Beruflich Dokumente
Kultur Dokumente
of
Power Electronics & Drives
Jayalakshmi Kedarisetti
Dinesh Gopinath
Mridula Jain
December 2, 2006
Contents
3.2.1 Implementation . . . . . . . . . . . . . . . . . . . . . . 38
3.2.2 FPGA Design Files . . . . . . . . . . . . . . . . . . . . 39
3.3 Closed Loop Control of DC Machine . . . . . . . . . . . . . . 41
3.4 Transfer Functions Of The Subsystems . . . . . . . . . . . . . 42
3.4.1 DC Motor and Load . . . . . . . . . . . . . . . . . . . 42
3.4.2 Chopper . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.4.3 Current and Speed Controllers . . . . . . . . . . . . . . 43
3.4.4 Current Feedback . . . . . . . . . . . . . . . . . . . . . 43
3.4.5 Speed Feedback . . . . . . . . . . . . . . . . . . . . . . 43
3.5 Simulation of DC Motor Drive . . . . . . . . . . . . . . . . . . 44
3.5.1 DC Motor Equations . . . . . . . . . . . . . . . . . . . 44
3.5.2 Speed-Feedback Filter . . . . . . . . . . . . . . . . . . 44
3.5.3 Current-Feedback Filter . . . . . . . . . . . . . . . . . 44
3.5.4 Speed Controller . . . . . . . . . . . . . . . . . . . . . 46
3.5.5 Current Controller . . . . . . . . . . . . . . . . . . . . 46
3.5.6 Chopper . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.6 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.7 FPGA Design Files . . . . . . . . . . . . . . . . . . . . . . . . 48
3.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1.1 Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2 PU Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Parameters of the Buck converter . . . . . . . . . . . . . . . . 16
2.2 Base Values for the Buck converter . . . . . . . . . . . . . . . 16
2.3 Per Unit Values . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4 Parameters of the Boost converter . . . . . . . . . . . . . . . . 23
2.5 Base Values for the Boost converter . . . . . . . . . . . . . . . 23
2.6 Parameters of the Buck-Boost converter . . . . . . . . . . . . 30
2.7 Base Values for Buck-Boost converter . . . . . . . . . . . . . . 30
3.1 Details of the DC Motor . . . . . . . . . . . . . . . . . . . . . 39
3.2 Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.3 PU Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.4 Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.1 Details of the Induction Motor . . . . . . . . . . . . . . . . . . 55
4.2 Parameters of the Induction Motor . . . . . . . . . . . . . . . 56
4.3 Base Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.4 PU Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.5 Equations for implementing in Digital Domain . . . . . . . . . 58
6.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.2 Torque . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.3 Torque (contd..) . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.4 Switch Control Signals . . . . . . . . . . . . . . . . . . . . . . 92
6.5 θ1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.6 θ2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.7 Details of the Switched Reluctance Motor . . . . . . . . . . . 96
6.8 Base Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
6.9 PU Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.1 Switch Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
2 LIST OF TABLES
Chapter 1
Real-Time Simulation Of
Dynamic Systems
1.1 Introduction
Any system can be represented by a mathematical model. Dynamic systems
are represented by differential equations or difference equations. Simulation of
dynamic systems require the solving of these differential/difference equations.
In off-line simulation tools, the solution is carried out in a non-real time man-
ner. For example, if it is required to study the response of a system for a few
seconds, it may take minutes or even hours to complete the simulation. This
depends on the complexity and simulation parameters. This is because the
actual time involved in the calculation of the variables is more. On the other
hand in Real-Time Simulation, the results are produced almost instantly. This
is possible if the system model is implemented by an electronic circuit. Histor-
ically, analog computers were used to solve differential equations. But analog
circuits are less flexible in simulating complex systems. Real-time simulation
can be done with digital circuits or microprocessors also. Here, the systems
equations need to be translated to difference-equations. However, conventional
microprocessors and Digital Signal Processors (DSPs) suffer from an inherent
bottleneck in performing calculations. They process the data in a sequential
manner. So, there is a limit to the minimum simulation step time that is
possible for a fixed clock frequency. This limitation can be overcome by par-
alleling processors. Such a scheme is scalable and very useful for simulating
very large dynamic systems with several variables. For real-time simulation of
comparatively less-complex systems, a less costly alternative is required.
An FPGA is a suitable platform for implementing such systems. The basic
advantage of an FPGA is that it can be programmed to process data in parallel.
Thus the implementation of system equations on an FPGA results in very short
execution time. The system model is realized as a combination of sequential
and combinational logic elements. This digital circuit is then programmed
in to the FPGA. The minimum time required to calculate the present state
4 Real-Time Simulation Of Dynamic Systems
Calculate State
Variables
CSV
Update State
Variables
USV
∆τ
variables from previous state variables is the sum of propogation delay of all
logical elements and setup and hold times of sequential elements. A clock
signal drives the digital circuit. During each clock cycle, the present states of
the system are calculated. These calculations are split into two stages as shown
in Fig. 1.1. In the first stage the present states of the system are calculated
using the previous states of the system. In the second stage the values are
updated.
e n−1
yn
Multiplier ADDER D−Flipflop
∆τ
USV
yn−1
D−Flipflop
CSV
D-Flipflop is used for delaying the input until the next clock pulse is given.
With the rising edge of the clock pulse CSV, yi (n − 1) is latched. Error
function ei (n − 1) which is dependent on the previous state variables
yi (n − 1) is calculated and given as input to the integration block. Inputs
to the ADDER block are yi (n − 1), ei (n − 1) ∗ ∆t as in the eq. (1.2). State
variables are updated with the rising edge of the clock USV.
B. Heun’s Method
The curve ei is approximated as a straight line within the interval (n −
1)∆t to n∆t. The area under this straight line is an estimate of the
integral of ei between the limits (n − 1)∆t and n∆t. So
ei (n − 1) + ei (n)
yi (n) = yi (n − 1) + ∆t (1.3)
2
Implementation of the Heun’s method with logical elements is similar
to Eulers method except the input ei (n − 1) is replaced by the average
of ei (n − 1) and ei (n). With the given previous state variable yi (n − 1),
ei (n − 1) is calculated. As said before, FPGA is a parallel device. so
ei (n − 1) has to be latched (Latch) for calculating the average of error
functions. Multiplexer (MUX) selects either ei (n − 1) or average of error
functions depending upon the clock ‘Avg’. So Heun’s method requires two
clock cycles to calculate the present state variables Fig. 1.3. The dotted
line shows division of the on period CSV into two halfs. Obviously, heun’s
method of integration appropriates more number of logic elements when
compare to Eulers method.
6 Real-Time Simulation Of Dynamic Systems
err
D−Flipflop ADDER MUX
2
e n−1 Latch
Avg
err
Multiplier yn
ADDER D−Flipflop
∆τ
USV
yn−1 D−Flipflop
yn−1
MUX CSV
yn−pred
D−Flipflop
Avg Avg
(a)
yn−1
CSV
e n−1
Latch
yn−pred
Avg
yn USV
∆τ
(b)
Figure 1.3: (a) First Order Integrator (Heun’s Method) (b) Triggering Timing.
1.2 Integration Methods 7
err
__
1 MUX
2
e n−1 K1 / K2
err
Multiplier yn
ADDER D−Flipflop
∆τ
USV
yn−1 D−Flipflop
yn−1
MUX CSV
yn−pred
D−Flipflop
Avg K1 / K2
(a)
yn−1
CSV
e n−1
K1 / K2
yn USV
∆τ
(b)
Figure 1.4: (a) Second Order Runge-Kutta Method (b) Triggering Timing.
yi (n) = yi (n − 1) + k2 ∗ ∆t
where k1 = ei yi (n − 1), t(n − 1) (1.4)
1 1
k2 = ei yi (n − 1) + k1 ∆t, t(n − 1) + ∆t
2 2
This is the improved polygon method. This method also requires two
clock cycles to calculate the present state variables fig. 1.4.
Eventhough there are different integration methods to solve (1.1), the only
difference would be the formulation of subroutine to compute the slopes. The
higher order techniques are always the methods of preference [1]. However,
other factors such as programming costs and the accuracy requirements of the
problem must also be considered when choosing a solution technique.
8 Real-Time Simulation Of Dynamic Systems
Ra L
+
i
Vg C Vc
_
Vg = 100 V; Ra = 10 Ω ; L = 20 mH; C = 4 µ H
i
Vg R i L d ib vc
= + + (1.6a)
Vb Rb ib Rb dt Vb
vc
i d Vb
= CRb (1.6b)
ib dt
where ib = Vb /Rb .
With the following abbrevations,
Vg i R vc L
= Vg∗ , = i∗ , = R∗ , = vc∗ , = τLR , CRb = τCR ,
Vb ib Rb Vb Rb
a nondimensional equation results
di∗ −R ∗
−1
∗
i 1
τLR dt Vg∗
dvc∗ = + (1.7)
τCR 1 0 vc∗ 0
dt
These first order linear differential equations can be solved using any numerical
methods as mentioned in section 1.2.
1.4 Implementation 9
1.4 Implementation
A. Parameters : The Table. 1.1 gives the base values for voltage, current and
the values of other quantities.
B. PU System Followed : The Table 1.2 shows the digital equivalent for the
pu values. The bit length of the digital word is not limited in an FPGA.
Inorder to incorporate the signed arithmetic, the digital equivalent for a
negative pu value is chosen as the 1’s compliment of digital equivalent of
its corresponding positive pu value.
1.6 Conclusion
In this chapter, different integration methods are explained. Simple RLC
circuit has been implemented in FPGA based controller.
12 Real-Time Simulation Of Dynamic Systems
Chapter 2
2.1 Introduction
DC-to-DC converters convert electrical power provided from a source at a
certain dc voltage to electrical power at a different dc voltage. Linear regula-
tors are most common converters used for this purpose. They are simple to
analyze and design, also provide very high quality output voltage. But the
major drawback of linear regulators is their poor efficiency[11]. The losses in
such converters appear as heat in the series and shunt elements. The linear
regulators are therefore used only for low power levels. For the applications,
where the efficiency is very important, linear regulators are not suitable. In
such applications, switched mode power converters are standard.
Switched mode power converters use power electronics semiconductor de-
vices which operate in ON and OFF states. Because there is a small power
loss (ideally zero) in those states (very small voltage across a switch in the ON
state and zero current through a switch in OFF state), switching converters
can achieve high energy conversion efficiency.
Vo(t)
on off
+ R
Vg
−
The average value of output voltage depends upon the position of switch.
The ON position connects the source Vg to the output. In the OFF position,
output is totally isolated from the input. The switch is operated at a switching
period of Ts
For a fraction [dTs ] of the switching period, the switch is kept ON. For the
rest of the time [(1 − d)TS ], switch is kept OFF. The fraction d is defined as
duty ratio of the switch[12]. The average output voltage can be calculated in
terms of d and is found to be:
Z TS Z Ton Z Ts
1 1 1 Ton
Vo = Vo dt = Vg dt + 0 dt = Vg = Vg d (2.1)
Ts 0 Ts 0 Ts Ton Ts
on L
iL
+ off
Vg Vc C R
−
The dynamic equations and output equations of the buck converter are
obtained as follows.
During ON time:
di
L = Vg − Vc (2.2a)
dt
dvc Vc
C = il − (2.2b)
dt r
During OFF time:
di
L = −Vc (2.3a)
dt
dvc Vc
C = il − (2.3b)
dt r
2.3 Buck Converter 15
di
L = Vg (d) − Vc (2.4a)
dt
dvc Vc
C = il − (2.4b)
dt r
vo = v c (2.5)
Vo = dVg (2.6)
Ig = dIo (2.7)
The above equations describe the complete modelling of the buck converter.
The Eqn 2.4 is the dynamic equation of the converter. Eqn 2.6 and Eqn 2.7
give the relation between input-output voltage and current respectively.
2.3.2 Implementation
Quantity Value
Vg 13.33V
L 0.64mH
R 5Ω
C 40µF
Tlr 0.00013
Trc 0.0002
d .75
Ts 102.4µsec
4t 3.2µsec
C. PU System Followed : The Table 2.3 shows the digital equivalent for the
pu values. The digital equivalent for a negative pu value is chosen as the
1’s compliment its corresponding positive pu value.
di
L = Vg (d) − Vc (2.13a)
dt
dvc Vc
C = il − (2.13b)
dt r
Dividing the above equations by ib , and Vb respectively,
Z
1
iL(pu) = Vg(pu) d − Vc(pu) dt
Tlr
Z
1
Vc(pu) = iL(pu) d − Vc(pu) dt
Trc
Using the Eulers Method of integration [1], the equation in digital domain
can be written as
2.3.5 Waveforms
The program developed (using the equations 2.14 and 2.15 ) for direct online
simulation in Quartus II tool are downloaded into FPGA board. The wave-
forms are recorded and are compared with the off-line simulation waveforms
as shown in Figs 2.5 and 2.6.
18
Real-Time Simulation Of SMPC
Figure 2.3: FPGA Design File of Buck Converter for PWM wave generation and DAC Interfacing
2.3 Buck Converter
Figure 2.4: FPGA Design File of Buck Converter for Vc and iL
19
20 Real-Time Simulation Of SMPC
1.5
Vg 1
0.5
0
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Vo
0.5
0
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
1.5
iL 1
0.5
0
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
t (sec)
Figure 2.5: Vg, Vo and iL waveforms (off-line simulation) for Buck Converter
Scale: y-axis: all quantities in per unit
Figure 2.6: Vg, Vo and iL waveforms (real time simulation) for Buck Converter
Scale: y-axis: 1 division = 5V (1 per unit = 5V )
2.4 Boost Converter 21
iL L off
+ on
Vg Vc C R
−
The dynamic equations and output equations of the Boost converter are
obtained as follows.
During ON time:
diL
L = Vg (2.16a)
dt
dvc Vc
C =− (2.16b)
dt r
During OFF time:
diL
L = Vg − Vc (2.17a)
dt
dvc Vc
C = il − (2.17b)
dt r
Combining equations 2.16 and 2.17 , following equations are obtained
diL
L = Vg − Vc (1 − d) (2.18a)
dt
dvc Vc
C = iL (1 − d) − (2.18b)
dt r
vo = v c (2.19)
22 Real-Time Simulation Of SMPC
Vg
Vo = (2.20)
(1 − d)
Io
Ig = (2.21)
(1 − d)
The above equations describe the modelling of the boost converter. The
Eqn 2.18 is the dynamic equation of the converter. Eqn 2.20 and Eqn 2.21
give the relation between input-output voltage and current respectively.
2.4.2 Implementation
A. Parameters : The parameters of the Boost Converter used in the simula-
tion program are shown in the Table 2.4.
B. Base Values for Different Quantities : As the implementation is digitally
realized, there arises the need for following a pu system for simple under-
standing. For perunitization of different quantities, the base values are
required, which can be chosen as per convenience. The Table. 2.5 shows
the base values for voltage, resistance and current. The other bases are
calculated from the above mentioned base quantities.
2.4 Boost Converter 23
Quantity Value
Vg 2.5V
L 0.32mH
R 5Ω
C 102.4µF
Tlr 0.000064
Trc 0.000512
d 0.25
Ts 102.4µsec
4t 3.2µsec
2.4.5 Waveforms
The program developed (using the equations 2.27 and 2.28 ) for direct online
simulation in Quartus II tool are downloaded into FPGA board. The wave-
forms are recorded and are compared with the off-line simulation waveforms
as shown in Figs 2.10 and 2.11.
2.4 Boost Converter
Figure 2.8: FPGA Design File of Boost Converter for PWM wave generation and DAC Interfacing
25
26
Real-Time Simulation Of SMPC
Figure 2.9: FPGA Design File of Boost Converter for Vc and iL
2.4 Boost Converter 27
1
Vg
0
−1
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
0.8
0.6
Vo
0.4
0.2
0
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
1.5
1
iL
0.5
0
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
time (sec)
Figure 2.10: Vg, Vo and iL waveforms (off-line simulation) for Boost Converter
Scale: y-axis: all quantities in per unit
Figure 2.11: Vg, Vo and iL waveforms (real time simulation) for Boost Con-
verter
Scale: y-axis: 1 division = 2V (1 per unit = 5V )
28 Real-Time Simulation Of SMPC
on off
+
Vg Vc C R
− L
iL
diL
L = Vg (2.29a)
dt
dvc Vc
C =− (2.29b)
dt r
During OFF time:
diL
L = Vc (2.30a)
dt
dvc Vc
C = −il − (2.30b)
dt r
Combining equations 2.29 and 2.30 , following equations are obtained
diL
L = Vg (d) + Vc (1 − d) (2.31a)
dt
dvc Vc
C = −iL (1 − d) − (2.31b)
dt r
vo = v c (2.32)
2.5 Buck-Boost Converter 29
d
Vo = −Vg (2.33)
(1 − d)
d
Ig = I o (2.34)
(1 − d)
The above equations describe the modelling of the Buck-Boost converter.
The Eqn 2.31 is the dynamic equation of the converter. Eqn 2.33 and Eqn 2.34
give the relation between input-output voltage and current respectively.
2.5.2 Implementation
A. Parameters : The parameters of the Buck-Boost Converter used in the
simulation program are shown in the Table 2.6.
B. Base Values for Different Quantities : As the implementation is digitally
realized, there arises the need for following a pu system for simple under-
standing. For perunitization of different quantities, the base values are
required, which can be chosen as per convenience. The Table. 2.7 shows
the base values for voltage, resistance and current. The other bases are
calculated from the above mentioned base quantities.
30 Real-Time Simulation Of SMPC
Quantity Value
Vg 7.5V
L 0.64mH
R 5Ω
C 102.4µF
Tlr 0.000128
Trc 0.000512
d .25
Ts 102.4µsec
4t 3.2µsec
diL
L = Vg (d) + Vc (1 − d) (2.39a)
dt
dvc Vc
C = −iL (1 − d) − (2.39b)
dt r
Dividing the above equations by ib , and Vb respectively,
Z
1
iL(pu) = dVg(pu) + Vc(pu) (1 − d) dt
Tlr
Z
1
Vc(pu) = − iL(pu) (1 − d) + Vc(pu) dt
Trc
Using the Eulers Method of integration, the equation in digital domain can
be written as
where
1
e (n − 1) = − (1 − d)iL (n − 1) − Vc (n − 1)
Trc
2.5.5 Waveforms
The program developed (using the equations 2.40 and 2.41 ) for direct online
simulation in Quartus II tool are downloaded into FPGA board. The wave-
forms are recorded and are compared with the off-line simulation waveforms
as shown in Figs 2.15 and 2.16.
32
Real-Time Simulation Of SMPC
Figure 2.13: FPGA Design File of Buck-Boost Converter for PWM wave generation and DAC Interfacing
2.5 Buck-Boost Converter
Figure 2.14: FPGA Design File of Buck- Boost Converter for Vc and iL
33
34 Real-Time Simulation Of SMPC
1
Vg
0
−1
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
−0.1
Vo −0.2
−0.3
−0.4
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
0.8
0.6
i 0.4
L
0.2
0
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
time (sec)
Figure 2.16: Vg, Vo and iL waveforms (real time simulation) for Buck-Boost
Converter
Scale: y-axis: 1 division = 2V (1 per unit = 5V )
2.6 Conclusion 35
2.6 Conclusion
In this chapter Real Time Simulation of basic Switching Power Converters
is presented. The Real time simulation is done by using Quartus II soft-
ware (version 5.0). Results are also verified with MATLAB Simulation re-
sults. MATLAB models used for offline simulation are placed in folder named
simulinkprogramfiles.
36 Real-Time Simulation Of SMPC
Chapter 3
Real-Time Simulation of
Separately Excited DC Machine
3.1 Introduction
The Direct Current machine have been dominating the field of adjustable speed
drives for over a century; they are still the most common choice if a controlled
electric drive operating over a wide speed range is specified. This is due to
their excellent operational properties and control characteristics.
+ M
L
Ia
Ra Ia
1 1 ω
kφ
J S
Va La
+ B
E = k φω E
kφ
_ −
Any Electrical machine can be represented in the form of state equations (3.2).
A generalized user machine block(Fig 3.2) is developed such that any machine
equations can be solved by knowing the state matrix and input matrix.
xn−1
Multiply e xn−1 xn
A11
yn−1
Multiply
A12 Integration
Va ADD
Multiply Block xn−1
B11
mL ∆τ
Multiply τe
B12
xn−1
Multiply e yn−1 yn
A21
yn−1
Multiply
A22 Integration
Va ADD
Multiply Block yn−1
B21
mL ∆τ
Multiply τm
B22
A= System Matrix B= Input Matrix ∆τ= Step Size
3.2.1 Implementation
A Parameters : The details of the separetly excited DC machine and the
values of other quantities are given in table. 3.1 and 3.2
3.2 Open Loop Control of DC Machine 39
VOLTAGE 460 V
RATED CURRENT 690 A
POWER 300 kW
SPEED 500 rpm
B PU System Followed : The Table 3.3 shows the digital equivalent for the
pu values. The digital equivalent for a negative pu value is chosen as the
1’s compliment its corresponding positive pu value.
1.5
1
(1)
0.5
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
1.5
1
(2) 0.5
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
10
5
(3) 0
−5
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
(4)
1
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Time
(a) (b)
Fig. 3.4(a) (OFF-Line Simulation Waveforms ) and Fig. 3.4(b) (Real Time
Simulation Waveforms) shows the waveforms obtained for th DC machine in
open loop.
3.3 Closed Loop Control of DC Machine 41
ωm
vω m−ref (s) vω m−error (s) GN(s) v ia−ref (s) v ia−error (s) Gc (s) v ref (s) v a (s)
+ + Gch (s) DC MOTOR i
Limiter Limiter
Chopper
Speed Current
Controller Controller
v ia−fb (s)
vω m−fb (s)
G2 (s)
Current
Feedback
G1 (s)
Speed
Feedback
where Ta = La /Ra is electrical time constant and Tem = JRa /(cφ)2 is elec-
tromechanical time constant.
The transfer function between armature current and input voltage is given by
1
ia (s) ωm (s) ia (s) cφ Js
F2 = = =
Va (s) Va (s) ωm (s) 1 + sTem (1 + sTa ) cφ
(3.4)
1 sTem
=
Ra 1 + sTem (1 + sTa )
Similarly the transfer function between speed and load torque is given by
ωm (s) Ra 1 + sTa
F3 = =− 2
(3.5)
mL (s) (cφ) 1 + sTem (1 + sTa )
The transfer function between armature current and load torque is given by
1
ia (s) cφ
F4 = = (3.6)
mL (s) 1 + sTem (1 + sTa )
3.4 Transfer Functions Of The Subsystems 43
3.4.2 Chopper
The chopper is modeled as first-order lag with a gain of G. The time delay Td
corresponds to the average conduction time. The transfer function is then
Va (s) G
Gch (s) = = (3.7)
vref (s) 1 + sTd
∗ ∆t ∗ ∗
via−ref (n) = KN vwm−error (n − 1) + y(n − 1) + KN vwm−error (n) (3.20)
TN
where vwm−error
∗
= vwm−ref
∗
− vwm−f
∗
b , y(n − 1) is previous value of I-Controller.
In order to maintain the drive system in safe operating region, the current
refernce via−ref
∗
is limited to allowable maximum limits determined by the
motor peak current capabilities. This reference limit is integrated into the
simulation as
∗ ∗
0 ≤ via−ref ≤ via−ref −max (3.21)
∗ ∆t ∗ ∗
vref (n) = Kc via−error (n − 1) + y(n − 1) + KN via−error (n) (3.24)
Tc
where via−error
∗
= via−ref
∗
− via−f
∗
b , y(n − 1) is previous value of I-Controller.
3.5 Simulation of DC Motor Drive 47
3.5.6 Chopper
The transfer function of the chopper is given by equation (3.7). At high
frequency of chopper, the transfer function becomes a simple gain.
Va
= G (3.25)
vref
This equation is normalized and resulted in the following form:
Va∗ = G vref
∗
(3.26)
Discretizing the above equation results in
Va∗ (n) = G vref
∗
(n) (3.27)
3.6 Implementation
A. Closed Loop Parameters : The details of the separetly excited DC ma-
chine and the values of other quantities are given in table. 3.4.
B. PU System Followed : The Table 3.3 shows the digital equivalent for the
pu values. The digital equivalent for a negative pu value is chosen as the
1’s compliment its corresponding positive pu value.
Fig. 3.7 and Fig. 3.8 are the FPGA design files respectively, to implement
filter, PI controller and saturation block.
Figure 3.9 shows the waveforms obtained for th DC motor in close loop.
Initially,the motor is running at no load along with full load refernce and 50
percent torque is applied after some time. Motor rotates in reverse direction
when speed reference has changed from positive to negitive.
3.8 Conclusion
DC Motor open loop as well as closed loop has implemented in FPGA based
controller. Both Real-time and offline simulation results are presented.
50 Real-Time Simulation of Separately Excited DC Machine
0.05
ωref 0
−0.05
0 0.5 1 1.5 2 2.5
1
mL 0.5
0
0 0.5 1 1.5 2 2.5 3
2
ia
0
−2
0 0.5 1 1.5 2 2.5
2
ωm
0
−2
0 0.5 1 1.5 2 2.5
time
Real-Time Simulation of
Induction Machine
4.1 Introduction
In the past, mostly DC motors alone were used for variable speed applications.
With the advances in power electronics, induction motors are replacing DC
Motors in variable speed drives. The 3-phase induction motors are most pre-
ferred for drive applications because of ruggedness, smaller size and low cost.
The basic theory of the squirrel cage induction motor is reviewed here.
Is I 'r
Rs Lls Io L'lr
Ic Im 1:n
R 'r
Rm Lm Vm E 'r = n s V m
Vs
ws wr
the rotor conductors. The interaction of the rotor currents and flux produces
torque and the rotor begins to turn. The rotor attempts to catch up with the
flux. However since this would result in disappearance of torque, an equlibrium
is reached where the rotor runs at a speed such that the relative motion of the
flux is sufficient to produce enough torque to sustain the rotor speed. If the
rotor runs at a speed of N revolutions per minute, the relative speed of the
flux with respect to the rotor is given by
Nr = N s − N (4.2)
bs
br
isb ar
is2 ir1 ira
irb
ε ε
ir2
as
is1 Space Vector isa
is3
Transformation
ir3
ψsa (t) = Ls isa (t) + M ira (t) cosε (t) − M irb (t) sinε (t) (4.6)
ψsb (t) = Ls isb (t) + M ira (t) sinε (t) + M irb (t) cosε (t) (4.7)
where
Ls is self inductance of stator coils,
M is the maximum value of mutual inductance between stator and rotor coils.
Combining equations 4.6 and 4.7 to form the stator flux space phasor.
dis (t) d
v s (t) = vsa + jvsb = Rs is (t) + Ls + M [ir (t) ejε(t) ] (4.14)
dt dt
dir (t) d
v r (t) = vra + jvrb = Rr ir (t) + Lr + M [is (t) e−jε(t) ] (4.15)
dt dt
For a squirrel cage induction motor v r (t) = 0. With the above two equa-
tions, the electrical behaviour of the machine is obtained.
It can also be shown that the torque developed by the machine is given by
2P
Md (t) =M Im[is (t) [ir (t) ejε(t) ]∗ ] (4.16)
32
Transformation of the rotor current equation 4.9 to the stator stationary
coordinates can be done by multiplying ejε(t) ,
isr (t) = ir (t) ejε(t) (4.17)
Simplifying the above equations and separating out the real and imaginary
parts, the following equations can be obtained.
disa (t) M 2 M
Rs isa (t) + σLs − ωisb (t) − Rr isra (t) − M ωisrb (t) = vsa (t) (4.18)
dt Lr Lr
4.4 Implementation 55
disb (t) M 2 M
Rs isb (t) + σLs + ωisa (t) + M ωisra (t) − Rr isrb (t) = vsb (t) (4.19)
dt Lr Lr
dis ra (t) M M
Rr is ra (t) + σLr − Rs isa (t) + M ωisb (t) + Lr ωis rb (t) = − vsa (t)
dt Ls Ls
(4.20)
dis rb (t) M M
Rr is rb (t) + σLr − M ωisa (t) − Rs isb (t) − Lr ωis ra (t) = − vsb (t)
dt Ls Ls
(4.21)
dωm 2P
J = M [isra (t) isb (t) − isrb (t) isa (t)] − Ml (4.22)
dt 32
P dε (t)
ωm = ω = (4.23)
2 dt
h 2
i
where σ = 1 − LMr Ls
ωm - rotor speed in mechanical rad/sec
ω-rotor speed in electrical rad/sec
J-moment of inertia
Ml -Load torque
The above equations describe the complete dynamic behaviour of the in-
duction machine. The Eqns. 4.18 and 4.19 serves the purpose of determning
the stator voltage v s (t) and the eqns. 4.20 and 4.21 corresponding to the
rotor, determine the machine electrical dynamic behaviour. The Eqn. 4.22
determines the mechanical dynamic behaviour of the machine.
4.4 Implementation
A. Parameters : The details of the induction motor is presented in Table. 4.1.
The parameters of the Induction motor used in the simulation program
are shown in the Table 4.2.
VOLTAGE 240 V
CURRENT 40 A
POWER 1.5 KW
POLES (P) 4
SPEED 3000 rpm
Quantity Value
Rs O.2Ω
Lls 7.2mH
Rr 0.4011Ω
Llr 3.6mH
M 88mH
Ls = M + Lls 95.2mH
Lr = M + Llr 91.6mH
J 0.2Kg − m2
B 0.003
C. PU System Followed : The Table 4.4 shows the digital equivalent for the
pu values. The digital equivalent for a negative pu value is chosen as the
1’s compliment its corresponding positive pu value.
pu value Equivalent digital Value Equivalent decimal value
2 pu 7F F FH 32767d
1 pu 3F F FH 16383d
0 pu 000H 0d
-1 pu(16bit) C000H 49152d
-2 pu(16bit) 8000H 32768d
4.7 Waveforms
The program developed (using the equations in table. 4.5) for direct online
control of Induction machine in Quartus II tool are downloaded into FPGA
board. The waveforms are recorded and are compared with the off-line simu-
lation waveforms.
58 Real-Time Simulation of Induction Machine
√
Vsa (n) 1.5 ∗ 2 pu Vs1 (n)
√
Vsb (n) 0.866 ∗ 2 pu (Vs2 (n) − Vs3 (n))
+0.00093puisrb (n − 1) + 0.00288puVsa (n − 1)
−0.001puisrb (n − 1) − 0.00276puVsb (n − 1)
ωm (n) ωm (n − 1) + [0.0038pu (isra (n) isb (n) − isrb (n) isa (n))] − 0.00074puMl (n − 1)
ω (n) ωm (n)
59
60 Real-Time Simulation of Induction Machine
61
62 Real-Time Simulation of Induction Machine
63
64 Real-Time Simulation of Induction Machine
0
Vsa
−1
−2
−3
0.5 0.51 0.52 0.53 0.54 0.55 0.56 0.57 0.58 0.59 0.6
Vsb 0
−1
−2
−3
0.5 0.51 0.52 0.53 0.54 0.55 0.56 0.57 0.58 0.59 0.6
Time
Isa 0
−1
0.5 0.6 0.7 0.8 0.9 1 1.1
1
Isb 0
−1
0.5 0.6 0.7 0.8 0.9 1 1.1
1
0
Isra
−1
0.5 0.6 0.7 0.8 0.9 1 1.1
1
Isrb 0
−1
0.5 0.6 0.7 0.8 0.9 1 1.1
Time
1.2
0.8
0.6
ωm(pu)
0.4
0.2
−0.2
0 0.2 0.4 0.6 0.8 1
0.06
0.04
0.02
M *0.2
g 0
−0.02
−0.04
0 0.2 0.4 0.6 0.8 1
Time
4.8 Conclusion
The principle of direct on line start Induction Motor has been explained. It is
implemented in FPGA based controller. Both Real-time and offline simulation
results are presented.
Chapter 5
5.1 Introduction
Variable Speed Drives (VSDs) are increasingly becoming popular in the indus-
try. Many of the speed control applications such as Fan/Pump drives do not
require very good dynamic performance. In such appications the most suitable
method of speed control of Induction Motors is V/F control. In this Chapter,
the implementation of a V/F Drive controller in FPGA is explained.
f1 f2 f3 f4
T max
Field weakening
region
Torque
Speed
keeping the flux constant. To some extent the drop in torque ability at low
speeds can be counterbalanced by giving a boost to the stator voltage Vs at
low frequencies above the constant Vfs value. Fig. 5.2 shows one example of
a v/f relation. In a V/F drive, the controller produces the required reference
V s(pu)
1.0
0.1
f s(pu)
0.1 1.0
Figure 5.2: V/F relation including low frequency voltage boost
frequency command in accordance with the v/f relation. This command is then
used by a modulator to produce the gating pulses for the inverter. The inverter
5.3 Sine-Triangle Modulation 71
1 pu
PWM
Tr
Su Sv Sw
DC link
U
V
W
Su Sv Sw
• PWM generation
• Slow Start
For sine wave generation 1024 samples per cycle are used. For each sample,
the angle θ is incremented in steps of ∆θ. Calculation of ∆θ is as follows:
Z
θ = ωdt (5.1)
∆θ = 2πf Ts (5.2)
5.4 Implementation of V/F Sine-Triangle Modulator in FPGA 73
Per-unit values of maximum sine value and maximum frequency are the same
(3FFFh). Hence, scaling of sine value for a particular reference frequency can
be done by simply multiplying the sine values by the reference frequency with
proper scaling. Fig. 5.7 shows the entire scheme in block diagram.
74 V/F Control of Induction Motor Drive
freq_ref[15..0]
16
∆ θ[31..0] ∆ θ [29..14]
θ (n)[15..0]
16
84H
θ (n−1)[15..0]
D
Q
θ (n)[15..0]
Counter10
clk
address_theta[15..0]
16383D
3FFF H 1 pu
0
o o
180 360
C000 H −1 pu
ROM
counter10 sine_val_R[31..0]
addr_theta_R[15..0] clk
addr_theta_Y[15..0] freq_ref[15..0]
A−B
5461d addr_theta_Y[15..0]
counter10 ROM
addr_theta_B[15..0] sine_val_Y[31..0]
A−B clk
freq_ref[15..0]
10922d
addr_theta_B[15..0]
counter10 ROM
sine_val_B[31..0]
clk
freq_ref[15..0]
Sub
A−B
12 Bit Up
Counter 12 Triangle
clock
MUX
carrier
2048D
count[11..0]
count11
Sub
7FFh count[11..0]
count11 0
1
triangle
carrier
0 102.4 µ s 204.8 µ s
sine_val_R[15..4]
shifted_sine_R[11..0] 12 bit
Comparator
1024d PWM_R
triangle_carrier[11..0]
To keep the motor currents within safe limits during starting, the speed refer-
ence has to be incremented slowly in a ramp fashion. The required ramping
time depends on the motor full load slip. The slow-starting scheme is described
in Fig. 5.12. Fig. 5.13 shows the slow rise of the frequency reference signal and
the corresponding sinusoidal reference produced by the v/f signal generator.
Counter
a<b up/down
slow_start_clock clk freq_cmd[15..0] freq_ref[15..0]
MUX
enable count enable
MUX
freq_input[15..0]
D
Q
clk clk
a=b enable
en
a<b
a=b
a>b
freq_ref[15..0]
a a=b
compare a<b
freq_cmd[15..0]
b a>b
can be expressed in terms of the switch states and DC link voltage, as:
16 SUB SUB
0 Vv
16 Vwu
Su 16 5431d
VDC
Vvw
16 Vv
Vvw Vvn
MUX
16 SUB SUB
0 Vw Vuv
16
Sv 16 5431d
VDC
Vwu
16 Vw
MUX
Vwn
16 Vwu
0 SUB SUB
Vvw
Vu 16
Sw 5431d
16
Clock Vdc(p.u)
Clock
(a) (b)
5.6 Conclusion
Basic principles of V/F control scheme for Induction Motor are explained.
Developing a v/f controller has been described. The modulation used in the
controller is on Sine-triangle modulation. Implementation details of modulator
and the v/f controller in the FPGA controller board is also discussed. Finally,
an entire open-loop v/f drive consisting of a VSI and an Induction motor is
simulated in real-time and the results are given. This shows the capability
of the FPGA controller board to work as a full-fledged control and real-time
simulation platform.
82 V/F Control of Induction Motor Drive
Chapter 6
Real-Time Simulation Of
Switched Reluctance Motor
6.1 Introduction
The SR motor differs from the conventional motors in many respects. The
motor has saliency both in stator and rotor. The excitation is limited to
stator only. The torque is developed by the tendency of the magnetic circuit
to adopt a configuration of minimum reluctance. The excitation currents are
unidirectional and discontinuous in nature. The stator phases are sequentially
excited to obtain continuous rotation. So the SR motor cannot be operated
with either AC or DC regular power sources. It has to be necessarily operated
from a switching power converter.
In SR motor, when a particular phase is excited, the flux in that phase will
try to align the nearest rotor pole along the corresponding stator pole axis.
The aligned position between the stator and the rotor pole provides minimum
reluctance. This reluctance torque mechanism can be explained with the help
of the diagram in Fig. 6.1. The figure shows that the rotor pole is lying along
the axis OA. For a counter clockwise rotation, this position of the rotor is
referred as unaligned position (0◦ ) with respect to Ph1. After 30◦ movement
of the rotor in the same direction, the same rotor pole will lie along OB, which
is referred as the aligned position (30◦ ) for Ph1. In Fig. 6.1, the rotor is shown
in aligned position with respect to Ph3 and in unlaigned position with respect
to Ph1, whereas it is at 15◦ position with respect to Ph2. On the account of
the saliency in the rotor, the inductance(L) of the exciting coil is a function
of rotor position.
84 Real-Time Simulation Of Switched Reluctance Motor
d(Li)
v = Ri + (6.2)
dt
L is a function of the rotor position. Equation 6.2 canbe decomposed as
di dL
v = Ri + L + iω (6.3)
dt dθ
ω = rotorspeed
θ = rotorposition
dL
Tg = i2 /2 (6.6)
dθ
Equation 6.5 gives the power flow in the machine. The input power is made
up of resistive loss (i2 R), rate of change of stored energy(d(Li2 /2)/dt), and the
converted mechanical power(Tg ω).
Equation 6.6 condenses the torque generating mechanism in the machine.
Torque generation requires an excitation current in the coil and inductance
gradient in the coil. the polarity of the generated torque is the independent of
the polarity of the excitation current. The polarity of the generated torque is
dependent only on the polarity of the inductance gradient(dL/dθ).
3. pole widths of the stator and the rotor are the same and
4. the stator has a pole width to pitch of 0.5.
The inductance of each phase repeats periodically at 60◦ as shown in Fig. 6.2.
The phase difference between two consecutive phases is 15◦ . The winding can
be excited either in the positive or negative gradient region of the inductance
profile. If the magnetic circuit is assumed to have no mutual coupling between
phases and the magnetic characteristics is linear, the torque developed in SR
motor can be expressed as given by Eqn. 6.6.
By causing excitation during the positive or negative slope region of the
inductance profile, the machine can be made to develop positive or negative
torque. By sequential switching of the phases, the motor can be driven in
either direction. In this set-up, the switching sequence for forward rotation
(anticlockwise from the shaft end) is Ph1-Ph2-Ph3-Ph4-Ph1. The sequence
for the reverse rotation (clockwise) will be Ph4-Ph3-Ph2-Ph1-Ph4. It can
be concluded that the motor can be run in either direction and for both the
directions of rotation, the torque can be either positive or negative. Therefore,
all four quadrant operation is possible using this motor.
Each phase is switched on and switched off at least once in every 60◦ . This
is called the fundamental switching frequency for the SR motor. During each
fundamental switching period all the phases are excited once and the interval
between the excitation of two consecutive phases are called the stroke angle.
The total number of strokes per revolution (n), stroke angle (qstroke ) and
fundamental switching frequency (f1) in one one phase can be expressed as,
n = qNr (6.7)
2π
θstroke = (6.8)
qNr
r.p.m
f1 = Nr (6.9)
60
where, q is number of phases and Nr is the number of rotor poles. For an
8/6 motor, strokes per revolution is 24, stroke angle is 15◦ and fundamental
switching frequency for rated speed (1500 rpm) is 150Hz. Since switching of
phases are done in accordance with the information of rotor position, a position
sensor and a switching power converter are essential for an SR motor drive.
where v is the instantaneous voltage across the phase winding; R is its resis-
tance and i the current. The flux-linkage ψ is,
Z
ψ = (v − Ri)dt (6.11)
The flux-linkage, can be computed for different values of current with the
help of Eqn. 6.11. Flux-linkage is measured for a set of rotor positions spanning
from 0◦ to 30◦ at the step of 1◦ . Since the inductance profile is symmetric with
respect to the aligned position of a particular phase the flux-linkage charac-
teristics of a phase will also be symmetric with respect to the aligned position
(30◦ ). Flux-linkage characteristics of a 4 kW, 8/6 pole SRM are given in
Fig. 6.3.
The static flux-linkage characteristics can be approximated by the following
mathematical expression.
i = K1 (θ)ψ + (ψ > ψ1 )K2 (θ)(ψ − ψ1 )2 + (ψ > ψ2 )K3 (θ)(ψ − ψ2 )3 (6.12)
where ψ1 ,ψ2 are constants; K1 (θ),K2 (θ), K3 (θ) are functions of position. The
above parameters K1 (θ),K2 (θ), K3 (θ) are stored in a look-up table for every
2◦ . Linear interpolation can be used for intermediate values. The said values
are given in tabular form in Table 6.1.
Table 6.1:
Position(θ) K1 K2 K3 ψ1 ψ2
0 ◦
68 0 0 0.2 0.45
2 ◦
66.75 0 0 0.2 0.45
4 ◦
65 0 0 0.2 0.45
6 ◦
60 0 0 0.2 0.45
8 ◦
52.5 0 0 0.2 0.45
10 ◦
42 20 0 0.2 0.45
12 ◦
29 48 0 0.2 0.45
14◦ 22 40 200 0.2 0.45
16◦ 17.5 25 400 0.2 0.45
18◦ 14.5 13.5 300 0.2 0.45
20◦ 13 5.5 200 0.2 0.45
22◦ 12 1.5 150 0.2 0.45
24 ◦
11 0 125 0.2 0.45
26 ◦
10 0 110 0.2 0.45
28 ◦
8.9 0 105 0.2 0.45
30 ◦
8.5 0 110 0.2 0.45
40
30
20
10
−10
−20
−30
−40
0 10 20 30 40 50 60
The summation of all the phase torques at any instant gives the instantaneous
torque of the motor. Torque characteristics of a 4 kW, 8/6 pole SRM are given
in Fig. 6.4.
6.5 Torque Characterstics 89
For an SR motor drive, switching on and switching off the phases are syn-
chronised with the rotor position. The test motor is provided with four dis-
crete sensors. The position signals made available from the position sensors.
In real-time simulation, these position signals are generated continuously in
92 Real-Time Simulation Of Switched Reluctance Motor
the simulation itself and are given in Fig. 6.5. In this figure, the idealised
inductance profiles of the four phases are also given.
High Speed Single Pulsed operation: In the high speed range, back-emf
is considerable and it exceeds the supply voltage. Phase currents no longer
can be controlled by chopping method. The individual phases are controlled
with two parameters defined as T-on and T-off. These turn-on time, T-on and
turn-off time, T-off are illustrated in Fig. 6.7. The raw enable signal for this
mode (shown in Table 6.4) is mixed with a controlled T-on and controlled T-
off angles. Hence control of torque in this region is realised by controlling the
dwell angle. Normally the T-on angle is defined as a function of shaft speed
and demanded torque, whereas T-off angle is defined as a function of speed
only. At any particular speed, the desired average torque may be obtained by
more than one combination of these angles. Among these, the combinations
which give minimum torque ripple and minimum peak current are desired.
Generally, exhaustive search method is employed in simulation to calculate
the optimum value of these parameters. These are stored in a look up table as
a function of load torque and speed and referred by the controller as desired.
The angle θ1 is a function of shaft speed as well as the desired torque. The
angle θ2 is a function of the shaft speed. The per unit shaft speed may vary
in the range of 0.5 to 1 in either direction. The desired per unit torque may
vary in the range of -1 to +1. For the controller the values of θ1 and θ2 are
stored in a look-up Table 6.5, 6.6 and referred to as desired.
94
Table 6.5: θ1
Torque -1.00 -0.90 -0.80 -0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Speed
-1.00 55.4 54.7 53.85 53.05 52.1 51 49.85 48.5 46.6 44 41 38.9 41.5 43.2 44.7 45.7 46.7 47.6 48.5 49.1 49.8
-0.90 53.65 53.05 52.3 51.5 50.7 49.8 48.75 47.5 45.9 43.5 40.5 38.4 40.8 42.4 43.6 44.6 45.5 46.3 47 47.8 48.4
-0.80 51.8 51.3 50.7 50.05 49.3 48.5 47.55 46.5 45 42.7 40 38 40 41.5 42.6 43.55 44.4 45.1 45.7 46.35 46.9
-0.70 50.05 49.6 49.1 48.5 47.9 47.2 46.4 45.4 44.05 42 39.5 37.6 39.5 40.9 41.9 42.7 43.5 44.05 44.55 45 45.6
-0.60 48.25 47.9 47.45 47 46.45 45.8 45.1 44.25 43.1 41.25 39 37.2 39 40.1 41 41.7 42.3 42.85 43.3 43.8 44.2
-0.51 46.4 46.05 45.7 45.3 44.9 44.4 43.8 43 42 40.4 38 36.5 38.2 39.2 39.9 40.4 40.9 41.4 41.9 42.2 42.5
0.51 17.5 17.8 18.1 18.6 19.1 19.6 20.1 20.8 21.8 23.5 21.5 19.6 18 17 16.2 15.6 15.1 14.7 14.3 13.95 13.6
0.70 14.4 15 15.45 15.95 16.5 17.3 18.1 19.1 20.5 22.4 20.2 18 15.95 14.6 13.6 12.8 12.1 11.5 10.9 10.4 9.95
0.80 13.1 13.65 14.3 14.9 15.6 16.45 17.4 18.5 20 22 19.6 17.3 15 13.5 12.45 11.5 10.7 9.95 9.3 8.7 8.2
0.90 11.6 12.2 13 13.7 14.5 15.4 16.4 17.6 19.2 21.6 19 16.5 14.1 12.5 11.25 10.2 9.3 8.5 7.7 6.95 6.35
1.00 10.2 10.9 11.5 12.4 13.3 14.3 15.3 16.8 18.5 21.1 18.5 16 13.4 11.5 10.15 9 7.9 6.95 6.15 5.3 4.7
Table 6.6: θ2
Torque -1.00 -0.90 -0.80 -0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Speed
-1.00 30 30 30 30 30 30 30 30 30 30 41 26.9 26.9 26.9 26.9 26.9 26.9 26.9 26.9 26.9 26.9
-0.90 30.50 30.50 30.50 30.50 30.50 30.50 30.50 30.50 30.50 30.50 40.50 27.00 27.00 27.00 27.00 27.00 27.00 27.00 27.00 27.00 27.00
-0.80 30.80 30.80 30.80 30.80 30.80 30.80 30.80 30.80 30.80 30.80 40.00 27.20 27.20 27.20 27.20 27.20 27.20 27.20 27.20 27.20 27.20
-0.70 31.20 31.20 31.20 31.20 31.20 31.20 31.20 31.20 31.20 31.20 39.50 27.50 27.50 27.50 27.50 27.50 27.50 27.50 27.50 27.50 27.50
-0.60 31.50 31.50 31.50 31.50 31.50 31.50 31.50 31.50 31.50 31.50 39.00 27.80 27.80 27.80 27.80 27.80 27.80 27.80 27.80 27.80 27.80
-0.51 31.80 31.80 31.80 31.80 31.80 31.80 31.80 31.80 31.80 31.80 38.00 28.00 28.00 28.00 28.00 28.00 28.00 28.00 28.00 28.00 28.00
0.51 32.00 32.00 32.00 32.00 32.00 32.00 32.00 32.00 32.00 32.00 21.50 28.20 28.20 28.20 28.20 28.20 28.20 28.20 28.20 28.20 28.20
0.60 32.20 32.20 32.20 32.20 32.20 32.20 32.20 32.20 32.20 32.20 20.80 28.50 28.50 28.50 28.50 28.50 28.50 28.50 28.50 28.50 28.50
0.70 32.50 32.50 32.50 32.50 32.50 32.50 32.50 32.50 32.50 32.50 20.20 28.80 28.80 28.80 28.80 28.80 28.80 28.80 28.80 28.80 28.80
0.80 32.80 32.80 32.80 32.80 32.80 32.80 32.80 32.80 32.80 32.80 19.60 29.20 29.20 29.20 29.20 29.20 29.20 29.20 29.20 29.20 29.20
0.90 33.00 33.00 33.00 33.00 33.00 33.00 33.00 33.00 33.00 33.00 19.00 29.50 29.50 29.50 29.50 29.50 29.50 29.50 29.50 29.50 29.50
1.00 33.10 33.10 33.10 33.10 33.10 33.10 33.10 33.10 33.10 33.10 18.50 30.00 30.00 30.00 30.00 30.00 30.00 30.00 30.00 30.00 30.00
6.8 Implementation 95
6.8 Implementation
A. Parameters : The details of the switched reluctance motor is presented
in Table. 6.7.
B. Base Values for Different Quantities : The Table. 6.8 shows the base
values for voltage (chosen as per phase rated voltage of SR Motor), current
(chosen rated current) and frequency (chosen as rated frequency). The
other bases are calculated from the above mentioned base quantities.
C. PU System Followed : The Table 6.9 shows the digital equivalent for the
pu values. The signed arithmetic, the digital equivalent for a negative
pu value is chosen as the 1’s compliment of its corresponding positive pu
value.
96 Real-Time Simulation Of Switched Reluctance Motor
VOLTAGE 280 V
CURRENT 18 A
POWER 4kW
Resistance O.2Ω
1 pu 3F F FH 16383d
0 pu 000H 0d
6.11 Waveforms
The program developed for switched reluctance motor in Quartus II tool
are downloaded into FPGA by using Byte-blaster cable. The waveforms are
recorded and are compared with the off-line simulation waveforms.
98 Real-Time Simulation Of Switched Reluctance Motor
Figure 6.11: (a) FPGA Design File for Generation of Enable Signals
102 Real-Time Simulation Of Switched Reluctance Motor
Figure 6.12: (b) FPGA Design File for Generation of Enable Signals
6.11 Waveforms 103
1.5
1
Enable
signal
0.5
0
0.74 0.742 0.744 0.746 0.748 0.75 0.752 0.754 0.756 0.758 0.76
1.5
1
Amp
0.5
0
0.74 0.742 0.744 0.746 0.748 0.75 0.752 0.754 0.756 0.758 0.76
1.5
1
Nm
0.5
0
0.74 0.742 0.744 0.746 0.748 0.75 0.752 0.754 0.756 0.758 0.76
time
Figure 6.13: (i) Enable Signal, (ii) Phase Current, (iii) Average Torque
[Scale: y-axis: 1 per unit = 5V , ωref = 1500 rpm (1 PU) and Mref =
32.1Nm (1 PU)]
1.5
1
Enable
signal
0.5
0
1.065 1.07 1.075 1.08 1.085 1.09 1.095 1.1 1.105
1.5
1
Amp
0.5
0
1.065 1.07 1.075 1.08 1.085 1.09 1.095 1.1 1.105
1.5
1
Nm
0.5
0
1.065 1.07 1.075 1.08 1.085 1.09 1.095 1.1 1.105
time
Figure 6.14: (i) Enable Signal, (ii) Phase Current, (iii) Average Torque
[Scale: y-axis: 1 per unit = 5V , ωref = 600 rpm (0.4 PU) and Mref =
32.1Nm (1 PU)]
1.5
1
Enable
signal
0.5
0
1.65 1.7 1.75
1.5
1
Amp
0.5
0
1.65 1.7 1.75
1.5
1
Nm
0.5
0
1.65 1.7 1.75
time
Figure 6.15: (i) Enable Signal, (ii) Phase Current, (iii) Average Torque
[Scale: y-axis: 1 per unit = 5V , ωref = 90 rpm (0.06 PU) and Mref =
32.1Nm (1 PU)]
104 Real-Time Simulation Of Switched Reluctance Motor
1.5
1
Enable
signal
0.5
0
0.645 0.65 0.655 0.66 0.665 0.67
1.5
1
Amp
0.5
0
0.645 0.65 0.655 0.66 0.665 0.67
1.5
1
Nm
0.5
0
0.645 0.65 0.655 0.66 0.665 0.67
time
Figure 6.16: (i) Enable Signal, (ii) Phase Current, (iii) Average Torque
[Scale: y-axis: 1 per unit = 5V , ωref = -1500 rpm (-1 PU) and Mref =
32.1Nm (1 PU)]
1.5
1
Enable
signal
0.5
0
0.88 0.882 0.884 0.886 0.888 0.89 0.892 0.894 0.896 0.898 0.9
1.5
1
Amp
0.5
0
0.88 0.882 0.884 0.886 0.888 0.89 0.892 0.894 0.896 0.898 0.9
−0.5
Nm
−1
−1.5
0.88 0.882 0.884 0.886 0.888 0.89 0.892 0.894 0.896 0.898 0.9
time
Figure 6.17: (i) Enable Signal, (ii) Phase Current, (iii) Average Torque
[Scale: y-axis: 1 per unit = 5V , ωref = -1500 rpm (-1 PU) and Mref =
-32.1Nm (-1 PU)]
1.5
1
Enable
signal
0.5
0
1.395 1.4 1.405 1.41 1.415
1.5
1
Amp
0.5
0
1.395 1.4 1.405 1.41 1.415
−0.5
Nm
−1
−1.5
1.395 1.4 1.405 1.41 1.415
time
Figure 6.18: (i) Enable Signal, (ii) Phase Current, (iii) Average Torque
[Scale: y-axis: 1 per unit = 5V , ωref = 1500 rpm (1 PU) and Mref =
-32.1Nm (-1 PU)]
6.11 Waveforms 105
1.5
1
Enable
signal
0.5
0
1.845 1.85 1.855 1.86 1.865
0.8
0.6
Amp
0.4
0.2
0
1.845 1.85 1.855 1.86 1.865
0.8
0.6
0.4
Nm
0.2
0
1.845 1.85 1.855 1.86 1.865
time
Figure 6.19: (i) Enable Signal, (ii) Phase Current, (iii) Average Torque
[Scale: y-axis: 1 per unit = 5V , ωref = 1500 rpm (1 PU) and Mref = 16Nm
(0.5 PU)]
106 Real-Time Simulation Of Switched Reluctance Motor
6.12 Conclusion
The principle of switched reluctance motor has been explained. It was im-
plemented in FPGA based controller. Both Real-time and offline simulation
results are presented.
Chapter 7
7.1 Introduction
Matrix Converters are direct AC-AC power converters without a DC link ca-
pacitor. Matrix Converters are now receiving considerable research attention
due to a number of advantageous features. In a Matrix Converter, output
voltages of desired frequency and amplitude (subjected to certain constraints)
are directly synthesized from the input voltages. Matrix Converters can be de-
signed to be more compact than the conventional AC-AC converters. In this
chapter the implementation of modulators for Matrix Converters is discussed.
The converters are also modelled in the FPGA and a real-time simulation
of the same is given. Results are given for the direct on-line starting of an
Induction Motor fed from the Matrix Converter.
The FPGA platform is capable of implementing complex converter con-
trols. It is possible to simulate the entire drive system consisting of controller,
converter and the machine in real-time.
one switch connected to any output phase (u,v or w) will short circuit the
respective input phases. Also, if at-least one switch connected to each output
phase is not ON, the current path will be broken. This will cause overvolt-
ages to appear across the devices. The basic constraints for a three-phase to
three-phase Matrix Converter can be stated as follows:
SAu + SBu + SCu = 1 (7.1)
SAv + SBv + SCv = 1 (7.2)
SAw + SBw + SCw = 1 (7.3)
where, Sij = 1 or 0, representing the state of switch connecting the in-
put phase i to output phase j. Due to the switching delays in the devices,
commutation schemes are needed to avoid the violation of these constraints.
Four-quadrant switches can be made with anti-series or anti-parallel connec-
tion of IGBTs or MOSFETs.
7.3 Direct and Indirect Matrix Converters 109
For obtaining a desired low frequency averaged waveform, the switches are
modulated at a switching frequency much higher than both the input and
output frequencies. The averaged output voltages can then be represented
in terms of the individual switch duty ratios instead of the switch states in
Eqn. 7.4. The output voltages are represented as follows:
vu (t) dau dbu dcu va (t)
=
vv (t) dav dbv dcv vb (t) (7.5)
vw (t) daw dbw dcw vc (t)
Or,
v uvw = M (t) × vabc (7.6)
where, M (t) is the modulation (duty ratio) matrix in Eqn. 7.5
The duty ratios are determined based on the desired amplitude and fre-
quency of the output voltages. There are several modulation methods to
110 Control and Real-Time Simulation of Matrix Converters
In the Indirect Matrix Converter, the AC-AC conversion is done in two stages.
The input converter is a current source type rectifier. This converter uses
bidirectional switches as in the case of Direct Matrix Converter. The switch-
count can be reduced if regeneration is not needed. A number of new topologies
in this direction have been reported. The output converter is a voltage source
inverter. Here the switches used are IGBTs with anti-parallel diodes. The
DC link does not have a capacitor. The control can be separated between the
two converters. The modulation of the Converters is done in an independent
manner. Fig. 7.2 shows the topology. The conversion involves two stages. In
idc
vpn
p
Su
1 3 5
a u
b Su Sv
n v v
dc
c Sw
Sv
w
4 6 2
Sw
n
vnn
the first stage, the input converter is modulated to get a dc link voltage. In the
second stage, the output converter is modulated to get the AC output from
the dc link voltage. Both the modulations are independently done. There is an
explicit DC link present in this topology. All modulation schemes applicable
to VSI can be used here. Input converter is the dual topology of VSI. The DC
link voltage produced by the input converter can be expressed as:
The above expression gives the instantaneous voltages as the switch states
change. The currents can also be represented in a similar manner.
The output voltages are obtained by the switching action of the output
converter. The relation between the DC link voltages vpn (t) & vnn (t) and the
output voltages {u, v, w} can be shown to be:
vu (t) Su Su
vpn (t)
vv (t) = Sv Sv (7.10)
vnn (t)
vw (t) Sw Sw
Comparing the direct and indirect conversion, it is clear (Eqns. 7.4, 7.9 and
7.10) that both are related.
vu (t) Sau Sbu Scu va (t)
vv (t) = Sav Sbv Scv vb (t)
vw (t) Saw Sbw Scw vc (t)
Su Su va (t)
S1 S3 S5
= Sv Sv
vb (t)
(7.11)
S4 S6 S2
Sw Sw vc (t)
vuvw = I × R × vabc (7.12)
In the input converter, there are 6 bidirectional switches. Fig. 7.3 shows the
current phasors obtained for each switching combination. For each sector one
switch is held ON during the sector, while the return phases are modulated.
For example, in sector 1, the switch S1 is kept ON, while switches S6 and S2 are
modulated. If the two switches connected to the same phase are closed, there
is no current in that particular phase. Such switch combinations are the zero
states of the converter. The switch states S1S4, S3S6, and S5S2 are the 3 zero
states in this converter. The modulation index is kept at 1. For achieving a
p 2, 3
i dc
1 5 III II 1, 2
3, 4
a 3
vdc
b
c IV I
4 6 2 4, 5
V VI 6, 1
n
5, 6
unity input power factor, the reference current space phasor should be in phase
with the input voltage phasor. In each sector, one switch is kept ON, and the
7.4 Modulation of Matrix Converters 113
other two are modulated. The duty ratio of each combination of switch-states
can be calculated from Fig. 7.4. These are given in Eqn. 7.13. The α and β
directions show the active vectors, which arise from the switch combinations.
In sector 1 (Fig.7.4), the switch combination S1S6 and S1S2 corresponds to
the active vectors in α and β directions respectively. The switch combination
S1S4 corresponds to the zero vector.
S1 S2 6 I 2 3 4 5 6
a b c
0
i(t)
iβ
θCSC
iα S1 S6
π
dα = sin − θCSC
3
dβ = sin (θCSC )
d0 = 1 − d α − d β (7.13)
Carrier
1
m2
m1 /m3
m0
T0 Tα Tβ 2T0 Tβ Tα T0
2 Ts
S0
Sα
Sβ
1 − d α − dβ
m0 =
2
= (0.5 − 0.5mSin(60o − θCSC ) − 0.5mSin(θCSC )) (7.14)
m1 = (m0 + dα )
= (0.5 + 0.5mSin(60o − θCSC ) − 0.5mSin(θCSC )) (7.15)
m2 = (m1 + dβ )
= (0.5 + 0.5mSin(60o − θCSC ) + 0.5mSin(θCSC )) (7.16)
m3 = m0 + d β
= (0.5 − 0.5mSin(60o − θCSC ) + 0.5mSin(θCSC )) (7.17)
where m is the modulation index. From Figs. 7.3 and 7.5 we can obtain
the relation between the raw switch signals and the actual switch control
signals.(See Table 7.1). The actual switch control signals derived from Table
7.1 are given in Eqns. 7.18.
7.4 Modulation of Matrix Converters 115
Output converter is a normal Voltage Source Inverter (VSI). Hence the famil-
iar modulation schemes can be applied to get the required output waveforms.
Here, the conventional space vector PWM scheme is explained. The modula-
tion functions for the conventional space vector PWM are the same as those
described for input modulation. The VSI configuration and the voltage space
phasors that are derived from the switch states are shown in Fig. 7.6. The
duty ratios are given in Eqn. 7.19- 7.21.
001 c 101
Sector 1 1
*
Vβ V
0 a b c
θ VSI Vdc
111
011 a
000 Vα 100
VSI
010 b 110
Figure 7.6: VSI and Space vectors
116 Control and Real-Time Simulation of Matrix Converters
π
dα = m × sin − θV SI (7.19)
3
dβ = m × sin (θV SI ) (7.20)
d0 = 1 − d α − dβ (7.21)
Carrier
1
m2
m1 /m3
m0
T0 Tα Tβ 2T0 Tβ Tα T0
2 Ts
S0
Sα
Sβ
Sv
Sw
A fixed point scaling is used to represent the different variables in the FPGA
program. For example, 1 p.u is represented by a 14 bit number 3F F F . Duty
ratios and angles follow this p.u system. For calculating the duty ratios the
sine table is stored in a ROM in the FPGA.
7.5 Implementation of Modulators in FPGA 119
Ref. phase
d0 m0
PWM
Duty dα Modulation m1
Mod.Index
Carrier Ratio Calc Functions
Clock
dβ m2
Gen./Timing
Sync. Circuit
Carrier
• PWM generation
1 p.u/ 3FFFh
3554h
2AAAh
p.u. angle
2000h
1555h
AAAh
o
60o 360
Sect 1
Sect 2
Sect 3
Sect 4
Sect 5
Sect 6
θ vsi
Carrier Generation
A triangular carrier is required for the calculation of different dwell times (or
the ’raw switch signals’). This can be generated as described in Ch. 5.
7.5 Implementation of Modulators in FPGA 121
For a functional simulation, an idealized model can be derived from the Con-
verter equations. In this section idealized models of both Direct and Indirect
Converters are presented. Offline and real-time simulation can be done based
on these models. Since the Converters doesnot contain any energy storage
elements, the equations are simple algebraic equations.
7.6 Modelling of Direct and Indirect Matrix Converters 123
Figure 7.17: PWM signals for Sau and Sbu of Direct Converter
The overall scheme of simulation for the indirect converter is shown in Fig. 7.20.
Fig. 7.21 shows the output line-line voltage and Fig. 7.22 shows the output
phase-neutral voltage.
Ref. Phase
d0 m0
Clock Carrier PWM
Gen./Timing Duty dα Modulation m1
Sync. Circuit Mod.Index Ratio Calc Functions
dβ m2
Carrier
in real time. One can observe the nature and form of the output variables in
a an Oscilloscope. For implementing the converters, the models described in
Eqn. 7.24, 7.25 and 7.4 are used. These equations can be implemented in
digital form using adders and subtracters.
To illustrate how the converter equations are implemented in FPGA, the
input converter model is shown in Fig. 7.26.
While implementing the converter models, the per-unit system adopted is
same as that of modulators. Sinusoidal sources are also generated in the FPGA
with sine tables.
Fig. 7.27 shows the DC link voltage of the indirect converter simulated in
real-time in the FPGA. Fig. 7.28 shows the output phase voltage. These
waveforms exactly match that of the offline simulation results presented in
section 7.6. Fig. 7.29 shows the direct online starting characteristics of an
Induction Motor driven by the Direct Matrix Converter. Here, the converter,
modulator and the machine are modelled together in the FPGA. This shows
the capability of the FPGA platform to simulate in real-time, even a complex
drive system as a Matrix Converter Drive.
130 Control and Real-Time Simulation of Matrix Converters
Figure 7.29: DOL starting of Induction Motor fed by Matrix Converter : Real-
time simulation
7.8 Conclusion
The principle of Direct and Indirect Matrix Converters has been explained.
Modulation of Matrix Converters is shown and implemented in FPGA based
controller. Both Real-time and offline simulation results are presented.
132 Control and Real-Time Simulation of Matrix Converters
Bibliography