Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
1
EFFICIENCY OPTIMIZATION OF A DIRECT TORQUE CONTROL (DTC)
INDUCTION MOTOR DRIVE
SIM SY YI
UNIVERSITI TUN HUSSEIN ONN MALAYSIA
2
UNIVERSITI TUN HUSSEIN ONN MALAYSIA
STATUS CONFIRMATION FOR DOCTORAL THESIS
EFFICIENCY OPTIMIZATION OF A DIRECT TORQUE CONTROL (DTC)
INDUCTION MOTOR DRIVE
ACADEMIC SESSION : 2015/2016
I, SIM SY YI, agree to allow this Doctoral Thesis to be kept at the Library under the following terms: 1. This Doctoral Thesis is the property of the Universiti Tun Hussein Onn Malaysia. 2. The library has the right to make copies for educational purposes only. 3. The library is allowed to make copies of this report for educational exchange between higher
educational institutions. 4. ** Please Mark (√)
CONFIDENTIAL (Contains information of high security or of great importance to Malaysia as STIPULATED under the OFFICIAL SECRET ACT 1972) RESTRICTED (Contains restricted information as determined by the Organization/institution where research was conducted) FREE ACCESS
Approved by, (WRITER’S SIGNATURE) (SUPERVISOR’S SIGNATURE) Permanent Address: NO 4, LORONG J, JALAN MASJID, 83000, BATU PAHAT, JOHOR. Date : ___________________________ Date: _____________________________ NOTE: ** If this Doctoral Thesis classified as CONFIDENTIAL or RESTRICTED,
please attach the letter from the relevant authority/organization stating reasons and duration for such classifications.
√
i
This thesis has been examined on date 14 December 2015 and is sufficient in fulfilling
the scope and quality for the purpose of awarding the Degree of Doctor of Philosophy.
Chairperson:
Prof. Madya Dr. Mohamad Nor bin Mohamad Than Faculty of Electrical and Electronic Engineering University Tun Hussein Onn Malaysia
Examiners:
Prof. Dr. Nasrudin bin Abd Rahim Deputy Vice Chancellor (Research & Innovation) University of Malaya
Dr. Shamsul Aizam bin Zulkifli Faculty of Electrical and Electronic Engineering University Tun Hussein Onn Malaysia
i
EFFICIENCY OPTIMIZATION OF A DIRECT TORQUE CONTROL (DTC)
INDUCTION MOTOR DRIVE
SIM SY YI
A thesis submitted in
fulfilment of the requirement for the award of the
Doctor of Philosophy
Faculty of Electrical and Electronic Engineering
Universiti Tun Hussein Onn Malaysia
JANUARY 2016
ii
I hereby declare that the work in this project report is my own except for quotations
and summaries which have been duly acknowledged
Student : …………………………………….……….
SIM SY YI
Date : …………………………………….……….
Supervisor : …………………………………….……….
DR WAHYU MULYO UTOMO
Co Supervisor : …………………………………….……….
ASSOC. PROF. DR .ZAINAL ALAM BIN
HARON
iii
Specially dedicated to my beloved family
iv
ACKNOWLEDGEMENT
First of all, I would like to take this opportunity to express my deepest gratitude to my
research supervisors; Dr. Wahyu Mulyo Utomo and Associate Professor. Dr .Zainal
Alam Bin Haron, who has persistently and determinedly assisted me during the whole
course of this research. It would have been very difficult to complete this research
without the enthusiastic support, insight and advice given by my supervisors.
My utmost thank also goes to my family. Words cannot be expressed for how
grateful I am to my parents for all the support that gave me throughout my academic
years. Without them, I might not be the person I am today.
Last but not least, my special gratitude is extended to all members of the F1
research group that has been my guidance and giving the support for my research. It is
my greatest thanks and joy that I have met these research group. Thank you.
v
ABSTRACT
Typical Direct Torque Control (DTC) Induction Motor (IM) drive has good efficiencies
while operating at its rated condition. Nevertheless, a motor drive operates far from its
rated operating condition, which impairs the efficiency. Therefore, a conventional DTC
IM drive system should be associated with a loss minimization strategy to maximize the
drive efficiency for a wider range of operation. For a given operating condition, the
losses in an IM can be minimized by adjusting to the appropriate flux level instead of
conventionally keeping constant at its rated flux value. Consequently, this research
proposed an online learning Artificial Neural Network Efficiency Optimization
(ANNEO) controller with the aim to generate an adaptive flux level to optimize the
efficiency of any different operating points, especially at low speed and low torque
condition. ANNEO is implemented with an online learning second order Levernberg-
Marquardt algorithm. The proposed controller uses the input power of the drive system
as the objective function and minimizes it. In order to further improve the system
dynamic response, an Artificial Neural Network speed controller (ANNSC) of DTC is
included. In ANNSC, the back propagation algorithm is employed. The goal of this
research is to achieve the efficiency optimization of DTC induction motor drives system
while preserve the superior speed performance of IM application over a wide speed
range. The entire DTC drive system along with the proposed adaptive flux ANN based
controller has been modeled in Simulink/Matlab and validated experimentally by the
ControlDesk 5.1 with dSPACE DS1103 Real-time Digital Signal Controller. An
efficiency improvement as high as 14.6% and 12.75% have achieved for simulation and
experimental testing respectively at low speed and low load condition. The simulation
and experimental results show that the proposed technique managed to generate adaptive
optimum flux level, hence improves the efficiency of the DTC induction motor drive
system while at the same time, preserving excellent speed performance of the systems.
vi
ABSTRAK
System pemacu motor aruhan (IM) Direct Torque Control (DTC) mempunyai
kecekapan yang baik ketika beroperasi pada keadaan kadar. Walau bagaimanapun,
kebanyakan masa motor beroperasi jauh daripada titik pengendalian yang kadar. Oleh
itu, kecekapan motor berkurang. Lantarannya, sistem pemacu DTC yang konvensional
seharusnya bersekutu dengan strategi peminimuman kerugian untuk memaksimumkan
kecekapan pemacu bagi pelbagai jenis operasi. Ia adalah satu fakta yang diketahui
bahawa pada satu titik pengendalian yang diberikan, kerugian di dalam satu motor
aruhan boleh dikurangkan dengan melaraskan fluks kepada tahap fluks yang sesuai.
Oleh yang demikian, penyelidikan ini mengamalkan Artificial Neural Network
Efficiency Optimization (ANNEO) yang pembelajaran online dengan matlamat untuk
menjana satu nilai fluks yang adaptif untuk mengoptimumkan kecekapan pada setiap
titik pengendalian yang berbeza, terutamanya apabila pada kelajuan motor yang
perlahan dan keadaan kilas yang rendah. ANNEO telah dilaksanakan dengan
pembelajaran online yang peringkat dua, iaitu Levernberg-Marquardt algoritma.
Pengawal yang dicadangkan untuk penyelidikan ini menggunakan kuasa input sistem
pemacu sebagai fungsi objektif dan mengurangkannya. Untuk memajukan lagi sistem
gerak balas dinamik, Artificial Neural Network Speed Controller (ANNSC) DTC telah
ditambahkan pada sistem. Dalam ANNSC, algoritma penyebaran belakang telah
dilaksanakan. Matlamat kajian ini adalah untuk mencapaikan kecekapan optimum pada
sistem pemacu motor aruhan DTC dan pada masa yang sama mengekalkan prestasi
kelajuan motor yang unggul pada julat kelajuan yang luas. Seluruh sistem pemacu DTC
bersama-sama dengan pengawal fluks adaptif ANN telah dimodelkan dalam Simulink
/ Matlab dan disahkan secara eksperimen oleh ControlDesk 5.1 dengan DSpace
DS1103 Real-time Digital Signal Controller. Peningkatan kecekapan sebanyak 14.6%
dan 12.75% telah dicapaikan oleh ujian simulasi dan eksperimen masing masing
dikeadaan kelajuan motor yang perlahan dan keadaan kilas yang rendah Keputusan
ujian simulasi dan eksperimen menunjukkan bahawa teknik yang dicadangkan berjaya
vii
menjana tahap fluks optimum yang adaptif serta meningkatkan kecekapan sistem
pemacu motor aruhan DTC dan pada masa yang sama kekalkan prestasi cemerlang
kelajuan sistem.
viii
TABLE OF CONTENTS
TITLE i
DECLARATION ii
DEDICATION iii
ACKNOWLEDGEMENT iv
ABSTRACT v
ABSTRAK vi
TABLE OF CONTENTS viii
LIST OF TABLES xiii
LIST OF FIGURES xiv
LIST OF SYMBOLS AND ABBREVIATIONS xxi
LIST OF APPENDICES xxiii
CHAPTER 1 INTRODUCTION
1.1 Introduction 1
1.2 Research Background 1
1.3 Problem Statements 3
ix
1.4 Research Objectives 4
1.5 Research Scope 4
1.6 Thesis Outline 5
CHAPTER 2 LITERATURE REVIEW
2.1 Introduction 7
2.2 Variable Speed Drive (VSD)-Induction Motor
Drive System
7
2.2.1 Scalar Control 10
2.2.2 Vector Control 11
2.2.2.1 Field Oriented Control
(FOC)
12
2.2.2.2 Direct Torque Control
(DTC)
14
2.3 Efficiency Optimization of Induction Motor 17
2.4 Efficiency Optimization Technique 21
2.4.1 Search Control Technique (Physics-
based)
23
2.4.2 Loss Model Technique (Model based
technique)
23
2.4.3 Hybrid Control Technique 24
2.5 Review of Existing Previous Work 25
2.5.1 Vector Control of Induction Motor 25
x
2.5.2 Efficiency Optimization of Vector
Control Induction Motor Drives
System
29
2.5.3 Result Comparison 39
2.6 Gap of Study 42
CHAPTER 3 RESEARCH METHODOLOGY
3.1 Introduction 44
3.2 Control Strategy of the Proposed Efficiency
Optimization of DTC IM Drive System
44
3.2.1 Flux and Torque Estimator Block 45
3.2.2 Flux and Torque Controller 47
3. 3 Proposed ANN Efficiency Optimization
Control Strategy
48
3.4 Learning Algorithm of Proposed ANNEO
Controller
52
3.5 Proposed ANN Speed Controller 58
CHAPTER 4 RESEARCH DESIGN AND RESULTS
4.1 Introduction 62
4.2 Overview of Systems Development 63
4.3 Simulation Model with the Implementation of
MATLAB/ SIMULINK
64
4.3.1 Direct Torque Controller Block 66
4.3.2 Space Vector Pulse Width
Modulation Inverter Block
67
4.3.3 Input Power and Efficiency Evaluation
Block
67
xi
4.3.4 Proposed ANN Efficiency Optimization
Controller Block
68
4.3.5 Proposed ANN Speed Controller Block 69
4.4 Hardware Development and Implementation 69
4.4.1 The Real-time Interface Integration
with dSPACE
70
4.4.1.1 Feedback of the Motor Speed
Signal
72
4.4.1.2 Feedback of the Motor
Current Signal
72
4.4.2 The Experiment Set-up 73
4.5 Results and Discussion 74
4.5.1 Simulation Tests for Efficiency
Improvement of the Proposed
ANNEO Controller.
75
4.5.2 Experimental Verification of the
Efficiency Improvement of the
Proposed ANNEO Controller
84
4.5.3 Results Discussion on Efficiency
Improvement of the Proposed
ANNEO Controller tests
93
4.4.4 Simulation Tests for Step Speed
Variation
104
4.5.5 Experimental Verification of Speed
Variation
106
4.5.6 Results Discussion on Speed Variations
Tests 108
4.5.7 Simulation Tests for Constant Speed
with Load Disturbance
109
4.5.8 Experimental Verification of Constant
Speed with Load Disturbance
112
xii
4.5.9 Results Discussion on Constant Speed
with Load Disturbance Tests 114
CHAPTER 5 CONCLUSION AND FUTURE WORK
5.1 Conclusion 116
5.2 Future Work 117
LIST OF PUBLICATIONS 119
REFERENCES 123
APPENDIX 135
VITAE 175
xiii
LIST OF TABLES
2.1 Gap of study for the efficiency optimization IM
drives from several recent works where the input
power is used as the objective function of the EO
controller
43
4.1 Resulting flux level determined by the adaptive
ANNEO for different operating points ( Simulation )
93
4.2 Resulting flux level determined by the adaptive
ANNEO for difference operating points
( Experiment )
93
4.3 Efficiency of the DTC with Constant Flux and
Optimum Flux reference corresponding to different
speeds and torques ( Simulation )
94
4.4 Efficiency of the DTC with Constant Flux and
Optimum Flux reference corresponding to different
speeds and torques ( Experiment )
95
4.5 Settling time comparison for the speed variation on
simulation and experimental tests. 109
4.6 Speed deviation comparison for the sudden load
disturbance on simulation and experimental tests 115
xiv
LIST OF FIGURES
2.1 Block diagram of a variable speed induction motor drive
system
9
2.2 General classification of IM control methods 9
2.3 Indirect vector control method 13
2.4 Direct vector control method 13
2.5 Hysteresis controlled of DTC induction motor drive 15
2.6 Block Diagram of a SVPWM-DTC IM drive system 17
2.7 Torque production with three different flux level in the light
load condition. (a)-nominal flux, (b) medium flux (optimum
flux) and (c) low flux
19
2.8 Types of losses of converter drive system with the flux
variation
20
2.9 Efficiency improvement by the flux program at variable
torque with constant speed
21
2.10 Categories of efficiency optimization control 22
2.11 Induction motor drive operated under real-time Efficiency
Optimization Technique
22
2.12 ωr and Te as the input of the optimization control block 30
2.13 ωr ,Te, Pin as well as the ∆ω as the inputs of the optimization
control block
33
2.14 Utilize the Te as the input to the optimization control block 35
2.15 Input to the optimization control block accordingly to the
input power, Pin and output power, Pout
38
3.1 Block Diagram of an efficiency optimization control system
of DTC IM drives
45
xv
3.2 Block Diagram of an efficiency optimization control system
of DTC IM drives
47
3.3 Architecture of proposed ANNEO controller 50
3.4 Architecture of proposed ANN speed controller 59
4.1 The general testing diagram of the proposed ANNEO DTC 64
4.2 Complete Simulink block of the proposed efficiency
optimization control for the direct torque control induction
motor drive system
65
4.3 Simulink circuit of the Direct Torque Control block 66
4.4 Simulink circuit of the SVPWM control technique for the
three phase’s inverter
67
4.5 Simulink block for the efficiency and input power evaluation. 68
4.6 Simulink circuit for the proposed ANN efficiency optimization controller
69
4.7 Simulink circuit for the proposed ANN speed controller 69
4.8 The connection between Matlab and dSPACE 70
4.9 Real-time interface model with dSPACE I/O Blocks 71
4.10 The experimental set-up of proposed efficiency optimization
control for the direct torque control induction motor drive
system
73
4.11 Simulation results of the system performance for the speed
at 500rpm and a 0.2Nm load torque applied where the
proposed ANNEO controller is activated at the time =0.5s:
(a) before the ANNEO controller is activated; (b) after the
ANNEO controller is activated
76
4.12 Simulation result of the d-q stator flux circle at 500rpm and
a 0.2Nm load torque applied, before and after the proposed
ANNEO controller activated
76
4.13 Simulation results of the system performance for the speed
at 500rpm and a 0.8Nm load torque applied where the
proposed ANNEO controller is activated at the time =0.5s:
(a) before the ANNEO controller is activated; (b) after the
ANNEO controller is activated
77
xvi
4.14 Simulation result of the d-q stator flux circle at 500rpm and
a 0.2Nm load torque applied, before and after the proposed
ANNEO controller is activated
77
4.15 Simulation results of the system performance for the speed
at 800rpm and a 0.2Nm load torque applied where the
proposed ANNEO controller is activated at the time =0.5s:
(a) before the ANNEO controller is activated; (b)after the
ANNEO controller is activated
78
4.16 Simulation result of the d-q stator flux circle at 500rpm and
a 0.2Nm load torque applied, before and after the proposed
ANNEO controller is activated
78
4.17 Simulation results of the system performance for the speed
at 800rpm and a 0.8Nm load torque applied where the
proposed ANNEO controller is activated at the time =0.5s:
(a) before the ANNEO controller is activated; (b) after the
ANNEO controller is activated
79
4.18 Simulation result of the d-q stator flux circle at 800 rpm and
a 0.8Nm load torque applied, before and after the proposed
ANNEO controller is activated
79
4.19 Simulation results of the system performance for the speed
at 1100rpm and a 0.2Nm load torque applied where the
proposed ANNEO controller is activated at the time =0.5s:
(a) before the ANNEO controller is activated; (b) after the
ANNEO controller is activated
80
4.20 Simulation result of the d-q stator flux circle at 1100rpm
and a 0.2Nm load torque applied, before and after the
proposed ANNEO controller is activated
80
4.21 Simulation results of the system performance for the speed
at 1100rpm and a 0.8Nm load torque applied where the
proposed ANNEO controller is activated at the time =0.5s:
(a) before the ANNEO controller is activated; (b) after the
ANNEO controller is activated
81
xvii
4.22 Simulation result of the d-q stator flux circle at 1100rpm and
a 0.8Nm load torque applied, before and after the proposed
ANNEO controller is activated
81
4.23 Simulation results of the system performance for the speed
at 1400rpm and a 0.2Nm load torque applied where the
proposed ANNEO controller is activated at the time =0.5s:
(a) before the ANNEO controller is activated; (b) after the
ANNEO controller is activated
82
4.24 Simulation result of the d-q stator flux circle at 1400rpm and
a 0.2Nm load torque applied, before and after the proposed
ANNEO controller is activated
82
4.25 Simulation results of the system performance for the speed
at 1400rpm and a 0.8Nm load torque applied where the
proposed ANNEO controller is activated at the time =0.5s:
(a) before the ANNEO controller is activated; (b) after the
ANNEO controller is activated
83
4.26 Simulation result of the d-q stator flux circle at 1400rpm and
a 0.8Nm load torque applied, before and after the proposed
ANNEO controller is activated
83
4.27 Experimental results of the system performance for the
speed at 500rpm and a 0.2Nm load torque applied where the
proposed ANNEO controller is activated at the time =10s
(a): before the ANNEO controller is activated; (b): after the
ANNEO controller is activated
85
4.28 Experimental results of the system performance for the
speed at 500rpm and a 0.8Nm load torque applied where the
proposed ANNEO controller is activated at the time =10s.
(a): before the ANNEO controller is activated; (b): after the
ANNEO controller is activated
86
4.29 Experimental results of the system performance for the
speed at 800rpm and a 0.2Nm load torque applied where the
proposed ANNEO controller is activated at the time =10s.
87
xviii
(a): before the ANNEO controller is activated; (b): after the
ANNEO controller is activated
4.30 Experimental results of the system performance for the
speed at 800rpm and a 0.8Nm load torque applied where the
proposed ANNEO controller is activated at the time =10s.
(a): before the ANNEO controller activated; (b): after the
ANNEO controller activated
88
4.31 Experimental results of the system performance for the
speed at 1100rpm and a 0.2Nm load torque applied where
the proposed ANNEO controller is activated at the time
=10s. (a): before the ANNEO controller is activated; (b):
after the ANNEO controller is activated
89
4.32 Experimental results of the system performance for the
speed at 1100rpm and a 0.8Nm load torque applied where
the proposed ANNEO controller is activated at the time
=10s. (a): before the ANNEO controller is activated; (b):
after the ANNEO controller is activated
90
4.33 Experimental results of the system performance for the
speed at 1400rpm and a 0.2Nm load torque applied where
the proposed ANNEO controller is activated at the time
=10s. (a): before the ANNEO controller is activated; (b):
after the ANNEO controller is activated
91
4.34 Experimental results of the system performance for the
speed at 1400rpm and a 0.8Nm load torque applied where
the proposed ANNEO controller is activated at the time
=10s. (a): before the ANNEO controller is activated; (b):
after the ANNEO controller is activated
92
4.35 The efficiency comparison for rated flux and optimal flux
value for the speed of (a) 1400rpm, (b) 1100rpm, (c) 800rpm
and (d)500rpm corresponding to different torques value
(simulation)
97
4.36 The efficiency comparison for rated flux and optimal flux
value for the speed of (a) 1400rpm, (b) 1100rpm, (c) 800rpm
99
xix
and (d)500rpm corresponding to different torques value
(experimental)
4.37 The efficiency comparisons for rated flux and optimal flux
value for different load torques, (a) 0.2Nm, (b) 0.4Nm, (c)
0.6Nm and (d) 0.8Nm corresponding to different speed value
(simulation)
101
4.38 The efficiency comparisons for rated flux and optimal flux
value for different load torques, (a) 0.2Nm, (b) 0.4Nm, (c)
0.6Nm and (d) 0.8Nm corresponding to different speed value
(experimental)
103
4.39 The speed, current and torque response for a step speed
variation under the 0.2Nm load torque applied: (a) online
learning ANNSC, (b) offline learning ANNSC
105
4.40 The speed, current and torque response for a step speed
variation under the 0.8Nm load torque applied: (a) online
learning ANNSC, (b) offline learning ANNSC
106
4.41 The speed and current response for a step speed variation
under the 0.2Nm load torque applied: (a) online learning
ANNSC, (b) offline learning ANNSC
107
4.42 The speed and current response for a step speed variation
under the 0.8Nm load torque applied: (a) online learning
ANNSC, (b) offline learning ANNSC
107
4.43 The speed, current and torque response for a constant speed
of 500rpm with load disturbance of 0.2Nm and 0.8Nm at the
time of 0.7s, and 1s: (a) online learning ANNSC, (b) offline
learning ANNSC
110
4.44 The speed, current and torque response for a constant speed
of 800rpm with load disturbance of 0.2Nm and 0.8Nm at the
time of 0.7s, and 1s: (a) online learning ANNSC, (b) offline
learning ANNSC
110
4.45 The speed, current and torque response for a constant speed
of 1100rpm with load disturbance of 0.2Nm and 0.8Nm at
111
xx
the time of 0.7s, and 1s: (a) online learning ANNSC, (b)
offline learning ANNSC
4.46 The speed, current and torque response for a constant speed
of 1400rpm with load disturbance of 0.2Nm and 0.8Nm at
the time of 0.7s, and 1s: (a) online learning ANNSC, (b)
offline learning ANNSC
111
4.47 The speed and current responses at constant speed of
500rpm with load disturbance of 0.2Nm and 0.8Nm at the
time of 6s, and 10s: (a) online learning ANNSC, (b) offline
learning ANNSC
112
4.48 The speed and current responses at constant speed of
800rpm with load disturbance of 0.2Nm and 0.8Nm at the
time of 6s, and 10s: (a) online learning ANNSC, (b) offline
learning ANNSC
113
4.49 The speed and current responses at constant speed of
1100rpm with load disturbance of 0.2Nm and 0.8Nm at the
time of 6s, and 10s: (a) online learning ANNSC, (b) offline
learning ANNSC
113
4.50 The speed and current responses at constant speed of
1400rpm with load disturbance of 0.2Nm and 0.8Nm at the
time of 6s, and 10s: (a) online learning ANNSC, (b) offline
learning ANNSC
114
xxi
LIST OF SYMBOLS AND ABBREVIATIONS
f - Frequency
Idq - d-q Axis Current
Ls - Stator Self Inductances
Lr - Rotor Self Inductances
Lm - Mutual Inductances
p - Pole
Pin - Input Power
Pout - Output Power
Rs - Stator Resistance
Rr - Rotor Resistance
Te - Torque
Vdq - d-q Axis Voltage
ѰrefANN - Optimum Flux reference
Ѱref0 - Rated Flux reference
η - Efficiency
xxii
ANNEO - Artificial Neural Network Efficiency Optimization
ANNSC - Artificial Neural Network speed controller
DTC - Direct Torque Control
FOC - Field Oriented Control
GA - Genetic Algorithms
IM - Induction Motor
LMC - Loss-Mode-Based Control
PWM - Pulse Width Modulation
SC - Search Control
SVPWM - Space Vector Pulse Width Modulation
VC - Vector Control
VSD - Variable Speed Drive
xxiii
LIST OF APPENDICES
A Modeling of induction Motor 135
B Space Vector Pulse Width Modulation technique ( SVPWM )
141
C Motor Parameters 153
D C++ Program for the Proposed Drive System 154
E DS1103 Controller Board 174
1
CHAPTER 1
INTRODUCTION
1.1 Introduction
In this chapter, the introduction of this research will be explained in detail which
consists of the research background towards the focusing of the research study,
problem statements, research objectives, research scopes and research outline.
1.2 Research Background
The increasing emphasis on energy saving highlights the importance of attaining
higher motor efficiency under all operating conditions especially for industrial
applications. According to International Energy Agency (IEA) 2011, it was estimated
that electric motor drive account for between 43% and 46% of all global electricity
consumption, giving rise to about 6040Mt of CO2 emissions. By 2030, without
comprehensive and effective energy‐efficiency policy measures, energy consumption
from electric motors is expected to rise to 13360 TWh per year and CO2 emissions to
8570Mt per year. In fact, 64% of it was consumed by the industrial sector, and about
15% of final energy use in industry worldwide (IEA 2007), which is reported around
2
4488TWh per year of electricity consumption. It is estimated that a full implementation
of efficiency improvement options could reduce worldwide electricity demand by
about 7% (IEA 2008). The end‐users now spend USD 565 billion per year on
electricity used in the electric motor drive, by 2030, that could rise to almost USD 900
billion [1], [2].
Electric motors drive both core industrial processes, they are utilized throughout
all industrial branches though main applications vary. With only some exceptions,
electric motors are the main source for the provision of mechanical energy in industry.
In recent years, many studies have identified large energy efficiency potentials in
electric motors and motor systems with many saving options showing very short
payback period and high cost effectiveness [3].
Induction Motor (IM), is diffusely used in electrical devices and particularly
consume a large fraction of all electric power. Thus, they are responsible for more
energy consumed by the electric motor and as a prime target for efficiency
improvement. According to IEA’s estimates, about 68% of electricity consumed by
electric motors are used by the medium size motor, which rated from 0.75kW to
375kW, the vast majority of which are induction motors. Therefore, efficiency
improvement in IM has been getting much attention in recent years. Increasing
efficiency on IM gives significant to energy saving. The efficiency optimization is
important not only from the viewpoint of energy saving but also from broad
perspective of environmental pollution control [1].
In order to realize the efficiency improvement, there has been enhancements in
using high-quality design, construction techniques, and materials. However, expert
control algorithm contributes the most in improving drive performance especially
when a motor operates at light load [4]. It is commonly known that the IM has good
efficiencies while operating at full load. However, at lighter load, which is a condition
that many machines experience for a significant portion of their service life, the
efficiency decreases to a large extent. Therefore, it is important to maximize the
efficiency of the motor drive system while operating in adjustable speed applications.
Generally, the efficiency of an IM drives can be improved by reducing flux level
when it operates under light load conditions. Existing loss minimization techniques
have different approaches and are selected accordingly to an application or drive type.
At present, the optimal efficiency control techniques for IM drives are commonly
classified into two categories namely 1) Search Control, (SC) and 2) Loss-Mode-Based
3
Control, (LMC) [5]–[13]. Primary idea of both techniques is to make the flux amplitude
change with the change of IM operating conditions, even though these two types have
dissimilar achievement and implementation way [14]–[16].
1.3 Problem Statements
Basically, when an IM is operated at a rated condition, i.e. rated load torque and rated
speed, the efficiency of motor is high and gives the best transient response. However,
in many applications, a motor operates far from its rated operating point, particularly
at light load, which is a condition that many machines experience for a significant
portion of their service life. At light load, the reference flux magnitude is held to its
initial value, and this cause the efficiency decreased to a large extent. At this light load
condition, rated flux operation causes excessive core loss, thus impairing the drive
efficiency due to the imbalance between iron and copper losses.
It is well known that for a given operating point, the efficiency of IM can be
improved by minimizing its losses. This can be achieved by reducing appropriate level
of magnetic flux or by programming the flux to obtain the balance between the copper
and iron losses.
Primary challenge of the efficiency optimization control of the IM drive system at
variable load operation is to obtain the optimum flux value. This optimum flux value
ensures a minimum input power that leading to maximum efficiency of the drives
systems, at any operating point over the centire torque and speed range. Meanwhile, it
is also important to remain the good primeval characteristics of speed response. In
addition, the IM drive is normally claimed as unreliable and sensitive to the
environmental changes due to its non-linear characteristics. Thus, designing a robust
efficiency optimization control needs to take into account of temperature variation,
magnetic saturation and load disturbances.
Therefore, the challenge of this research is not limited to predict the extent to
which the flux can be reduced, at any operating points over the entire torque and speed
range, which maximizes the efficiency, while at the same time, preserving and
enhancing the system performance to deal with the uncertainties of the system, such as
the dynamic response load disturbances.
4
1.4 Research objectives
An investigation to determine the system efficiency as well as the drive performance
over a wide speed range has been done. The proposed control strategy should be able
to improve the overall efficiency while at the same time, preserving the superior
transient performance of the drive. The aim of this research is to develop an efficiency
optimization of a DTC induction motor drive system.
The objectives of this research are:
i. to design and simulate the proposed DTC induction motor drive using
Matlab/Simulink simulation package.
ii. to develop a controlled program of optimum DTC induction motor drive along
with the proposed ANNEO and ANNSC controller by using dSPACE DS1103
Real-time Digital Signal Controller.
iii. to evaluate the effectiveness and robustness of the proposed controllers
towards the system’s efficiency improvements and the system performances
by simulation and experimentally.
1.5 Research Scope
The scope of this research is to focus on the efficiency optimization of the DTC
induction motor drives while at the same time preserves the excellent speed
performances. The scope of this research will be conducted according to the following
stage:
i Simulation model development
- The Matlab Simulink is used in order to model and simulate the DTC IM
drive to identify the system performances.
5
ii Controller program development
- A controller for the efficiency optimization DTC motor drive system to
improve its efficiency is designed.
- The speed controller to deal with the uncertainties of the system such as the
speed variation and dynamic response load disturbance is designed.
iii Hardware development
- The prototype of the proposed efficiency optimization of a DTC induction
motor drive system is built.
- The interface of dSPACE DS1103 Real-time Digital Signal Controller board
for the proposed DTC motor drive is constructed.
- The performances of the developed optimization DTC motor drive is
evaluated.
1.6 Thesis Outline
The outlines of the structure for this thesis are given as follow:
Chapter 2 presents the overview of control method for the induction motor, the DTC
control technique are further discussed. The basic principle of the efficiency
optimization control of the IM drives as well as the concept of various efficiency
improvement algorithm is provided. In addition, the reviews of previous and currently
research conducted in efficiency optimization vector control of induction motor drives
systems are presented.
Chapter 3 devotes the developments of the complete drive system with the proposed
algorithm. The prospective of the proposed neural network control implemented on
the efficiency optimization control and the speed control is discussed. Levernberg-
Marquardt learning scheme is implemented for ANNEO controller. The EO controller
is designed based on the Artificial Neural Network search algorithm as a predictor of
the optimum flux in order to produce an adaptive flux level. On the other hand, instead
6
of the conventional PID based speed controller (SC), back propagation ANN based
speed controller, ANNSC is proposed in order to allow better performance. Generally,
the weight updating can be done in two primary ways, namely the online and offline
training. Both online and offline training will be applied to the proposed ANNSC in
this research to verify its validity. The control algorithm of the neural network is given
in details.
Chapter 4 provides the minutiae explanation for both simulation and hardware
implementation in this research. The proposed controllers are simulated by using
Matlab-Simulink. The experimental hardware setup is controlled through the dSPACE
DS1103 Real-time Digital Signal Controller and interfaced by the ControlDesk 5.1. In
addition, the promising performance of the proposed controllers is obtained through
simulation and then verified by the relevant experiment results. Further analysis and
discussion on the obtained result are provided.
Chapter 5 presents the summaries of the contribution of this research and the
recommendation for future research direction.
7
CHAPTER 2
LITERATURE REVIEW
2.1 Introduction
This chapter presents an overview of efficiency optimization for variable speed
induction motor drive, followed by its theoretical background. The fundamentals
knowledge of induction motor modeling with the involved transformation is first
described. Then, the fundamental efficiency optimization control approach is
described. Background studies of the available control method that have been
implemented in IM drive system are discussed further. Review of previous work for
various types of efficiency optimization control method for a variable speed drive as
well as the advantages and disadvantages are also presented.
2.2 Variable Speed Drive (VSD) - Induction motor drive system
The development in the field of electric drives has never stopped since the first
inception of the first principle of the electrical motor by Michael Faraday in 1821. The
world dramatically changed after the first induction machine was patented (US Patent
381968) by Nikola Tesla in 1888.
8
Before semiconductor and fast microprocessor or digital signal processor
become available, DC motors have been used extensively for the past decade. Even
though DC motor is more expensive, it provided simpler control structure due to its
inherent decoupled control of torque and flux. Meanwhile, the induction motors were
commonly applied as fixed speed machine due to their fixed connection to fixed
frequency and voltage supply. However, main drawbacks of DC motor were the
presence of brushes and commutators and thus signify the limitation of its workplace
[17]–[19].
Compared with DC Motor, the induction motor has distinct advantages, such as
having no commutator and brushes required, thus, leading to the facts as a
maintenance-free motor, ruggedness, lower rotor inertia, simpler protection,
economical, compactness [20]–[23]. The main drawback that makes AC motors retreat
from the industry was the inherent coupling between torque and flux.
Developing of new control principle, algorithm, and modern control have led to
a new generation of variable speed induction motor drives. The advent of power
electronic converters with forced commutation in 1960s and later with the power
semiconductors (BJT, GTO, and IGBT) made possible the use of the induction motor
as a variable speed drive in the last two decades [24], [25]. Variable speed drives have
facilitated the revolution of industrial automation leading to higher productivity and
better quality in various industries and home appliances.
Variable speed drives are established when an electric motor is combined with a
power electronic converter. By introducing variable speed to a driven load, it is
possible to optimize the efficiency of the entire systems and makes the greatest
efficiency gain possible. The variable speed induction motor drive system consist of
few component, such as the driver controller, a controllable power converter, an
electric motor which drives a mechanical load at any adjustable speed, as shown in
Figure 2.1 [25]–[27].
9
Figure 2.1: Block diagram of a variable speed induction motor drive system [25]–
[27].
The development of variable speed AC drives was tardy until the 1980s when
the rapid improvement in power electronics as well as the microprocessors revolution
has made the complex control algorithm possible. These advanced control methods
have made the induction motor possible for high performance and comparable with
that of the DC machine. This made the AC machines the dominating machine in the
drives market and seeing wider use in variable-speed applications [28], [29], [30].
Plenty of control methods for induction motor in the field of electric motor drives
have been drawing the attention of the researcher. Various control methods for IMs
have been proposed. Generally, IM control methods can be divided mainly into open
and closed loop control technique, which are also known as scalar and vector control
[28], [31]–[33]. The general classification of IM control methods is presented in Figure
2.2.
Figure 2.2: General classification of IM control methods [28], [31]–[33].
IM Control methods
Scalar based controller
Vector based controller
Field oriented control - FOC
Direct torque control - DTC
Controller
Power Converter IM Mechanical
Load
Commands
Power Supply
10
IM control methods are divided depending on what quantities they are control.
If magnitude of parameters is inspected and controlled, the control technique is
assigned as a scalar control. This technique is mainly implemented through direct
measurements of the machine parameter. It is usually employed in low-performance
variable speed drives. For high-performance variable speed drive systems, vector
control is implemented, in which the control variables consider the inspection and
controlled by the instantaneous values of positions of the controlled parameter, thus
permitting high dynamic performance of the drive. These techniques are realised in
both direct measurements and estimation of the machine parameter [32], [34], [35].
2.2.1 Scalar Control
At the beginning, a study on efficiency controller has been done for scalar control,
sometimes known as constant Volts/Hertz method. Scalar Control is the simplest
control method for the controlling the IM. Optimization of this method is based on the
relationship valid for steady states, with assumption that rated flux is proportional to
the voltage over constant frequency. The philosophy is to control the amplitude and
frequency of the stator voltage in order to keep the stator flux constant through motor
speed range [33].
Main advantages of this control method are that it is not complicated, low cost
and it is commonly used without position control requirements or the need for speed
feedback. Scalar control offers good steady-state response. However, this scheme
cannot perform decoupling between input and output, thus result in the problem of
independent control of outputs, for example, torque and flux. In addition, this
controller does not attain fine precision in either speed or torque response especially
in low speed since stator flux, and torque is not diametrically controlled [36]–[38].
Due to the uncontrolled transient and unknown position angle of control variables
resulted by lack of feedback, the system is unable to permit the best dynamic response
performance, which limits the application of this control method [39].
11
2.2.2 Vector Control
To obtain better dynamic and high precision, scalar control has been phased out by
more effective control methods. At present, the highlight is on vector control in order
to achieve decoupling in high-performance IM drives which offer far better dynamic
performance than those with scalar control. Vector control (VC) are aimed for
independent control of the machine torque and flux producing stator currents [40].
The vector control is based on the relationship valid for dynamic states, not only
the magnitude and frequency (angular speed), but also its instantaneous positions of
voltage and current are controlled. Thus, the vector control system acts on the position
of the space vectors and provides their correct orientation for both steady state and
transient conditions. This ensure dynamically decouples fast flux and torque control
and belongs to high-performance control implemented in a closed-loop fashion [28],
[31], [39], [41], [42]. Vector control presents several benefits, such as precise speed
regulation, a wide range of speed control and fast dynamic response.
Essentially, the control problem is reformulated to resemble the control of a DC
motor. Thus, the advantage of the DC motor control of being able to decouple the flux
and torque is thereby opened up. The vector control algorithm is based on two
fundamental ideas. The first is the flux and torque producing currents. An induction
motor can be modeled and controlled most simply using two quadrature currents rather
than the familiar three phase currents applied to the motor. These two currents are
called direct (Id), and quadrature (Iq), and they are responsible for producing flux and
torque respectively in the motor. By definition, the Iq current is in phase with the stator
flux, and Id is at right angles. Of course, the actual voltages applied to the motor and
the resulting currents are in the familiar three-phase system. The move between a
stationary reference frame and a reference frame, which is rotating synchronously with
the stator flux, becomes a problem then. This leads to the second fundamental idea
behind vector control. The second fundamental idea is that of reference frames. The
idea is on reference frame is to transform a quantity that is sinusoidal in one reference
frame, to a constant value in a reference frame, which is rotating at the same frequency.
Once a sinusoidal quantity is transformed to a constant value by careful choice of
reference frame, it becomes possible to control that quantity with appropriate
controller [32],[43].
12
These two main control methods can be further divided into a few number of
different control strategies depending on their functionality [39]. The vector control
can be implemented in many different ways but only several basic schemes that are
offered on the market. The most popular strategies among them are Field Oriented
Control (FOC) and Direct Torque Control (DTC) and Direct Torque Control - Space
Vector Pulse Width Modulation (DTC-SVM) [29], [31], [33], [34], [39].
In early 1970s, the appearance of the Field oriented control (FOC) allowed a
considerable increase of dynamic performance of the induction motors [44].
Theoretically, FOC that based on Fleming's law [45] makes the control performance
of induction motor as good as the DC motor’s where torque and flux are decoupled
and hence could be controlled independently. However, during the practical practice
of engineering application, the actual performance of vector control will be worse than
predicted due to the effect of factors such as inaccurate control model and variable
motor parameters [46]. Several methods are investigated to inquire into this problem
and some improved techniques such as flux observer, rotor resistance identification
are adopted in order to reduce the effect of this variation so that the control
performance of FOC can be satisfied in most of applications [44], [45]. The Direct
Torque Control was first introduced by Takahashi around the mid-1980s has found
great success with the notion to reduce the dependence on parameters of induction
motor and increase the precision and the dynamic of flux and torque response [47].
2.2.2.1 Field Oriented Control (FOC)
FOC have made possible the application of induction motors for high-performance
applications where traditionally only DC drives were applied. The field oriented
scheme enables the control of the induction motor in the same way as separately
excitation DC motors. One of the methods used in variable frequency drives or variable
speed drives to control the torque and thus the speed of three-phase AC electric motors
by controlling the current. As in the DC motor, with FOC torque control of induction
motor is achieved by controlling the torque current component and flux current
component independently.
13
The basic schemes of indirect and direct methods of FOC are shown in Figures
2.3 and 2.4. The direct FOC method depends on the generation of unit vector signals
from the stator or air-gap flux signals. The air-gap signals can be measured directly or
estimated from the stator voltage and current signals. The stator flux components can
be directly computed from stator quantities. In these systems, the rotor speed is not
required for obtaining rotor field angle information. In the indirect FOC method, the
rotor field angle and thus the unit vectors are indirectly obtained by summation of the
rotor speed and slip frequency [28].
Figure 2.3: Indirect vector control method [28].
Figure 2.4: Direct vector control method [28].
The FOC consists of controlling the stator currents represented by a vector. This
control is based on projections which transform a three phase time and speed
dependent system into a two coordinates, d and q coordinates time invariant system.
These projections lead to a structure similar to that of a DC machine control.
Field orientated controlled machines need two constants as input references
which are the torque component that aligned with the q coordinate and the flux
V I
Inverter
IM FOC
Slip
1/s
Ψ* Te*
θ
+ +
FOC Inverter IM
Ψ Vector Measurement /Estimation
θ
V I Ψ* Te*
ω
ω ω_cal
14
component that aligned with d coordinate. As FOC is simply based on projections, the
control structure handles instantaneous electrical quantities. This makes the control
accurate in every working operation, and independent of the limited bandwidth
mathematical model.
Fundamental requirements for the FOC are the knowledge of two currents and
the rotor flux position. Knowledge of the rotor flux position is the core of the FOC. In
fact, if there is an error in this variable the rotor flux is not aligned with d-axis and the
current components are incorrectly estimated. In the induction machine, the rotor
speed is not equal to the rotor flux. The basic method is the use of the current model.
Thanks to FOC it becomes possible to control, directly and separately, the torque and
flux of the induction motors. Field oriented controlled induction machines obtain every
DC machine advantage is instantaneous control of the separate quantities allowing
accurate transient and steady state management.
2.2.2.2 Direct Torque Control (DTC)
Direct Torque Control, DTC is a suitable technology for the high performance electric
drive system and is characterized by simple algorithms, fast dynamic response, and
strong robustness. It was introduced in the middle of 1980’s and was considered as an
alternative technique to the FOC. It was first introduced for IM by Takahashi and
Noguchi [47] in 1984 and the Direct Self Control method by Depenbrock [46], [48],
[49] in 1985.
DTC scheme is very simple in its basic, which consists of DTC controller, torque
and flux calculator, and Voltage Source Inverter (VSI). In recent years, the use of DTC
strategies has become more universal and popular for induction motor drives and
seems to have very rapid growth and development. The methods were characterized
by their simplicity, good performance, and robustness. The control philosophy of DTC
is to control directly the inverter state to maintain the stator flux and torque within
hysteresis band limits which requires no current regulator loops while similar
performance of the FOC can be achieved at the same time or much better.
Unlike the FOC method, DTC works without any external measurement of rotors
mechanical position. It has many advantages compared to FOC, such as less machine
15
parameter dependence, simpler implementation, and quicker dynamic torque response.
It only needs to know the stator resistance and terminal quantities (v and i) in order to
perform stator flux and torque estimations. The configuration of DTC is simpler than
FOC system due to the absence of frame transformer, current controlled inverter and
position encoder, which introduces delays and requires mechanical transducer.
The configuration of the conventional DTC drive as proposed by Takashi is
shown in Figure 2.5. It consists of two loops, the magnitude of stator flux and the
torque respectively. Hence, DTC performs the separate control of the stator flux and
torque, which is known as decouple control.
Figure 2.5: Hysteresis controlled of DTC induction motor drive [47].
The three level hysteresis comparator take the output error between the estimated
torque T and the reference torque Tref while the error from the estimated stator flux Ψ
and reference flux magnitude Ψref is fed into the two-level hysteresis comparator [50].
The estimated values are calculated by means of the adaptive motor model [51]. The
pair of hysteresis comparators is used to minimize torque and flux errors to zero.
Besides, it also determines the appropriate voltage vector selection and the period of
voltage vector selected.
The position of the stator flux is sync with the resulting torque error and flux
error of the hysteresis block are used as the input for the selection table. The position
of the stator flux is divided into six different sections. The decent voltage vector is
selected based on the selection table. The selection table blocks are responsible for
proper inverter switching state selection at each sampling time in order to confirm the
torque and flux errors lays within hysteresis band [52].
As a matter of fact, the fundamental concept of DTC is to produce the appetent
torque by diametrically manipulate the stator flux vector. The instantaneous value of
Hysteresis Comparator
Hysteresis Comparator
Voltage Source Inverter (VSI)
Stator Flux and
Torque Estimator
T
Terr
θ
Tref*
Ѱref
Ѱ
Ѱerr
Sa
0 IM Voltage vector
Selector
V
Sb Sc
I
16
the stator flux and torque are calculated from the stator variable by utilizing a close
loop torque and flux calculator. By selecting appropriate inverter state, independent
and direct control of stator flux and torque can be achieved. Meanwhile, the stator
voltage and currents are indirectly controlled and therefore, currents feedback loops
are unnecessary [53], [54].
The major problems in conventional hysteresis-based DTC was the performance
of the system directly dependent on estimation of stator flux and torque, thus,
improper estimation will result in incorrect voltage vector selection. In addition,
conventional hysteresis-based DTC also produces high torque ripples, stator current
distortion in terms of low-order harmonics and variable switching frequency due to its
hysteresis comparators [18]. Variety methods have been proposed to overcome these
issues, space vector modulation technique was one of the enticing candidates.
Although DTC is gaining its popularity, there are some drawbacks that can be
rectified to improve the performance. The space vector depends on the reference
torque, and flux is used to solve this problem [55]. Reference voltage vector is then
realized using a voltage vector modulator. DTC based on SVPWM not only preserve
its transient virtue but also yields the superior performance in the steady state over a
wide speed range.
The SVPWM technique is somewhat similar to the Sine 3rd harmonic PWM
technique, but the means of implementation is different. It is used to emerge desire
voltage or current for motor phase signal. This control method is progressively
welcomed for the AC drive motor applications since it establish smallest harmonic
currents and produce largest output voltage with same DC bus voltage. Commonly,
PWM control technique compares the three phase’s sinusoidal waveforms with a
triangle carrier to produce switching position patterns. The space vector theory denotes
some extra enhancements for both harmonic copper loss and output crest voltage. The
maximum output voltage produced is 32 =1.155 times larger based on space vector
theory as compared to traditional sinusoidal modulation. This makes possibilities for
higher voltage to feed the motor than the easier sub-oscillation modulation method.
Higher torque at high speed with high efficiency can be achieved by this modulator
[56], [57].
In conventional hysteresis based DTC systems, next switching condition of the
VSI is generated directly from torque and flux errors. Nonetheless, the SVPWM-DTC
17
import a constant switching frequency signal to VSI by the stator reference voltage
vectors which produce from the flux and torque errors [58]. Block scheme of the DTC-
SVPWM is presented in Figure 2.6.
Figure 2.6: Block Diagram of a SVPWM-DTC IM drive system [56]–[58].
As shown in Figure 2.6, the motor speed ω is compared with the desired speed
reference ωref , the resulting speed errors is fed into the proposed controller to generate
the torque reference value, Tref. The output errors, Terr between the estimated torque T
and the reference torque Tref, as well as the errors from estimated stator flux Ψ and
reference flux magnitude Ψref, is fed into the torque and flux controller respectively.
The estimated values are calculated by means of the adaptive motor model.
The 3 phase’s current, Iabc is required to transform to the 2 phase’s current, Idq by
the Clarke transformation. The two phase’s voltage, Vdq is generated through the torque
and flux controller from respective errors and convert back to the three phase’s voltage
to feed the SVPWM blocks.
2.3 Efficiency Optimization of Induction Motor
Induction motors, particularly those are widely used in electrical devices, consume
large fraction of electric power. Thus, they responsible for most energy consumption
and as prime target for improvement in efficiency [59]. In order to realize the efficiency
improvement, some researches enhanced it by using high-quality materials [60]–[62],
design and construction techniques. However, there is another promising algorithm that
can be applied directly to a drive controller, which is the expert control algorithm. It
0
Torque controller
Flux controller
SVPWM
IM
Flux and Torque Estimator
Speed controller
T
Terr ωref
ω
ωerr Tref*
Ѱref
Ѱ
Ѱerr
VDC
2ϕ 3ϕ
Vabc Vds Vqs
2ϕ
3ϕ
Iabc Ids Iqs
18
contributes the most to improve drive performance, especially when the motor operates
at non-rated condition, which is low speed and light load [5], [63]. The foundation of
such a control can be described as follow.
Basically, when an induction motor operates at rated condition, i.e. rated load
torque and speed, the efficiency of the motor is quite high and gives the best transient
response. However, in many applications, a motor operates far from the rated operating
point, particularly at light load where the reference flux magnitude is held on its initial
value, and this causes problems. At light load, rated flux operation causes excessive
core loss, thus impairing the efficiency of the drive due to imbalance between iron and
copper losses. For a given operating point, the induction motor efficiency can be
improved by minimizing losses and reducing the magnetic flux level appropriately or
by programming the flux to obtain balance between the copper and iron losses [6],
[25], [52], [64]–[66]. Due to the fact that electromagnetic losses in a machine is a direct
function of magnetic flux, thus, with proper adjustment of flux, the appropriate balance
between iron and copper losses can then be achieved [7], [63], [66]–[68]. In this
condition, the motor flux is more than necessary for the development of required
torque. Therefore, to improve the motor efficiency, its air gap flux must be reduced.
The strategy to reduce drive losses to minimum or in other words, to obtain
minimum power input by adjusting flux level according to motor load, is called
energy-optimal control [69], [70]. This control strategy is also known as efficiency
optimization control or loss minimization controls, some named it as part load
optimization as well.
The fundamental control strategy of efficiency optimization control is hereafter
explained. Electromagnetic torque of the induction motor can be approximated by
[43], [71].
rme IIkT = (2.1)
Where : Te= electromagnetic torque.
K=constant.
Im=magnetizing current.
Ir=rotor current.
19
For a given load torque, the motor’s electromagnetic torque can be obtained by
the combination of magnetizing current and torque producing rotor current. Therefore,
it is possible to achieve same torque with different combination of flux and current
value. The motor is normally designed to work with optimum near rated load.
However, for every speed and load condition, there exists an optimum flux level where
maximum efficiency can be achieved [67], [72].
Generally, low efficiency achievement during low load condition is due to
inappropriate flux level selection. The efficiency improvement can be achieved by
proper flux adjustment. Illustration of relationship between flux level and
electromagnetic torque is shown in Figure 2.7.
Figure 2.7: Torque production with three different flux level in the light load
condition. (a)-nominal flux, (b) medium flux (optimum flux) and (c) low
flux [67], [72].
As depicted in Figure 2.7, a vector diagram of low loaded motor at three different
rotor flux levels, which are nominal, medium and low are presented. The developed
electromagnetic torque, represented by the shaded areas, is proportional to ΨrIr, for all
the three cases. At nominal flux, as shown in Figure 2.7 (a), the rotor current, Ir is small
but the stator current, Is and magnetizing current, Im is large. Thus, rotor copper losses
are low but both stator copper and core losses are high.
When rotor flux is reduced to half of its nominal value, as shown in Figure 2.7
(b), the magnetizing current, Im and stator current, Is are reduced. However, the rotor
current, Ir is doubled to create same amount of electromagnetic torque. This reduces
core and stator copper losses considerably but increases rotor copper losses. As a
result, motor losses in Figure 2.7 (b) are smaller than the one in Figure 2.7 (a).
Ψr
Im Ir Is torque
Ir
Ψr
Im
Is torque
Ψr
Im
Is Ir
torque
(a) (b) (c)
20
However, if rotor flux is reduced even more, which is shown in Figure 2.7 (c), the core
losses will still decrease but as both rotor and stator copper losses increase again,
hence, it increases total losses again. A clearer depiction of the relationship between
flux level and losses are show in Figure 2.8 [71].
Figure 2.8: Types of losses of converter drive system with the flux variation [71].
For certain steady state, in light load condition and at certain speed, typical losses
in a converter drive system and its variation with the adjustable flux is presented in
Figure 2.8. When flux, Ψr is reduced from rated value, core loss will decrease, but
copper and converter losses increase. However, the total loss decreases to a minimum
which leads to maximum efficiency. However, one can observe that total losses
increase again after optimum value. Thus, it is desirable to set rotor flux at Ψr
(optimum) so that overall drive system efficiency is optimum.
Figure 2.9 shows typical optimum flux program for variable torque and constant
speed and compares corresponding efficiency with that of rated flux. Note that at rated
torque condition, the flux should be set close to rated value, and there is no significant
efficiency improvement. The improvement of efficiency becomes larger as the torque
is decreased.
Time
Torque Speed Total Loss
Copper loss
Converter loss Core loss
Flux (Ψr) →Decreasing ↑ Ψr (rated)
↑ Ψr (optimum)
21
Figure 2.9: Efficiency improvement by the flux program at variable torque with
constant speed [71].
On the basic of the flux reduction in the three cases in Figures 2.7 and 2.8, it can
be concluded that, for a given load, minimum losses can be achieved by the proper
adjustment of flux to obtain optimum flux level.
Nevertheless, flux reduction has a lot of limitations. For a given operating point,
when the flux is reduced, stator frequency will increase, hence, the pull-out torque of
the motor will be reduced and make it more sensitive to sudden load disturbances.
Thus, the speed for vector controlled motor drive will drop significantly and only when
the motor field has been restored, the motor speed can be recovered. Therefore, when
designing efficiency optimization control, it is important to ensure that the drive can
withstand load disturbances with optimum flux obtained.
2.4 Efficiency Optimization Technique
A number of strategies have been published on efficiency optimization control for IM
drives system, especially at light load. In general, the optimal efficiency control
techniques or Loss Minimization Technique for IM drives in real time are commonly
classified by two categories namely, (1) Search Control (SC), and (2) Loss-Mode-
Based Control (LMC) as shown in Figure 2.10. The primary idea is to make flux
amplitude varies with the change of IM operating conditions even though these two
Ter Torque (Te) 0
Flux program for
ƞ optimization
Ψr (rated)
ƞ at programmed flux
ƞ at rated flux Ψr
Effic
ienc
y (ƞ
)
Flux
(Ψr)
22
types have dissimilar achievement and implementation way [5]–[9], [14], [15], [73]–
[78]. Afterward, the third technique, (3) Hybrid, which uses both techniques
mentioned above was invented. It improves each technique's disadvantages to achieve
a better performance [14], [15], [42]. On the other hand, certain replace the SC
approach with others such as Minimum Stator Current (MSC) [79], estimation of iron
losses [80] , and power-factor improvement method [4], [64]. However, this method
is suboptimal and cannot produce maximum efficiency. Figure 2.11 show a typical IM
motor drive operates under control of real-time efficiency optimization control
technique.
Figure 2.10: Categories of efficiency optimization control [5]–[9].
Figure 2.11: Induction motor drive operated under real time Efficiency Optimization
Technique [14].
Control Algorithm
Inverter
+ DC bus -
Sensors/ Estimators
Induction Machine
Commands
Efficiency Optimization Vector Control of IM
Search Control (SC)
Hybrid Control
Loss Model- Based Control (LMC)
Efficiency Optimization
Technique
23
2.4.1 Search Control Technique (Physics-based technique)
SC utilized measured IM input power, Pin or DC-link power in optimization process.
It drives the control input to minimize Pin from the current and voltage measurements,
regardless of the motor rating or parameter. For a given output power, Pin=Po+Ploss,
minimizing Pin is equivalent to minimizing Ploss for a given Po [81].
This approach based on varying the flux up to the point where the measuring
power input is minimum for one point of operation [42]. More detail, at steady state,
for a given load torque and speed, the flux level (or its equivalent variables) iteratively
search until minimum input power is achieved. This method has the merit of robustness
to parameter variation while the effects of parameter changes are expressed in loss
model and hybrid strategies. In addition, it does not require any model of the system
and thus search process is simple and is applicable universally to any motors [15], [79].
The drawback of this approach is relatively long response time, flux convergence
to its optimal value is too slow and its performance is dependent on the quality of
power input measurement. In addition, noise and disturbances are not measured
accurately in measured signal. Thus, the optimal flux cannot be assured, and maximum
efficiency cannot be obtained. Moreover, the torque and flux ripples will be produced
because in real-time drive system, the stable state will not be achieved when search
process is slow. In simple way, the search controller utilizes feedback to measure input
power and iteratively change flux level until minimum input power is detected [5],
[14], [15], [42], [79], [82], [83].
2.4.2 Loss Model Technique (Model based technique)
Loss model controller, LMC utilizes the model of all important losses occurring in IM
during operation to compute optimum flux for a given load and speed that minimize
losses, thus they depend on motor parameters. It does not include closed-loop power
measurement but might implement feedback [14]. This approach does not produce
torque ripples [5], with fast response time and high speed of convergence in searching
minimum loss point and stability in the drive when great change in torque or speed is
24
demanded [6], [8], [42], [82], [84]–[86]. Since the LMC do not rely on any hardware
modification on existing drives, makes it more popular and universal than minimum
power input (search) based efficiency controller [42].
However, the performance of LMC depends on the accuracy of modeling of
motor drive and losses. There is always a trade-off between accuracy and complexity
in the development of loss model [9]. In addition, this approach is not robust to
parameter variation due to temperature changes and magnetic circuit saturation, it
requires exact knowledge of parameter for controlled machine to achieve true optimal
operation, the power loss modeling and calculation of the optimal control can be very
complex [15]. Considering that, both fundamental losses and harmonic losses
produced by inverter should be taken into consideration. Incorporating with the afore
mentioned parameter variation phenomena, therefore, the model of the drive system is
complex, and simplified model will result in suboptimal output [79].
In a compendious way, loss-model-based controllers use a functional loss model
to compute losses and to select an optimum flux level that minimizes this loss. This
method is faster than search methods but sensitive to parameter variations [82], [87].
Among loss minimization algorithm for an induction motor, a loss model based
approach has the advantages of fast response and high accuracy [9].
2.4.3 Hybrid Control Technique
The hybrid Control technique combines the features from both search and loss model
control techniques, which proposes mixing good characteristics of two optimization
strategies. Hybrid techniques require a motor or system model to search for minimum
power loss, and then use electromechanical principles and mathematical
characteristics to achieve optimality. It combines both advantages of the mentioned
approach but also rectifies the disadvantages. The two main examples of hybrid
techniques among several possible combinations of search and loss model based control
characteristics are: (i) applying Perturb and Observe (P&O) on Ploss model, and (ii) using
a parameter-dependent estimator with a search control technique [81].
The hybrid technique is not a straight forward process, each control technique
performs control strategy individually. During transient process LMC is used to