Chapter 7

Chapter 7
DC-to-AC Converters
“Introduction to Modern Power Electronics”, 3rd Ed., John Wiley 2015

by
Andrzej M. Trzynadlowski
Chapter 7 DC-AC Converters 1

CONTENT
• 7.1 Voltage-Source Inverters 276
• 7.1.1 Single-Phase VSI 277
• 7.1.2 Three-Phase VSI 286
• 7.1.3 Voltage Control Techniques for PWM Inverters 295
• 7.1.4 Current Control Techniques for VSIs 306
• 7.2 Current-Source Inverters 315
• 7.2.1 Three-Phase Square-Wave CSI 315
• 7.2.2 Three-Phase PWM CSI 319
• 7.3 Multilevel Inverters 322
• 7.3.1 Diode-Clamped Three-Level Inverter 324
• 7.3.2 Flying-Capacitor Three-Level Inverter 327
• 7.3.3 Cascaded H-Bridge Inverter 329
• 7.4 Soft-Switching Inverters 333
• 7.5 Device Selection for Inverters 341
• 7.6 Common Applications of Inverters 344

Voltage-source inverter
supplied from a diode rectifier
RECTIFIER INVERTER
DC LINK
AC LINE
LOAD
Fig. 7.1

Single-phase voltage-source inverter
ii
SA SB
io
Vi
vo
SA' SB'
Fig. 7.2

Circuit diagram of a Z-source
FROM RECTIFIER
TO INVERTER
Fig. 7.3

Each inverter leg can assume two states only: either the upper (common-anode) switch is on and the lower
(common-cathode) switch is off, or the other way around. Thus, two switching variables a and b can be assigned to
the inverter legs, and defined as
0 𝑖𝑓 SA 𝑖𝑠 OFF 𝑎𝑛𝑑 SA′ 𝑖𝑠 ON
𝑎=ቊ
1 𝑖𝑓 SA 𝑖𝑠 ON 𝑎𝑛𝑑 SA′ 𝑖𝑠 OFF
0 𝑖𝑓 SB 𝑖𝑠 OFF 𝑎𝑛𝑑 SB ′ 𝑖𝑠 ON
𝑏=ቊ
1 𝑖𝑓 SB 𝑖𝑠 ON 𝑎𝑛𝑑 SB ′ 𝑖𝑠 OFF
An inverter state is designated as ab2. Four states are possible. The output voltage, vo, of the inverter can be
expressed as
𝑣𝑜 = 𝑉𝑖 (𝑎 − 𝑏)
and the voltage can assume three values only: Vi, 0, and -Vi, corresponding to state 2, states 0 or 3, and state 1,
respectively.
The basic version of the so-called square-wave operation mode of the inverter is described by the following control
law:
1 𝑓𝑜𝑟 0 < 𝜔𝑡 ≤ 𝜋
𝑎=ቊ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1 𝑓𝑜𝑟 𝜋 < 𝜔𝑡 ≤ 2𝜋
𝑏=ቊ
where ω is the fundamental output radian frequency of the inverter. Only states 1 and 2 are used.

States, switching variables, and waveforms
of output voltage and current in a single-phase VSI
in the basic square-wave mode
Fig. 7.4

The total harmonic distortion of the output voltage can be minimized by interspersing states 1 and 2 with states
0 and 3 lasting in the ωt domain 0.81 rad (46.5o) each, as shown in Figure 7.5. Then, in comparison with the
basic square-wave operation, the fundamental output voltage decreases by 8%, to 0.828Vi, but the total
harmonic distortion is reduced by as much as 40%, to 0.29. The control law yielding the optimal square-wave
mode is
1 𝑓𝑜𝑟 𝛼𝑑 < 𝜔𝑡 ≤ 𝜋 + 𝛼𝑑
𝑎=ቊ
1 𝑓𝑜𝑟 𝜋 + 𝛼𝑑 < 𝜔𝑡 ≤ 2𝜋 − 𝛼𝑑
𝑏=ቊ
where αd = 0.405 rad (23.2o).
The quality of operation of the inverter can be improved further by pulse width modulation. The single-phase
inverter has only one output voltage, so the space vector PWM technique is not applicable. Here, for
illustration purposes, a simple PWM strategy based on a sinusoidal modulating function, F(m,ωt) = msin(ωt), is
assumed.
As in all PWM converters, the operating time is a sequence of short switching cycles. Denoting the duty ratios
of switching variables a and b in the nth switching cycle by dan and dbn, the control law is
1
𝑑𝑎𝑛 = 1 + 𝐹(𝑚, 𝛼𝑛 )
2
1
𝑑𝑏𝑛 = 1 − 𝐹(𝑚, 𝛼𝑛 )
2
where m denotes the modulation index and αn is the phase angle of the output voltage in the center of the
switching cycle.

States, switching variables, and waveforms
of output voltage and current in a single-phase VSI
in the optimal square-wave mode
Fig. 7.5

Waveforms of output voltage and current
in a single-phase VSI in the PWM mode, N = 10:
(a) m = 1, (b) m = 0.5
Fig. 7.6

in a single-phase VSI in the PWM mode, N = 20:
(a) m = 1, (b) m = 0.5
Fig. 7.7

Harmonic spectra of output voltage
in a single-phase VSI: (a) basic square-wave mode,
(b) optimal square-wave mode
Fig. 7.8
Harmonic spectra of output voltage
in a single-phase VSI in the PWM mode (m = 1):
(a) N = 10, (b) N = 20
Fig. 7.9

Harmonic spectra of output current
in a single-phase VSI: (a) basic square-wave mode,
(b) optimal square-wave mode
Fig. 7.10

Harmonic spectra of output current
in a single-phase VSI in the PWM mode (m = 1):
(a) N = 10, (b) N = 20
Fig. 7.11

Synthesis of a Sinusoidal Output by PWM
Copyright © 2003 Chapter 8 Switch-Mode DC-

by John Wiley & Sons, Inc. Sinusoidal AC Inverters
Details of a Switching Time Period
• Control voltage can be assumed constant during a

switching time-period

Waveforms of input current
in a single-phase VSI: (a) optimal square-wave
mode, (b) PWM mode (m = 1, N = 20)
Fig. 7.12

Harmonic spectra of input current
in a single-phase VSI: (a) optimal square-wave mode,
(b) PWM mode (m = 1, N = 20)
Fig. 7.13

Analysis assuming Fictitious Filters
• Small fictitious filters eliminate the switching-

frequency related ripple

DC-Side Current
• Bi-Polar Voltage switching

Output Waveforms:
Uni-polar Voltage
Switching
• Harmonic components
around the switching
frequency are absent

DC-Side Current in a Single-Phase Inverter
• Uni-polar voltage switching

Three-phase voltage-source inverter
ii
SA SB SC
Vi
SA' SB' SC'
v AB vBC
vCA
A B C
iA iB iC
vAN v BN vCN
N
Fig. 7.14

It is easy to show that in the three-phase inverter the instantaneous line to line output voltages, vAB, vBC, and
vCA, are given by
𝑣𝐴𝐵 1 −1 0 𝑎
𝑣𝐵𝐶 = 𝑉𝑖 0 1 −1 𝑏
𝑣𝐶𝐴 −1 0 1 𝑐
In a balanced three-phase system, the instantaneous line-to-neutral output voltages, vAN, vBN, and vCN, can be
expressed as
𝑣𝐴𝑁 −1 𝑣𝐴𝐵
1 1 0
𝑣𝐵𝑁 = −1 1 0 𝑣𝐵𝐶
𝑣𝐶𝑁 3
0 −1 1 𝑣𝐶𝐴
which yields
𝑣𝐴𝑁 −1 𝑎
𝑉𝑖 2 −1
𝑣𝐵𝑁 = −1 2 −1 𝑏 .
𝑣𝐶𝑁 3
−1 −1 2 𝑐
The line-to-line voltages can only assume three values, 0 and ±Vi, while the line-to-neutral voltages can
assume five values, 0, ±Vi/3, and ±2Vi/3.
If the 5 – 4 – 6 – 2 – 3 – 1 – ... sequence of states is imposed, each state lasting one-sixth of the desired period
of the output voltage, the individual line-to-line and line-to-neutral voltages acquire waveforms shown in Figure
7.15. This is the square-wave mode of operation, in which each switch of the inverter is turned on and off once
within the cycle of output voltage. The peak value, VLL,1,p, of the fundamental line-to-line output voltage equals
approximately 1.1 Vi and that, VLN,1,p, of the line-to-neutral voltage, 0.64 Vi. Both voltages have the same total
harmonic distortion, THD, of 0.31. As in the square-wave single-phase inverter, the magnitude control of the
output voltage must be realized on the dc supply side.

TABLE 7.1 States and Voltages of the Three-Phase Voltage-Source Inverter
______________________________________________________________________________
State abc 𝑣𝐴𝐵 𝑉𝑖 𝑣𝐵𝐶 𝑉𝑖 𝑣𝐶𝐴 𝑉𝑖 𝑣𝐴𝑁 𝑉𝑖 𝑣𝐵𝑁 𝑉𝑖 𝑣𝐶𝑁 𝑉𝑖
____________________________________________________________________________________________________________________
0 000 0 0 0 0 0 0
1 001 0 -1 1 -1/3 -1/3 2/3
2 010 -1 1 0 -1/3 2/3 -1/3
3 011 -1 0 1 -2/3 1/3 1/3
4 100 1 0 -1 2/3 -1/3 -1/3
5 101 1 -1 0 1/3 -2/3 1/3
6 110 0 1 -1 1/3 1/3 -2/3
7 111 0 0 0 0 0 0
______________________________________________________________________________

Switching variables and waveforms of output voltages in a
three-phase VSI in the square-wave mode
Fig. 7.15

Waveforms of output voltage (line-to-neutral) and current in
a three-phase VSI in the square-wave mode (RL load)
Fig. 7.16

Waveform of input current in a three-phase VSI
in the square-wave mode
Fig. 7.17

Output current waveform
 Voltage-sourced inverter, VSI
 Inverter is a voltage source => output voltage
waveform
 Outpurt current depends on the impedance of the
load
 Fundamental component and harmonics can be analysed
separately in linear circuits, superposition principle
 Current-sourced inverter, CSI
 Output current is defined by the inverter
 Output voltage depends on load
30
Equivalent circuit of an
induction machine
RS jwLS jwLR
US U jwLm w RR = RR
wR s
 f = supply frequency
 w = 2πƒ angular frequency
 wm = rotor mechanical speed
 wM = p wm rotor electrical speed
 wR = w – wM = sw angular frequency of induced currents in
the rotor
 p = number of poles
 s = slip
31
High frequency impedance
 Harmonic n of voltage, angular frequency wn
and wRn wn = nw
w Rn  nw  w  w R    n  1w w`R   n  1w
 Plus sign e.g with 5th harmonic

 Minus sign e.g. with 7th harmonic
 Magnetising reactance
 High when compared to rotor resistance and
reactance
 Can be neglected
 High frequency impedance is
n
Z n  Rs  RR  jnw  L S  L R   jnw  L S  L R 
n 1
32
5th harmonic
 Fundamental, phases A, B, C
 At 5th harmonic phases, A, C, B, order reversed
 => 5th harmonic system rotates in reverse direction
1
0.5
-0.5
-1
0 1 2 3 4 5 6
0.5
-0.5
-1
0 1 2 3 4 5 6
0.5
-0.5
-1
0 1 2 3 4 5 6
33
7th harmonic
 Fundamental, phases A, B, C
 At 7th harmonic phases, A, B, C, same as fudamental
 => 7th harmonic system rotates in the same direction
1
0.5
-0.5
-1
0 1 2 3 4 5 6
0.5
-0.5
-1
0 1 2 3 4 5 6
0.5
-0.5
-1
0 1 2 3 4 5 6
34
Harmonic current
 When induction machine is at stand still connected to
nominal supply UN, starting current Is is
UN
Zk   jw N  L S  L R 
Is
 High frequency impedance can be estimated with

starting impedance w
Z n  nZ k
wN
 Harmonic current Un Un Is wN Un Is
In = = n =
Zn UN w w UN n
wN
 (w/wN) UN is the wanted output voltage U1at frequency w
 Harmonic current when compared to the starting
current In = Un
Is n U1
35
Example
 5th voltage harmonic is 10%
 I5 = 0,1/5* Is = 0,02 Is
 Starting current of the motor is Is = 5…7 IN
 I5 = 0,1…0,14 IN fifth harmonic is therefore larger than the
corresponding voltage harmonic
 Harmonic currents are the higher
 the higher the starting current of the motor is
 i.e. the smaller leakage inductances are
 Motors with higher power
 Leakage inductances are getting smaller
 Harmonic currents are higher than in smaller machines
36
Current harmonics in CSI
 Output current is 120° wide positive and
negative pulses separated by 60° zero
 Harmonics In = I1 , n = 6k ± 1, k = 1, 2,...
n
 Harmonics depend on the fundamental
component, not on the structure (impedances) of
the machine
 VSI
 Current harmonics depend on motor leakage
inductance and not on the motor load
37
Current in induction machine
 a) output voltage and
fundamental component
 b) harmonics of the voltage
 c) harmonics of the current
 integral of voltage harmonics

  u  u1 dt   un  t dt
i  n2
L S  L R L S  L R
 d) current at no load
 e) current at load
38
Side effects of harmonics
 In the inverter
 Peak values of current are higher, higher current
rating for the power semiconductor devices
needed
 Motor
 RMS value of current includes harmonics
 Fundamental current component is reduced and
torque production is lower
 More losses, magnetising and winding losses
 Motor rating must be higher than with sinusoidal
current
39
Torque harmonics
 Torque harmonics caused by current harmonics are
often small but not in six step (square wave) operation
 Fundamental component of airgap flux and current
harmonics
 Torque harmonics
 5th and 7th current harmonic are producing 6th torque harmonic
 Small inertia
 Angular speed starts to change, oscillate
 Also mechanical resonances possible
40
Switching variables and waveforms of output voltages in a
three-phase VSI in the PWM mode
Fig. 7.18

Waveforms of output voltage and current in an RL load of a
three-phase VSI in the PWM mode:
(a) load angle of 30𝑜 , (b) load angle 60𝑜
Fig. 7.19

Waveforms of input current in a three-phase VSI in the
PWM mode: (a) load angle of 30𝑜 , (b) load angle 60𝑜
Fig. 7.20

The principle of the so-called carrier-comparison method is illustrated in Figure 7.21. Reference waveforms, rA, rB,
and rC, given by
𝑟𝐴 𝜔𝑡 = 𝐹(𝑚, 𝜔𝑡)
2
𝑟𝐵 𝜔𝑡 = 𝐹 𝑚, 𝜔𝑡 − 𝜋
3
4
𝑟𝐶 𝜔𝑡 = 𝐹 𝑚, 𝜔𝑡 − 𝜋
3
where 𝐹(𝑚, 𝜔𝑡) denotes the modulating function employed, are compared with a unity-amplitude triangular
waveform y. Values of the switching variables, a, b, and c, change from 0 to 1 and from 1 to 0 at every sequential
intersection of the carrier and respective reference waveforms.
The sinusoidal modulating function, 𝐹 𝑚, 𝜔𝑡 = 𝑚 sin 𝜔𝑡 , is simple, but the voltage gain of the inverter can
significantly be increased using a non-sinusoidal modulating function, many of which were developed over the
years. All those functions consist of a fundamental and triple harmonics, which are not reflected in the three-phase
output voltages and currents of the inverter. The popular third-harmonic modulating function, shown in Figure 7.22
with m = 1, is given by
2 1
𝐹 𝑚, 𝜔𝑡 = 𝑚 sin 𝜔𝑡 + sin(3𝜔𝑡)
3 6
having thus only the fundamental and third harmonic. At m = 1, the fundamental equals 2 3 ≈ 1.15, which
represents a 15% increase in the voltage gain at no changes to the inverter.

Carrier-comparison PWM technique (N = 12, m = 0.75)
Fig. 7.21

Overmodulation
 When ma < 1, linea area
 Harmonics around
multiples of switching
frequency + output
frequency
 Output voltage is not
reaching its maximum
 Overmodulation, ma > 1
 Also lower frequency
harmonics
 Output voltage depends
also on frequency ratio of
output/switching, mf
46
Effect of switching frequency
 Same amplitude of reference but with lower fs
only one pulse/half cycle
 Fundamental of output is higher
 Harmonics at lower frequencies and amplitudes
higher
1.5 1.5
1 1
0.5 0.5
0 0
-0.5 -0.5
-1 -1
-1.5 -1.5
0 1 2 3 4 5 6 0 1 2 3 4 5 6
ma = 1,5 mf = fo/fs = 5 ma = 1,5 mf = 15

47
Adding third harmonic
 No effect in load
 Eliminated in line-to-line voltages
 In wye or delta connection loads have no path for
zero-sequence current
 Amplitude of third harmonic can be 16.7 %
 Flattens the sinusoid at its peak
 Modulator is not saturating
 If added amplitude were higher than 16.7 %
the result would be higher than 1 at p/3
48
Modulating function when 16,7 % third
harmonic added
1.5
0.5
-0.5
-1
-1.5
0 1 2 3 4 5 6
49
ma = 1,167 mf = 15
 Adding third harmonic prevents pulses from
merging together
 Modulator is linear in 0 - 1,167
1.5 1.5
1 1
0.5 0.5
0 0
-0.5 -0.5
-1 -1
-1.5 -1.5
0 1 2 3 4 5 6 0 1 2 3 4 5 6
No third hamonic added Added third harmonics is

16,7 %
50
Third-harmonic modulating function
and its components at m = 1
Fig. 7.22

Digital modulation
 Previous methods were based on analog
technologies
 Ideal sinusoid easy to realise
 Often called natural sampling
 Digital electronics
 Everything is replaced by sampled signals
 Symmetrical sampling (also called uniform,
regular)
 Asymmetrical sampling
52
Natural sampling
 Comparison is done on both edges of
the pulse, leading and trailing
 Result contains much imformation
53
Symmetrical sampling
Modulationg
function
Sampled
modulating
Carrier
Fundamental of
output
54
Symmetrical sampling
 Pulse width depends on
equal distance samples Sampling points
 => symmetric
 Distance between pulse

centers is constant
 Pulse width is calculated
from
Tc
t p  1  ma sin wmt1
2
55
Comparison
 Symmetric
 Pulse edges modulated
similarly
 Natural
 Pulses are not symmetric
around the center points
21.1.2015 S-81.3110 Suuntaajatekniikka, Jorma Kyyrä 56

Asymmetrical
sampling
 Sampling at 2*fk
 Every pulse edge is
modulated separately
 Contains more
information than
symmetric sampling
 Harmonics are reduced
 Pulse width is calculated
from
T  ma 
t p  1   sin wmt1  sin wmt3 
2 2 
57
Sampling and harmonics
 Modulating
function is replaced by a
sampled waveform
 Harmonics of output voltage are changed
when compared to natural sampling
 Fundamental component is not any more
equal to the amplitude of modulating
function
58
Space-vector modulation

The voltage space vector plane with the reference vector 𝑣Ԧ ∗ of line-to-neutral voltages is shown in Figure 7.23(a)
and repeated in Figure 7.23(b) in the per-unit format, with the maximum available magnitude of that vector as the
base. The modulation index, m, constitutes the magnitude of per-unit 𝑣Ԧ ∗ . Neglecting the voltage drops in the
inverter, the highest available peak value of the output line-to-line voltage equals the dc input voltage, 𝑉𝑖 . Thus
∗
𝑉𝐿𝐿,𝑝
𝑚=
𝑉𝑖
∗
where 𝑉𝐿𝐿,𝑝 denotes the reference peak line-to-line voltage. With m = 1 the maximum available peak value of the
fundamental line-to-line output voltage equals the dc input voltage.
In the steady-state, when the fundamental output voltage and current maintain fixed magnitude and frequency, m
is constant and 𝑣Ԧ ∗ rotates with a constant speed. However, the space vector PWM technique allows synthesis of
an instantaneous voltage vector, which may change in magnitude and speed from one switching cycle to another.
The angular position, 𝛽, of the reference vector allows determination of the sector of the complex plane in which
the vector is located within the given sampling cycle of the digital modulator. Specifically,
3
𝑆 = 𝑖𝑛𝑡 𝛽 +1
𝜋
where 𝛽 is expressed in radians and S is the sector number (I to VI). The in-sector position, 𝛼, of 𝑣Ԧ ∗ is then given
by
𝜋
𝛼=𝛽− 𝑆−1
3

Voltage space vector plane of a three-phase VSI:
(a) in volts, (b) per unit
jq jq
_
j
_
V 3
V
V2 2 i V6 V2 j1
V6
III II III II
   
v* I v* I _2_
V* m
Vi V3
V3 V4 V3 V4
IV VI IV VI
V V
V1 V5 V1 V5
(a) (b)
Fig. 7.23

Control of the voltage
 Induction machine is a typical load for VSIs
 Inductances are smoothening current
 Voltage is caused by a changing flux
us  d  s dt
 Flux is an integral of voltage

t
 s   us dt   s 0
0
 Resistances assumed to be small
 Describes the air gap flux of an induction machine
 More detailed models needed in accurate control,
discussed more in courses on electric drives
62
Requirements
uˆs jw t  π 2 
 Idealvoltage vector causes flux s  e
w1
 In VSI
 Only six non-zero voltage vectors
 Nevertheless, output voltage integral
should be similar to ideal
 This is true
 When flux has constant amplitude
 Rotates smoothly with the wanted angular
frequency
63
Flux vector Method where two
space-vectors
separated by zero
vector is tapa
usedB
 Circle is created by s Flux generated
by sinusoidal
sinusoidal voltage sinimuotoisen
voltage
jännitteen luoma vuo
 VSI v3    v2   
 Integral stops when zero
vector is used
v4    v1   
 Non-zero vectors move
flux with constant speed
in the direction of the v6   
v5   
voltage vector
tapa A
Method where one
space-vector
separated by zero
vector is used
64
SVM, Space vector modulator
 Every 60° wide sector is divided either to
 Constant angle slices ∆
 Or constant duration time segments ∆T
 Flux change is same with ideal voltage or with
VSI
T T
 s   us   dt   uˆs e jw t dt
0 0
65
The revolving reference voltage vector is synthesized from stationary active (non-zero)
vectors,𝑉𝑋 and 𝑉𝑌 , framing the sector in question, and a zero vector, 𝑉0 or 𝑉7 . Durations, TX, TY, and
TZ, of states generating those vectors are given by:
𝑇𝑋 = 𝑚𝑇𝑠𝑤 sin 60o − 𝛼

𝑇𝑌 = 𝑚𝑇𝑠𝑤 sin 𝛼
𝑇𝑍 = 𝑇𝑠𝑤 − 𝑇𝑋 − 𝑇𝑌 .
Times TX, TY, and TZ indicate only how long a given state should last in the given switching cycle, but
how the cycle is divided between the employed states must also be specified. The two most
commonly used state sequences can be called a high-quality sequence and a high-efficiency
sequence. The high-quality sequence is
𝑋 − 𝑌 − 𝑍1 − 𝑌 − 𝑋 − 𝑍2 …
where each state in the sequence lasts half of the allotted time. states Z1 and Z2, complementarily 0
and 7, are placed in such an order that a transition from one state to another involves switching in
one inverter leg only. The number of commutations can further be reduced, at the expense of slightly
increased distortion of output currents, when the high-efficiency state sequence
𝑋−𝑌−𝑍−𝑌−𝑋…
is employed. Now, states X and Y last TX/2 and TY/2 seconds respectively, and state Z lasts TZ
seconds. Moreover, Z = 0 in the even sectors (II, IV, and VI) and Z = 7 in the odd sectors (I, III, and
V).
With this state sequence, the average number of pulses of a switching variable per cycle of output
voltage is 2N/3 + 1  2N/3. As a result, the switching losses decrease by about 30% in comparison
with the high-quality state sequence,

Example high-quality space sequence
STATE: X Y Z1 Y X Z2
(4) (6) (7) (6) (4) (0)
100 110 a 111 110 100 000
1
0
b
1
0
c
1
0
T_X T_Y T_Z T_Y T_X T_Z
2 2 2 2 2 2
Tsw
Fig. 7.24

Example high-efficiency space sequence
STATE: X Y Z Y X
(4) (6) (7) (6) (4)
100 110 a 111 110 100
1
0
b
1
0
c
1
0
T_X T_Y T_Y T_X
2 2 TZ 2 2
Tsw
Fig. 7.25

Switching pattern with half- and quarter-wave symmetries
Fig. 7.26

Optimal primary switching angles
as functions of the magnitude control ratio (K = 5)
Fig. 7.27

Harmonic spectrum of line-to-neutral voltage
with the harmonic-elimination technique (K = 5, M = 1)
Fig. 7.28

Switching patterns and voltage and current waveforms:
(1) carrier-comparison PWM with sinusoidal reference,
(2) space vector PWM with high-efficiency state sequence,
(3) programmed PWM with harmonic elimination
Fig. 7.29

Waveforms of output current in a three-phase VSI: (a)
regular PWM, (b) random PWM
Fig. 7.30

Frequency spectra of the line-to-neutral output voltage
in a three-phase VSI: (a) regular PWM, (b) random PWM
Fig. 7.31
Comparison of random PWM techniques
with the regular PWM
SAMPLING & SWITCHING CYCLES

n n+1 n+2 n+3 n+4 ...
t
PWM
SAMPLING & SWITCHING CYCLES

n n+1 n+2 n+3 n+4 ...
t
RPWM
Tsmp
SAMPLING CYCLES
n n+1 n+2 n+3 n+4 ...
t
SWITCHING CYCLES
n-1 n n+1 n+2 n+3 n+4 ...
t
VD-RPWM
t del Tsw,n
Fig. 7.32

Hysteresis current control scheme
B
N
a b c iC
 iC
i C* iB
 iB
i B* iA
 iA
i A*
Fig. 7.33

Characteristic of the hysteresis current controller
a
1
0
 iA
h h
2 2
Fig. 7.34

Waveforms of output currents in a VSI
with hysteresis current control:
(a) 20% tolerance, (b) 10% tolerance
Fig. 7.35
Waveform of output currents in a VSI with hysteresis
current control at a rapid change in the magnitude,
frequency, and phase of the reference current
Fig. 7.36

Space vector version
of the hysteresis current control scheme
B
N
 id zd a b c
i d*
id SWITCHING
 iq zq TABLE
i q*
iq
iA
dq
iB
abc
Fig. 7.37
Best control effects are obtained when state 0 or 7 is imposed

for (zd, zq) = (0, 0), state 1 for (0, 1) and (1, 1), state 2 for (1, -1),
state 3 for (1 ,0), state 4 for (-1, 0), state 5 for (-1, 1), and state 6
for (-1, -1) and (0 ,-1).

Characteristic of a current controller
for the space vector version
of the hysteresis current control scheme
zd
1
h h
2 0 4
 id
h h
4 2
-1
Fig. 7.38

Ramp-comparison scheme for current-controlled VSI
RAMP GENERATOR
CONTROLLER y COMPARATOR
iA iA zA
iA* a
iA
Fig. 7.39

Waveforms of the output current in a VSI
with the ramp comparison current control:
(a) 𝑓𝑟 𝑓1 = 10, (b) 𝑓𝑟 𝑓1 = 20
Fig. 7.40
Current-regulated delta modulation scheme
for a current-controlled VSI
iA
iA* S&H a
iA
Fig. 7.41

Linear current control scheme for a VSI
B
N
C
LINEAR
CONTROLLERS
 id v*d a b c
i d*
PULSE
id WIDTH
 iq v*q MODULATOR
i q*
iq
iA
dq
iB
abc
Fig. 7.42

CSI, Current Source
Inverter

Current-source inverter
supplied from a controlled rectifier
RECTIFIER INVERTER
AC LINE DC LINK LOAD
CURRENT FEEDBACK
Fig. 7.43

Three-phase current-source inverter
ii
SA SB SC
Ii
SA' SB' SC'
vAB vBC
Note: vCA
Symmetrical power A B C
semiconductors like punch- iA iB iC
through IGBTs don’t require
series connected diodes
vAN vBN vCN
N
Fig. 7.44

Switching variables in a three-phase CSI
Fig. 7.45

Idealized waveforms of output currents
in a three-phase CSI in the square-wave mode
Fig. 7.46

in a three-phase CSI in the square-wave mode:
(a) RL load, (b) LE load
Fig. 7.47

Three-phase PWM current-source inverter
RECTIFIER INVERTER
DC LINK LOAD
A i'A iA
iAB
AC LINE
B i'B iB
N
iBC
i CA
C i'C iC
CURRENT FEEDBACK
Fig. 7.48

Carrier-comparison method for the PWM CSI
Fig. 7.49

Optimal switching pattern for the PWM CSI
with two primary switching angles
Fig. 7.50

Waveforms of the output current, capacitor current, and
output voltage in a three-phase PWM CSI
(wye-connected RL load, P = 9)
Fig. 7.51
Generic five-level inverter
S1
V1
Vi
C1
4
S2
V2
Vi
C2
4
S3
V3
Vi
C3 Vi
4
S4
V4 vo
C4 Vi
4
S5
V5
Fig. 7.52

Half-bridge voltage-source inverter
ii
io
Vi
vo
Fig. 7.53

Three-level neutral-clamped inverter
ii
S1
C1
S2
Vi G
S3
C2
S4
vAB vBC
vCA
A B C
iA iB iC
vAN vBN vCN
N
Fig. 7.54

In practice only three states of a leg are used, which makes for the total of twenty-seven states of
the three-level inverter. The limiting condition is that two and only two adjacent switches must be
ON at any time. A ternary switching variable can thus be assigned to each inverter phase and, for
phase A, defined as
0 𝑖𝑓 S3 & S4 𝑎𝑟𝑒 𝑂𝑁
𝑎 = ൞1 𝑖𝑓 S2 & S3 𝑎𝑟𝑒 𝑂𝑁
2 𝑖𝑓 S1 & S2 𝑎𝑟𝑒 𝑂𝑁
Switching variables b and c for the other two phases are defined analogously. It is easy to see
that the potential a given output terminal of the inverter with respect to the "ground" (inverter's
neutral), G, can be expressed in terms of the associated switching variable and input voltage. For
instance, the voltage, vA, at terminal A is
𝑎−1
𝑣𝐴 = 𝑉𝑖 .
2
Consequently, the output line-to-line voltages are given by
𝑣𝐴𝐵 0 𝑎
𝑉𝑖 1 −1
𝑣𝐵𝐶 = 0 1 −1 𝑏
𝑣𝐶𝐴 2
−1 0 1 𝑐
and the line-to-neutral voltages by
𝑉𝑖 2 −1
𝑣𝐵𝑁 = −1 2 −1 𝑏
𝑣𝐶𝑁 6
−1 −1 2 𝑐
Listing all possible values of the line-to-line and line-to-neutral voltages it can be seen that they
can assume five and nine values, respectively. Generally, these numbers in an l-level inverter are
2l – 1 and 4l – 3.

Voltage space vectors
of a three-level neutral-clamped inverter
jq
_
V3
_
V6 j 2 Vi V15 V24
V7 V3 = V16 V12 = V25 V21
V8 V4 = V17 V9 = V22 V18

d
Vi
V5 V1 = V14 V10 = V23

V19
V2 V11 V20
Fig. 7.55

States, switching variables, and waveforms of output
voltages in a three-level neutral-clamped inverter
Fig. 7.56

in a three-level neutral-clamped inverter
Fig. 7.57

Three-level flying-capacitor inverter
ii
S1
C1
S2
Vi G C3
S3
C2
S4
vA B vB C
vC A
A B C
iA iB iC
vA N vB N vC N
0 𝑖𝑓 𝑆3&𝑆4 𝑎𝑟𝑒 𝑂𝑁
𝑎 = ൞1 𝑖𝑓 𝑆1&𝑆3 𝑜𝑟 𝑆2&𝑆4 𝑎𝑟𝑒 𝑂𝑁
2 𝑖𝑓 𝑆1&𝑆2 𝑎𝑟𝑒 𝑂𝑁
Fig. 7.58

Cascaded H-bridge inverter:
(a) block diagram, (b) constituent bridge
S1 S3
Vdc v o,k
vAB vBC
v CA
A B C
iA iB iC
S2 S4
vAN v BN vCN
(a) (b)
Fig. 7.59

An H-bridge can generate three voltage levels between its output terminals, namely –Vdc, 0, and Vdc. The
number N of H-bridges in an l-level inverter is (l - 1)/2, thus N bridges form an inverter with 2N + 1
levels. In the constituent H-bridge two and only switches can be ON at any time. For the k-th bridge in a
leg of a three-phase inverter, the ternary switching variable, ak, is defined as follows:
0 𝑖𝑓 𝑆2&𝑆3 𝑎𝑟𝑒 ON
𝑎𝑘 = ൞1 𝑖𝑓 𝑆1&𝑆3 𝑜𝑟 𝑆2&𝑆4 𝑎𝑟𝑒 𝑂𝑁
2 𝑖𝑓 𝑆1&𝑆4 𝑎𝑟𝑒 𝑂𝑁.
The output voltage, vo,k, of the bridge is then given by
𝑣𝑜,𝑘 = (𝑎𝑘 − 1)𝑉𝑑𝑐 .
The voltage of terminal A, vA, with respect to the inverter’s neutral, G, is a sum of output voltages of all
the bridges. Consequently, a switching variable of phase A of the inverter can be defined as
𝑁
𝑎=෍ 𝑎𝑘
𝑘=1
where, depending on control of the individual bridges, a can assume any integer value from the 0 to 2N
range. Then,
𝑣𝐴 = (𝑎 − 𝑁)𝑉𝑑𝑐 .

Switching variables b and c are defined analogously, yielding the
following equations for the line-to-line and line-to-neutral output
voltages of the inverter:
𝑣𝐴𝐵 1 −1 0 𝑎
𝑣𝐵𝐶 = 𝑉𝑑𝑐 0 1 −1 𝑏
𝑣𝐶𝐴 −1 0 1 𝑐
and
𝑉𝑑𝑐 2 −1
𝑣𝐵𝑁 = −1 2 −1 𝑏
𝑣𝐶𝑁 3
−1 −1 2 𝑐

Approximation of a sinewave by a stepped waveform
in the H-bridge cascaded inverter
Fig. 7.60

Endpoints of line-to-neutral vectors
of two-bridge cascaded inverter
Fig. 7.61

Cells with diode rectifiers of two-bridge cascaded inverter:
(a) single-phase, (b) three-phase
(a) (b)
Fig. 7.62

Cells with PWM rectifiers for ac-supplied cascaded inverter:
(a) single-phase, (b) three-phase
(a) (b)
Fig. 7.63

7.4 Soft-switching inverters
Switched network for illustration
Note: additional reading, not of the operating principle of a resonant dc link
required in the exam
R L
ii
iC iS iD
Vi C S D vo Io
Fig. 7.64

Waveforms of voltage and current
in the resonant dc link
Fig. 7.65
Three-phase resonant dc link inverter with an active clamp
C cl D S
Vi C
A B C
N
Fig. 7.66

Waveforms of line-to-line output voltages
in a resonant dc-link inverter
Fig. 7.67

Auxiliary resonant commutated pole inverter:
(a) one phase with the auxiliary circuit, (b) the entire inverter
Ci S1 Cr
AC
Lr
Vi
S
Lf
Ci S2 Cr
(a)
AC
AC
AC
A B C
(b)
Fig. 7.68

Idealized line-to-neutral voltage and line current
waveforms in a VSI in the square-wave mode
Fig. 7.69

Block diagram of a photovoltaic utility interface
P V A RR A Y INV E RT E R T RA N SF OR MER D IOD E R EC T IF IE R C ON T ROLL ED RE C T IFIE R

FILT E R
D C LIN K
GRID
Fig. 7.70

Block diagram of an active power filter
NONLINEAR LOAD
AC LINE
iA i'A
A
B
C
i"A
INVERTER
OUTER FEEDBACK
INNER FEEDBACK
CONTROL SYSTEM
Fig. 7.71

Waveforms of voltage and current
in an active power filter
Fig. 7.72

UPS System
S T A T IC SW IT C H
R EC T IF IE R D C LIN K INV E RT E R
L IN E LOAD
FILT E R FILT E R
LOAD
GRID
B AT T ER Y
Fig. 7.73

Block diagram of an ac drive system
with scalar speed control
RECTIFIER INVERTER
DC LINK
MOTOR
SPEED
AC LINE
SENSOR
LOAD
v* w*
VOLTAGE CONTROLLER pp
w*syn
w* wM
w*sl
SLIP CONTROLLER
 wM
wM
w*M
Fig. 7.74

Use of the modular frequency changer of Figure 2.24 in
an ac drive: (a) system with a braking resistor,
(b) system with a step-up chopper
BRAKING
RECTIFIER RESISTOR INVERTER
AC LINE
TO MOTOR
(a)
RECTIFIER STEP-UP CHOPPER INVERTER
BOOST
INDUCTOR
AC LINE
TO MOTOR
(b)
Fig. 7.75

PWM rectifier-inverter cascades for bidirectional power flow in ac
motor drives: (a) current-type rectifier, inductive dc link, and
current-source inverter, (b) voltage-type-rectifier, capacitive dc
link, and voltage-source inverter
RECTIFIER DC LINK INVERTER
LINE MOTOR
FILTER
AC LINE
(a)
RECTIFIER DC LINK INVERTER
LINE
FILTER MOTOR
AC LINE
(b)
Fig. 7.76

Switching pattern of the inverter in Example 7.2
STATE: X Y Z Y X
(1) (5) (7) (5) (1)
001 101 a 111 101 001
1
0 t ( s)
b
1
0 t ( s)
c
1
0 t ( s)
0 32 92.5 157.5 218 250
Fig. 7.77

Per-unit voltage vectors of the three-level inverter
in Example 7.3
jq
V24
j 23
V15
j 43 V12 = V25 V21
v*
 V9 = V22 V18
d
0 0.5 1
Fig. 7.73

Chapter 7

Uploaded by

Copyright:

Available Formats

Chapter 7

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 7

Uploaded by

Copyright:

Available Formats

Chapter 7

“Introduction to Modern Power Electronics”, 3rd Ed., John Wiley 2015

Chapter 7 DC-AC Converters 1

Chapter 7 DC-AC Converters 2

Chapter 7 DC-AC Converters 3

Chapter 7 DC-AC Converters 4

Chapter 7 DC-AC Converters 5

Chapter 7 DC-AC Converters 6

Chapter 7 DC-AC Converters 7

where αd = 0.405 rad (23.2o).

Chapter 7 DC-AC Converters 8

Chapter 7 DC-AC Converters 9

Chapter 7 DC-AC Converters 10

Chapter 7 DC-AC Converters 11

Chapter 7 DC-AC Converters 13

Chapter 7 DC-AC Converters 14

Chapter 7 DC-AC Converters 15

Copyright © 2003 Chapter 8 Switch-Mode DC-

• Control voltage can be assumed constant during a

Copyright © 2003 Chapter 8 Switch-Mode DC-

Chapter 7 DC-AC Converters 18

Chapter 7 DC-AC Converters 19

• Small fictitious filters eliminate the switching-

Copyright © 2003 Chapter 8 Switch-Mode DC-

• Bi-Polar Voltage switching

Copyright © 2003 Chapter 8 Switch-Mode DC-

• Uni-polar voltage switching

Copyright © 2003 Chapter 8 Switch-Mode DC-

SA' SB' SC'

Chapter 7 DC-AC Converters 24

Chapter 7 DC-AC Converters 25

State abc 𝑣𝐴𝐵 𝑉𝑖 𝑣𝐵𝐶 𝑉𝑖 𝑣𝐶𝐴 𝑉𝑖 𝑣𝐴𝑁 𝑉𝑖 𝑣𝐵𝑁 𝑉𝑖 𝑣𝐶𝑁 𝑉𝑖

1 001 0 -1 1 -1/3 -1/3 2/3

2 010 -1 1 0 -1/3 2/3 -1/3

3 011 -1 0 1 -2/3 1/3 1/3

4 100 1 0 -1 2/3 -1/3 -1/3

5 101 1 -1 0 1/3 -2/3 1/3

6 110 0 1 -1 1/3 1/3 -2/3

Chapter 7 DC-AC Converters 26

Chapter 7 DC-AC Converters 27

Chapter 7 DC-AC Converters 28

Chapter 7 DC-AC Converters 29

 Plus sign e.g with 5th harmonic

 High frequency impedance can be estimated with

Chapter 7 DC-AC Converters 41

Chapter 7 DC-AC Converters 42

Chapter 7 DC-AC Converters 43

Chapter 7 DC-AC Converters 44

Chapter 7 DC-AC Converters 45

ma = 1,5 mf = fo/fs = 5 ma = 1,5 mf = 15

No third hamonic added Added third harmonics is

Chapter 7 DC-AC Converters 51

 Distance between pulse

21.1.2015 S-81.3110 Suuntaajatekniikka, Jorma Kyyrä 56

Chapter 7 DC-AC Converters 59

Chapter 7 DC-AC Converters 60

Chapter 7 DC-AC Converters 61

 Flux is an integral of voltage

𝑇𝑋 = 𝑚𝑇𝑠𝑤 sin 60o − 𝛼

Chapter 7 DC-AC Converters 66