Вы находитесь на странице: 1из 44

MODULATION

TECHNIQUES IN
INVERTERS
Advanced Power Electronics and Applications
ECEG 6309
Organization
I. Voltage Source Inverter (VSI)
A. Six-step VSI
B. Pulse Width Modulated VSI
II. PWM Techniques
A. Sine PWM
B. Hysteresis (Bang-bang)
C. Space Vector PWM
III. References
I. Voltage Source Inverter (VSI)
A. Six-Step VSI (1)
 Six-Step three-phase Voltage Source Inverter

Fig. 1 Three-phase voltage source inverter.


I. Voltage Source Inverter
(VSI)
A. Six-Step VSI (2)
• Gating signals, switching sequence and line to negative voltages

 Each switch conducts for


180 degree.
 There is switching at
every 60 degree.

Fig. 2 Waveforms of gating signals, switching sequence, line to negative voltages


for six-step voltage source inverter.
I. Voltage Source Inverter (VSI)
A. Six-Step VSI (3)
Switching Sequence:
561 (V3)  612(V4)  123 (V5)  234 (V6)  345 (V1)  456 (V2)  561 (V1)
where, 561 means that S5, S6 and S1 are switched on

V3= [0 1 0] V4= [0 1 1]

V5= [0 0 1] V6= [1 0 1]

V1= [1 0 0]
V2= [1 1 0]

Fig. 3 Six inverter voltage vectors for six-step voltage source inverter.
I. Voltage Source Inverter (VSI)
A. Six-Step VSI (4)
Line to line voltages (Vab, Vbc, Vca) and line to neutral voltages (Van, Vbn, Vcn)
Line to line voltages
 Vab = VaN - VbN

 Vbc = VbN - VcN

 Vca = VcN - VaN

 Phase voltages

 Van = 2/3VaN - 1/3VbN - 1/3VcN

 Vbn = -1/3VaN + 2/3VbN - 1/3VcN

 Vcn = -1/3VaN - 1/3VbN + 2/3VcN

Fig. 4 Waveforms of line to neutral (phase) voltages and line to line voltages
for six-step voltage source inverter.
I. Voltage Source Inverter (VSI)
A. Six-Step VSI (5)
.  Amplitude of line to line voltages (Vab, Vbc, Vca)

 Fundamental Frequency Component (Vab)1

3 4 Vdc 6
(Vab )1 (rms)   Vdc  0.78Vdc
2 2 

 Harmonic Frequency Components (Vab)h


: amplitudes of harmonics decrease inversely proportional to their harmonic order
0.78
(Vab )h (rms)  Vdc
h

where, h  6n  1 (n  1, 2, 3,.....)
I. Voltage Source Inverter (VSI)
A. Six-Step VSI (6)

.  Characteristics of Six-step VSI

 It is called “six-step inverter” because of the presence of six “steps”


in the line to neutral (phase) voltage waveform

 Harmonics of order three and multiples of three are absent from


both the line to line and the line to neutral voltages
and consequently absent from the currents

 Output amplitude in a three-phase inverter can be controlled

by only change of DC-link voltage (Vdc)


I. Voltage Source Inverter (VSI)
B. Pulse-Width Modulated VSI (1)

 Objective of PWM

 Control of inverter output voltage

 Reduction of harmonics

 Disadvantages of PWM
 Increase of switching losses due to high PWM frequency

 Reduction of available voltage

 EMI problems due to high-order harmonics

9
I. Voltage Source Inverter (VSI)
B. Pulse-Width Modulated VSI (2)

 Pulse-Width Modulation (PWM)

Fig. 5 Pulse-width modulation.


10
I. Voltage Source Inverter (VSI)
B. Pulse-Width Modulated VSI (3)

 Inverter output voltage

 When vcontrol > vtri, VA0 = Vdc/2

 When vcontrol < vtri, VA0 = -Vdc/2

 Control of inverter output voltage

 PWM frequency is the same as the frequency of v tri

 Amplitude is controlled by the peak value of vcontrol

 Fundamental frequency is controlled by the frequency of v control

 Modulation Index (m)

vcontrol peak of (V A0 )1
m   ,
vtri Vdc / 2
where, (VA0 )1 : fundamenta l frequecny component of VA0
11
II. PWM METHODS
A. Sine PWM (1)

 Three-phase inverter

Fig. 6 Three-phase Sine PWM inverter.

12
II. PWM METHODS
A. Sine PWM (2)
 Three-phase sine PWM waveforms
vtri vcontrol_A vcontrol_B vcontrol_C
 Frequency of vtri and vcontrol

 Frequency of vtri = fs

VA0
 Frequency of vcontrol = f1

VB0
where, fs = PWM frequency
f1 = Fundamental frequency

VC0
 Inverter output voltage

VAB
 When vcontrol > vtri, VA0 = Vdc/2

VBC
 When vcontrol < vtri, VA0 = -Vdc/2

where, VAB = VA0 – VB0


VCA

VBC = VB0 – VC0


t
VCA = VC0 – VA0
Fig. 7 Waveforms of three-phase sine PWM inverter.

13
II. PWM METHODS
A. Sine PWM (3)

 Amplitude modulation ratio (ma)

peak amplitude of vcontrol peak value of (V A0 )1


 ma   ,
amplitude of vtri Vdc / 2
where, (VA0 )1:fundamental frequecny component of VA0

Frequency modulation ratio (mf)


fs
mf  , where, f s  PWM frequency and f1  fundamenta l frequency
f1
 mf should be an odd integer
 if mf is not an integer, there may exist sunhamonics at output voltage
 if mf is not odd, DC component may exist and even harmonics are present at output voltage

 mf should be a multiple of 3 for three-phase PWM inverter


 An odd multiple of 3 and even harmonics are suppressed

14
II. PWM METHODS
B. Hysteresis (Bang-bang) PWM (1)

 Three-phase inverter for hysteresis Current Control

Fig. 8 Three-phase inverter for hysteresis current control.

15
II. PWM METHODS
B. Hysteresis (Bang-bang) PWM (2)

 Hysteresis Current Controller

 based on deriving the switching signals from the


comparison of the current error with a fixed
tolerance band.
 control is based on the comparison of the actual
phase current with the tolerance band around the
reference current associated with that phase.
 The goal is to keep the actual value of the currents
within their hysteresis bands all the time.

Fig. 9 Hysteresis current controller at Phase “a”.

16
II. PWM METHODS
B. Hysteresis (Bang-bang) PWM (3)

 Characteristics of hysteresis Current Control

 Advantages

 Excellent dynamic response

 Low cost and easy implementation

 Drawbacks

 Large current ripple in steady-state

 Variation of switching frequency

 No intercommunication between each hysterisis controller of three phases


and hence no strategy to generate zero-voltage vectors.
As a result, the switching frequency increases at lower modulation index and
the signal will leave the hysteresis band whenever the zero vector is turned on.

 The modulation process generates subharmonic components


17
II. PWM METHODS
C. Space Vector PWM (1)

 Output voltages of three-phase inverter (1)

Fig. 10 Three-phase power inverter.

where, upper transistors: S1, S3, S5


lower transistors: S4, S6, S2
Pole voltages: Va, Vb, Vc

18
II. PWM METHODS
C. Space Vector PWM (2)

 Output voltages of three-phase inverter (2)

 S1 through S6 are the six power transistors that shape the ouput voltage
 When an upper switch is turned on (i.e., S1, S3 or S5 is “1”), the corresponding lower
switch is turned off (i.e., S4, S6 or S2 is “0”)
 Eight possible combinations of on and off patterns for the three upper transistors (S 1, S3, S5)

 Line to line voltage vector [Vab Vbc Vca]t in terms of pole voltages

Vab  1  1 0 Va 
V   V 0 1  1 Vb , where switching variable vector [a b c]t
 bc  dc   
Vca   1 0 1 Vc 
 Line to neutral (phase) voltage vector [V an Vbn Vcn]t in terms of pole voltages
Van  2  1  1 Va 
V   1 V  1 2  1 Vb 
 bn  3 dc   
Vcn   1  1 2 Vc 
19
II. PWM METHODS
C. Space Vector PWM (3)

Definition: SVM is one of the preferred real-time


modulation techniques and is widely used for digital control of
voltage source inverters.
• Switching States: The operating status of the switches in the
two-level inverter in Fig. shown right can be represented by
switching states.

– Switching state ‘1’ denotes that the upper switch in an


inverter leg is on and the inverter terminal voltage (VAN,
VBN, or VCN) is positive (+Vd) while
– Switching state ‘0’ indicates that the inverter terminal
voltage is zero due to the conduction of the lower switch.
• There are eight possible combinations of switching states in the
Table: Definition of Switching States
two-level inverter as listed in Table below (next slide).
– The switching state [1 0 0], for example, corresponds to
the conduction of S1, S6, and S2 in the inverter legs A, B,
and C, respectively.
• Among the eight switching states, [1 1 1] and [O O O] are zero 1
states and the others are active states. 0

20
II. PWM METHODS
C. Space Vector PWM

 Output voltages of three-phase inverter (4)


 The eight combinations, phase voltages and output line to line voltages

VAN= VA0+ V0N ………………… (1)


VBN= VB0+ V0N ………………… (2) S1 S3 S5 VA0 VB0 VC0
VCN= VC0+ V0N ……………(3)
By adding the three equations

VAN +VBN +VCN = VA0+ VA0 + VA0 +3V0N


For balanced loads,
VA0 (t) + VA0(t) +VA0 (t) =0
Hence,
V0N =( VAN +VBN +VCN)/3
From (1),
VA0 = VAN - V0N =(2 VAN – VBN - VCN )/3

Similarly, we can calculate VB0 and VC0


21
II. PWM METHODS
C. Space Vector PWM
 Principle of Space Vector PWM
 Treats the sinusoidal voltage as a constant amplitude vector rotating at constant
frequency
 This PWM technique approximates the reference voltage Vref by a combination

of the eight switching patterns (V0 to V7)


 Coordinate Transformation (ABC reference frame to the stationary α-β frame): A
three-phase voltage vector is transformed into a vector in the stationary α-β
coordinate frame which represents the spatial vector sum of the three-phase
voltage
 The vectors (V1 to V6) divide the plane into six sectors (each sector: 60 degrees).
 Vref is generated by two adjacent non-zero vectors and two zero vectors
22
Cont…
Table : Space Vectors, Switching States, and On-State Switches

010 110

[0 0 0]
[1 1 1]
[1 0 0]

[1 1 0] 011 100

[0 1 0]

[0 1 1] 001 101

[0 0 1] Fig: Space vector diagram for the two-level


inverter.
[1 0 1]

The active and zero switching states can be represented by active and zero space
vectors, respectively. A typical space vector diagram for the two-level inverter is
shown in Fig. right, where the six active vectors V1 to V6 form a regular hexagon with
six equal sectors (I to VI). The zero vector V0 lies on the center of the hexagon
Each sector (1 to 6): 60 degrees
09/30/2020
Cont…
• To derive the relationship between the
space vectors and switching states, refer to
the two-level inverter in Figure 13.
• Assuming that the operation of the
inverter is three-phase balanced, we have
vAO(t) + vBO(t) + vCO(t) = 0
• where vAO, vBO, and vCO are the
instantaneous load phase voltages.
• From mathematical point of view, one of
the phase voltages is redundant since
given any two phase voltages, the third
one can be readily calculated ( x+y+z=0 is
equation of a plane). Figure 13 : Simplified two-level inverter for high-power applications

• Therefore, it is possible to transform the


three-phase variables to equivalent two-
phase variables

09/30/2020
Transformations of the Space Vectors

 
= –

= –)

(b)
(a)
Cont…
Space Vectors: Transformation

The coefficient 2/3 is somewhat arbitrarily chosen. The commonly used value is 2/3
or 2 / 3 . The main advantage of using 2/3 is that the magnitude of the two-phase
voltages will be equal to that of the three-phase voltages after the transformation.
A space vector can be generally expressed in terms of the two-phase voltages in the α-β plane

substituting the above matrix equations, we have

where ejx = cosx + jsinx and x = 0, 2π/3 or 4 π /3.


For active switching state [1 0 0],
The corresponding space vector, denoted as V1, can be obtained by substituting this value
in the equation
Following the same procedure, all six active vectors can be
derived
09/30/2020
Cont…
Table: Space Vectors, Switching States, and On-State Switches

[0 0 0]
[1 1 1]

[1 0 0]

[1 1 0]

[0 1 0]

[0 1 1]

[0 0 1]

[1 0 1]

The zero vector has two switching states [PPP] and [OOO], one of which seems redundant. As will be seen later, the
redundant switching state can be utilized to minimize the switching frequency of the inverter or perform other useful
functions.
Note that the zero and active vectors do not move in space, and thus they are referred to as stationary vectors.
09/30/2020
Cont…
• the reference vector Vref in Fig. shown in slide #23 rotates in space at an angular velocity

• where f1 is the fundamental frequency of the inverter output voltage.

• The angular displacement between Vref and the α-axis of the plane can be obtained by

• For a given magnitude (length) and position, Vref can be synthesized by three nearby
stationary vectors, based on which the switching states of the inverter can be selected and
gate signals for the active switches can be generated.
• When Vref passes through sectors one by one, different sets of switches will be turned on
or off.
• As a result, when Vref rotates one revolution in space, the inverter output voltage varies
one cycle over time.
 The inverter output frequency corresponds to the rotating speed of Vref, while its output
09/30/2020 voltage can be adjusted by the magnitude of Vref.
II. PWM METHODS
C. Space Vector PWM (9)

 Realization of Space Vector PWM

 Step 1. Determine Vα, Vβ, Vref, and angle (θ)

 Step 2. Determine time duration Ta, Tb, T0

 Step 3. Determine the switching time of each transistor (S 1 to S6)

26
II. PWM METHODS
C. Space Vector PWM (10)

 Step 1. Determine Vα, Vβ, Vref, and angle (θ) V  VA0  VB0  cos60  VC0  cos60
1 1
 VA0  VB0  VC0
 Coordinate transformation 2 2
: VA0 VB0VC0 to αβ
V  0  VB0  cos30  VC0  cos30
3 3
 VB0  VC0
β axis 2 2
 1 1  V
 A0 
Vβ 1  
VB0 V  2  2 
2  
     VB0 
V
  3  3 3
0  VC0 
 2 2   
θ
Vα VA0 α axis V ref  VdV2 α2 +V
Vqβ22
Vq
α   tan 1 ( )ω
Θs t= 2πf s t
Vd
VC0
(where, f s  fundamental frequency)
Fig. 13 Voltage Space Vector and its components in (α, β).

27
Step 2 : Dwell Time Calculation
• As mentioned earlier, the reference Vref can be synthesized by three stationary vectors.
• The dwell time for the stationary vectors essentially represents the duty-cycle time (on-state or off-
state time) of the chosen switches during a sampling period Ts of the modulation scheme.
• The dwell time calculation is based on ‘volt-second balancing’ principle, that is, the product of the
reference voltage Vref and sampling period Ts equals the sum of the voltage multiplied by the time
interval of chosen space vectors.
• Assuming that the sampling period Ts is sufficiently small, the reference vector Vref can be
considered constant during Ts. Under this assumption, Vref can be approximated by two adjacent
active vectors and one zero vector

where Ta, Tb, and T0 are the dwell times for the vectors V1, V2 and V0, respectively.
The space vectors in the above equation can be expressed as

Substituting this into the above equation and then splitting the resultant equation into
the real (α-axis) and imaginary (β -axis) components in the αβ – plane, we have
Solving the above equation together
with Ts = Ta + Tb + T0 yields
Fig:

09/30/2020 For ‘k’ sector


II. PWM METHODS
C. Space Vector PWM (13)

 Step 2. Determine time duration Ta, Tb, T0 (3)


 Switching time duration at any Sector

3  Ts  V ref    n 1  
 Ta   sin       
Vdc   3 3 
3  Ts  V ref  n 
  sin    
Vdc  3 
3  Ts  V ref  n n 
  sin  cos  cos  sin  
Vdc  3 3 

3  Ts  V ref   n 1  
 Tb   sin    
Vdc   3  
3  Ts V ref  n 1 n 1 
   cos  sin   sin   cos 
Vdc  3 3 

 where, n  1 through 6 (that is, Sector1 to 6) 


 T0  Ts  Ta  Tb ,  
 0  α  60 
Space Vector Modulation Index
The dwell time equation can be rewritten if
a the modulation index is defined like

The maximum magnitude of the reference vector, Vref,max,


110
corresponds to the radius of the largest circle that can be
010 inscribed within the hexagon shown in Fig. Since the hexagon is
formed by six active vectors having a length of 2Vd/3, Vref,max
can be found from

011 100 If this substituted in the index equation, gives the maximum
modulation index: ma,max = 1
from which the modulation index for the SVM scheme is in the
range of
001 101

Figure : Space vector diagram for the two-level inverter.

09/30/2020
II. PWM METHODS
C. Space Vector PWM (7)

 Comparison of Sine PWM and Space Vector PWM (1)

Fig. 12 Locus comparison of maximum linear control voltage


in Sine PWM and SV PWM.

24
II. PWM METHODS
C. Space Vector PWM (8)

 Comparison of Sine PWM and Space Vector PWM (2)

 Space Vector PWM generates less harmonic distortion


in the output voltage or currents in comparison with sine PWM

 Space Vector PWM provides more efficient use of supply voltage


in comparison with sine PWM

 Sine PWM
: Locus of the reference vector is the inside of a circle with radius of 1/2 V dc

 Space Vector PWM


: Locus of the reference vector is the inside of a circle with radius of 1/3 Vdc

 Voltage Utilization: Space Vector PWM = 2/3 times of Sine PWM

25
Step 3. Determine Switching Sequence
• In general, the switching sequence design for a given Vref is not unique, but
it should satisfy the following two requirements for the minimization of the
device switching frequency:
(a) The transition from one switching state to the next involves only two
switches in the same inverter leg, one being switched on and the other
switched off.
(b) The transition for moving from one sector in the space vector
diagram to the next requires no or minimum number of switching.
• Figure in the next slide shows a typical seven-segment switching sequence
and inverter output voltage waveforms for in sector I, where is
synthesized by .
• The sampling period Ts is divided into seven segments for the selected
vectors

09/30/2020 Advanced Power Electronics & Applications ECEG-6309 36


Cont…

000 100 110 111 110 100 000

Figure : Seven-segment switching sequence for Vref in sector I.


09/30/2020 Advanced Power Electronics & Applications ECEG-6309 37
Switching Sequence (Comparison)

P stands for 1,
and O stands
for 0.

Let us now examine a case given in these Fig. where the vectors V1 and
Seven-segment switching sequence for Vref in sector I. V2 in lower Fig. are swapped. Some switching state transitions, such as
the transition from [OOO] to [PPO], are accomplished by turning on and off
four switches in two inverter legs simultaneously. As a consequence, the
total number of switching's during the sampling period increases from six in
the previous case to ten. Obviously, this switching sequence does not
satisfy the design requirement and thus should not be adopted.
It is interesting to note that the waveforms of VAB in Figs(shown in the left).
produced by two different switching sequences seem different, but they are
essentially the same. If these two waveforms are drawn for two or more
consecutive sampling periods, we will notice that they are identical except
for a small time delay (Ts/2). Since Ts is much shorter than the period of the
inverter fundamental frequency, the effect caused by the time delay is
negligible.
Table shown above gives the seven-segment switching sequences for Vref
residing in all six sectors. Note that all the switching sequences start and
end with switching state [OOO], which indicates that the transition for Vref
Undesirable seven-segment switching sequence moving from one sector to the next does not require any switching. The
09/30/2020 switching sequence design requirement (b) is satisfied.
Spectrum Analysis (SVM)

The simulated waveforms for the


inverter output voltages and load
current are shown in this Fig. . The
inverter operates under the condition
of f1 = 60 Hz, Ts = 1/720 s, fsw = 720
Hz, and ma = 0.8 with a rated three-
phase inductive load. The load power
factor is 0.9 per phase. It can be
observed that the wave form of the in-
shown inverter line-to-line voltage VAB
is not half-wave symmetrical, that
is,

Therefore, it contains even-order


harmonics, such as 2nd, 4th, 8th, and
10th, in addition to odd-order
harmonics
Figure : Inverter output waveforms produced by SVM scheme with f1 = 60 Hz,
fsw = 720 Hz, and ma = 0.8.
09/30/2020 Advanced Power Electronics & Applications ECEG-6309 39
Cont…
NB: Although the low-order harmonics, such as 2nd, 4th,
5th, and 7th, are not eliminated, they have very low
magnitudes. The maximum fundamental line-to-line voltage
(rms) occurs at ma = 1 and can be found from

(a) Even-order harmonics

which is around 15.5% higher than that given in for the


SPWM scheme.

(b) Odd-order harmonics


09/30/2020 Advanced Power Electronics & Applications ECEG-6309 40
Even-Order Harmonic
Elimination
• In the inverter-fed medium-voltage drives, these harmonics may not have a significant impact on the
operation of the motor. However, when the converter used as a rectifier, its line current THD should comply
with harmonic standards such as IEEE 519-1992. standards about rectifier have more stringent
requirements on even-order harmonics than on odd-order ones.

when Vref falls into sector IV

180 degree
even-order harmonics

None even-order harmonics

09/30/2020 41
Cont…
• Examining the two switching sequences in Fig. in the above slide, we can find
out that the type-A sequence starts and ends with [OOO] while the type-B
sequence commences and finishes with [PPP].
• The waveforms of VAB generated by both sequences seem different. However,
they are essentially the same except for a small time delay (Ts/2), which can
be clearly observed if these two waveforms are drawn for two or more
consecutive sampling periods.
• To make the three-phase line-to-line voltage half-wave symmetrical, type-A
and type-B switching sequences can be alternatively used.
• In addition, each sector in the space vector diagram is divided into two
regions as shown in Fig. in the following slide.
• Type-A sequence is used in the non-shaded regions, while type-B sequence
is employed in the shaded regions.
• The detailed switching sequence arrangements as well are given in Table

09/30/2020 Advanced Power Electronics & Applications ECEG-6309 42


Cont… Table : Switching Sequence of the Modified SVM for Even-
Order Harmonic Elimination

Figure : Alternative use of two switching sequences for


even-order harmonic elimination.

It can be observed from the table that the


transition for Vref moving from region a to b
causes additional switchings.
This implies that the even-order harmonic
elimination is achieved at the expense of an
increase in switching frequency.
The amount of switching frequency increase can
be determined by

where f1 is the fundamental frequency of the


09/30/2020inverter output voltage.
Cont…
• The inverter output waveforms measured from a laboratory two-level inverter with modified
SVM scheme are shown in Fig. below.
• The inverter operates under the condition of f1 = 60 Hz, Ts = 1/720 s, and ma = 0.8. The
waveforms of the inverter line-to-line voltage VAB and load phase voltage VAO are of half-wave
symmetry.

(b) Spectrum (500 Hz/div)

Compared with the harmonic spectrum given in Fig. unmodified SVM, the
magnitude of the 5th and 7th harmonics in VAB is increased while the THD
essentially remains the same.

09/30/2020 Advanced Power Electronics & Applications ECEG-6309 44

Вам также может понравиться