Professional Documents
Culture Documents
Hak cipta terpelihara. Tiada dibenarkan mengeluar ulang mana-mana bahagian artikel,
ilustrasi, dan isi kandungan buku ini dalam apa juga bentuk dan cara apa jua sama ada
dengan cara elektronik, fotokopi, mekanik, atau cara lain sebelum mendapat izin bertulis
daripada Timbalan Naib Canselor (Penyelidikan dan Inovasi), Universiti Teknologi
Malaysia, 81310 Skudai, Johor Darul Ta’zim, Malaysia. Perundingan tertakluk kepada
perkiraan royalti atau honorarium.
All rights reserved. No part of this publication may be reproduced or transmitted in any
form or by any means, electronic or mechanical including photocopy, recording, or any
information storage and retrieval system, without permission in writing from Universiti
Teknologi Malaysia, 81310 Skudai, Johor Darul Ta’zim, Malaysia.
Modelling and control of power converters and drives / editor Nik Din Muhamad.
Includes index
ISBN 978-983-52-0648-1
1. Electric current converters. I. Nik Din Muhamad.
621.313
CONTENTS
Preface vii
PREFACE
1
CONTROL LOOP DESIGN OF DC-DC
CONVERTER SYSTEMS USING PSPICE
1.1 INTRODUCTION
In recent years, there are many SPICE-based models developed
for simulating dc-dc converters. In general, three types of model
can be identified: the detailed model, the large-signal averaged
model, and the small-signal model. All three models are valuable
in analysis and design of dc-dc converter systems. The small-
signal model is required to design the control system of a dc-dc
converter. There are a number of well-documented techniques and
guidelines for designing the control system of a dc-dc converter [1-
6]. Nevertheless, the design of a feedback compensator is still not a
simple task, especially for a new designer. The task involves the
tedious, mechanical and human error prone computation of the
transfer functions, and repetitive fine-tuning of compensation
network component values. Moreover, the designer’s judgment
and experience are often required in the design process.
In this paper we demonstrate that the design guidelines can be
programmed in compact form in PSpice. For this purpose, an
option available in PSpice called Analog Behavioral Modeling
(ABM) is used. We choose to program the control loop design
procedure [4] due to its generic, systematic and a widely adopted
procedure [7-10]. The chosen design procedure is made as a
subcircuit model and stored in PSpice’s library. In this manner,
2 Modeling and Control of Power Converters and Drives
⎛ s ⎞
⎜⎜1 + ⎟
⎝ ω zESR ⎟⎠
= g co
vˆo
(1)
1+ +
s2
Qω o ω o2
vˆc s
where
Control Loop Design of DC-DC Converter Systems Using PSpice 3
g co = ωo = , ω zESR = , and Q =
Vg
ω oL
1 1 R
,
Vp LC rC
+
Vg + C R
- vo
Driver
-
PWM Modulator
d
vc
+
- Vp
Sawtooth
Comparator Generator
ω co ⎛ K ⎞
⎜ s + 1⎟⎟
2
⎜
Aco K ⎝ ω co
Gc ( s ) = ⎠
⎛ s ⎞
(3)
s⎜⎜ + 1⎟
2
⎟
⎝ K ω co ⎠
where ωco is the desired crossover frequency, K is the pole
frequency and zero frequency control factor. The value of K can be
adjusted depending on the phase boost (Øboost) required to make
the phase compensation.
φ boost = PM − 90 o − φ co
provided:
(4)
and the K-factor is calculated from equations 2 and 3 as:
⎛φ ⎞
K = tan 2 ⎜ boost + 45 o ⎟
⎝ 4 ⎠
(5)
C3
C2 R2 C1 R1
R3
+ vo
vc
-
Rbias
Vref
= (−1) ×
vo 1
vc s (C 2 + C 3 ) R3
(1 + s( R1 + R3 )C1 )(1 + sR2 C 2 )
⎛ ⎛ C C ⎞⎞
(5a)
ω (1 + s ω z1 )(1 + s ω z 2 )
= (−1) i
s (1 + s ω p1 )(1 + s ω p 2 )
vo
(5b)
vc
ωi =
1
R3 (C 2 + C 3 )
(7)
ω z1 =
1
( R1 + R3 )C1
(8)
ω z2 =
1
(9)
( R2 C 2 )
ω p1 =
1
(10)
R1C1
C 2 + C3
ω p2 = (11)
R2 C 2 C 3
II. By knowing Aco and Øco, and specifying the desired phase
margin, PM, the block of K-factor can be employed to
Vg L C R Vp r
Power Stage
Gp(s) fco
Aco Øco
K-factor
PM
ωi ωp12 ωz12
Vo Compensator
Vref
R3
C1 C2 C3 R1 R2
Figure 2 Flowchart for the proposed approach
⎛ ω ⎞
1 + ⎜⎜ ⎟⎟
2
⎝ ω zesr ⎠
Gco
= A=
vo
⎛ ⎛ ω ⎞2 ⎞ ⎛ ω ⎞2
(12)
⎜1 − ⎜ ⎟ ⎟ + ⎜ ⎟
vc
⎜ ⎜⎝ ω o ⎟⎠ ⎟ ⎜⎝ Qω o ⎟⎠
⎝ ⎠
calculated as
ω
ω Qω o
φ = tan −1 − tan −1
ω zesr ⎛ω ⎞
(13)
1 − ⎜⎜ ⎟⎟
2
⎝ ωo ⎠
To implement (12) and (13) in PSpice, the frequency is assigned as
a variable, and other parameters are assigned as constants. The
.PARAM statement is used for this purpose. By using the
.PARAM statement we can create parameters and assign algebraic
mathematical expressions to it.
and Øco, and speciying the phase margin, PM, the phase boost
required can be calculated as
φ boost = PM − 90 o − φ co (14)
ω z12 = K ω co (16)
The integrator gain of the compensator is
ω co
ωi = (17)
Aco K
(14)-(17) are also implemented in PSpice by using the .PARAM
statement.
block are ωp12, ωz12, and ωi that were obtained earlier from K-
zeros to component values of the error amplifier. The inputs to this
factor block. Besides, the user must provide reference voltage, Vref,
and Rbias.The conversions occur as follows:
C3 =
(ω i R3 f p12 )
f z12
(18)
⎛ f p12 ⎞
C 2 = C 3 ⎜⎜ − 1⎟⎟
⎝ f z12 ⎠
(19)
R2 =
2πf z12 C 2
1
(20)
R1 =
R3
⎛ f p12 ⎞
(21)
⎜⎜ − 1⎟⎟
⎝ f z12 ⎠
C1 =
2πf p12 R1
1
(22)
Rbias =
Vref
Vo − Vref
R3 (23)
Control Loop Design of DC-DC Converter Systems Using PSpice 11
the overall loop. Typical PMs range 45° to 75°. Lower PM, like
Another parameter to choose is the phase margin, PM, desired in
24.42V K
WI R1 427.0V
40.25KV
R3 = 10K
PM = 60 FCO = 16.6667KHZ
From the bode plot, we can see that the crossover phase is –119.6°,
PSpice AC analysis. Bode plot of loop-gain is shown in Fig. 5.
(PM = 60.0° and fco = 16.667kHz) and the results of PSpice are
frequency is 16.623 kHz. The small differences between our target
17.40V K
217.4KV WI R1 609.7V
R3 = 10K
PM = 60 FCO = 33.333kHz
5.2V
5.1V
5.0V
4.9V
4.8V
4.7V
7.9ms 8.0ms 8.1ms 8.2ms 8.3ms 8.4ms 8.5ms
V(OUT
Time
REFERENCES
[1] Lloyd H. Dixon, “Closing the feedback loop,” Unitrode
Power Supply Design Seminar Handbook: SEM 700A, 1990.
[2] W. Tang, F. C. Lee and R. B. Ridley, “Small-Signal
Modeling of Average Current-Mode Control,” IEEE Trans.
Power Electronics, Vol. 8, no. 2, Apr. 1993, pp. 112-119.
[3] B. Holland, “Modeling, Analysis and Compensation of the
Current-Mode Converter,” Proceeding of the Powercon 11,
1984, pp. I-2-1-I-2-6.
[4] H. Dean Venable, “The k-factor: A New mathematical Tool
for Stability, Analysis, and Synthesis,” Proceeding of
Powercon 10, San Diego, CA, March 22-24, 1983.
[5] Lloyd H. Dixon, “Average Current Mode Control of
Switching Power Supplies,” Unitrode Application Note,
1999.
[6] J. Sun, R. M. Bass, “Modeling and Practical design issues for
Average Current Control,” Proc. of APEC, Vol. 2, pp. 980-
986,1999.
[7] S.A. Chickamenahalli et. al., “Effect of target impedance and
control loop design on VRM stability,” Proc. Of APEC,
2002, Vol. 1.
[8] Abraham I. Pressman, Switching Power Supply Design,
16 Modeling and Control of Power Converters and Drives
McGrawHill, 1998.
[9] N. Mohan, T. M. Undeland, W. P. Robins, Power
Electronics: Converters, Applications and Design, John
Wiley, 1995.
[10] C. M. Liaw, T. H. Chen, W.L. Lin, “Dynamic modelling and
control of a step up/down switching-mode rectifier,” IEE
Proc. - Electric Power Applications, Vol. 146 Issue: 3, May
1999, pp. 317–324.
APPENDIX 1
* source ERROR-AMP
.SUBCKT Type-3-all Aco C1 C2 C3 k Pboost Pco R1 R2
+ Rbias wi wp12 wz12 PARAMS:
+ Vg=12 Resr=20m R3=10k Vo=12V Vp=3V Vref=3V L=100uH fco=10kHz
+ PM=50 C=400uF R=5
.PARAM C2={C3*((wp12/wz12)-1)} w={2*pi*fco} R1={R3/(k-1)} C3=
+ {wz12/(wi*R3*wp12)} mag1={SQRT((1+(w*w)/(wzesr*wzesr)))} R2=
+ {1/(wz12*C2)} Pboost={PM-90-Pco} wi={w/(Aco*k)} wz12={w/sqrt(k)}
+ Rbias= {Vref*R3/(Vo-Vref)}
+ k={(Tan(((Pboost/4)+45)*pi/180))*(Tan(((Pboost/4)+45)*pi/180))}
+phase={-atan((w/(Q*wo))/(1-(w*w)/(wo*wo)))*180/pi}
+Mcomp={(1/mag)} Pco={Pcomp+phase1}
+ phase1={atan(w/wzesr)*180/pi} wo={1/SQRT(L*C)}
+ Aco= {Mcomp*mag1*Vg/Vp}
+ Pcomp={IF((1-((w*w)/(wo*wo)))<0,phase-180,phase)}
+ pi=3.14159 wp12={w*sqrt(k)} wzesr={1/(Resr*C)} Q={R/(wo*L)}
+ mag={SQRT(((w/(Q*w))*(w/(Q*w)))+(1-(w*w/(wo*wo)))*(1-
+(w*w/(wo*wo))))} C1={1/(wp12*R1)}
****************************************************************
***
E_ABM5 R1 0 VALUE { R1 }
E_ABM3B1 R2 0 VALUE { R2 }
E_ABM5B1 C1 0 VALUE { C1 }
E_ABM9B1 RBIAS 0 VALUE { Rbias }
E_ABM6 PBOOST 0 VALUE { Pboost }
E_ABM6B1 WI 0 VALUE { wi }
E_ABM7B1 WZ12 0 VALUE { wz12 }
E_ABM1B1 C3 0 VALUE { C3 }
E_ABM1X K 0 VALUE { k }
E_ABM4A1 ACO 0 VALUE { Aco }
Control Loop Design of DC-DC Converter Systems Using PSpice 17
E_ABM2B1 C2 0 VALUE { C2 }
E_ABM8B1 WP12 0 VALUE { wp12 }
E_ABM5A1 PCO 0 VALUE { Pco }
.ENDS Type-3-all
****************************************************************
***
Symbol R=5 RESR = 20M
VG = 12V L = 100UH
C = 400UF VO = 12V
VREF = 3V VP = 3V
PBOOST C1
ACO U6 C2
PCO Type-3-all C3
WI R1
WZ12 R2
WP12 RBIAS
R3 = 10K
PM = 50 FCO = 10KHZ
18 Modeling and Control of Power Converters and Drives
2
SIMULATION OF POWER CONVERTERS
WITH SLIDING MODE CONTROL USING
PSPICE
2.1 INTRODUCTION
Design controllers for power converter systems present
interesting as well as difficult challenges. This is due to the fact
that power converters are non-linear time-varying systems. One
approach is to use a linear control. With this approach, non-linear
power converters have to be transformed into their linear
equivalent, so that classical control theory could be applied to the
design. The main disadvantage of this approach is that the stability
of control design is only guaranteed at a specific operating point.
Another approach is to use nonlinear control. One of these
approaches is sliding mode control. Sliding mode control uses
variable structure nature of power converter operation. This
approach improves the robustness against parameter, line and load
variation. This explains the wide use of sliding mode control in
power electronics converter systems lately.
The theory of sliding mode control to power converters has
widely investigated in literatures. Most of the papers have focused
on the theoretical aspect [1]-[3] and practical implementation [4]-
[6]. Hardware design engineers are discovering that increasing the
use of circuit simulation on their designs minimizes the number of
prototypes required, and eliminate many repetitive verification
Simulation of Power Converters with Sliding Mode Control Using 19
PSpice
2.2.1 Background
In sliding mode control, the trajectory of the system is
constrained to slide along a predetermined sliding surface in the
state space. Such mode, known as sliding mode, is completely
robust and independent of parametric variations and disturbances.
Consider a system with two state variables and several control
inputs as described by the following state equations:
x& (t ) = f ( x, t ) + g ( x, t )u (t ) (1)
S ( x) = c1 x1 + c 2 x 2 = C T x = 0 (3)
umax
u y
umin Eq. (1)
section.
The control law equation described in (3) computes the
instantaneous state-variable trajectory S(x). In practice, this is
realized through an analog or digital computer. In Pspice, this is
easily realized through an ABM part. The realization of (3)
through the ABM parts in PSpice rarely gives numerical instability
due to the continuity of the state variables.
The plant equation described by (1) is in general. For power
electronic circuits, the plant usually consists of a voltage source, a
resistor, an inductor, a capacitor, fully-controlled switches such as
MOSFET and diodes. There are several ways to implement the
plant in PSpice. The easy way is to implement it as it is. However,
this tends to raise convergence problem especially for power
electronics circuits that consist of more than one fully-controlled
switch, such as inverter circuits. In addition, it also requires a big
size of memory space to store the data and takes a long time to run
the simulation. In this paper, due to its simplicity, the buck
converter circuit will be implemented as it is. But, for the inverter
circuit, a system level simulation will be employed. This means
that neither fully-controlled switch nor diode is directly used in the
circuit. Therefore, the convergence problems related to the use of
fully controlled switch and diode are eliminated.
2.3 COMPARATOR
vo
+ +
o
- v+- v-
-V
v+ =
R1
R1 + R 2
vo (6)
R2
R1
+
o
0 vi -
vo
+V
v+ - v-
-V
HB
Since the output voltage vo can assume any two values, so does
24 Modeling and Control of Power Converters and Drives
v+. When the output voltage vo = +V, the positive terminal voltage
is
v+ = V ≡ VUB
R1
R1 + R 2
(7)
v+ = − V ≡ V LB
R1
R1 + R 2
(8)
(3) and (4) define the upper boundary (UB) and lower boundary
(LB) of the hysteresis band (HB). The hysteresis band, HB is given
by
VUB − V LB = 2V ≡ HB
R1
R1 + R 2
(9)
ABM2
+
-
IF(V(%IN1)>V(%IN2), 10, -10)
By using ABM2 part which has two inputs and one output, we
can write the expression for the comparator as shown in Figure 4.
As usual, we use the value of V = 10 Volt.
To implement the hysteresis comparator, the only difference is
that the terminal positive of the comparator is not independent; it
depends on the output voltage as in (10). However, this can be
easily added in PSpice by using ABM1 part which has a single
input and a single output as shown in Figure 5.
ABM1 ABM2
+
0.05*HB*V(%IN) IF(V (%IN1)>V(%IN2), 10, -10)
-
PA RAM ET E RS:
HB = 1
Notice that ABM1 part takes the output of ABM2 as its input,
does the computation and outputs the value to the positive terminal
of ABM2. Parameter HB is used in ABM2 and it is defined by
PARAM part. HB is the only parameter needs to be changed to suit
various applications.
2.3.3 Simulation
The above described behavioral model of the hysteresis
comparator has been simulated using PSpice to demonstrate its
validity. A sine wave signal with frequency of 500 kHz and
amplitude of 10 V is applied to the inverting input of the hysteresis
comparator. The hysteresis band HB is set as a default value i.e. 1.
The simulation parameters (.option) are set as their default values
too.
26 Modeling and Control of Power Converters and Drives
(a)
(b)
default maximum step size (b) Using maximum step size at 0.001 µs.
Figure 6 Simulation results of the hysteresis comparator (a) Using
The simulation results for the input and the output are shown in
Figure 6a. From the results we can see that the rise and the fall of
the output waveform are less sharp. It means that the accuracy of
the simulation result is quite poor when the default values of
simulation parameter (.option) are used. To make the accuracy of
the output waveform to be acceptable, we have to set just one
simulation parameter (.option) i.e. the maximum step size. From
our experience, the maximum step size should be set at least 1/100
IRF150 0.15 L1
out
110u F
RL IC = 0 V
V1 C1
+
24V D4 100uF Rload
-
Dbreak IC = 0 6
+
-
+
-
E1 H1 H
Ic
+
GAIN = 12 GAIN = 1 -
0
c b a
SMC
(a)
28 Modeling and Control of Power Converters and Drives
ABM2a a b
1
3
2
ABM1 ABM2b
0.05*HB*V(%IN)
+ c
IF(V(%IN1)>V(%IN2), 10, -10)
-
Hysteresis Comparator
Figure 7(b) shows the circuit that implements the sliding mode
controller (SMC). ABM2a part implements a standard sliding
mode control law for buck converter which is given by
S (v o , i c ) = (V ref − βv o ) − i c [3]-[4].
βR L
1
(11)
that no overshoot occurs during start-up event for both the inductor
current and the output voltage, verifying the theory.
Rl L
Rload
1 2
+ Resr
E
- - E
+
determine the value of k for this system due to the fact that the load
step changes in this system results in unbounded disturbance. In
[5] k is determined solely by simulation study and experimental
investigation.
0.15 L1
Vc
6m
IC = 0
360*V(%IN+, %IN-) C1
E1 100uF
IN+ OUT+ IC = 0 RL
IN- OUT-
20
GAIN = 10k
10m
EVALUE Ic
+
- H
c b 0 H1 a
SMC
ABM1 ABM2b
+ c
0.05*HB*V(%IN)
IF(V(%IN1)>V(%IN2), 10, -10)
-
PARAM ET ERS:
HB = 50
PARAM ET ERS: ABM2a
k = 1m
b a
k*(V(Vc_dot_ref)-V(%IN1)) 1
3
+(V(Vc_ref)-V(%IN2)) 2
Vc_dot_ref Vc_ref
VAMPL = 0.891 V1 V2
FREQ = 50 VAMPL = 280
PHASE = +90 FREQ = 50
REFERENCES
3
DYNAMIC EVOLUTION CONTROL FOR
STEP DOWN DC-DC CONVERTER
3.1 INTRODUCTION
In recent years, the application of power electronics converter
has grown extremely. Some applications that are increasingly
being dominated by power electronics are: 1) switched-mode
power supplies; 2) adjustable speed motor drives; 3) efficient
control of heating and lighting; 4) efficient interface for
photovoltaic; 5) fuel cell and high voltage dc system for efficient
transmission of power, etc.
Much power electronic application operated with parameter
variation, non linearity, load disturbance, etc. Therefore, there is an
increasing need for a good controller design to perform tight
regulation under high unpredictable load variation. In designing of
classical control theory, e.g. PID controller, small signal linear
approximations have been applied to the nonlinear system. This
approach enables the designer to use a simple linear controller to
keep the system stable. But the main disadvantage using this type
of control is that it is applicable for operation only near a specified
operating point.
Since power converters are non-linear time-varying systems,
the design of controllers must have capability to cover up the
nonlinearity and time-varying properties of the system.
38 Modeling and Control of Power Converters and Drives
Y = Ce − mt (1)
Y = Yo.e− mt (2)
= − m.Yo.e − mt
dY
dt
40 Modeling and Control of Power Converters and Drives
= − m.Y
dY
dt
+ m.Y = 0; m>0
dY
(3)
dt
Vo = Vg .α − L
diL
(5)
dt
.
The derivative of Y is given by:
=k
dY dVerr
(7)
dt dt
+ m.k .Verr = 0
dVerr
k
dt
The expression for duty cycle α is the control action for the
converter controller.
Control law (9) forces the state error function (Y) to satisfy
equation (6). According to this equation, the state error function
(Y) forced to make evolution follow equation (2) and decrease to
zero (Y = 0) with a decrease rate m. so, the state error function (Y)
satisfy the equation
Y = k .Verr = 0
Consequently state error of the converter will converges to
zero.
Verr = 0 (11)
Vo = Vref (12)
TABLE I
SIMULATION MODEL PARAMETERS
Parameter Value
converter output voltage when m=6000 is better than the other, and
trajectory of dynamic evolution control is depicted in figure 6.
12
10
Output Voltage
8
Inductor Current
6
4
Ampere
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
Time (sec)
4
x 10 Trajectory of Dynamic Evolution Control
4
2
Start Point
1
change of error
-1
-2
-3
-4
-10 -5 0 5 10
error
REFERENCES
[1] Raviraj, V.S.C.; Sen, P.C.; Comparative study of proportional
integral, sliding mode, and fuzzy logic controllers for power
converters, Industry Applications, IEEE Transactions on
Volume 33, Issue 2, March-April 1997 Page(s):518 – 524
4
DIGITAL PI CONTROLLER DESIGN
FOR POWER INVERTER USING
MATLAB-SIMULINK
Shahrin Md. Ayob
Naziha Ahmad Azli
Zainal Salam
4.1 INTRODUCTION
In dc-ac conversion equipments, it is utmost important for
inverter to be able to maintain a stable and a clean ac voltage
waveform regardless of type of loads connected to it. In addition,
the inverter should recover from disturbance with excellent transient
response and low distortion. Those requirements are oftenly
difficult to achieve due to more harmonics are injected into the
power line. These harmonics are commonly originated from the
modern loads, which employed high-efficiency power converters
such as computers and power supplies. To achieve such
requirements, feedback control technique is employed.
Proportional-Integral (PI) controller is one of prominent feedback
control technique used in industries [1, 2, and 3]. The controller
yields a very simple control structure and its development cost is
very low. It can be implemented either using analogue circuitry or
digital processors. In addition, the performance and the stability of
the controller are predictable via control metrics such as bandwidth,
phase margin and gain margin.
Despite of the advantages, PI controller is known to have several
limitations. It is a model-based controller where, its control
parameters are designed based on the mathematical model of the
Digital PI Controller Design for Inverter Regulation Using 51
Matlab-Simulink
L iL
s1 s3
iC iO
vdc va C vo R
s4 s2
k PWM =
vdc
(2)
vˆtri
In equation (2), vdc is the dc voltage value of the inverter while vˆtri is
the peak value of the triangular carrier. It follows that the dynamic
of the system is now mainly determined by the LC filter. The
simplified equivalent circuit for the output filter for an inverter is
shown in Figure 2.
iL iO
+
L
iC
va
−
C vo R
vcont . − iL
+ + −
va 1 1 iC vo
k PWM
Ls Cs
iO
1
R
Figure 3. Linear model of single-phase inverter system
⎡ ( K p / Ki )s + 1 ⎤
C (s) = Ki ⎢ ⎥
⎣ ⎦
(4)
s
⎡ z − 1⎤
s= ⎢⎣ z + 1 ⎥⎦
2
(5)
Ts
54 Modeling and Control of Power Converters and Drives
mz + n
C ( z) =
z −1
(6)
Where the parameters m and n are given by:
⎡ K p Ts ⎤
m = Ki ⎢ + ⎥
⎣ Ki 2 ⎦
(7)
⎡T K p ⎤
n = Ki ⎢ s − ⎥
⎣ 2 Ki ⎦
(8)
Where e(k) is error signal sampled at kth interval, while e&(k ) is the
change of error, e(k ) , which can be computed as:
u& (k ) u (k )
−n
e&(k )
u (k − 1)
e(k − 1)
Digital PI Controller Design for Inverter Regulation Using 55
Matlab-Simulink
u (k ) = u& (k ) + u (k − 1) (11)
L
s1 s3
vdc va iL C
R vo
s4 s2
s1 s2 s3 s4 vref
iLref
PWM PI I PIV
Parameter Value
Bandwidth and phase margin are analysed using bode plot. Bode
plot is used since it offers a convenient way to design linear
controllers. For an efficient design, sisotool from Matlab is
employed. The tool provides an easy way of designing linear
controller via its Graphic User Interface (GUI). Figure 6 shows the
snapshot of sisotool dialogue box. Using this tool, one can
simultaneously analyse the effects of changing the poles and zeros
of the compensator on the system performance and stability. In
Digital PI Controller Design for Inverter Regulation Using 57
Matlab-Simulink
Transfer function of
the compensator
K PWM =
vdc
(12)
vˆtri
Where vdc is the input dc voltage and vˆtri is the peak value of
triangular carrier. In this work, vdc is 100V while vˆtri is 1 units.
Therefore, KPWM can be computed to have,
K PWM = = = 100
vdc 100
(13)
vˆtri 1
vo
iLref + vcont . va − iL + iC 1 vo
+
PI I K PWM 1
− −
Ls Cs
iL
iO 1
R
iLref + vcont . va Cs + R iL
LCs 2 + RLs + L
PI I K PWM
−
iL
The open-loop transfer function for the inner current loop, GolI can
be written as:
⎡ K s + Ki ⎤
⎥ [ K PWM ] ⎢
⎡ Cs + R ⎤
Gol1 ( s ) = ⎢ p
⎣1424 3⎦ ⎣ LCs + RLs + L ⎥⎦
2
(14)
s
PI1
-20
-30
-40
0
-45
-90
1 2 3 4 5
10 10 10 10 10
Frequency (Hz)
60 Modeling and Control of Power Converters and Drives
0.114s + 8628
C ( s) I = (15)
s
0.222z − 0.0063
C ( z )1 =
z −1
(16)
Current loop
vref + iLref iL + iC 1 vo
−
PIV
−
1
Cs
vo
io 1
R
Current loop
vref + iLref iL 1 vo
Cs + R
PIV
−
1
vo
Figure 11. Outer voltage loop after simplification using Mason rules
The open-loop transfer function for the outer voltage loop, Gol2 is
written as:
⎡ K s + Ki ⎤
⎥ [{1] ⎢
⎡ 1 ⎤
GolV = ⎢ p
⎣1424 3⎦ loop ⎣ Cs + R ⎥⎦
(17)
s
current
PIV
Bode Diagram
5
0
Magnitude (dB)
-5
-10
Closed-loop bandwith =100 Hz
-15
-20
-25
0
-45
-90
1 2 3
10 10 10
Frequency (Hz)
From sisotool, the transfer function of the compensator for the outer
voltage loop, C(s)V is obtained as:
0.415s + 28000
C ( s )V = (18)
s
0.765z − 0.065
C ( z )V =
z −1
(19)
Pulse
Generator
c
i 2
Universal Bridge +
- 1 +
v
-
g L Breaker
+ Vo
A
R
C
-
B Zero-Order
Vdc Hold1
Discrete Vref
PWM Generator
iLref
Pulses Signal(s) Controller Controller
Zero-Order
Hold
1 m+n 1
Ref Control signal
1
-n z
Unit Delay
1
z
Unit Delay1
feedback
[A2]
Zero-Order
Hold1
Figure 15 shows the simulation result for the output voltage vo.
A step-load disturbance is subjected to the system at 5ms. As can
be seen, the transient revealed no oscillation and a zero steady-state
error. Thus, the result verified the design that has been conducted in
this work.
Digital PI Controller Design for Inverter Regulation Using 65
Matlab-Simulink
100
80
60
40
Output voltage (V)
20
-20
-40
-60
-80
-100
0 2 4 6 8 10 12 14 16 18 20
time(ms)
REFERENCES
501.
[5] Duan, Y. and Jin, H. 1999. Digital Controller Design for
Switchmode Power Converters. The 14th Annual Applied
Power Electronics Conference and Exposition, United States of
America, pp. 967 – 973.
[6] Mohan, N. Electric Drive: An Integrative Approach. 2001.
John Wiley Publication.
Overview of Direct Torque Control of Induction Machines 67
5
OVERVIEW OF DIRECT TORQUE
CONTROL OF INDUCTION MACHINES
Nik Rumzi Nik Idris
•
assumed to contain the 3-phase windings of negligible depth.
Permeability of the stator and rotor iron is assumed to be
infinity and the saturation and iron loss are assumed to be
•
negligible.
The symmetrical stator windings are Y-connected with the
neutral electrically isolated. The rotor is of a squirrel cage
•
type.
The space harmonics of the air-gap mmf and flux density are
negligible. The real axis of the stator winding coincides with
•
the phase A winding axis
The rotation is said to be positive in the counterclockwise
direction.
68 Modeling and Control of Power Converters and Drives
x=
2
3
(xa + axb + a 2 xc ) , (1)
Overview of Direct Torque Control of Induction Machines 69
where a = ej2π/3
The coefficient 2/3 in (1) corresponds to the non-power invariant
or amplitude invariant definition of the space phasor [3]. In terms
of its d and q axes, the space phasor can be conveniently expressed
as follows:
x = xd + jxq (2)
⎡2
( )
xd = Re[x ] = Re ⎢ xa + axb + a 2 xc ⎥ = ⎜ xa − xb − xc ⎟
⎤ 2⎛ 1 1 ⎞
⎣ ⎦ ⎝ ⎠
(3)
3 3 2 2
⎡2
(
xq = Im[x ] = Im ⎢ xa + axb + a 2 xc ⎥ =
⎤
)
1
( xb − x c )
⎣3 ⎦
(4)
3
Te = ψ s × is
3p g g
(9)
dωm 2 dωr
22
J =J = Te − Tload (10)
dt p dt
reference frame (ωg = 0), and re-arranging them into matrix form,
equations (5) – (8) to their equivalent d-q axis forms in stationary
⎡v sd ⎤ ⎡ Rs + sLs 0 ⎤ ⎡isd ⎤
⎢ ⎥ ⎢ ⎢ ⎥
sLm ⎥⎥ ⎢isq ⎥
0 sLm
⎢ vsq ⎥ = ⎢ 0 Rs + sLs
⋅
⎢v ⎥ ⎢ sLm ωr Lr ⎥ ⎢ird ⎥
0
ωr Lm Rr + sLr
,
⎢ rd ⎥ ⎢ ⎥ ⎢ ⎥
⎢⎣ vrq ⎥⎦ ⎣ − ωr Lm sLm − ωr Lr Rr + sLr ⎦ ⎢⎣irq ⎥⎦
(11)
⎡i&sd ⎤ ⎡ R s Lr − ω r L2m i sq − Rr Lm − ω r Lm Lr ⎤ ⎡i sd ⎤ ⎡− Lr 0⎤
⎢& ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ 0
⎢i sq ⎥ = ⎢ ω r Lm ω r Lm L r − Rr Lm ⎥ ⎢i sq ⎥ ⎢ Lr ⎥⎥ ⎡v sd ⎤
⋅ + ⋅⎢ ⎥
2
⎢i&rd ⎥ L − L L ⎢ − R L ω r Lr Ls ⎥ ⎢i rd ⎥ L2m − Lr Ls ⎢ Lm 0 ⎥ ⎣ v sq ⎦
1 R s Lr 1
ω r Lm L s
⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
2
Rr Ls
⎣⎢i rq ⎦⎥ ⎣⎢− ω r Lm Ls − R s Lm − ω r Lr L s Rr Ls ⎦⎥ ⎣⎢i rq ⎦⎥ ⎣ 0 Lm ⎦
m r s s m
&
(12)
Te =
3 p
ψ s × is =
3p
(Ls is + Lm ir )× is (13)
22 22
This section reviews the principles of DTC. The 3-phase VSI and
its voltage space vectors are briefly reviewed. It will be shown that
by utilising the eight possible switches configurations in the 3-
phase VSI, the direct flux and torque control of induction machine
can be established. The independent control of stator flux and
torque is obtained by selecting voltage vectors that will satisfy both
the flux and torque demand. Various methods used to estimate the
stator flux are also discussed.
( )
phase voltage space vectors of the 3-phase VSI are given by (15).
(15)
Sa Sb Sc k
0 0 0 0
1 0 1 1
1 0 0 2
1 1 0 3
0 1 0 4
0 1 1 5
0 0 1 6
1 1 1 7
( k − 2 )π
zero space phasor is the same (i.e. 2/3Vd) with a phase of
(where k = 1, 2,…6).
3
Overview of Direct Torque Control of Induction Machines 73
∆ψ s = vs ∆t (16)
Te = − ψ sg × ψ rg = ψ sg ψ rg sin θ sr
3 Lm 3 Lm
2 σL s Lr 2 σL s Lr
(17)
The relation between the rotor and the stator flux linkages in the
rotor flux reference frame, is given by (18).
ψ rr = ψ sr
Lm
1 + pστ r
Ls
(18)
decreases θsr
selected voltage vectors. Choosing voltage vectors that increases or
causes the torque to increase or decrease
respectively.
Assuming that rotor is rotating counterclockwise with the stator
decrease.
Case 4. Voltage vectors, vs,k-1 and vs,k-2, which are
referred as reverse active voltage vectors are chosen.
rapidly.
∫
frame.
ψ s = ( v s − Rs i ) dt (19)
This estimation technique, however, suffers from initial value and
drift problems [12]-[14]. Even a small dc offset present in the
measured current or voltage can cause the integrator to saturate. At
low to zero speeds, with low stator back emf, the stator resistance
drop becomes significant and a small deviations in its value can
rather than stator flux − for the Direct Self Control (DSC) scheme
of induction motor drive. In this way, the problem of stator
∫
⎛ L i − Ψr ⎞
ψ r = ⎜⎜ m s − jωr Ψr ⎟⎟ dt
⎝ τr ⎠
(20)
ψs = ψ r + σLs is
Lm
(21)
Lr
In discrete implementation, the integrator in (20) uses the previous
value of rotor flux, i.e. a closed-loop integrator, as opposed to the
open-loop integrator used in (19). For this reason, the stability and
−which is valid for steady state conditions. It was found that the
parameter variations utilising the frequency function response
bandwidth, the current model and the voltage model dominates the
estimation, respectively. The closed-loop bandwidth can be set
using the controller parameters. In [15], the voltage and current
models are used alternately and the stator and rotor resistance are
periodically updated to improve the estimation accuracy. Other
variations of closed-loop observer includes Gopinath’s reduced
order observer [23], Luenberger observer [13] and Kalman
filter[24].
Although the performance of the closed-loop flux observer is
attractive, it has a drawback of requiring the knowledge of speed or
position of the rotor. This is regarded as unsuitable in some
applications where the installation of the speed or position sensors
are not favourable; speed sensors results in higher cost, increased
size of the drive machine, the drive is less robust and reduced
reliability. To resolve this problem, some researchers have used an
estimated speed to calculate or estimate the flux. In [25], the speed
is estimated using the Model Reference Adaptive System (MRAS)
incorporating the mechanical system model. In a more recent
sensorless DTC drive [26], the rotor position used in the current
model is obtained from the estimated stator flux which is used as
the implicit reference for the closed loop flux observer. The speed
is estimated using the MRAC. However as pointed out in [22], the
zero or low speed operation problems remain unsolved, because by
doing so, one actually attempts to solve three unknowns using two
equations. Without heavy filtering, the positive feedback inherent
in this type of control can result in instability. The three unknowns
are the two flux estimates and the rotor speed while the two
equations are the voltage and current models. It is shown in [22]
that comparable speed-sensorless flux estimation can be achieved
by simply modifying the pure integrator used in the voltage model
estimator. The method replaces the pure integrator with a low-pass
filter with its cutt-off frequency implemented entirely using
software. The suitable cut-off frequency therefore can be
dynamically change depending on the operating conditions. The
modified voltage model-based estimator was also applied to a
SVM based DTC drive of an induction machine [27] and has been
86 Modeling and Control of Power Converters and Drives
Where Iphase and I1,phase are the r.m.s values of the phase current and
the fundamental harmonic component of the phase current,
respectively.
For a fixed torque hysteresis band, the THD increases with the flux
hysteresis band. As the flux band is further increased, the flux
locus approaches that of hexagonal shape, similar to that of the six-
step inverter-fed drive [29]. A simulation of the DTC drive was
carried out (machine parameters as given in Chapter V) with the
torque hysteresis band varied from 5% to 25% of the torque limit
allowed during acceleration/deceleration; the torque limit in this
case is set to 1 N-m. The stator flux band is also varied from 2.5%
to 15% of the command flux which is set to the rated value of 0.45
Wb. The THD of the phase current is calculated and the results are
Overview of Direct Torque Control of Induction Machines 87
(a)
Overview of Direct Torque Control of Induction Machines 89
(b)
torque limit is set to 1 N-m. The stator flux band also varied from
2.5% to 15% of the command flux which is set to the rated value
of 0.495 Wb. The motor speed is set to run at 25 rad/s, 100 rad/s,
and 301 rad/s. The switching frequency is calculated and a three
dimensional graph is plotted (Figure 12) which shows that the
switching frequency is mainly affected by the width of the torque
hysteresis band.
Overview of Direct Torque Control of Induction Machines 91
(a)
(b)
92 Modeling and Control of Power Converters and Drives
(c)
Figure 12 Switching frequency variation with torque and flux
hysteresis bands : (a) at 25 rad/s, (b) at 100 rad/s and (c) at 301
rad/s
dψ s
ψ s ,n+1 = ψ s ,n + ∆t (22)
dt n
Overview of Direct Torque Control of Induction Machines 93
dψ r
ψ r ,n+1 = ψ r ,n + ∆t (23)
dt n
dψ s ,n
= − s ψ s ,n + s m ψ r ,n + v s ,n
R RL
σLs σLs Lr
(24)
dt
dψ r ,n Rr Lm ⎛ ⎞
= ψ s ,n + ⎜⎜ jωr − r ⎟⎟ψ r ,n
R
σLs Lr ⎝ σLr ⎠
(25)
dt
[ ]
written as:
[(v ]
)• jψgr,n ∆t
⎛ 1 1 ⎞
Te,n +1 = Te,n − Te,n ⎜ + ⎟∆t + − jω r ψs,n
3 p Lm
⎝ στ s στ r ⎠ 2 2 σL sL r
g
s,n
[
(vs,n − jωr ψs,ng )• jψgr,n ]
Te,n +1 − Te,n ⎛ 1 1 ⎞ 3 p Lm
= −Te,n ⎜ + ⎟+
∆t ⎝ στs στ r ⎠ 2 2 σL sL r
(27)
[(vs − jωr ψ s )⋅ jψ r ]
Te ,n+1 − Te ,n
slope + ≡ = − e ,n +
T 3 p Lm (28)
∆t στ sr 4 σLs Lr
[(− jωr ψ s )⋅ jψ r ]
Te ,n+1 − Te ,n
slope − ≡ = − e ,n +
T 3 p Lm
∆t στ sr 4 σLs Lr
(29)
based on torque and stator flux errors− required to drive the torque
and stator flux to their reference values within one sampling instant
was presented. The demand stator voltage was implemented using
SVM technique. Consequently, the switching frequency of the
device is totally controlled by the sampling frequency. The method
however increases the complexity of methematics involved, thus
require faster and powerful prosessor. Other works which are
based on predictive control scheme were presented in [32][28].
The duty ratio of the non-zero and zero voltage vectors was
calculated for every sampling instant and hence results in a
constant switching frequency. The non-zero voltage vectors are
selected based on the torque and stator flux errors, as determine in
the conventional DTC scheme. The approach which include the
use of fuzzy controller to reduce the torque ripple are given in
[34][33], which at the same time resulted in a constant switching
frequency.
REFERENCES
6
STUDY ON STABILITY AND
PERFORMANCES OF DTC DUE TO
STATOR RESISTANCE VARIATION
6.1 INTRODUCTION
dψ s
v s = Rs i s + (1)
dψ r
dt
0 = Rr i r + − jω r ψ r (2)
dt
ψ s = Ls i s + Lm i r (3)
ψ r = Lr i r + Lm i s (4)
( )
reference frame that
(
ψˆ sq = ∫ vsq − Rˆ s isq dt ) (5b)
where vsd ( isd ) and vsq ( isq ) are the respective d and q-axis of
stator voltage (current) components. R̂s is the estimated stator
resistance. Thus the magnitude of stator flux is calculated by
⎛ ψˆ sq ⎞
φ = arctan⎜⎜ ⎟⎟
⎝ ψˆ sd ⎠
(7)
Tˆe = ψˆ s i s sin (α − φ )
3 p
(8)
22
their respective estimated values. The errors are fed to a two level
comparator, in the case of flux, and a three level comparator in the
case of torque. The flux comparator output is defined as
∆ψ s
dψ s = 1 for ψˆ s ≤ ψ s −
*
(9a)
∆ψ s
2
dψ s = 0 for ψˆ s ≥ ψ s +
*
(9b)
2
* dTe
Te
dψ s
ψs
*
φ
ψˆ e
Tˆe
The output of the comparators and the stator flux angle are used
to index a look up table of optimum voltage vectors as proposed in
[Takahashi and T. Noguchi 1986], in order to determine the
suitable voltage vectors, which is tabulated in Table 1(a). The
sector of the stator flux as illustrated in Figure 2 is divided into six
sectors. Figure 2, indicates that, the appropriate voltage vector (is
taken from the table of optimum voltage vectors) should be chosen
in a particular sector, either to increase stator flux or to decrease
stator flux and either to increase torque or to reduce torque. The
selection of voltage vector is made so as to restrict the errors of the
stator flux and torque within their respective hysterisis bands.
Consequently the fastest torque response and highest efficiency at
every instant can be obtained [Takahashi and T. Noguchi 1986].
Counter Sector
dψ s
clockwise
dTs I II III IV V VI
106 Modeling and Control of Power Converters and Drives
IV III
VI I
value, which is used in the controller, must match to its real value.
The mismatch between the estimated stator resistance value and its
actual value somehow can deteriorate the torque and flux control
especially when the estimated stator resistance is higher than its
actual value as reported in [B. S. Lee and R. Krishnan 1998].
However, the explanations on how the unstable of the performance
occurs were not discussed in details. This section will perform a
study on the effect of the higher estimated stator resistance than its
actual value on DTC. On the other hand, the discussion of the root
cause of the instability, will also be presented. All the analysis will
be performed through simulations using induction machine
parameters as given in Table 1(b).
Rs
is
Te
ψs
T̂e
ψˆ s
the estimated stator flux, just after a step actual stator resistance
change is applied at t=2 s.
The divergence between the actual and estimated stator flux may
be reasoned as follows. Firstly, let us study the change of the
actual and estimated stator flux vectors in the particular sector as
illustrated in Figure 5. Using (1), the change of the estimated stator
flux vector, ∆Ψ ˆ s and actual stator flux vector, ∆Ψ can be
s
( )
expressed as given in (12a) and (12b), respectively.
∆Ψ
ˆ = V s , k − I s Rˆ .∆T
= (V R ).∆T
(12a)
∆Ψs −Is
s s
s,k s (12b)
ψs
t ≥ 2s
ψˆ s
t ≤ 2s t ≥ 2s
ωe
ωe
ψˆ s =ψ s
ψs
ψˆ s
(a) (b)
Figure 8 Effect to the stator current, actual and estimated stator fluxes
corresponds to the incorrect voltage vector selection as shown in
Figure 7.
Study on Stability and Performances of DTC Due to Stator Resistance 115
Rs
is
Te
ψs
T̂e
ψˆ s
REFERENCES
1999.
D. A. Paice, “Motor thermal protection by continuous monitoring
of winding resistance,” IEEE Trans. Ind. Electron. Instr.,
vol. IECI-27, pp. 137-141, Aug. 1980.
L. A. S.Ribeiro, C. B.Jacobina, and A. M. N.Lima, “Linear
parameter estimation for induction machines considering
the operating conditions,” IEEE Trans. Power Electron.,
vol. 14, pp. 62-73, Jan. 1999.
S. Mir, M. E. Elbuluk, and D. S. Zinger, “PI and fuzzy estimators
for tuning the stator resistance in direct torque control of
induction machines,” IEEE Trans. On Power Electronics,
Vol. 13, No.2, pp.279-287, March 1998.
Bose, B. K. and Patel, N.R. (1998). "Quasi-fuzzy estimation of
stator resistance of induction motor," IEEE Trans. Power
Electronics, Vol. 13, No. 3, pp. 401-409.
T. G. Habetler, F. Profumo, G. Griva, M. Pastorelli, and A. Bettini,
“Stator resistance tuning in a stator flux field oriented drive
using an instantaneous hybrid flux estimator,” Conf.
Record, EPE Conf., Brighton, UK, V01.4, pp.292-299,
1993.
Takahashi and T. Noguchi, “A new quick-response and high
efficiency control strategy of an induction machine”, IEEE
Trans. Ind. Applicat. 22, 820–827 (1986).
118 Modeling and Control of Power Converters and Drives
7
A QUICK DYNAMIC TORQUE
CONTROL FOR DIRECT TORQUE
CONTROL HYSTERESIS-BASED
INDUCTION MACHINES
7.1 INTRODUCTION
dψ s
v s = rs i s + (1)
dψ r
dt
0 = rr i r − jω r ψ r + (2)
dt
ψ s = L s i s + Lm i r (3)
ψ r = Lr i r + Lm i s (4)
Te = ψ s i s sin δ
3 p
(5)
22
σT
Torque Hysteresis
comparator
+ q(t)
Te* 0
1
Voltage
Sa
-1
- Sb Voltage
σψ
Vector
ψ
Flux Hysteresis Source
+
Selection IM
ψ s*
Comparator
Table Sc Inverter
+
θs
1
0
-
ψ + Vdc -
Stator Flux and
Te Torque Estimator
(a) (b)
1.0 1.0
Torque (p.u)
Torque (p.u)
0.5 0.5
0 0
t1 t2
Time (s) Time (s)
(c) (d)
1.0 1.0
Torque (p.u)
0 0
t3 t4
Time (s) Time (s)
1.0
(a)
Torque (p.u)
(b)
0.5
(c)
t3
Time (s)
σT
Torque Hysteresis
comparator
+ q(t)
Te* 0
1
Voltage Sa
ψ
-1
- Voltage
σψ
Vector Sb
−
Flux Hysteresis Source
Selection IM
ψ
Comparator
Inverter
+
Modified Table Sc
ψ+
flux error
θs
* 1
s 0 status
-
ψ + Vdc -
Stator Flux and
Te Torque Estimator
ct o
r4
Se
6
tor
Su
Sec
b s ec
to
r
∆ta
∆tb
1.0
(b)
(a)
Torque (p.u)
0.5
2/3Vdc ∆ta
Phase voltage (V)
0
(a)
-2/3Vdc
2/3Vdc ∆tb
Phase voltage (V)
(b) 0
-2/3Vdc
1.00
Stator flux (p.u)
0.75
0.3 0.305 0.31 0.315 0.32 0.325 0.33 0.335
Time (s)
1.0
T e*
Torque (p.u)
0.5
Te
0
1.00
0.75
0.3 0.305 0.31 0.315 0.32 0.325 0.33 0.335
Time (s)
1.0
T e*
Torque (p.u)
0.5
Te
0
1.0
T e*
Torque (p.u)
0.5
Te
0
0.26 0.28 0.3 0.32 0.34 0.36 0.38 0.4 0.42 0.44
Time (s)
Phase voltage (V)
2/3Vdc
-2/3Vdc
0.26 0.28 0.3 0.32 0.34 0.36 0.38 0.4 0.42 0.44
Time (s)
REFERENCES
8
IMPROVED TORQUE CAPABILITY
THROUGH OVERMODULATION FOR
DIRECT TORQUE CONTROL
HYSTERESIS-BASED INDUCTION
MACHINES
Auzani Bin Jidin
Nik Rumzi Bin Nik Idris
Abdul Halim Bin Mohd Yatim
8.1 INTRODUCTION
dψ s
v s = rs i s + (1)
dt
Improved Torque Capability through Overmodulation for Direct 135
Torque Control Hysteresis-based Induction Machines
dψ r
0 = rr i r − jω r ψ r + (2)
dt
ψ s = Ls i s + Lm i r (3)
ψ r = Lr i r + Lm i s (4)
Te = ψ s i s sin δ
3 p
(5)
2 2
Te = P ' m ψ r ψ s sin γ
3 L '
(6)
2 Ls Lr
where γ is the angle between the stator and rotor flux space
vectors. Since the rotor flux vector rotation is continuous in
counter clockwise (lagging the stator flux vector), the
instantaneous output torque can be quickly controlled by changing
the position of stator flux vector. Thus, the angle γ in (6) plays a
vital role in controlling the output torque. For instance, an active
voltage vector (either u2 or u3) is selected to increase the angle γ as
well as output torque to reach its demand. When the torque error
touches the upper band, zero voltage vector (either u7 or u0) is
selected, which ideally stops the stator flux. This in turn reduces
the output torque as the angle γ decreases. In such ways, the output
torque and stator flux magnitudes are restricted within their
respective hysteresis bands.
Improved Torque Capability through Overmodulation for Direct 137
Torque Control Hysteresis-based Induction Machines
σT
Torque Hysteresis
comparator
* + q(t) Sa
Te 0
1
Voltage
-1
- Voltage
σψ
Vector Sb
ψ +
Flux Hysteresis Source
Selection IM
ψ
Comparator
Table Sc Inverter
* + 1
θs
s 0
-
ψs
+ Vdc -
Stator Flux and
Te Torque Estimator
ψs
Sector 3
γ ψr
u0,u7 (FH, TS)
Sector 2
than the stator voltage limit, v s , lim . In this situation, the DTC
algorithm produces sufficient and appropriate voltage vectors v s1
to control both stator flux and torque, simultaneously. At base
with dashed line). On the other hand, the flux producing current
component iψsds is still under controlled in regulating the stator flux
at its rated value.
To enhance the output torque capability, a flux weakening
method is commonly applied. For example, [S. H. Kim 1995 and
G. Griva et al. 1998] consider voltage and current limits condition
to obtain maximum torque capability in the field weakening
region. However, this strategy requires a precise current
regulation, reference frame transformation and knowledge of
machine parameters to derive the optimal controllable currents iψsds
and iψsqs . Recently, a robust field weakening strategy utilizing
hysteresis-based direct torque control was proposed in [D. Casadei
et al. 2007]. The main idea of this scheme is to adjust the flux
reference on the basis of torque error. In this way, the appropriate
flux level is spontaneously determined and maximum torque
capability is obtained as well. This proposed method is less
dependent on machine parameters, allowing a satisfying operation
in the whole speed range.
jω e 2ψ s
jωe1ψ s
i s Rs i s Rs
vs2
ψs v s1
isd is
v s ,lim
ψs
ψs isq
jω e3ψ s i ' R
jω e 2ψ s s s
i s Rs
v s2
ψs '
isd i s is
v s ,lim
ψs
ψs isq
σT
Torque Hysteresis
comparator
+ q(t)
Te* 0
1
Voltage Sa
ψ−
-1
- Voltage
σψ
Vector Sb
Flux Hysteresis Source
Selection IM
ψ s*
Comparator
Inverter
+
Modified Table Sc
ψ+
flux error
θs
1
0 status
-
ψ + Vdc -
Stator Flux and
Te Torque Estimator
No Yes
Ψ- = Ψ+ Overmodulation? Ψ- ≠ Ψ+
q(t) always 1 for
∑(∆θs)=π/12 rad.
Get stator flux
Get flux error angle, θs
status, Ψ+
Calculate θs within a
Stator flux increases/ sector,
decreases 0 ≤ θs ≤ π/3
Ψ- = Ψ+
1.5
ωreference 1.5
ωreference
Speed (p.u)
Speed (p.u)
1
ωactual ωactual
1
0.5 0.5
0.2 0.25 0.3 0.35 0.4 0.45 0.2 0.25 0.3 0.35 0.4 0.45
Time (s) Time (s)
1.0
Treference 1.0
Treference
Torque (p.u)
Torque (p.u)
0
Tactual 0
Tactual
0.2 0.25 0.3 0.35 0.4 0.45 0.2 0.25 0.3 0.35 0.4 0.45
Time (s) Time (s)
2.0 2.0
Stator currents (p.u)
0 0
-1.0 -1.0
-2.0 -2.0
0.2 0.25 0.3 0.35 0.4 0.45 0.2 0.25 0.3 0.35 0.4 0.45
Time (s) Time (s)
2/3Vdc 2/3Vdc
Phase voltage (V)
0 0
-2/3Vdc -2/3Vdc
0.2 0.25 0.3 0.35 0.4 0.45 0.2 0.25 0.3 0.35 0.4 0.45
Time (s) Time (s)
(a) (b)
Figure 7 Waveforms of rotor speed, output torque, d-q stator current
components, phase stator voltage during speed acceleration, for (a)
DTC without overmodulation (b) DTC with the proposed
overmodulation
1.5 1.5
Speed at Speed at
1 0.75 p.u 1 0.75 p.u
q-Stator flux axis (p.u)
q-Stator flux axis (p.u)
0.5 0.5
Speed at Speed at
0
1.5 p.u 0
1.5 p.u
-0.5 -0.5
-1 -1
ω ω
-1.5 -1.5
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5
d-Stator flux axis (p.u) d-Stator flux axis (p.u)
(a) (b)
Figure 8 Stator flux locus during acceleration speed from 0.75 to 1.5
p.u. for (a) DTC without overmodulation (b) DTC with the proposed
overmodulation
Improved Torque Capability through Overmodulation for Direct 147
Torque Control Hysteresis-based Induction Machines
REFERENCES
9
THE MODELING AND SIMULATION
OF IMPROVED DTC OF PMSM DRIVES
USING MATLAB/SIMULINK
Tole Sutikno
Nik Rumzi Nik Idris
9.1 INTRODUCTION
The Permanent Magnet Synchronous Motor (PMSM) offers
many advantages over the induction machine, DC motor and
synchronous motor. The PMSM has lower inertia, higher
efficiency and power density, smaller losses and more compact
motor size. Because of the advantages, PMSM are indeed excellent
for use in many applications.
Usually, the AC machine control technique can be classified as
scalar control and vector control. The AC machine vector
controlled drives are widely used in many drive applications since
they offer high dynamic performance. The most popular vector
controlled are known as field oriented control (FOC) and direct
torque control (DTC) [1], [2], [3]. The DTC method has more
advantages such as lower parameter dependency and lower
complexity of control structure compared to the FOC.
The DTC method was proposed by Takahashi [4] and
Depenbrock [5] for induction motor drives in the middle of 1980’s;
it also was developed in 1990’s for PMSM drive [6]. The DTC is
not use current controller and not depend motor parameters. The
DTC method provides a very quick and precise torque response
without any complex field orientation system and current
controller.
The modeling and Simulation of Improved DTC of PMSM Drives 149
Using Matlab/Simulink
Figure 2 Voltage vectors of the 3-phase VSI and the six stator
flux position sectors
speed_ref
Speed Set
Novel DTC of IPMSM Drive
u*
Continuous
y [Torque _ref ]
Torque _ref s
-
+
B
g
flux _r
PMSM
Flux is_abc
theta uq
m
Hysteresis
Ref Flux
m wm
Flux _ref
Switching Table
angle (0-2*phi) Te
id
Est Flux
d-axis Machines Scope 1
Flux _est stator current Measurement
Theta id id Demux Torque
Iabc To Workspace4
Flux_est
Iq iq
Torque_est
d-axis flux [speed ]
iq ABC-DQ
Est Torque ud ud Goto
fd Flux_d q-axis
q-axis flux d-axis speed_act
Torque _est stator current
voltage
fq Flux_q uq uq
q-axis
Torque & Flux
voltage
estimator
(1)
(2)
(3)
(4)
Next, the stator flux linkage trace from four different sets of
DTC simulations for PMSM is show in Figure 13. The
comparison of the flux response simulation under the basic DTC
was using switching table 1, table 2, table 3, and under proposed
DTC scheme. It is shown that the proposed DTC system for
PMSM can also reduce flux ripples, so the stator flux linkage is
controlled better at the required value.
The modeling and Simulation of Improved DTC of PMSM Drives 159
Using Matlab/Simulink
DTC clearly and correctly. The simulation result has shown that
the proposed method can reduce the ripple of torque and flux
greatly. By this way, torque and flux ripple is smaller than those of
basic DTC.
REFERENCES
[1]. G. Diamantis and J. M. Prousalidis, "Simulation of a ship
propulsion system with DTC driving scheme," in Power
Electronics, Machines and Drives, 2004. (PEMD 2004).
Second International Conference on (Conf. Publ. No. 498),
2004, pp. 562-567 Vol.2.
[2]. B. Bose, "Power Electronics and Motor Drives: Advances
and Trends," Elsevier Inc., 2006.
[3]. C. L. Ferreira and R. W. G. Bucknall, "Modelling and real-
time simulation of an advanced marine full-electrical
propulsion system," in Power Electronics, Machines and
Drives, 2004. (PEMD 2004). Second International
Conference on (Conf. Publ. No. 498), 2004, pp. 574-579
Vol.2.
[4]. I. Takahashi and T. Noguchi, "A New Quick-Response and
High-Efficiency Control Strategy of an Induction Motor,"
IEEE Transactions on Industry Applications, vol. Vol.IA-22,
No.5, pp. 820-827, Sept/Oct 1986.
[5]. M. Depenbrock, "Direct self control (DSC) of inverter-fed
induction machine," IEEE Trans. on Power Electronics, vol.
3 (4), pp. 420–429, 1988.
[6]. L. Zhong, M. F. Rahman, W. Y. Hu, and K. W. Lim,
"Analysis of direct torque control in permanent magnet
synchronous motor drives," Power Electronics, IEEE
Transactions on, vol. 12, pp. 528-536, 1997.
[7]. L. Zhuqiang, S. Honggang, H. L. Hess, and K. M. Buck,
"The modeling and simulation of a permanent magnet
synchronous motor with direct torque control based on
Matlab/Simulink," in Electric Machines and Drives, 2005
IEEE International Conference on, 2005, p. 7 pp.
The modeling and Simulation of Improved DTC of PMSM Drives 161
Using Matlab/Simulink
10
MODELING AND SIMULATION OF A
NETWORK OF ADJUSTABLE SPEED
DRIVES
Makbul Anwari
M. Imran Hamid Taufik
10.1 INTRODUCTION
The single phase input ASD employs single phase rectifier mainly
in bridge connection using controlled or uncontrolled switch. This
ac/dc converter is well known creates rich odd and triple
harmonics current. The input voltage and current shape of this type
of ASD using full bridge diode rectifier are shown in Fig. 2. For
each half cycle of input voltage it is created one pulse on their
input current. A distortion on voltage shape appears during
commutation period between rectifier switch. For input current as
5.54 A, its draw distortion as 3.59 % fundamental voltage total
harmonics distortion (THDV) and 58.38 % fundamental current
total harmonics distortion (THDI) with dominated by the 3rd, 5th,
7th, 11th and 13th harmonic order.
Modeling and Simulation of a Network of Adjustable Speed Drives 165
The above three types of ASDs have widely used and tend to
increase along with the need of drive demand in many application.
As have shown, during operation each type of ASD draw their
typical harmonics content and distort the line where they are
connected. Along with the increase of number ASD used,
distortion on a certain location on the line (PCC) is also increase
which harmonics content depend on ASD type that used. For
example if there are many single phase input ASD are used, then
the line will rich with the 3rd, 5th, 7th, 11th and 13th harmonic order.
The magnitude of these harmonics increase as the number of ASDs
connected. In this condition, if harmonics filter will be placed to
reduce harmonics problem, relatively high capacity with high cost
filter must be placed.
As describes previously, each type ASD will give typical and different
harmonics order on their voltage and current spectrum. Figure 2 – 4
168 Modeling and Control of Power Converters and Drives
show that current shape are more significantly different between ASD,
meanwhile the voltage shape relatively identical with small distortion in
form of notch caused by switching commutation. The current
magnitude distortion will become greater when several ASD from same
type are operated together. One method to reduce the harmonics
distortion created from several ASDs is by operating some ASDs in
different type in a same network. Basically the method based on
harmonics reduction method with addition of nonlinear component to a
distorted network. By this method, several ASDs in different type are
connected on a PCC of the network so that the harmonics from each
type of ASD are interact each other and give other voltage and current
shape with lower distortion content.
VL
Load
Inverter Gate
Circuit 3
Rpm _1
Gate
bc [ASD1]
A L1
ScopeA
PWM
multi winding transformer 2 IGBT Inverter 4
a bc IL 1 1
com
A A1
A a g Tm
5 2
10
rpm
9 +
3
b aB -K-
A <Rotor speed (wm)> (rpm)
A2
7 4 A A
C8
B B b
2 5 m
c b B 6 6
A4 <Electromagnetic torque Te (N*m)>
B B
4 7
11
C C c 3 -
8 OutputMech 1
C C
n2 c A0
C C
8 9
M4 rectifier -bridge 2
Tm _1
LINE SUPPLY
MULTI-PULSE ASD Inverter Gate
Load 1
Circuit 1
Rpm _
Gate
C2
PWM
Rectifier 2 IGBT Inverter 2
g Tm
[ASD2]
A + rpm 1
+
c
-K-
2 <Rotor speed (wm)> (rpm)
1 A A
C
Breaker
m
c B -
2 -
OutputMech 2
1
C C
Breaker1
Tm _2
Rpm _2
[ASD1] Gate
C3
PWM
Rectifier 3
IGBT Inverter 1
com
a g
A Tm
+ rpm 2
+
A -K-
<Rotor speed (wm)> (rpm)
A A
b B C1
m
B
B <Electromagnetic torque Te (N*m)>
B
-
C
c -
C OutputMech 3
C
C
Tm _
Scope 3
i
+ - c
A
aB
B
b
C
c
C
node 10
REFERENCES
INDEX