HARMONICS IN POWER SYSTEM

A harmonic is any voltage or current whose frequencies are

integral multiples of f. For example a set of sine waves whose
frequencies are 50, 150, 250, 450 Hz is said to possess the
following components:

Fundamental frequency 50 Hz (the lowest frequency)

Third harmonic: 150 Hz (3 x 50 Hz)
Fifth harmonic: 250 Hz (5 x 50 Hz)
Ninth harmonic: 450 Hz (9 x 50 Hz)

The distortion of a voltage or current can be traced to the

harmonics it contains. This distortion can be produced by
magnetic saturation in the core of transformers or by the
switching of thyristors or IGBTs in electronics drive.
1
 All periodic signals of frequency “f" can be represented in the
form of a composite sum:
1. of a sinusoidal term at frequency “f": the FUNDAMENTAL
(H1).
2. of sinusoidal terms of which frequencies are integer multiples
of fundamental H1: the HARMONICS (Hn).
3. of a possible continuous component (DC component)

y(t) = h1(t) + h3(t)

2
Harmonics :Order and Spectrum
Order:
The order of the harmonic is the
value of the integer which
determines its frequency.
Example: harmonic of order 5,
frequency = 250 Hz
(when fundamental f is 50 Hz)

Spectrum:
The spectrum of a signal is the
graph representing amplitudes
of the harmonics as a function
of their frequency.

3
To summarize: the harmonics are nothing less than the
components of a distorted waveform and their use allows us
to analyse any periodic non-sinusoidal waveform through
different sinusoidal waveform components.
Figure below shows a graphical representation of this
concept.

Non-sinusoidal waveform Third harmonic

4
First harmonic (fundamental) Fifth harmonic
 Two sinusoidal sources connected in series

U1max = 60 V @ 50 Hz U1 U2
U2max = 20 V @ 150 Hz
U3

fundamental and third harmonic

80
60
40
20 fundamental
0 third harmonic
-20 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 total
-40
-60
-80

5
 EXAMPLE OF THE HARMONICS
harmonics analyzis

25

20
15

10 Basic harmonic
5 3th harmonic
Ampers

5th harmonic
0 7th harmonic
9th harmonic
-5 11th harmonic
total curent
-10
-15
-20
-25

6
 fundamental + harmonics

25
20
15
10
5
0
-5 1 22 43 64 85 106 127 148 169 190 211 232 253 274 295 316 337 358
-10
-15
-20
-25

fundamental all harmonics total current

7
 How harmonics are generated?

Harmonics are generated by nonlinear loads. When we apply

a sinusoidal voltage to a load of this type, we shall obtain a
current with non-sinusoidal waveform.
i i

t u t u

t t
Linear load Nonlinear load 8
 Linear and non-linear loads
A load is said to be linear when there
is a linear relationship between
current and voltage. In simpler terms,
a linear load absorbs a sinusoidal
current when it is supplied by a
sinusoidal voltage: this current may
be displaced by an angle ϕ
compared with voltage.

When this linear relationship is not

verified, the load is termed non-
linear. It absorbs a nonsinusoidal
current and thus harmonic
currents, even when it is supplied
by a purely sinusoidal voltage current absorbed by
a non-linear load. 9
 GENERATING HARMONICS IN TRANSFORMER
MAGNETIZING CURRENT

The excitation current Ie is split into two components: the

magnetizing current I µ and IFe , proportional to the core power
losses. These currents are displaced from each other by an
angle Π /2. This displacement can be explained by means of
excitation current waveform. If the coil is supplied with
sinusoidal voltage the flux Φ must be sinusoidal too. Since
the magnetizing characteristic B-H is nonlinear, and has a
hysteresis loop, the current waveform obtained from
magnetizing curve is far from sinusoidal.

10
11
 Definition and characteristic quantities related to harmonics

Joseph FOURIER proved that all non-sinusoidal periodic

functions can be represented by a sum of sinusoidal terms, the
first one of which, at the recurrence frequency of the function,
is said to be fundamental, and the others, at multiple
frequencies of the fundamental, are said to be harmonic. A DC
component may complete these purely sinusoidal terms.
n=
Fourier's formula: y (t) = Yo + Σ Yn √2 sin (nω t –
n=1
ϕ n)
where:
- Yo: DC component value, generally nil and would not be
considered
- Yn: rms value of the nth harmonic component,
- ω : angular frequency of the fundamental, 12
th
 Harmonics: Effective (RMS - Root Mean Square) Value

The effective value of a non-sinusoidal periodic value is equal to:

T n=
1
Yrms =
T
2
y (t)dt = Σn=1 Y2n
0

Effective value = 2 2 2 2 2
Y1 + Y2 + Y3 + Y4 +…..+Y n

Y1 = fundamental component; Y2,..,Yn = harmonic components.

Calculation of effective current absorbed by single-phase load:

I fund. = 56.2A ; Ih3 = 27.2A ; Ih5 = 2.7A ; Ih7 = 9.2A ; Ih9 = 7.8A

2 2 2 2 2
Irms = 56.2 + 27.2 + 2.7 + 9.2 + 7.8 = 63.6 A 13
 Total harmonic distortion

Total harmonic distortion is a parameter globally defining

distortion of the alternating quantity.
This is the ratio of the RMS
value of the harmonics over
the RMS value of the
fundamental:

2 2 2 2 2
Y 2 + Y3 + Y4 + Y5 +…..+Y n
THD = 100
Y1

There is another definition which replaces the fundamental

Y1 with the total rms value Yrms .
This definition is used by some measuring instruments. 14
 Distortion factor vs. THD

1
(distortion factor) =
2
1 + (THD)

15
 Individual harmonic ratio

This quantity represents the ratio of the value of an harmonic

over the value of the fundamental (Y1), according to the
standard definition or over the value of the alternating quantity
(Yrms ).

(Frequency) spectrum

Representation of harmonic amplitude as a function of their

order: harmonics value is normally expressed as a
percentage of the fundamental.

16
 Power factor (PF) and Displacement Power Factor (DPF)

It is important not to confuse these two terms when harmonics

are present, as they are equivalent only when currents and
voltages are completely sinusoidal.

The power factor (λ ) is the The displacement power

ratio between active power P factor (cos ϕ 1) relates to
and apparent power S: fundamental quantities, thus:
λ =P/ cos ϕ 1 = P1 / S1
S
In pure sinusoidal waveform: cos ϕ 1 = cos ϕ = λ

Distortion factor
The IEC 146-1-1 defines this factor as the ratio between the
power factor and the displacement power factor cos ϕ 1 :
17
ν = λ / cos ϕ 1
 Peak factor

The ratio of peak value over rms value of a periodic quantity.

Fc = Ypeak / Yrms

Some peak factor examples:

• Linear load Fc = 1.41
• IT load Fc = 2 to 2.5
• Micro computing load Fc = 2.2 to 3

18
The current drawn by non-linear loads passes through all of the
impedance between the system source and load. This current
produces harmonic voltages for each harmonic as it flows
through the system impedance. These harmonic voltages sum
and produce a distorted voltage when combined with the
fundamental. The voltage distortion magnitude is dependent on
the source impedance and the harmonic voltages produced.

19
A non linear load is effectively drawing current from the power
source at the fundamental frequency, and generating current
back at higher frequencies. This results in a distorted current
waveform as shown previous. Current harmonics disturb the
supply voltage and this also results in a distorted voltage at
the point of common coupling. Example: Consumer A and B
are fed from the same line. The non linear loads of consumer
A will distort the voltage of consumer B even if the latter has
only linear loads.
Point of common coupling

A B

system
impedance
20
 Voltage and current total harmonic distortion

A non-linear load generates harmonic voltage drops in the

circuits supplying it. In actual fact all upstream impedances
need to be taken into consideration right through to the
sinusoidal voltage source.
Consequently a load absorbing harmonic currents always has
a non-sinusoidal voltage at its terminals. This is characterized
by the voltage total harmonic distortion:

where Zn is the total source impedance at the frequency of

harmonic n, and In the rms value of harmonic n.
21
Output impedance of the various sources as a function of
frequency.

22
 Sources of Harmonics
There are many sources of power system harmonics. Some
examples of harmonic producing devices are:
Transformers: Third harmonic currents are present in the
magnetizing current (a small portion of the transformer full load
current). If the transformer saturates (due to over-voltage), the
harmonic distortion level of the current increases substantially.
Fluorescent Lamps: These devices produce a predominantly
third order harmonic current
on the order of 20% to 30% of the fundamental current.
Electronic ballasts have slightly
different characteristics but exhibit similar levels of harmonics.
Pulse-Width Modulated Converters: These devices use an
external controller for switching the input transistors allowing
the current waveform to be shaped more desirably. However,
these converters are limited in power and typically used in
23
applications less than a few hundred kilowatts.
Switched Mode Power Supplies: Typically found in single-
phase electronic devices such as computers and other
business and consumer electronics, these devices use a
switching regulator to precisely control the DC voltage. The
input of these power supplies normally consists of a full-wave
bridge rectifier and a DC filter capacitor which produces
an alternating pulse current waveform rich in third harmonic.
Though they are not used in large power applications, the
cumulative effects of many devices may create concerns,
particularly for 400/230 Volt Y systems.

24
Wave shape of current absorbed by some non-linear loads.

Light dimmer or heating regulator

H3 H5 H7 H9 H11 H13 H15 H17 H19
54 18 18 11 11 8 8 6 6

Switch mode power supply rectifier

H3 H5 H7 H9 H11 H13 H15 H17 H19
75 45 15 7 6 3 3 3 2

25
Three-phase rectifier with front end capacitor

H3 H5 H7 H9 H11 H13 H15 H17 H19

0 80 75 0 40 35 0 10 5

Three-phase rectifier with DC filtering reactor

H3 H5 H7 H9 H11 H13 H15 H17 H19
0 25 7 0 9 4 0 5 3

26
 Harmonic currents in three phase systems

Neutral conductor

Harmonics get more complicated in three phase applications.

Here not only do we have to deal with phase conductors, but
also the neutral conductor, triplen (odd multiples of 3 i.e. 3rd ,
9th , 15th etc,) harmonics, and sequence harmonics. The
triplen harmonics are the major cause of heat because they
add together in the neutral conductor. The magnitude of the
harmonic current produced by the triplens can approach twice
the phase current. This causes the neutral conductor to
overheat because neutral conductors were historically
designed with the same ampacity as the phase conductors.

27
 With, for example, an harmonic 3 of 75%, the current flowing
in the neutral is 2.25 times the fundamental. The current in
each phase is only SQR (1+ 0.752 ) = 1.25 times the
fundamental.
neutral current

150

100

50

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37
-50

-100

-150

Ia Ib Ic Ia 3rd Ib 3rd Ic 3rd neutral current

28
 Induction motor

A situation that produces abnormal amounts of heat in

motors is the combination of positive and negative
sequenced harmonics. The positive sequenced harmonics
are the fundamental, 7th , 13th , 19th , etc. They tend to apply
an additional forward force in the direction of the motor
rotation. The negative sequenced harmonics are the 5th ,
11th , 17th , etc. They present a force that opposes the motor
rotation and tries to make the motor rotate in the opposite
direction. The force of these harmonics acting upon each
other creates heat which leads to premature failure.
Harmonic voltage distortion causes increased eddy current
losses in the motors, in the same way as seen for
transformers. 29
 Transformers

The effects of harmonics inside the transformers involve

mainly three aspects:

• a) increase of iron losses (or no-load losses)

• b) increase of copper losses
• c) presence of harmonics circulating in the windings

a) The iron losses are due to the hysteresis phenomenon and

to the losses caused by eddy currents; the losses due to
hysteresis are proportional to the frequency, whereas the
losses due to eddy currents depend on the square
of the frequency.

30
b) The copper losses correspond to the power dissipated
by Joule effect in the transformer windings. As the
frequency rises (starting from 350 Hz) the current tends to
thicken on the surface of the conductors (skin effect); under
these
circumstances, the conductors offer a smaller cross section
to the current flow, since the losses by Joule effect
increase.

These two first aspects affect the overheating which

sometimes causes a de-rating of the transformer

31
c) The third aspect is relevant to the effects of the triple-N
harmonics (homopolar harmonics) on the transformer
windings. In case of delta windings, the harmonics flow
through the windings and do not propagate upstream
towards the network since they are all in phase; the delta
windings therefore represent a barrier for triple-N harmonics,
but it is necessary to pay particular attention to this type of
harmonic components for a correct dimensioning of the
transformer. The triplen harmonics are trapped and circulate
in the delta primary of the transformer. Since most loads
produce high levels of the 3rd harmonic (one of the triplens),
the harmonic content reflected back to the source is reduced.
The circulating harmonics in the primary of the transformer
creates heat because of their higher frequencies. For this
reason, a transformer that can handle the excess heat is
needed. This transformer is called a K-rated transformer
32
 Induction motor

A situation that produces abnormal amounts of heat in

motors is the combination of positive and negative
sequenced harmonics. The positive sequenced harmonics
are the fundamental, 7th , 13th , 19th , etc. They tend to apply
an additional forward force in the direction of the motor
rotation. The negative sequenced harmonics are the 5th ,
11th , 17th , etc. They present a force that opposes the motor
rotation and tries to make the motor rotate in the opposite
direction. The force of these harmonics acting upon each
other creates heat which leads to premature failure.
Harmonic voltage distortion causes increased eddy current
losses in the motors, in the same way as seen for
transformers. 33
 Skin effect losses

The transmission distribution grid was designed to carry the

fundamental 50 Hertz frequency. A problem exists with higher
frequencies (harmonics), that is, they do not fully penetrate the
conductor. They travel on the outer edge of the conductor. This
is called skin effect. When skin effect occurs, the effective
cross sectional area of the conductor decreases; increasing the
resistance and the I2R losses, which in turn heats up the
conductors and anything connected to them. This heating effect
can cause circuit breakers to trip, neutral and phase conductors
to heat up to critical flash over temperatures, and premature
failure of motors and transformers.
This is costly in terms of downtime, loss of production, repair,
and possible reconstruction.
34
In the presence of high-order harmonics, it is necessary to take
skin effect into account, because it affects the life of cables. In
order to overcome this problem, it is possible to use multiple
conductor cables or busbar systems formed by more
elementary isolated conductors.

35
 Harmonic currents in three-phase four-wire networks

n=
uan (t) = Um cos ω t ia(t) = Σ Ian cos (nω t – ϕ an )
n=1
n=
ubn (t) = Um cos (ω t-120) ib(t) = Σ Ibn cos [n(ω t – 120) - ϕ bn ]
n=1
n=
ucn (t) = Um cos (ω t-120) ic(t) = Σ Icn cos [n(ω t + 120) - ϕ36 cn ]
n=1
 Neutral current

If the load is unbalanced, then the neutral connection may

contain currents having spectrum similar to the line currents.

n=
iN(t) = Σ { Ian cos (nω t – ϕ an ) + Ibn cos [n(ω t – 120) – ϕ bn ]
n=1
+ Icn cos [n(ω t + 120) – ϕ cn ] }

In the balanced case, Ian = Ibn = Icn = In and ϕ an =ϕ bn =ϕ cn

= ϕ n , for all n; i.e., the harmonics of the three phases all
have equal amplitudes and phase shifts. The neutral current
is then n=
iN(t) = Σ 3 In cos (nω t –
n=3.6.9..
ϕ n) 37
 n=
iN(t) = Σ 3 In cos (nω t –
n=3.6.9..
ϕ n)
Fundamental and most harmonics cancel out
Triplen (triple-n, or 3, 6, 9, ...) harmonics do not cancel out,
but add.
n=
I2 n
rms neutral current is INrms = 3 Σ
n=3,6,9.. 2

38
39
 Example

A balanced nonlinear load produces line currents containing

fundamental and 20% third harmonic: ian (t) = I1 cos(ω t – ϕ 1)
+ 0.2 I1 cos(3ω t – ϕ 3). Find the rms neutral current, and
compare its amplitude to the rms line current amplitude.
n=
I 2n ( 0.2 I1 )2 I1
INrms = 3 Σ =3
2
= 0.6
n=3,6,9.. 2 2
INrms = 60% of I1rms

The triplen harmonics in the three phases add, such that 20%
third harmonic leads to 60% third harmonic neutral current.
Significant unexpected neutral current flows.
40
 n=
I2 n
Generally the rms current = Σ
n=1 2

n=
I2 n ( I 1 )2 ( I 3 )2
I1rms = Σ =
2
+
2
n=1 2

( I1 )2 (0.2 I1 )2 I1 I1
I1rms = + = 1 + 0.04
2 2 2 2

Yet the presence of the third harmonic has very little effect
on the rms value of the line current.

41
Y-connected nonlinear load, no neutral connection:

If the load is balanced, then it is still true that

n=
iN(t) = Σ 3 In cos (nω t –
n=3.6.9..
ϕ n) 42
But iN (t) = 0, since there is no neutral connection and the ac
line currents cannot contain triplen harmonics.

What happens?

A voltage is induced at the load neutral point, that causes the

line current triplen harmonics to become zero.
The load neutral point voltage contains triplen harmonics.

With an unbalanced load, the line currents can still contain

triplen harmonics

43
 Delta-connected load

There is again no neutral connection, so the ac line currents

contain no triplen harmonics.
The load (phase) currents may contain triplen harmonics: with
a balanced nonlinear load, these circulate around the delta.

44
 Harmonic currents in power factor correction capacitors

PFC capacitors are usually not intended to conduct significant

harmonic currents.
Heating in capacitors is a function of capacitor equivalent
series resistance (esr) and rms current. The maximum
allowable rms current then leads to the capacitor rating:

Irms
rated rms voltage Urms =
2π fC

I2rms
rated reactive power QC =
2π fC

45
46
 Average power

Voltage and current as Power per cycle

Fourier series:
T

u (t) = Σ Un cos (nω t – ϕ n) Pcycle = v(t) i(t) dt

n=1
0

i (t) = Σ In cos (nω t – θ n)

n=1 This is related to average
power as follows:
influence of harmonics on
T
average power: Pcycle 1
Pav = = v(t) i(t) dt
T T
0
T
1
Pav = [Σ Un cos (nω t – ϕ n)]Σ[ In cos (nω t – θ n)] dt
T n=1 n=1
47
0
 Integrals of cross-product terms are zero
T
[Un cos (nω t – ϕ n)] [In cos (mω t – θ m)] dt
0
0 if n = m So net energy is

{
= Un In
2
cos (ϕ n –θ m) if n = m
transmitted to the load
only when the Fourier
series of u(t) and i(t)
contain terms at the
Expression for average power same frequency. For
becomes example, if the voltage
Un In and current both
Pav = Σ cos (ϕ n – θ n) contain third harmonic,
n=1 2 then they lead to the
average power:
U3 I3
cos (ϕ 3 – θ 3)48
2
 Example 1
u (t) i (t)

Voltage: fundamental only

Current: third harmonic only

Power: zero average

49
 Example 2
u (t), i (t)
Voltage: third harmonic only
Current: third harmonic only,
in phase with voltage

Power: nonzero average

50
 Example 3

Fourier series:
u(t) = 1.2 cos (ω t) + 0.33 cos (3ω t) + 0.2 cos (5ω t)
i(t) = 0.6 cos (ω t + 30°) + 0.1 cos (5ω t + 45°) + 0.1 cos (7ω t + 60°)

Average power calculation:

(1.2)(0.6) (0.2)(0.1)
Pav = cos (30°) + cos (45°) = 0.32 W
2 2

51
Voltage: 1st, 3rd, 5th
Current: 1st, 5th, 7th

Power: net energy is

transmitted at
fundamental and fifth
harmonic frequencies

52
In AC circuits the fundamental current and fundamental voltage
together produce fundamental power. This fundamental power
is the useful power that cause motor to rotate and deliver work
on the rotor’s shaft or to make electrical heater to deliver heat.
The product of a harmonic voltage times the corresponding
harmonic current also produces a harmonic power. That one is
usually dissipated as a heat and does not do useful work.
Harmonic currents and voltages should be kept as small as
possible.

The product of a fundamental voltage and a harmonic current

yields zero net power.

53
 GENERATING HARMONICS WITH SWITCH
Irms = 70.7 A
Closed switch
Synchronous I = 1000/10 = 100 A
1000 V switch R
60 Hz 10 P = I2R = 1002x10
Ω = 100 kW

Operational switch
(half time opened)
1410 V
Dissipated power = 50 kW
141 A
I2 = P/R = 50000/10 = 5000
I = 70.7 A

2
P = I2 x R = 70.7 x 10
54
P = 50 kW
 Chopped current
The chopped current can be
1410 V Fundamental decomposed to fundamental
component and harmonics component.
141 A
84 A
The 10Ω resistor absorbs a
fundamental active power
32.50
P = I2 x R = 59.32 x 10 = 35.2 kW
Apparent fundamental power supplied
by source The difference of 50 – 35.2 = 14.8 kW
S = U x I = 1000 x 84/1.414 = 59.3 kVA goes to the harmonic power absorbed
by resistor
Active fundamental power supplied by
source 14.8kW
P = S x cosϕ = 59.3 x cos 32.5 = 50 IF = 59.3 ∠-32.5 A
kW
31.9kVAr
Reactive fundamental power supplied
14.8kW
by source 1 kV

Q = V S2 - P2 = V 59.32 – 502 = 31.9 50kW 35.2kW 55

kVAr 10
The switch carries a fundamental current of 59.3 A and it
absorbs 14.8 kW and 31.9 kVAr, it can be represented by
resistance and inductive reactance connected in series.
4.21Ω j9.07Ω
R = P / I2 = 14800 / 59.32 = 4.21 Ω
X = Q / I2 = 31900 / 59.32 = 9.07 Ω
1kV 10Ω
Effective value of the IF = 59.3A
harmonic current is
Equivalent circuit for the
IH = V I2 - IF2 = V 70.72 – 59.32 fundamental component
= 38.5 A

Consequently the voltage 385 V

10Ω
across the 10Ω resistor IH = 38.5A
is U = I x R = 38.5 x 10 =
385 V. Equivalent circuit for all
harmonic components 56
57
 Low-power harmonic limits

In a city environment such as a large building, a large fraction

of the total power system load can be nonlinear
• Example: a major portion of the electrical load in a building
is comprised of fluorescent lights, which present a very
nonlinear characteristic to the utility system.
• A modern office may also contain a large number of
personal computers, printers, copiers, etc., each of which
may employ peak detection rectifiers.
• Although each individual load is a negligible fraction of the
total local load, these loads can collectively become
significant.

58
 Short Term Effects

Over consumption of RMS current

Unwanted tripping of protections
Malfunction of sensitive applications
Interference of remote control and telecommunication systems
Abnormal vibration and noise (LV panels, motors, transformers)

Long Term Effects-Overheating

Overheating of capacitor banks

Overheating of transformers, alternators
Overheating of phases, particularly neutral

59
 Harmful Effects on Receivers

Cables:
Overheating of cables
Additional losses due to skin effect
Increase in dielectric losses of insulation

Induction motors:
Increase in core (stator) and Joule losses
Pulsating torques causing efficiency reduction,
abnormal vibration, rotor overheating

60
 General Solutions

Limit injected harmonic currents:

Install limitation induction coils for speed drives
Install specific rectifiers called active front end

Install anti-harmonics induction coils

Install filters to trap harmonics:

Passive filters
Active filters
Hybrid filters

Oversize equipment

61
 Harmonic Currents add in the Neutral

The 120°
phase shift
between linear
load currents
will result in
their balanced
portions
instantaneously
canceling in the
neutral.
With linear loads, the neutral can be the same size as the
phase conductors because the neutral current cannot be
larger than the largest phase current, even when the load
is completely unbalanced. 62
When the load is non-
linear however, the
current pulse on one
phase will not have a
pulse on either of the
other phases for
which to cancel. The
pulses are additive
which often leads to
heavier current on the
neutral conductor than
on any phase
conductor. The
frequency of this With non-linear loads, the neutral
neutral current is current generally exceeds the largest
primarily 150 Hz (3rd phase current, even when the loads
harmonic). are in perfect RMS current balance.63
The presence of harmonic currents can also lead to some
special problems in three-phase systems:
• In a four-wire three-phase system, harmonic currents can
lead to large currents in the neutral conductors, which may
easily exceed the conductor rms current rating
• Power factor correction capacitors may experience
significantly increased rms currents, causing them to fail

64
 K factor

Harmonic currents are generated whenever a non-linear load

is connected to the mains supply. The problems caused by
harmonic currents include overheating of cables, especially
the neutral conductor, overheating and vibration in induction
motors and increased losses in transformers.
Where power factor capacitors are fitted, harmonic currents
can damage them and care must be taken to avoid
resonance with the supply inductance.
Losses in transformers are due to stray magnetic losses in
the core, and eddy current and resistive losses in the
windings. Of these, eddy current losses are of most concern
when harmonics are present, because they increase
approximately with the square of the frequency. Before the
excess losses can be determined, the harmonic spectrum of
the load current must be known. 65
 The eddy current loss at a particular harmonic is given by:

Pn = Pf In2 n2

where
Pn is the eddy current loss at harmonic number n
Pf is the eddy current loss at the fundamental frequency f
In is the fraction of total rms load current at harmonic number n
The total eddy current loss is given by summing the losses for
the individual harmonics and the fundamental:

n = max
Pt = Pf Σn=1
In2 n2

where Pt is the total eddy current loss. 66

There are two distinct approaches to accounting for this
increased eddy current loss in selecting a transformer. The
first, devised by transformer manufacturers in conjunction with
Underwriters Laboratories in the United States, is to calculate
the factor increase in eddy current loss and specify
a transformer designed to cope; this is known as ‘K-Factor’.

The second method, used in Europe, is to estimate by how

much a standard transformer should be de-rated so that the
total loss on harmonic load does not exceed the fundamental
design loss; this is known as ‘factor K’. The figures produced
by each method are numerically different; ‘factor K’ is a total
rating factor while ‘K-factor’ is a multiplier (although a de-rating
factor can be derived from it). The fact that both methods use
K as a designation can lead to confusion when talking to
67
suppliers.
 K-Factor

In US practice, where dry-type transformers are often used,

the K-factor is the ratio of eddy current losses when driving
non-linear and linear loads:

Pt n = max
K= = Σ In2 n2
Pf n=1

This K-factor is read directly by many power meters (e.g.

Fluke 41 & 43). Once the K-Factor of the load has been
determined, it is a simple matter to specify a transformer
with a higher K-rating from the standard range of 4, 9, 13,
20, 30, 40, 50.

68
 Factor K

In Europe, the transformer de-rating factor is calculated

according to the formulae in BS 7821 Part 4. The factor K is
given by:
n=N
e I1 2 In 2 0.5
K = { 1+ [ ] Σ [ n (
q
) ]}
1+e I n=2 I1
where

e is the eddy current loss at the fundamental frequency divided

by the loss due to a dc current equal to the RMS value of the
sinusoidal current, both at reference temperature.
n is the harmonic order
I is the rms value of the sinusoidal current including all
harmonics given by:
69
 n=N n=N In
0.5
I= [ Σ ( In) ]
2
= I1[ Σ (
2 0.5
)]
n=1 n=1 I1

In is the magnitude of the nth harmonic

I1 is the magnitude of the fundamental current
q is an exponential constant that is dependent on the type
of winding and frequency. Typical values are 1.7 for
transformers with round or rectangular cross section
conductors in both windings and 1.5 for those with foil low
voltage windings.

70
A standardized empirical formula (NFC 52-114) is used to
calculate the de-rating factor k to be applied to a transformer.

For example where

H5 = 25% ; H7 = 14% ; H11 = 9% ; H13 = 8%,
the factor k is 0.91.

71
 K-Rated or De-Rated?

The great advantage of a ‘K-rated’ transformer is that it will

have been designed with harmonic loads in mind and care
will have been taken to keep losses low. For example, eddy
current losses will have been reduced by the use of stranded
conductors and magnetic losses will have been reduced by
the use of low loss steels. The neutral point connections are
usually brought out individually, so that the star point has a
300% current rating.

72
On the other hand, de-rating a standard transformer has a
number of disadvantages. Because the transformer is
oversized, the primary over-current protection level may be too
high to protect the secondary, but if the protection level is
reduced, the inrush current may cause tripping. A de-rated
transformer is less efficient; the excess losses are still being
generated and dissipated within the transformer, rather than
being designed out, and a larger core than necessary, with
larger losses, is being magnetized.

There is also a potential maintenance problem – long after

installation, changes in the needs of the facility may result in
additional load being added without reference back to the initial
de-rating. This may lead to overloading and consequent
failure.

73
 Typical calculation according to BS 7821 Part 4
(taking q as 1.7 and assuming that eddy current loss at
fundamental is 10% of resistive loss i.e. e = 0.1).
Harmonic RMS In/I1 (In/I1)
2
nq nq (In/I1)
2

No. current (In)

1 1 1 1
3 0.82 0.82 0.6724 6.473 4.3525
5 0.58 0.58 0.3364 15.426 5.1893
7 0.38 0.38 0.1444 27.332 3.9467
9 0.18 0.18 0.0324 41.900 1.3576
11 0.045 0.045 0.0020 58.934 0.1193
Σ =2.1876 Σ =14.9653

Total rms (I1/I)2 = e/(1+e) = K2=1+(0.091x0.457x de-rate74to

Typical calculation according to Underwriters’ Laboratories
method:

Harmonic RMS In/I1 (In/I1)

2
In/I (In/I)2 (In/I)2 x n2
No. current (In)

1 1 1 1 0.6761 0.4571 0.4571

3 0.82 0.82 0.6724 0.5544 0.3073 2.7663
5 0.58 0.58 0.3364 0.3921 0.1538 3.8444
7 0.38 0.38 0.1444 0.2569 0.0660 3.2344
9 0.18 0.18 0.0324 0.1217 0.0148 1.2000
11 0.045 0.045 0.0020 0.0304 0.0009 0.1120
Σ =2.1876 11.6138

Total rms (I) = K-factor 11.6138

75
1.479
 Failure of power factor capacitors

The impedance of a circuit dictates the current flow in that

circuit.
As the supply impedance is generally considered to be
inductive, the network impedance increases with frequency
while the impedance of a capacitor decreases.

Inductive reactance X(Ω ) XL

XL = 2 π f L (Ω )
XL = XC
f (hz)
Capacitive reactance
1 resonant
XC = (Ω ) frequency
2π fC XC
76
 1 1
XL = XC = X = 2 π f L = fr =
2π fC 2π LC

If this condition occurs on, or close to, one of the harmonics

generated by non linear load, then large harmonic currents
can circulate between the supply network and the capacitor
equipment. These currents are limited only by the damping
resistance in the circuit. Such currents will add to the harmonic
voltage disturbance in the network causing an increased
voltage distortion. This results in a higher voltage across the
capacitor and excessive current through all capacitor
components. Resonance can occur on any frequency, but in
general, the resonance we are concerned with is on, or close
to, the 5th, 7th, 11th and 13th harmonics for 6 pulse systems.
77
 Example of the resonant frequency calculation

3 phase 11/0.4 kV distribution transformer of 1.6 MVA, with

(uk) 5.82 % impedance, would have a short circuit current at
main panel of
S 1600
ISC = = = 39681 A
3 U2 uk 1.732 x 0.4 x 0.0582

and short circuit power of

SSC = 3 ISC U2 = 1.732 x 39681 x 0.4 = 27491 kVA

Assuming the size PFC capacitors of 230 kVAr

SSC 27491
System harmonic calculation: n = = = 10.9
QC 230 78
This calculation indicates that the resonant frequency can
develop if the distribution system contains the 11th harmonic
in any significant amount, causing the capacitor to overheat
and possibly, explode.
Using detuning reactors
This solution consists of protecting the capacitors, designed
to improve the displacement power factor by installing a
series reactor. This reactor is calculated so that resonance
frequency matches none of the harmonics present.
Typical tuning frequencies are for a 50 Hz fundamental: 135
Hz (order 2.7), 190 Hz (order 3.8) and 255 Hz (order 4.5).
Thus for the fundamental, the battery can perform its
displacement power factor improvement function, while the
high impedance of the reactor limits amplitude of the
harmonic currents. 79
PFC capacitors are usually not intended to conduct significant
harmonic currents.
Heating in capacitors is a function of capacitor equivalent
series resistance (esr) and rms current. The maximum
allowable rms current then leads to the capacitor rating:

Irms
rated rms voltage Urms =
2π fC

I2rms
rated reactive power QC =
2π fC

80
 Evaluating System Harmonics
In order to prevent or correct harmonic problems that could
occur within an industrial facility, an evaluation of system
harmonics should be performed if the facility conditions
meet one or more of the criteria below.
1.The application of capacitor banks in systems where 20% or
more of the load includes other harmonic generating equipment
or where background distortion exceeds 2%.
2.The facility has a history of harmonic related problems,
including excessive capacitor fuse operation.
3.Large single non-linear loads are being added greater than
about 10% of the transformer rating.
4.Many small identical non-linear loads are being added that
operate together.
In facilities where restrictive power company requirements limit
the harmonic injection back into their system to very small
81
magnitudes.
 Mitigation of Harmonics

There are many ways to reduce harmonics, ranging from

variable frequency drive designs to the addition of auxiliary
equipment. Following are some of the more common
methods used today for controlling power system harmonics.

Addition of a
downstream
reactor or
reduction in
upstream source
impedance
reduces voltage
THD at the point
considered.
82
 Carefully choosing the installation structure

Sensitive loads should not be parallel-connected with non-

linear loads

Solution to avoid Solution to recommended

Very powerful non-linear loads should preferably be

supplied by another MV/LV transformer.
83
 Passive harmonic filters
Passive or ‘trap’ filters employ ‘passive’ elements (capacitors
and inductors) to ‘trap’ or absorb harmonics. Passive filters
are generally configured to remove only one or two specific
harmonics. Passive filters are generally regarded as
unsuitable for filtering 3rd harmonics. For this reason, they are
best suited for applications in which 3rd harmonics are not an
issue, power factor correction is
required, and specific harmonics such as
5th or 7th are creating the problem.
Passive filters are ideal for systems that Non
have a high percentage of 6 pulse drives linear
load
and other linear loads. However, the
filters may need to be retuned for
changes in the power system. Filters can Harmonic
be designed for several nonlinear loads trap filter 84
or for an individual load,
85
Principle of compensation of harmonic components by “shunt-
type” active harmonic conditioner

The device should able

to inject at any time a
current where each
harmonic current has
the same amplitude as
that of the current in the
load and is in opposition
of phases, then
Kirchoff’s law at point A
guarantees that the
current supplied by the
source is purely
sinusoidal
86
Isolation Transformers: An isolation transformer provides
several advantages. First and foremost, it provides
impedance to the drive, which reduces current distortion. It
obviously resolves voltage mismatch between the supply
and the load. If the secondary is grounded, it isolates ground
faults and reduces common mode noise.

87
 Engineering recommendation G5/4

The Engineering Recommendation G5/3 has been in place for

decade and a half and has provided certain guidelines. This
include limits of harmonic currents that can be fed into the
system by customers, limits for harmonic distortion caused,
suggestion for whether a load can be connected and procedures
for measurement and assessment of new loads. The new G5/4
is more stringent than G5/3, requiring more careful assessment
of new loads and measurements to be taken at different voltage
levels. It also extends the range of harmonics covered and
includes inter harmonics and notching currents.
The original edition of G5/4 included certain limits which were to
be applied from1st January 2005. However, after discussion with
industry it was agreed to modify this requirement, and this has
resulted in G5/4-1:2005
88
The standard G5/4 published in March 2001,seeks to limit
harmonic voltage distortion levels on public networks at the
time of connection of new non-linear loads to ensure
compatibility of all connected equipment.

It does this by seeking data from the customer and then

making an assessment to see whether the planning limits are
likely to exceed at the time of connection

The enforcement of this is via the Electricity Supply

Regulations, the Grid and Distribution Codes, and the
connection agreements between NOCs and customers.

89
 Features of G5/4

G5/4 defines planning levels and introduces compatibility

levels for individual harmonics and THD over the voltage
range from 400V to 400 kV.

Emphasis placed on voltage distortion levels in Stage 2 and 3

assessment (compared with other standard)

The three stage assessment process of G5/3 retained

Information on harmonic impedance for use in network

modeling has been updated.

Description material and examples moved to the Application

Guide ETR 122.
90
Uses IEC standards wherever possible

Introduces specific emission requirements for number of

aggregated low voltage equipment

THD assessment required up to 50th harmonic

5th harmonic current emissions levels reduced

Harmonic combination rules clarified to account for harmonic

phase angles likely to be

Harmonic emissions modifiable relative to fault level

Introduces a flow chart to help users through the assessment

process
91
 What are the G5/4 limits?

Total harmonic distortion limits are recommended based upon

the type of installation. It is the defacto standard for power
utilities and is often required of large consumers and medium
voltage systems. This specification has also become
increasingly common in low voltage system.

Total voltage distortion not to exceed 5% in 400V systems and

4% in 6.6kV, 11kV and 20 kV systems, and 3% for the other
systems with voltage level above 20kV.

92
Compatibility levels in voltage systems are based on the
immunity of capacitors as they are susceptible to harmonic
voltage distortion and are common in use.

Gaps between compatibility levels and planning

levels for THD harmonics level
Voltage level Compatibility Planning level
(kV) level
0.4 8% 5%
6.6 – 20 8% 4%
>20 – 36.5 8% 3%
66 - 145 5% 3%
275 - 400 3.5 % 3%
93
Total current distortion is
limited on harmonic by
harmonic basis measured in
absolute RMS amps.
Table 7 is based on a system
fault level of 10MVA at 400V.
If the fault levels varies from
this base level, the figures in
Table 7 may be varied
prorate, as will the powers
that can be connected.

94
Voltage distortion limits

95
96
97
98
99
100
101
102