Mansoura University

Faculty of Engineering
Electrical Engineering Department

Power
Electronics (2)
Course Code: ELE 221

Assoc. Prof. Abdelfattah Eladl

1
 Chapter 4
Inverters
Converting DC to AC
V(t) AC output voltage
DC input voltage
Inverter
+
i
V
-
0 t

2
 Fourier Series Analysis

In the case of the square wave, the Fourier series contains

the odd harmonics and can be represented as:
∞
𝟒𝑽𝒅𝒄 𝟒𝑽𝒅𝒄
𝒗𝒐 𝒕 = ෍ 𝒔𝒊𝒏 𝐧 𝝎𝒐 𝐭 𝑽𝒏 =
𝐧𝝅 𝐧𝝅
𝐧=𝒐𝒅𝒅

𝟒𝑽𝒅𝒄 𝟏 𝟏
𝒗𝒐 𝒕 = 𝒔𝒊𝒏 𝝎𝒐 𝐭 + 𝒔𝒊𝒏 𝟑 𝝎𝒐 𝐭 + 𝒔𝒊𝒏 𝟓 𝝎𝒐 𝐭 + ⋯
𝝅 𝟑 𝟓
3
4
 𝟒𝑽𝒅𝒄
𝒗𝒐𝟏 = 𝒔𝒊𝒏 𝝎𝒐 𝐭 𝟒𝑽𝒅𝒄
𝝅 𝒗𝒐𝟑 = 𝒔𝒊𝒏 𝟑 𝝎𝒐 𝐭
𝟑𝝅
𝟒𝑽𝒅𝒄
𝒗𝒐𝟓 = 𝒔𝒊𝒏 𝟓 𝝎𝒐 𝐭
𝟓𝝅

𝟒𝑽𝒅𝒄
𝒗𝒐𝒏 = 𝒔𝒊𝒏 𝐧 𝝎𝒐 𝐭
𝐧𝝅
∞ 𝑽𝒏
𝑰𝒏 =
𝒊𝒐 𝒕 = ෍ 𝑰𝒏 𝒔𝒊𝒏 𝐧𝝎𝒐 𝐭 − 𝝓𝒏 𝒁𝒏
𝐧=𝒐𝒅𝒅

𝒏𝝎𝒐 𝑳
𝒁𝒏 = 𝑹𝟐 + 𝒏𝝎𝒐 𝑳 𝟐 𝝓𝒏 = tan−𝟏
𝑹

5
 Power absorbed by a load with a series resistance is
determined from Irms2 R, where the rms current can be
determined from the rms currents at each of the
components in the Fourier series by
∞ ∞ 2
𝐼𝑛 𝑉𝑛
𝐼rms = ෍ 𝐼2 𝑛,𝑟𝑚𝑠 = ෍ 𝐼𝑛 =
2 𝑍𝑛
𝑛=1 𝑛=1

and Zn is the load impedance at harmonic n.

power absorbed in the load resistor
∞ ∞

𝐏 = ෍ 𝑷𝒏 = ෍ 𝑰𝟐 𝐧,𝒓𝒎𝒔 𝐑
𝐧=𝟏 𝐧=𝟏

6
 Example
The H-bridge inverter has a switching sequence that
produces a square wave voltage across a series RL load.
For the inverter
(Vdc = 100 V, R =10 Ω, L = 25 mH, f = 60 Hz).

Determine (a) an expression for load current, (b) the

power absorbed by the load, and (c) the average current
in the dc source.
(d) the amplitudes of the Fourier series terms for the
square wave load voltage, the amplitudes of the Fourier
series terms for load current, and the power absorbed by
the load
7
■ Solution
(a) 𝑇 = 1Τ𝑓 = 1Τ60 = 0.0167 s
𝜏 = 𝐿Τ𝑅 = 25 × 10−3 Τ10 = 0.0025 s
𝑇Τ2𝜏 = 3.33

The maximum and minimum current.

𝑇Τ2𝜏
𝑉𝑑𝑐 1 − 𝑒 −
𝐼𝑚𝑎𝑥 = −𝐼𝑚𝑖𝑛 = 𝑇Τ2𝜏
𝑅 1 + 𝑒−

𝟏𝟎𝟎 𝟏 − 𝒆− 𝟑.𝟑𝟑
𝑰𝒎𝒂𝒙 = −𝑰𝒎𝒊𝒏 = = 𝟗. 𝟑𝟏 𝑨
𝟏𝟎 𝟏 + 𝒆− 𝟑.𝟑𝟑

8
 100 100
𝑖𝑜 𝑡 = + −9.31 − 𝑒 −𝑡Τ0.0025
10 10
−𝒕Τ𝟎.𝟎𝟎𝟐𝟓
𝟏
= 𝟏𝟎 − 𝟏𝟗. 𝟑𝟏 𝒆 𝒇𝒐𝒓 𝟎 ≤ 𝒕 ≤
𝟏𝟐𝟎
100 100
𝑖𝑜 𝑡 = − + 9.31 + 𝑒 −(𝑡−0.0167Τ2ሻΤ0.0025
10 10
−(𝒕−𝟎.𝟎𝟎𝟖𝟑𝟓ሻΤ𝟎.𝟎𝟎𝟐𝟓
𝟏 𝟏
= −𝟏𝟎 + 𝟏𝟗. 𝟑𝟏 𝒆 𝒇𝒐𝒓 ≤𝒕≤
𝟏𝟐𝟎 𝟔𝟎

(b) Power is computed from I2rms R, where Irms is

𝟏Τ𝟏𝟐𝟎

𝑰𝒓𝒎𝒔 = (𝟏𝟐𝟎ሻ න 𝟏𝟎 − 𝟏𝟗. 𝟑𝟏 𝒆−𝒕Τ𝟎.𝟎𝟎𝟐𝟓 𝟐 𝒅𝒕 = 𝟔. 𝟔𝟒 𝑨

𝟎
9
 Power absorbed by the load is

𝑷 = 𝑰𝒓𝒎𝒔 𝟐 𝑹 = 𝟔. 𝟔𝟒 𝟐 × 𝟏𝟎 = 𝟒𝟒𝟏 𝑾

(c) Average source current can also be computed by

𝑷𝒅𝒄 𝟒𝟒𝟏
𝑰𝒔 = = = 𝟒. 𝟒𝟏 𝑨
𝑽𝒅𝒄 𝟏𝟎𝟎

(d) Fourier analysis

The load voltage is represented as the Fourier series in Eq. (3-16).
The amplitude of each voltage term is
𝟒𝑽𝒅𝒄 𝟒 × 𝟏𝟎𝟎
𝑽𝒏 = =
𝒏𝝅 𝒏𝝅
10
The amplitude of each current term is determined from

𝑉𝑛 𝑉𝑛 400Τ𝑛 𝜋
𝐼𝑛 = = =
𝑍𝑛 𝑅2 + 𝑛𝜔𝑜 𝐿 2 102 + 𝑛 × 2𝜋 × 60 × 0.025 2

Power at each frequency is determined from

2
𝐼𝑛
𝑃𝑛 = 𝐼2 𝑛,𝑟𝑚𝑠 𝑅= 𝑅
2
Table 3-1 Fourier Series Quantities for Example 3-2
n fn (Hz) Vn (V) Zn (Ω) In (A) Pn (W)
1 60 127.3 13.7 9.27 429.3
3 180 42.4 30.0 1.42 10.0
5 300 25.5 48.2 0.53 1.40
7 420 18.2 66.7 0.27 0.37
9 540 14.1 85.4 0.17 0.14
11
 As the harmonic number n increases, the amplitude of
the Fourier voltage component decreases and the
magnitude of the corresponding impedance increases,
both resulting in small currents for higher-order
harmonics. Therefore, only the first few terms of the
series are of practical interest.

Power absorbed by the load is computed from

𝑷 = ෍ 𝑷𝒏 = 𝟒𝟐𝟗. 𝟑 + 𝟏𝟎. 𝟎 + 𝟏. 𝟒𝟎 + 𝟎. 𝟑𝟕 + 𝟎. 𝟏𝟒 + ⋯ ≈ 𝟒𝟒𝟏 𝐖

12
DC Side Current Ripple

The switching frequency is assumed to be very high,

approaching infinity. Therefore, to filter out the high-
switching-frequency components in vo and id, the filter
components L and C required in both ac- and de-side
filters approach zero.
13
This implies that the energy stored in the filters is
negligible. Since the converter itself has no energy storage
elements, the instantaneous power input must equal the
instantaneous power output.
Having made these assumptions, vo is a pure sine wave at
the fundamental output frequency ω1,
𝒗𝒐 = 𝑽𝒐𝟏 = 𝟐 𝑽𝒐 𝒔𝒊𝒏 𝝎𝟏 𝒕
If the load is inductive as shown in Fig., then the output
current would also be sinusoidal and would lag vo for an
inductive load such as an ac motor:
𝑰𝒐 = 𝟐 𝑰𝒐 𝒔𝒊𝒏 𝝎𝟏 𝒕 − 𝝓
where ϕ is the angle by which io lags vo .
14
Assuming that no energy is stored in the filters,

𝑷𝒊𝒏𝒑𝒖𝒕 = 𝑷𝒐𝒖𝒕𝒑𝒖𝒕
𝑉𝑑 𝑖𝑑∗ = 𝑣𝑜 𝑡 𝑖𝑜 𝑡 = 2 𝑉𝑜 sin 𝜔1 𝑡 × 2 𝐼𝑜 sin 𝜔1 𝑡 − 𝜙

But 1
sin 𝐴 × sin 𝐵 = cos 𝐴 − 𝐵 − cos 𝐴 + 𝐵
2

𝑽𝒐 𝑰𝒐 𝑽𝒐 𝑰𝒐
∴ 𝒊𝒅∗ (𝒕ሻ = 𝐜𝐨𝐬 𝝓 − 𝐜𝐨𝐬 𝟐𝝎𝟏 𝒕 − 𝝓
𝑽𝒅 𝑽𝒅

𝑖𝑑∗ 𝑡 = 𝐼𝑑 + 𝑖𝑑2
15
 𝑉𝑜 𝐼𝑜 𝑉𝑜 𝐼𝑜
Where 𝐼𝑑 = cos 𝜙 𝑖𝑑2 = cos 2𝜔1 𝑡 − 𝜙
𝑉𝑑
𝑉𝑑
𝒊∗𝒅 consists of a dc component ld , which is responsible for
the power transfer from Vd on the DC side of the
inverter to the AC side. Also, contains a sinusoidal
component at twice the fundamental frequency.

16
 Performance Parameters
1- Total Harmonic Distortion (THD)
The objective of the inverter is to use a dc voltage source to
supply a load requiring ac, it is useful to describe the quality of
the ac output voltage or current.

The quality of a nonsinusoidal wave can be expressed in terms

of total harmonic distortion (THD).

The THD is a measure of closeness in shape between a

waveform and its fundamental component, is defined as:

𝟐
σ∞
𝐧=𝟐 𝑽𝐧,𝒓𝒎𝒔 𝑽𝟐 𝒓𝒎𝒔 − 𝑽𝟐 𝟏,𝒓𝒎𝒔
𝐓𝐇𝐃 = =
𝑽𝟏,𝒓𝒎𝒔 𝑽𝟏,𝒓𝒎𝒔

17
 The THD of current is determined by substituting current
for voltage in the previous equation.

The THD of load current is often of greater interest than

that of output voltage.

The rms value of the square wave voltage is the same as

the peak value,
𝑽𝒓𝒎𝒔 = 𝑽𝒅𝒄

The fundamental frequency component is the first term

𝑽𝟏 𝟒 𝑽𝒅𝒄
𝑽𝟏,𝒓𝒎𝒔 = =
𝟐 𝟐𝝅
18
The total harmonic distortion for voltage in square wave inverters

2 2
𝑉 2 𝑟𝑚𝑠 − 𝑉 21,𝑟𝑚𝑠 − 4 𝑉𝑑𝑐 Τ 2 𝜋
𝑉𝑑𝑐
𝑇𝐻𝐷𝑣 = =
𝑉1,𝑟𝑚𝑠 4 𝑉𝑑𝑐 Τ 2 𝜋
= 0.483 = 48.3 %

The THD of the current is computed using the truncated Fourier

series which was determined
2
σ∞
𝑛=2 𝐼𝑛,𝑟𝑚𝑠
𝑇𝐻𝐷𝑖 =
𝐼1,𝑟𝑚𝑠
2 2 2 2
1.42Τ 2 + 0.53Τ 2 + 0.27Τ 2 + 0.17Τ 2
≈
9.27Τ 2
= 0.167 = 16.7 %
19
2- Harmonic Factor (HF)
The harmonic factor (of n th harmonic), is a measure of individual
harmonic contribution, is defined as
Harmonic factor is the ratio of the RMS value of the individual
harmonics to the RMS value of the fundamental.
𝑽𝒏,𝒓𝒎𝒔
𝑯𝑭𝒏 =
𝑽𝟏,𝒓𝒎𝒔

3- Distortion Factor ( DF )
The THD gives the total harmonic content, but it does not
indicate the level of each harmonic component. If a filter is
used at the output of inverters, the higher-order harmonics would
be attenuated more effectively. Therefore, knowledge of both the
frequency and magnitude of harmonic is important.
20
The distortion factor indicates the amount of harmonic
distortion that remains in a particular waveform after
the harmonics of that waveform have been subjected to
a second-order attenuation (i.e., divided by n2).

Thus DF is a measure of effectiveness in reducing

unwanted harmonics without having to specify the values
of a second-order load filter and is defined as:

∞ 𝟐
𝟏 𝑽𝐧,𝒓𝒎𝒔
𝐃𝐅 = ෍
𝑽𝟏,𝒓𝒎𝒔 𝒏𝟐
𝒏=𝟐,𝟑,..

21
The distortion factor of an individual (or nth) harmonics
component is defined as:
𝑽𝒏,𝒓𝒎𝒔
𝑫𝑭𝒏 = 𝟐
𝒏 𝑽𝟏,𝒓𝒎𝒔

4- Lowest-order Harmonic (LOH)

The LOH is that harmonic component whose

amplitude is closest to the fundamental one, and its is
amplitude is greater than or equal to 3% of the
fundamental component.

22
 Amplitude And Harmonic Control
The amplitude of the fundamental frequency for a 𝟒𝑽𝒅𝒄
𝑽𝟏 =
square wave output from of the full-bridge inverter is 𝝅
determined by the dc input voltage

A controlled output can be produced by modifying the switching scheme.

This output voltage can be controlled by adjusting the interval on
each side of the pulse where the output is zero.

where α is the angle of zero voltage on each end of the pulse.

23
24
 The rms value of the output voltage waveform is
𝝅−𝜶
𝟏
𝑽𝒓𝒎𝒔 = න 𝑽𝟐𝒅𝒄 𝒅 𝒘𝒕
𝝅
𝜶

𝟐𝜶
= 𝑽𝒅𝒄 𝟏−
𝝅

The Fourier series of the waveform is expressed as

∞

𝒗𝒐 𝒕 = ෍ 𝑽𝒏 𝐬𝐢𝐧 𝒏 𝝎𝒐 𝒕
𝒏=𝒐𝒅𝒅

𝟒𝑽𝒅𝒄
𝑽𝐧 = 𝐜𝐨𝐬(𝒏𝜶ሻ
𝒏𝝅
25
The amplitude of each frequency of the output is a function of α.
In particular, the amplitude of the fundamental frequency (n =1) is
controllable by adjusting α:
𝟒𝑽𝒅𝒄
𝑽𝟏 = 𝐜𝐨𝐬 𝜶
𝝅
Harmonic content can also be controlled
by adjusting α. If α =30, for example, V3 𝟒𝑽𝒅𝒄
= 0. This is significant because the third 𝑽𝟑 = 𝐜𝐨𝐬(𝟑𝜶ሻ
𝟑𝝅
harmonic can be eliminated from the
output voltage and current.

Other harmonics can be eliminated by choosing a value of α

which makes the cosine term in Eq. (3-20) to go to zero. Harmonic n
is eliminated if
𝟗𝟎
𝜶=
𝒏
26
Amplitude control and harmonic reduction may not be
compatible.

For example, establishing α at 30o to eliminate the third

harmonic fixes the amplitude of the output fundamental
frequency at V1 = (4Vdc/π) cos 30o = 1.1Vdc and removes
further controllability.

To control both amplitude and harmonics using this

switching scheme, it is necessary to be able to control
the dc input voltage to the inverter. A dc-dc converter
(Chaps. 2&3) placed between the dc source and the
inverter can provide a controlled dc input to the inverter.
27
The Half-bridge Inverter
The half-bridge converter can be used as an inverter
In this circuit, the number of switches is reduced to 2 by dividing the
dc source voltage into two parts with the capacitors

Each capacitor will be the same value

and the voltage Vdc/2 across it.
When S1 is closed, the load voltage
is -Vdc/2.
When S2 is closed, the load voltage
is +Vdc/2.

The voltage across an open switch is

twice the load voltage, or Vdc. By KV
Law
28
 The voltage across an open switch is twice the load
voltage, or Vdc. As with the full-bridge inverter, blanking
time for the switches is required to prevent a short circuit
across the source, and feedback diodes are required to
provide continuity of current for inductive loads.

29
Fourier Series of the output voltage

ao
vo = +  ( an cos( nt ) + bn sin(nt ) )
2 n =1

ao , an = 0

1  −Vs 
0
Vs
bn =   sin( nt )d (t ) +  sin( nt )d (t ) 
  − 2 0
2 
2Vs
bn = → n = 1, 3, 5,...
n

2Vs
vo =  sin( nt )
n =1,3,5,.. n

30
 Multilevel Inverters
The H (full) bridge inverter
previously illustrated produces
output voltages of Vdc, 0, and –Vdc.
These multilevel-output
voltages are more sinelike in
quality and thus reduce
harmonic content.
The multilevel inverter is
suitable for applications
including adjustable-speed
motor drives and interfacing
renewable energy sources
such as photovoltaics to the
electric power grid.
31
Multilevel Converters with Independent DC Sources
The total instantaneous voltage vo on the output of the
multilevel converter is any combination of individual bridge
voltages. Thus, for a two-source inverter, vo can be any of
the five levels +2Vdc, Vdc, 0, -Vdc, or -2Vdc.
Each bridge operates at a different delay angle α, resulting
in bridge and total output voltages like the following

32
Voltage output of each of the H bridges and the total voltage for the
two-source multilevel inverter

33
 The Fourier series for the total output voltage vo for the two-source
circuit contains only the odd-numbered harmonics and is
∞
𝟒𝑽𝒅𝒄 𝐬𝐢𝐧 𝒏 𝝎𝒐 𝒕
𝒗𝒐 𝒕 = ෍ 𝐜𝐨𝐬 𝒏 𝜶𝟏 + 𝐜𝐨𝐬 𝒏 𝜶𝟐
𝝅 𝒏
𝒏=𝟏,𝟑,𝟓,..

The Fourier coefficients for this series are

𝟒𝑽𝒅𝒄
𝑽𝒏 = 𝐜𝐨𝐬 𝒏 𝜶𝟏 + 𝐜𝐨𝐬 𝒏 𝜶𝟐
𝒏𝝅
The modulation index Mi is the ratio of the amplitude of the
fundamental frequency component of vo to the amplitude of the
fundamental frequency component of a square wave of amplitude
2Vdc, which is 2 (4Vdc/ π). for the same wave form of the 2 converter
𝑽𝟏 𝐜𝐨𝐬 𝜶𝟏 + 𝐜𝐨𝐬 𝜶𝟐
𝑴𝒊 = =
Τ
𝟐 𝟒𝑽𝒅𝒄 𝝅 𝟐
34
Some harmonics can be eliminated from the output voltage
waveform with the proper selection of α1 and α2 . For the two-
source converter, harmonic n can be eliminated by using delay
angles such that

𝒄𝒐𝒔 𝒏 𝜶𝟏 + 𝐜𝐨𝐬 𝒏 𝜶𝟐 = 𝟎

To eliminate the nth harmonic and also meet a specified

modulation index for the two-source inverter requires the
simultaneous solution

𝒄𝒐𝒔 𝜶𝟏 + 𝐜𝐨𝐬 𝜶𝟐 = 𝟐 𝑴𝒊

35
Example
For the two-source multilevel inverter with Vdc =100 V: (a)
Determine the Fourier coefficients through n =9 and the modulation
index for α1 =20o and α2 =40 o. (b) Determine α1 and α2 such that
the third harmonic (n =3) is eliminated and Mi =0.8.
■ Solution
(a) To evaluate the Fourier coefficients,
4𝑉𝑑𝑐
𝑉𝑛 = cos 𝑛 𝛼1 + cos 𝑛 𝛼2
𝑛𝜋
4 × 100
= cos 𝑛20𝑜 + cos 𝑛40𝑜
𝑛𝜋
resulting in V1 =217, V3 =0, V5 =28.4, V7 =10.8, and V9 = 0. Note
that the third and ninth harmonics are eliminated. The even
harmonics are not present.
36
 The modulation index Mi is evaluated from Eq. (3-25).
cos 𝛼1 + cos 𝛼2 cos 20𝑜 + cos 40𝑜
𝑀𝑖 = = = 0.853
2 2
The amplitude of the fundamental frequency voltage is therefore
85.3 % of that of a square wave of ±255 V.

(b) To achieve simultaneous elimination of the third harmonic

and a modulation index of Mi = 0.8 requires the solution to the
equations
cos 3𝛼1 + cos 3𝛼2 = 0
and
cos 𝛼1 + cos 𝛼2 = 2𝑀𝑖 = 1.6
Using an iterative method, α1 = 7.6o and α2 = 52.4o.
37
The preceding concept can
be extended to a multilevel
converter having several dc
sources.

38
Equalizing Average Source Power with Pattern Swapping

In the two-source inverter

the source and H bridge producing the voltage v1 supplies

more average power (and energy) than the source and H
bridge producing v2 due to longer pulse widths in both the
positive and negative half cycles.
If the dc sources are batteries, one battery will discharge
faster than the other. A technique known as pattern
swapping or duty swapping equalizes the average power
supplied by each dc source.
39
The principle of pattern swapping is to have each dc
source conduct for an equal amount of time on
average.
An alternate switching scheme for the two source
circuit is used the first source conducts for a longer
time in the first half-cycle while the second source
conducts for more time in the second half-cycle.

Over one complete period, the sources conduct

equally, and average power from each source is
the same.

40
For the five-source converter , a switching scheme to
equalize average power is used. Note that five half
cycles are required to equalize power.

41
 Thank You
And
Any Question?

42