.aspx
75

Vector Control of 3-Stage / Slice Axial Flux Permanent Magnet Synchronous Motor

Uploaded by

Copyright:

Available Formats

Vector Control of 3-Stage / Slice Axial Flux Permanent Magnet Synchronous Motor

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Vector Control of 3-Stage / Slice Axial Flux Permanent Magnet Synchronous Motor

Uploaded by

Copyright:

Available Formats

VECTOR CONTROL OF 3-STAGE / SLICE AXIAL

FLUX PERMANENT MAGNET SYNCHRONOUS

DEPARTMENT OF ELECTRICAL & ELECTRONICS ENGINEERING

File Name: PLAGCHECK_14131D5211.DOCX (2.44M)

Dissertation Title: VECTOR CONTROL OF 3-STAGE / SLICE AXIAL FLUX

Date and Time of Submission: 25-DEC-2016 07:24AM

Submission ID: 755978842

Similarity Index: 25%

Dean – Academic Programs (PG & Research)

This dissertation relates to simulation of vector control of 3-stage / slice axial

Maximum torque/ampere operation has been achieved by forcing the d-axis

I wish to express my sincere gratitude towards my supervisor Dr. SASTRY V.

I cherish to express my gratitude to HBL Power Systems Limited, Yapral, and

I would like to express my heartfelt thanks to Dr. G. V. E. Satish Kumar,

(Regd. No. 14131D5211)

Fig.1.1 Classification of axial flux permanent magnet synchronous 4

Fig.1.2 Single sided axial flux permanent magnet machine 5

Fig.1.3 Double-sided axial flux permanent magnet synchronous 6

Fig.1.4 Multi Stage axial flux permanent magnet synchronous 7

Fig.1.5 Block diagram for vector control 3-phase PMSM 8

Fig.2.1 Block diagram of vector control of 3-stage, 3-phase 12

Fig.2.3 Bi-polar sinusoidal PWM voltage switching 15

Fig.2.4 Uni-polar sinusoidal PWM voltage switching 16

Fig.3.1 Representation of Clarke’s and Park’s transformations 18

Fig.3.2 Clarke’s transformation 20

Fig.3.3 Park’s transformation 21

Fig.4.1 Simple PI controller closed loop system 29

Fig.4.2 Block diagram of permanent magnet synchronous motor: 31

Fig.4.3 Direct-axis current loop of PMSM 32

Fig.4.4 Block diagram of permanent magnet synchronous motor: 33

Fig.4.5 Quadrature-axis current loop of PMSM 33

d-axis current control loop

Fig.4.7 The PI controller along with feed forward compensation for 36

electromagnetic torue control for three phase PMSM

Fig.4.8 Outer speed loop of 3-phase permanent magnet 39

Fig.4.9 Simplified outer speed loop of 3-phase permanent magnet 40

Fig.4.10 Saber sketch schematic diagrams of vector control of 41

Fig.4.11 Parameters of PI controllers used in the simulation 42

electrical rad/sec] of three phase PMSM

Fig.4.13 Actual speed (in mechanical rad/ sec), of 3-phase PMSM 43

Fig.4.14 Load torque and shaft power of 3-phase PMSM 43

Fig.4.15 D-axis, q-axis currents and electromagnetic torque of 44

PMSM at steady state

Fig.5.1 The PI controllers along with feed forward compensations 46

for electromagnetic torue control for 3-stage, PMSM

Fig.5.2 Outer speed loop of 3-stage permanent magnet 50

Fig.5.3 Simplified outer speed loop of 3-stage permanent magnet 51

Fig.6.1 Saber sketch schematics of vector control of 3-stage, 3- 56

Fig.6.2 Parameters of PI controllers 57

Fig.6.4 Actual speed (in mechanical rad/ sec) of three stage 59

Fig.6.6 Shaft power of three stage axial flux permanent magnet 60

Fig.6.7 D-axis, q-axis currents and electromagnetic torque of slice- 60

Fig.6.8 D-axis, q-axis currents and electromagnetic torque of slice- 60

Fig.6.9 D-axis, q-axis currents and electromagnetic torque of slice- 61

Fig.6.10 Electromagnetic torque developed by each stage and total 61

Table 7.1 Summary table of slice-1 of 3-slice 67

Table 7.2 Summary table of slice-2 of 3-slice 68

Table 7.3 Summary table of slice-3 of 3-slice 68

PMSM : Permanent magnet synchronous motor