UNIT III

MITIGATION OF POWER SYSTEM HARMONICS

Introduction- principle of harmonic filters-Series tuned filters – Double band pass filters –
Damped filters – Detuned filters – Active filters- Power converters- Harmonic filter design –
Tuned filter – Second order Damped filter – Impedance plots for filter banks – Impedance
plots for a three branch 33 kv filter

3.1 Introduction
This unit is devoted to the reduction or suppression of power system harmonics.
Particular attention on how to attain this in converters, capacitor banks, transformers and
rotating machines is given.
Precautionary (Preventive) solutions are those policies sought for at discretion to avoid
harmonics and their consequences.
These include:
 Phase cancellation or harmonic control in power converters.
 Developing procedures and methods to control, reduce or eliminate harmonics in
power system equipment; mainly capacitors, transformers and generators.
Attempting at keeping harmonics at a low "damage-free" level, standards are
further developed setting limits on the level of individual frequency harmonics
and/or harmonic distortion factors.

Corrective (Remedial) solutions are those techniques recoursed to aiming at overcoming existing
harmonic problems.
They include:
 The use of filters.
 Circuit detuning which involves the reconfiguration of feeders or relocation of capacitor
banks to overcome resonance.
3.2 Harmonic Filters
Harmonics can be efficiently reduced through the use of a passive filter which consists, basically, of a
series combination of a capacitor and a reactor tuned to a specific harmonic frequency. Filters provide
a low impedancetrap” to a harmonic to which the filter is tuned. Theoretically, the filter has a zero
Impedance at the tuning frequency thus absorbing the harmonic of interest.
3.2.1 Series-Tuned Filters
 A series-tuned filter consists of a series combination of a capacitor and a Reactor and is tuned
to low harmonic frequencies.
 At the tuned harmonic, the Capacitor and the reactor have equal reactances and the filter has a
purely Resistive impedance.
 The filter’s impedance is capacitive for lower harmonics And inductive for higher harmonics,
a consequence of which is aggravating The impedance below the lowest tuned frequency.

3.2.2 Double Band-Pass Filters

 A Double band-pass filter is a series combination of a main capacitor, a main Reactor and a
tuning device which consists of a tuning capacitor and a Tuning reactor connected in parallel.
The impedance of such a filter is low at Two tuned frequencies.
3.2.3 Damped Filters
 Damped filters can be 1st, 2nd or 3rd_order. However, the most commonly Used is the
2nd-order. A 2nd-order damped filter consists of a capacitor in Series with a parallel
combination [4, 6] of a reactor and a resistor.
 It provides A low impedance for a moderately wide range of frequencies. When used
to eliminate high order harmonics (17th and above), a damped Filter is referred to as a
high-pass filter, providing a low impedance for high Frequencies but stopping low
ones.
 Damped filters have a low quality factor, 0.5 < Q < 5, and are usually Tuned to hn<
hr, that is 10.7, 16.5, …

3.2.4 Detuned (Anti-Resonant) Filters

A detuned filter is tuned below a characteristic harmonic (usually tuned To the fourth
harmonic), thus absorbing some of the harmonic but not asMuch as a higher tuned one.

Fig 3.1: Typical harmonic filters: (i) Series tuned (ii) Double band-pass (iii) 1 st ,2nd,3rd - order damped

3.2.5 Active Filters

Active filters are being developed to alleviate the disadvantages of conventional passive
filters, namely:
 The filtering characteristics being dependent on the source impedance.
 Aggravating the impedance below the lowest tuned harmonic.
 Being inadequate for filtering non-characteristic harmonics (different from the filter's
tuned frequency), such as those produced by cycloconverters.
3.3 Power Converters
 Harmonics of pulse converters constructed through the operation of lower pulse
number converters can be eliminated through the proper selection of phase shifts.
This is called phase cancellation or phase multiplication. Analysis revealed that 5th,
7th, 17th, 19th, · · · harmonics are eliminated in two six-pulse converters operating in
parallel or series with 0 and -30° phase shifts.
 This means that a twelve-pulse converter has a lower harmonic impact than two six-
pulse units of a comparable size. Grady further points out that two twelve-pulse
converters operating in parallel or series through +7.5° and -7.5° phase shifts
eliminate eleventh and thirteenth harmonics. A consequential conclusion is to use
converters with higher number of pulses.
3.4 Transformers
 Harmonics are reduced through the ingenuity of transformer connection.
 Delta-connected transformers prevent the flow of zero sequence triplen harmonics,
 thus acting as two-way filters protecting both the source and load sides of a power
system. .
 Harmonic currents can be damaging to transformers designed to operate at 50 / 60 Hz.
Oversizing the neutral conductor and derating the transformer represents a short-term
solution.
 K-rated transformers are specifically designed to tolerate harmonics and have the
following features :
 They have lower than normal flux densities and can, thus, tolerate overvoltages
coupled with circulating harmonic currents.
 They employ an electromagnetic shield between the primary and secondary windings
of each coil, thus attenuating higher frequency harmonics.
 They provide a neutral with twice the size of a phase conductor, to account for
increased neutral currents due to the flow of triplen harmonics
 Windings are designed with several smaller sizes parallel conductors, therefore
reducing skin effect at higher frequency harmonics.
 They use insulated and transposed conductors resulting in reduced losses.

The K-factor is a useful term for describing the additional heating that occurs in a transformer
supplying non-linear loads, and is defined as

(3.1)
Normalising by the rms current through the transformer, the K-factor becomes

(3.2a)

(3.2b)

Transformers specifically designed for use with non-linear loads are marked “Suitable for
non-sinusoidal current load with K-factor not to exceed … “ Where standardised K-factor
ratings are 4, 9, 13, 20, 30, 40, 50 .

Should the K-factor exceed 4, it becomes necessary to use a K-rated transformer or derate a
standard transformer.

The derating factor for standard non-harmonic rated transformers, per method 1 ofiEEE
C57.110-1986,

(3.3)

Example 3.1.
The current through a 3 MVA, 11000/690 V, 157.5/2510 A, -Yg,
, Z = 1 + j6.2 % transformer is given in the following table. Can the
transformer operate in this harmonic environment? Provide a plot for the
transformer current waveform neglecting phase angles.

Table 3.1.11000/690 V transformer current

Solution: Based on the harmonic current spectrum of Table 3.1, transformer current is
plotted in Fig. 3.2. Transformer current THD is further calculated to be 25.18 %.

Accordingly, should a standard non-harmonic rated transformer be used, the derating factor is

As the K-factor was calculated to be 4.171, one has one of two options:
• Use a standard transformer and make sure it is derated (no more than 70 % loaded).

• Use a K-rated transformer with K = 4. The transformer can then be fully loaded.
Fig. 3.2. Transformer current

3.5 Rotating Machines

Inspection of Equations shows that
kph<kp1< 1
kdh<kd1< 1

Eh < E1.
This makes it clear that, although distribution and chording result in reducing the
fundamental component of the induced emf, this negative effect is more than outweighed
when considering the purpose they were intended for, namely harmonics.
Obviously, they have the advantage of reducing harmonic voltages. Recalling that
(3.4)
Where
Kph is the pitch factor at the hth harmonic.
Kp1 is the pitch factor at the fundamental frequency, kp1 = sin(, /2) < 1 ,
is the coil span in °elect., , =s
s is the coil span in slots,
it is noticed that any one harmonic can be completely eliminated through selecting a coil span
(fractional pitch) that results in the respective pitch factor being zero.

(3.5)

That is to say, a coil span of 4/5 pole pitch (144 °elect.) results in eliminating the fifth harmonic.

The third harmonic is suppressed through using a coil Of 2/3 pole pitch (120 °elect.) span.
Furthermore, a coil span of 5/6 pole Pitch (150 °elect.) greatly reduces fifth and seventh harmonics

(kp5 = kp7 = 0.2588).

  Finally, damper (squirrel-cage) windings, consisting of copper bars that are short-circuited at
both ends, are introduced into the pole shoes of Rotating machines to, as the name implies
and among other functions, damp the effect of backward mmf.
 Bearing in mind that harmonic-generated fields Interact with the rotor field of rotating
machines to produce a pulsating Field, this pulsating field can be thought of as the
combination of two fields Rotating in the forward and backward directions.
 The backward component does not contribute to the energy transfer 1 but produces double-
frequency Eddy-currents in the pole shoes. Eddy-currents flowing in the damper winding Set-
up a flux that, by Lenz’s law, counterbalances the backward field thus Keeping it at a low
level.

3.6 Capacitor Banks

• Relocating capacitors changes the source-to-capacitor inductive reactance thus avoiding parallel
resonance with the supply.

• Varying the reactive power output of a capacitor bank will alter the resonant frequency.
Capacitors can be designed to trap a certain harmonic by employing a tuning reactor whose
inductive reactance is equal to the capacitive

1 For energy transfer between the electrical and mechanical channels, the following conditions
should be satisfied:
• mmf from both the stator and rotor sides should exist.

. • The two mmf's should be relatively at rest.

• The two mmf's should have the same number of poles.
reactance of the capacitor at the tuned frequency. Parallel resonance involving a capacitor and a
source inductance is achieved when

Equation 3.6 can be rewritten as

The above says that a capacitor with a reactance of X c1 = hr2XS1 excites resonance at the hrth harmonic
frequency. Tuning the capacitor to a certain harmonic, alternatively, designing the capacitor to trap
(filter) a certain harmonic, requires the addition of a reactor. At the tuned harmonic,
 (3.8)

(3.9)

Should the capacitor be tuned to the harmonic activating resonance, i.e. h n = hr, Equation 3.9 reduces
to XL1= XS1· Expressed differently, Equation 3.8 becomes

(3.10)

which clarifies that a reactor with XL 1 = 0.04Xc1 represents a fifth harmonic filter (hn = ..)1/0.04 =
5).

Example 5.2.
A 33 kV, 6.2 MVA capacitor bank is to be installed at a bus where the short-circuit capacity is 750
MVA. Investigate resonance and avoid the problem! Solution Interaction between the capacitor bank
and the system would cause resonance at hr = lr = !Xc = rscc = 11 lo YX: YTc kV2 332 Xc = - = - =
175.645 n Qc 6.2 114 5. Mitigation of Power System Harmonics Design the capacitor bank to trap the
eleventh harmonic by adding a reactor in series with XL= Xc = 175.645 _ 1 452 [l h2 112 - . n With Q =
60, the reactor would have a resistance R of Xn = JxLXc = 15.968 fl R = ~ = 0.266 fl The filter's
impedance, plotted in Fig. 5.3, is Rated filter size is kV2 h2 QFilter = Xc _XL = h~ ~ 1 · Qc = 6.252
MVAr.

5. 7 Harmonic Filter Design

Tuning a capacitor to a certain harmonic, alternatively, designing the capacitor to trap (filter) a certain
harmonic, requires the addition of a reactor.At the tuned harmonic

(3.11)

Being sensitive to peak voltages, the capacitor needs to be able to withstand the total peak voltage
across it. That is, it needs to have a voltage rating equal to the algebraic sum of the fundamental and
tuned harmonic voltages.
(3.12)

However, since it is likely that a capacitor tuned to a certain harmonic will absorb other harmonics, a
safety measure would be to let the capacitor have a voltage rating of

(3.13)

While the rms voltage is

(3.14)

(3.15a)
5.7.1 Series-Tuned Filters

(3.15b)

(3.16a)

(3.16b)

5.7.1 Series-Tuned Filters

A series-tuned filter is a capacitor designed to trap a certain harmonic by

adding a reactor with XL = XC at the tuned frequency fn.

(3.17)

(3.18)
 (3.19)

(3.20)

(3.21)
Example 3.3. A filter is tuned to the 13th harmonic. Given X C = 507 , calculate the filter elements
and plots its impedance.

Filter impedance is plotted in Fig. 3.4.

Example 3.4. What is the tuning order and the quality factor for a 33 kV series-tuned filter with X C =
544.5 , XL= 4.5 and R = 0.825 ?
Fig.3.4. Impedance of a series-tuned filter

The filter's impedance, ZF(h) = R + j(hXL- Xc/h), is plotted in Fig. 3.5.

Fig. 3.5. Eleventh-tuned 33 kV filter impedance

Example 3.5. Which harmonics are being trapped by a filter comprising four series-tuned circuits
with 3¢, 50 Hz, 400 V, Y -connected, lx30+3x20 kVAr capacitor banks and 0.779, 0.583, 0.233 and
0.166 mH reactors?

Table 3.2. Four-branch filter data- 400 V

The filter's impedance, which is the parallel equivalent of four series-tuned circuits with Z F(h) = R +
J(hXL - Xc/h), is plotted in Figs. 3.6 and 3.7.
Fig. 3.6. Four-branch 400 V filter impedance locus
Fig. 3. 7. Four-branch 400 V filter impedance as a function of frequency

3.7.2 Second-Order Damped Filters

Design steps: For a second-order damped filter tuned to the hn harmonic:

• Determine the capacitor size Qc in MVAr, say the reactive power requirement of the harmonic
source.

• The capacitor's reactance is

• To trap the hn harmonic, the reactor should have a size of

• The resistor bank has a size of

where Q is the filter's quality factor, 0.5 < Q < 5.

The characteristic reactance is given by

Filter size is then

(3.22)

For a second-order damped filter, the impedance at any harmonic h becomes

(3.23a)
 (3.23b)

(3.24)

(3.25a)

(3.25b)

Power loss in the resistor is

(3.26)

Example 5. 6. A second-order damped filter is tuned to hn> 17. Knowing

XC= 1.734 D, calculate the filter elements and plot its impedance.
The filter impedance is plotted in Fig. 3.8.

Fig. 3 .8. Impedance of a second-order damped filter

Solution Filter elements are calculated as follows and the impedance is plotted in Fig. 3.9.

Rated filter size is

Fig. 3.9. Second-order damped filter impedance

Example 5. 7. A 33 kV, 6.8 MVAr capacitor bank is to be used as a secondorder damped filter tuned to
hn 4. Find the elements of the filter.

Rated filter size is

Fig. 3.9. Second-order damped filter impedance

5.7.3 Impedance Plots for Filter Banks

In the following, we present plots for the impedance of three different filter

banks, namely:

• a three-branch 33 kV filter; 7th and 11th tuned plus a highpass (secondorder damped) branch for all
harmonics from the 17th and above.

• a four-branch 20 kV filter; 5th, 7th, nth and 13th tuned .

• a five-branch 690 v filter; 5th' 7th, 11th and 13th tuned plus a high pass

(second-order damped) branch for all harmonics from the 17th and above.

3.7.3.1 Impedance Plots for a Three-Branch 33 kV Filter

Consider the 33 kV filter of Table 3.3. Requested is to provide plots for the

filter impedance and calculate the filter rating.

Recalling that
filter elements are readily calculated and provided in Table 3.3. The filter rated size is calculated as

Ten plots for the filter impedance (magnitude and angle) are presented in Figs. 3.10 and 3.11.

Table 3.3. Filter data- 33 kV

Fig. 3.10. T hree-branch filter impedance and impedance locus
Fig. 3.11. Impedance plots for the three-branch filter

3.7.3.2 Impedance Plots for a Four-Branch 20 kV Filter

The 20 kV filter of Fig. 3.12 has four series-tuned branches; 5th-tuned with Q C = 2.4 MVAr, 7th-tuned
with Qc = 1.2 MVAr, 11th-tuned with Q C = 620 kVAr, and 13th-tuned with Q C = 620 kVAr. All filters
have a quality factor Q of 30. Requested is to plot the impedance for the different branches and the
filter as a whole.

The filter rating is calculated as

filter elements are calculated and provided in Table 3.4. Twelve plots for the filter impedance
(magnitude and angle) are presented in Figs. 3.13 and 3.14.

Fig. 3.12. A 20 kV filter

Table 3.4. Four-branch filter data- 20 Kv

Fig. 3.13. Four-branch filter impedance locus
Fig. 3 .14. Four-branch filter impedance as a function of frequency

3.7.3.3 Impedance Plots for a Five-Branch 690 V Filter

Consider the five-branch 690 V filter of Fig.3.15. Filter data is provided in Table 3.5. Requested is to
calculate its rated size and plot its impedance. Filter elements in Table 3.5 can be readily calculated
knowing
Impedance plots are provided in Figs. 3.16 and 3.17.

Fig. 3.15. Five-branch filter configuration

Table 3.5. Five-branch filter data - 690 V

Fig. 5.16. Impedance plots for the series-tuned branches
Fig. 3.17. Five-branch filter impedance and impedance locus

5.8 Active Filters

An active filter generates a harmonic current spectrum that is opposite in phase to the distorted
harmonic current it measures. Harmonics are thus cancelled and the result is a non-distorted
sinusoidal current.

In the following are five different examples to illustrate the principle of active filtering. Each example
presents a sample distorted current waveform and the corresponding current generated by the
active filter to cancel harmonics. The plots show the distorted current and the active filter current
compensation. The resultant is a sinusoidal waveform.
• Single-phase rectifier with a smoothing capacitor (6-pulse switched-mode power supply)

Fig. 3.18. Distorted current and active filter current

• Six-pulse three-phase diode rectifier

Fig. 3.19. Distorted current and active filter current

Six-pulse converter
Fig. 3.20. Distorted current and active filter current

Fig. 3.21. Distorted current and active filter current

\
Fig. 3.22. Distorted current and active filter current

Fig. 5.23. Distorted current and active filter current

Fig. 5.24. Distorted current and active filter curren

das