Search in the document preview
CHAPTER 2
SINGLE PHASE PULSE WIDTH MODULATED INVERTERS
2.1 Introduction
The dcac converter, also known as the inverter, converts dc power to ac power
at desired output voltage and frequency. The dc power input to the inverter is
obtained from an existing power supply network or from a rotating alternator through
a rectifier or a battery, fuel cell, photovoltaic array or magneto hydrodynamic
generator. The filter capacitor across the input terminals of the inverter provides a
constant dc link voltage. The inverter therefore is an adjustablefrequency voltage
source. The configuration of ac to dc converter and dc to ac inverter is called a dc
link converter.
Inverters can be broadly classified into two types, voltage source and current
source inverters. A voltage–fed inverter (VFI) or more generally a voltage–source
inverter (VSI) is one in which the dc source has small or negligible impedance. The
voltage at the input terminals is constant. A current–source inverter (CSI) is fed with
adjustable current from the dc source of high impedance that is from a constant dc
source.
A voltage source inverter employing thyristors as switches, some type of forced
commutation is required, while the VSIs made up of using GTOs, power transistors,
power MOSFETs or IGBTs, self commutation with base or gate drive signals for their
controlled turnon and turnoff.
16
A standard singlephase voltage or current source inverter can be in the half
bridge or fullbridge configuration. The singlephase units can be joined to have
threephase or multiphase topologies. Some industrial applications of inverters are for
adjustablespeed ac drives, induction heating, standby aircraft power supplies, UPS
(uninterruptible power supplies) for computers, HVDC transmission lines, etc.
In this chapter singlephase inverters and their operating principles are
analyzed in detail. The concept of Pulse Width Modulation (PWM) for inverters is
described with analyses extended to different kinds of PWM strategies. Finally the
simulation results for a singlephase inverter using the PWM strategies described are
presented.
2.2 Voltage Control in Single  Phase Inverters
The schematic of inverter system is as shown in Figure 2.1, in which the
battery or rectifier provides the dc supply to the inverter. The inverter is used to
control the fundamental voltage magnitude and the frequency of the ac output
voltage. AC loads may require constant or adjustable voltage at their input terminals,
when such loads are fed by inverters, it is essential that the output voltage of the
inverters is so controlled as to fulfill the requirement of the loads. For example if the
inverter supplies power to a magnetic circuit, such as a induction motor, the voltage
to frequency ratio at the inverter output terminals must be kept constant. This avoids
saturation in the magnetic circuit of the device fed by the inverter.
17
Battery or
Rectifier Inverter
dV
dC AC
Voltage
Figure 2.1: Schematic for Inverter System
The various methods for the control of output voltage of inverters can be classified as:
(a) External control of ac output voltage
(b) External control of dc input voltage
(c ) Internal control of the inverter.
The first two methods require the use of peripheral components whereas the third
method requires no external components. Mostly the internal control of the inverters
is dealt, and so the third method of control is discussed in great detail in the following
section.
2.2.1 Pulse Width Modulation Control
The fundamental magnitude of the output voltage from an inverter can be
controlled to be constant by exercising control within the inverter itself that is no
external control circuitry is required. The most efficient method of doing this is by
Pulse Width Modulation (PWM) control used within the inverter. In this scheme the
inverter is fed by a fixed input voltage and a controlled ac voltage is obtained by
18
adjusting the on and the off periods of the inverter components. The advantages of the
PWM control scheme are [10]:
a) The output voltage control can be obtained without addition of any external
components.
b) PWM minimizes the lower order harmonics, while the higher order
harmonics can be eliminated using a filter.
The disadvantage possessed by this scheme is that the switching devices used in the
inverter are expensive as they must possess low turn on and turn off times,
nevertheless PWM operated are very popular in all industrial equipments. PWM
techniques are characterized by constant amplitude pulses with different duty cycles
for each period. The width of these pulses are modulated to obtain inverter output
voltage control and to reduce its harmonic content. There are different PWM
techniques which essentially differ in the harmonic content of their respective output
voltages, thus the choice of a particular PWM technique depends on the permissible
harmonic content in the inverter output voltage.
2.2.2 SinusoidalPulse Width Modulation (SPWM)
The sinusoidal PWM (SPWM) method also known as the triangulation, sub
harmonic, or suboscillation method, is very popular in industrial applications and is
extensively reviewed in the literature [12]. The SPWM is explained with reference to
Figure 2.2, which is the halfbridge circuit topology for a singlephase inverter.
19
S11
S12
dV 2 dV
2 dV
+
+
C
C
oV
Figure 2.2: Schematic diagram for HalfBridge PWM inverter.
For realizing SPWM, a highfrequency triangular carrier wave is
compared with a sinusoidal reference of the desired frequency. The intersection of
and waves determines the switching instants and commutation of the modulated
pulse. The PWM scheme is illustrated in Figure 2.3 a, in which v is the peak value of
triangular carrier wave and v that of the reference, or modulating signal. The figure
shows the triangle and modulation signal with some arbitrary frequency and
magnitude. In the inverter of Figure 2.2 the switches and are controlled based
on the comparison of control signal and the triangular wave which are mixed in a
comparator. When sinusoidal wave has magnitude higher than the triangular wave the
comparator output is high, otherwise it is low.
cv
rv
cv rv
c
12
r
11S S
v > is on , r cv 11S 2 d
out V =V (2.1)
and
< is on , rv cv 12S 2 d
out V −=V (2.2)
20
(a)
(b)
Figure 2.3: SPWM illustration (a) SineTriangle Comparison (b) Switching Pulses after comparison.
21
The comparator output is processesed in a trigger pulse generator in such a
manner that the output voltage wave of the inverter has a pulse width in agreement
with the comparator output pulse width. The magnitude ratio of c
r
v v is called the
modulation index ( ) and it controls the harmonic content of the output voltage
waveform. The magnitude of fundamental component of output voltage is
proportional to . The amplitude of the triangular wave is generally kept
constant. The frequency–modulation ratio is defined as
im
im cv
fm
m
t f f
f m = (2.3)
To satisfy the Kirchoff’s Voltage law (KVL) constraint, the switches on the same leg
are not turned on at the same time, which gives the condition
+ = 1 (2.4) 11S 12S
for each leg of the inverter. This enables the output voltage to fluctuate between
2 dV and
2 dV− as shown in Figure 2.4 for a dc voltage of 200 V.
22
Figure 2.4: Output voltage of the HalfBridge inverter.
2.3 SinglePhase Inverters
A singlephase inverter in the full bridge topology is as shown in Figure
2.5, which consists of four switching devices, two of them on each leg. The full
bridge inverter can produce an output power twice that of the halfbridge inverter
with the same input voltage. Three different PWM switching schemes are discussed
in this section, which improve the characteristics of the inverter. The objective is to
add a zero sequence voltage to the modulation signals in such a way to ensure the
clamping of the devices to either the positive or negative dc rail; in the process of
which the voltage gain is improved, leading to an increased load fundamental voltage,
reduction in total current distortion and increased load power factor. In Figure 2.5, the
top devices are assigned to be S11 and S21 while the bottom devices as S12 and S22, the
voltage equations for this converter are as given in the following equations.
23
S21
S22
dV abV
S11
S12
a
b
2 dV
2 dV
+
+
C
C
o +

Figure 2.5: Schematic of a Single Phase FullBridge Inverter.
aonoan d VVVSSV =+=− )(
2 1211 (2.5)
bonobn d VVVSSV =+=− )(
2 2221 (2.6)
bnanab VVV −= (2.7)
The voltages and V are the output voltages from phases A and B to an
arbitrary point n, V is the neutral voltage between point n and the midpoint of the
DC source. The switching function of the devices can be approximated by the Fourier
series to be equal to
anV bn
no
)M1( + 2 1 where M is the modulation signal which when
compared with the triangular waveform yields the switching pulses [19]. Thus from
Equations 2.4, 2.5, and 2.6, the expressions for the modulation signals are obtained as
d
noan
V VV
M )(2
11 +
= (2.8)
. )(2
21 d
nobn
V VV
M +
= (2.9)
24
Equations 2.8 and 2.9 give the general expression for the modulation signals for
singlephase dcac converters. The various types of modulation schemes presented in
the literature can be obtained from these equations using appropriate definition for
, V and V . Making use of this concept different modulation schemes have been
proposed some of which are explained in detail in the following sections.
anV bn no
2.3.1 SPWM With Bipolar Switching
In this scheme the diagonally opposite transistors S11, S22 and S21 and S12 are
turned on or turned off at the same time. The output of leg A is equal and opposite to
the output of leg B. The output voltage is determined by comparing the control signal,
and the triangular signal, V as shown in Figure 2.6(a) to get the switching pulses
for the devices , and the switching pattern is as follows.
rV c
rV >V , Sc 11 is on => 2 Vd
ao =V and S22 is on => 2 Vd
bo −=V ; (2.10)
rV <V , Sc 12 is on => 2 Vd
ao −=V and S21 is on => 2 Vd
bo =V ; (2.11)
hence
)()( tVtV aobo −= (2.12)
25
(a)
(b)
(c)
Figure 2.6:Bipolar PWM (a) Sinetriangle comparison (b) Switching pulses for S11/S22 (c) Switching pulses for S12/S21
(a)
(b)
(c)
Figure 2.7: Bipolar PWM scheme (a) Modulation signal for leg ‘a’ (b) output lineline voltage (c) load current
26
The linetoline voltage is given as in Equation 2.13.
)V (2.13) (2)()()( tVtVtVt aoboaoab =−=
The peak of the fundamentalfrequency component in the output voltage is given as
[10]
diab VmV = ( ) (2.14) 0.1≤im
and
dabd VVV π 4 << (m ). (2.15) 0.1≥i
Since the voltage switches between two levels dV− and V , the scheme is called the
Bipolar PWM. The relationship between fundamental input and output voltage in the
overmodulating region is given as [10].
d
do MVV = (2.16)
where
)1(sin2 21 ααα π
−+= −i mM ,m 1>i
im/1=α .
For a fullbridge inverter with bipolar PWM scheme the output voltage is between
2 dV− and
2 dV . Figure 2.7 shows the modulation signal, output voltage, and the load
current for bipolar modulation scheme on a singlephase inverter with an RL load of
10 Ω and 0.125H.
27
For the bipolar PWM switching scheme there is only one modulation signal and the
switches are turned ‘on’ or turned ‘off’ according to the pattern given in Equations
2.10 and 2.11. The input dc voltage was 200 V and the modulation index (mi) was
taken to be 0.8. The switching frequency for the carrier, which is the triangle, is 10
kHz.
2.3.2 SPWM With Unipolar Switching
In this scheme, the devices in one leg are turned on or off based on the
comparison of the modulation signal V with a high frequency triangular wave. The
devices in the other leg are turned on or off by the comparison of the modulation
signal with the same high frequency triangular wave. Figure 2.8 and 2.9 show
the unipolar scheme for a single –phase full bridge inverter, with the modulation
signals for both legs and the associated comparison to yield switching pulses for both
the legs.
r
rV−
In Figure 2.8 the simulation results show the sine triangle comparison, the
switching pulses for S11 and S21 are shown. The switching for the other two devices is
obtained as S12 = 1 – S11 and S22 = 1 S21. Figure 2.9 shows the phase voltages , line
toline voltages obtained from a unipolar PWM scheme , also shown is the load
current. The simulation was carried out for an RL load of R = 10Ω and L = 0.125H.
The dc voltage is 200 V and the switching frequency is 10kHz. The modulation signal
has a magnitude of 0.8, i.e mi = 0.8.
28
(a)
(b)
(c)
Figure 2.8: Unipolar PWM voltage switching scheme (a) Sine triangle comparison (b) switching pulses for S11 (c) switching pulses for S21.
(a)
(b)
(c)
(d)
Figure 2.9: Unipolar PWM voltage switching scheme (a) phase voltage ‘a’ (b) phase voltage ‘b’ (c) line to line voltage Vab (d) load current
29
The logic behind the switching of the devices in the leg connected to ‘a’ is
given as
rV > V : is on and V = c 11S an 2 dV (2.17)
rV <V : is on and V = c 11S an 2 dV− (2.18)
and that in the leg connected to ‘b’ is given as
V >V : is on and V = r c 11S bn 2 dV (2.19)
V <V : is on and V = r c 11S bn 2 dV− (2.20)
Table 2.1 shows the switching state of the unipolar PWM and the
corresponding voltage levels. It can be observed from the table that when the two top
or the two bottom devices are turned on the output voltage is zero.
In Unipolar switching scheme the output voltage level changes between
either 0 to V or from 0 to +V . This scheme ‘effectively’ has the effect of doubling
the switching frequency as far as the output harmonics are concerned, compared to
the bipolar switching scheme. The voltage waveforms V and are 180o out of
phase from each other as seen in Figure 2.10. The output voltage V is as shown in
Figure 2.11 along with the load current.
d d
an bnV
ab
Since the harmonic components at the switching frequency in and
have the same frequency, this results in the cancellation of the harmonic
component at the switching frequency in the output voltage.
anV
bnV
30
Table 2.1. Switching state of the unipolar PWM and the corresponding voltage levels.
11S 12S 21S 22S AnV BnV BnAno VVV −=
ON   ON dV 0 dV
 ON ON  0 dV V d
ON  ON  dV dV 0
 ON  ON 0 0 0
The fundamental component of the output voltage is given as
dio VmV = ( ) (2.21) 0.1≤im
dod VVV π 4 << (m ). (2.22) 0.1>i
2.3.3 SPWM With Modified Bipolar Switching Scheme (MBPWM)[14]
In the inverter employing the bipolar switching scheme, switches are
operated in such a way that during the positive half of the modulation signal one of
the top devices in one of the switching leg is kept on and the two other switching
devices in the other leg are PWM operated, and during the negative half of the
modulation signal one of the bottom switching device is kept on continuously while
the other two switching devices in the other leg are PWM operated. The output
voltage is determined by comparing the control signal and the triangular wave. rV
31
The switching pattern along with the sinetriangle comparison is as shown in Figure
2.10. The switching pattern for positive values of modulating signal V is as given m
V > V , is on (2.23) r c 21S
and V <V , is on . r c 22S
(a)
(b)
(c)
(d)
(e)
Figure 2.10: Modified bipolar PWM (a) Sinetriangle comparison (b), (c), (d), and (e) switching pulses for devices S11, S12, S21 and S22.
32
(a)
(b)
Figure 2.11: Modified bipolar PWM scheme (a) linetoline voltage (b) load current
The switching pattern for negative values of the modulating signal V is given as m
V < V , is on (2.24) r c 21S
and V > V , is on . r c 22S
The output voltage is given as V )()()( tVtVt BnAno −= , as shown in Figure 2.11. The
load current is also shown in the same plot. The RL load has an R = 10 Ω and L =
0.125H. The modulation signal for the sinetriangle comparison is 0.8. The switching
pattern for the Modified Bipolar Switching Scheme is as given in Table 2.2.
33
Table 2.2. Switching state of the modified bipolar PWM and the corresponding voltage.
11S 12S 21S 22S AnV BnV BnAno VVV −=
ON   ON dV 0 dV
 ON ON  0 dV V d
ON  ON  dV dV 0
 ON  ON 0 0 0
From Table 2.2 it can be observed that when the two top or the two bottom devices
are turned on the output voltage is zero.
In the modified bipolar switching scheme the output voltage level changes
between either 0 to V or from 0 to +V . Since the sign of the modulation signal
decides the switching pattern the analysis of this switching scheme is complex. The
relationship between input and output voltage is given as [14],
d d
do mVV =
(2.25)
where )4(5.0 π += imm ( 0.1<im ) . (2.26)
Thus from the above equation it can be observed that the fundamental component of
the voltage as obtained from the MBPWM is the maximum when compared to the
other switching schemes even in the linear modulation region; that is when the
modulation index is less than unity.
34
2.3.4 Generalized Carrierbased PWM
In the inverter shown in Figure 2.5, the output voltage and the input current are
given as
noanaod VVVSSV +==− )(5.0 1211 (2.27)
nobnbod VVVSSV +==− )(5.0 2221 (2.28)
)( 2111 SSII ad −= (2.29)
bnanab VVV −= . (2.30)
The voltages V and V are the output voltages from phases ‘a’ and ‘b’ to a arbitrary
point while V is the neutral voltage between the point ‘n’ and the midpoint of the
DC source. The generalized carrierbased PWM scheme is obtained by defining the
quantity using the concept of qd Space Vector representation. A special qd
reference frame transformation to transform the two phase voltages to orthogonal qd
voltage components is defined as
an
no
bn
noV
)(5.0 bnanq VVV += (2.31)
)(5.0 bnand VVV −= (2.32)
where and are the qaxis and the daxis voltages in an orthogonal coordinate
system. The qd voltages for each of the possible switching instant are shown in
Table 2.3.
qV dV
35
Table 2.3. Switching state of the generalized carrier based PWM scheme.
11S 21S aoV boV qV dV
  dV5.0− dV5.0− nod VV −− 5.0 0
 ON dV5.0− dV5.0 noV− dV5.0−
ON  dV5.0 dV5.0− noV− dV5.0
ON ON dV5.0 dV5.0 nod VV −5.0 0
Figure 2.12 also shows the space vector representation of the output phase
voltages. To synthesize a given reference output voltage V or equivalently V , the
four vectors shown in the figure are averaged over one switching period for the
inverter
ab *
qd
qdddqdccqdbbqdaaqd VtVtVtVtV +++= * (2.33)
where are the normalized times for which the averaging vector spent in
each of the four quadrants. The normalized times should satisfy the condition that
. The normalized times t can be expressed as some equivalent
time t
dcba tttt ,,,
1=++ dc tt+ ba tt dc t,
o such that
odc ttt =+ (2.34)
or equivalently t can be written as tdc t, oc tγ= which implies t od t)1( γ−= , ]10[∈γ so
Equation 2.33 can be written as
qddoqdcoqdbbqdaaqd VtVtVtVtV )1( * γγ −+++= (2.35)
36
qdb d fVVno =−− ]0, 2
[
qdc d fVVno =− ] 2
,[
qda d fVnoV =− ]0, 2
[
qdd d fVVno =−− ] 2
,[
* qdf
Figure 2.12: Space vector representation of the voltages in a singlephase inverter.
The time is the actual time which the vector spends in the null state that is
when either both the top or both the bottom devices are off or on at the same time.
This time is split in to two time periods such that
ot
1=++ oba ttt ; let t then xba tt =+
xa tt ξ= and so xb tt )1( ξ−= , where ]1,0[∈ξ and ]1,0[∈γ . The quantities t re
the normalized times (with respect to the switching period of the converter). Solving
Equation 2.33 we can get the expression for the zero sequence voltage V in terms of
other known quantities as
oba tt ,
no
, a
*
*
2)12( )12(
)12(5.0 dd
dq dno VV
VV VV
−−
−− −−=
ξ ξ
γ (2.34)
Equations 2.8, 2.9, along with 2.34 constitute the generalized discontinuous PWM
scheme for the singlephase inverter. An infinite number of possibilities for the
discontinuous PWM exist depending on the choice of ξ and γ .
37
2.4 Bipolar and Modified Bipolar PWM Schemes with Zero Sequence Voltage
In the PWM modulation scheme with bipolar voltage switching, the
diagonally opposite switching devices are switched as switch pairs resulting in an
output voltage switching between V and . The zero sequence voltage expression
for the bipolar schemes is given as,
d dV
5.0 )12( −= γdno VV as the qaxis voltage is zero
(refer Equations 2.31 and 2.32). If γ is so chosen so as to locate the zero sequence
voltage to be centered about the peak of the modulation signal, we can achieve higher
fundamental component of the load voltage and less switching because the effect of
the zero sequence is to increase the modulation signal to more than unity. In which
case the comparison of the triangle and the modulation signal would yield continuous
‘on’ or ‘off’ of the switching device for a long period of time as when compared to
the regular sine triangle comparison.
2.5 Implementation of the Bipolar and the Modified bipolar PWM Schemes for
an RL load
The singlephase inverter in the fullbridge topology has been simulated in
Matlab/Simulink for a RL load with R = 10Ω and L = 0.05 H. The modulation signals
(for a modulation index of 0.8) for the switching devices have been obtained from the
TMSLF2407, Texas Instruments DSP. Figure 2.13 shows the simulation result of
bipolar PWM with the zerosequence voltage while without the zero sequence was
already shown in Figures 2.8 and 2.9. The simulation results for the modified bipolar
PWM scheme with the zerosequence voltage are as shown in Figures 2.14. In the
38
simulation the dc voltage was assumed to be 200 V and the modulation index to be
0.8.
(a)
(b)
(c)
(d)
(e)
Figure 2.13: Bipolar PWM scheme (a) modulation signal (b) &(c) switching pulses S11/S22 and S21/S12 respectively.
39
(a)
(b)
(c)
(d)
(e)
(f)
(g)
Figure 2.14: Modified bipolar PWM scheme (a) modulation signal (b), (c), (d) & (e) switching pulses S11, S12, S21 and S22 (f) linetoline voltage (g) load current.
40